From scoder at codespeak.net  Sat Apr  8 23:16:38 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr  8 23:16:39 2006
Subject: [Lxml-checkins] r25600 - in lxml/branch/xsltext/src/lxml: . tests
Message-ID: <20060408211638.0C4E91024B@code0.codespeak.net>

Author: scoder
Date: Sat Apr  8 23:16:35 2006
New Revision: 25600

Modified:
   lxml/branch/xsltext/src/lxml/etree.pyx
   lxml/branch/xsltext/src/lxml/nsclasses.pxi
   lxml/branch/xsltext/src/lxml/tests/test_xslt.py
   lxml/branch/xsltext/src/lxml/xslt.pxd
   lxml/branch/xsltext/src/lxml/xslt.pxi
Log:
API rewrite: XSLT elements are now classes that inherit from XSLTElementBase (i.e. they represent real elements), instances are callable to apply the element [currently not working]

Modified: lxml/branch/xsltext/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/xsltext/src/lxml/etree.pyx	(original)
+++ lxml/branch/xsltext/src/lxml/etree.pyx	Sat Apr  8 23:16:35 2006
@@ -768,7 +768,6 @@
         return XPathElementEvaluator(self, namespaces).evaluate(_path, **_variables)
 
 cdef _Element _elementFactory(_Document doc, xmlNode* c_node):
-    cdef _Element result
     cdef char* c_ns_href
     result = getProxy(c_node, PROXY_ELEMENT)
     if result is not None:
@@ -785,6 +784,11 @@
         element_class = _Comment
     else:
         assert 0, "Unknown node type: %s" % c_node.type
+    return _newElement(doc, c_node, element_class)
+
+cdef _Element _newElement(_Document doc, xmlNode* c_node,
+                          object element_class):
+    cdef _Element result
     result = element_class()
     result._tag = None
     result._doc = doc

Modified: lxml/branch/xsltext/src/lxml/nsclasses.pxi
==============================================================================
--- lxml/branch/xsltext/src/lxml/nsclasses.pxi	(original)
+++ lxml/branch/xsltext/src/lxml/nsclasses.pxi	Sat Apr  8 23:16:35 2006
@@ -28,7 +28,7 @@
         return __CLASS_NAMESPACE_REGISTRIES[ns_utf]
     except KeyError:
         registry = __CLASS_NAMESPACE_REGISTRIES[ns_utf] = \
-                   _ClassNamespaceRegistry(ns_uri)
+                   _ClassNamespaceRegistry(ns_uri, ElementBase)
         return registry
 
 def FunctionNamespace(ns_uri):
@@ -54,18 +54,20 @@
 def XSLTElementNamespace(ns_uri):
     """Retrieve the XSLT element namespace object associated with the given
     URI.  Creates a new one if it does not yet exist.  This namespace can only
-    be used to register functions for XSLT extension elements.
+    be used to register subtypes of XSLTElementBase for XSLT extension
+    elements.
 
-    The registered functions must have the signature
+    Instances of the registered element type must be callable with the
+    following signature:
 
-    result = function(_, subtree)
+    result = instance(subtree)
 
-    'subtree' is the current result tree from inside the element.  It may be a
-    string value, a node or a sequence of nodes.  The element function can
-    freely operate on this.
+    'subtree' is the context node of the current template.  It may be a string
+    value or a node.  The element function must not modify this value.
 
-    The function must return either a None, a node, a sequence of nodes or a
-    string value.
+    The function must return either None, a node, a sequence of nodes or a
+    string value.  It may call its own 'apply_templates' method on any of its
+    children (only one at a time).
     """
     if ns_uri is not None:
         ns_utf = _utf8(ns_uri)
@@ -73,14 +75,16 @@
         return __XSLT_ELEMENT_NAMESPACE_REGISTRIES[ns_utf]
     except KeyError:
         registry = __XSLT_ELEMENT_NAMESPACE_REGISTRIES[ns_utf] = \
-                   _FunctionNamespaceRegistry(ns_uri)
+                   _ClassNamespaceRegistry(ns_uri, XSLTElementBase)
         return registry
 
 cdef class _NamespaceRegistry:
     "Dictionary-like namespace registry"
     cdef object _ns_uri
+    cdef object _entries
     def __init__(self, ns_uri):
         self._ns_uri = ns_uri
+        self._entries = {}
 
     def update(self, class_dict_iterable):
         """Forgivingly update the registry. If registered values do not match
@@ -102,61 +106,56 @@
 
     cdef object _get(self, object name):
         cdef python.PyObject* dict_result
-        dict_result = python.PyDict_GetItem(self._classes, name)
+        dict_result = python.PyDict_GetItem(self._entries, name)
         if dict_result is NULL:
-            dict_result = python.PyDict_GetItem(self._extensions, name)
+            raise KeyError, "Name not registered."
+        return <object>dict_result
+
+    cdef object _getForString(self, char* name):
+        cdef python.PyObject* dict_result
+        dict_result = python.PyDict_GetItemString(self._entries, name)
         if dict_result is NULL:
             raise KeyError, "Name not registered."
         return <object>dict_result
 
+    def clear(self):
+        self._entries.clear()
+
 cdef class _ClassNamespaceRegistry(_NamespaceRegistry):
     "Dictionary-like registry for namespace implementation classes"
-    cdef object _classes
-    def __init__(self, ns_uri):
+    cdef object _base_type
+    cdef object _type_error
+    def __init__(self, ns_uri, base_type):
         _NamespaceRegistry.__init__(self, ns_uri)
-        self._classes = {}
+        self._base_type = base_type
+        self._type_error = "Registered item must be subtypes of %s." % self._base_type.__name__
 
     def __setitem__(self, name, item):
         if not python.PyType_Check(item) or \
-               not issubclass(item, ElementBase):
-            raise NamespaceRegistryError, \
-                  "Registered item must be subtypes of ElementBase."
+               not issubclass(item, self._base_type):
+            raise NamespaceRegistryError, self._type_error
         if name is not None:
             name = _utf8(name)
-        self._classes[name] = item
-
-    def clear(self):
-        self._classes.clear()
+        self._entries[name] = item
 
     def __repr__(self):
         return "Namespace(%r)" % self._ns_uri
 
 cdef class _FunctionNamespaceRegistry(_NamespaceRegistry):
-    cdef object _extensions
     def __init__(self, ns_uri):
         _NamespaceRegistry.__init__(self, ns_uri)
-        self._extensions = {}
 
     def __setitem__(self, name, item):
         if not callable(item):
-            raise NamespaceRegistryError, "Registered function must be callable."
+            raise NamespaceRegistryError, \
+                  "Registered functions must be callable."
         if name is not None:
             name = _utf8(name)
-        self._extensions[name] = item
-
-    cdef object _get(self, object name):
-        cdef python.PyObject* dict_result
-        dict_result = python.PyDict_GetItem(self._extensions, name)
-        if dict_result is NULL:
-            raise KeyError, "Name not registered."
-        return <object>dict_result
+        self._entries[name] = item
 
     def __repr__(self):
         return "FunctionNamespace(%r)" % self._ns_uri
 
-    def clear(self):
-        self._extensions.clear()
-
 cdef class _XPathFunctionNamespaceRegistry(_FunctionNamespaceRegistry):
     cdef object _prefix
     cdef object _prefix_utf
@@ -209,11 +208,11 @@
 
 cdef object _find_all_extension_elements():
     "Internal lookup function to find all extension elements for XSLT."
-    cdef _FunctionNamespaceRegistry registry
+    cdef _ClassNamespaceRegistry registry
     ns_extensions = {}
     for (ns_utf, registry) in __XSLT_ELEMENT_NAMESPACE_REGISTRIES.iteritems():
-        if registry._extensions:
-            ns_extensions[ns_utf] = registry._extensions
+        if registry._entries:
+            ns_extensions[ns_utf] = registry._entries
     return ns_extensions
 
 cdef object _find_extension_elements(namespaces):
@@ -228,26 +227,37 @@
             __XSLT_ELEMENT_NAMESPACE_REGISTRIES, ns_uri)
         if dict_result is NULL:
             continue
-        extensions = (<_FunctionNamespaceRegistry>dict_result)._extensions
+        extensions = (<_ClassNamespaceRegistry>dict_result)._entries
         if extensions:
             python.PyDict_SetItem(extension_dict, ns_uri, extensions)
     return extension_dict
 
+cdef object _find_extension_element_class(char* c_namespace_utf,
+                                    char* c_element_name_utf):
+    return _find_element_in_registries(
+        __XSLT_ELEMENT_NAMESPACE_REGISTRIES,
+        c_namespace_utf, c_element_name_utf)
+
 cdef object _find_element_class(char* c_namespace_utf,
                                 char* c_element_name_utf):
+    return _find_element_in_registries(
+        __CLASS_NAMESPACE_REGISTRIES,
+        c_namespace_utf, c_element_name_utf)
+
+cdef object _find_element_in_registries(registries,
+                                        char* c_namespace_utf,
+                                        char* c_element_name_utf):
     cdef python.PyObject* dict_result
     cdef _ClassNamespaceRegistry registry
     if c_namespace_utf is not NULL:
-        dict_result = python.PyDict_GetItemString(
-            __CLASS_NAMESPACE_REGISTRIES, c_namespace_utf)
+        dict_result = python.PyDict_GetItemString(registries, c_namespace_utf)
     else:
-        dict_result = python.PyDict_GetItem(
-            __CLASS_NAMESPACE_REGISTRIES, None)
+        dict_result = python.PyDict_GetItem(registries, None)
     if dict_result is NULL:
         return _Element
 
     registry = <_ClassNamespaceRegistry>dict_result
-    classes = registry._classes
+    classes = registry._entries
 
     if c_element_name_utf is not NULL:
         dict_result = python.PyDict_GetItemString(

Modified: lxml/branch/xsltext/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/branch/xsltext/src/lxml/tests/test_xslt.py	(original)
+++ lxml/branch/xsltext/src/lxml/tests/test_xslt.py	Sat Apr  8 23:16:35 2006
@@ -352,10 +352,11 @@
   </xsl:template>
 </xsl:stylesheet>''')
 
-        def myext(_, tree):
-            s = etree.Element("TEST")
-            s.append(tree)
-            return s
+        class myext(etree.XSLTElementBase):
+            def __call__(self):
+                s = etree.Element("TEST")
+                s[:] = map(self.apply_templates, self)
+                return s
 
         namespace = etree.XSLTElementNamespace('testextns')
         namespace['myext'] = myext
@@ -365,7 +366,7 @@
                           '<A><TEST><a><b>B</b></a></TEST></A>')
         etree.XSLTElementNamespace('testextns').clear()
 
-    def test_xslt_extension_elements_nodeset(self):
+    def _test_xslt_extension_elements_nodeset(self):
         tree = self.parse('<a><b>B</b><b>C</b></a>')
         style = self.parse('''\
 <xsl:stylesheet version="1.0"
@@ -378,11 +379,12 @@
   </xsl:template>
 </xsl:stylesheet>''')
 
-        def myext(_, tree):
-            print tree, len(tree)
-            s = etree.Element("TEST")
-            s.append(tree)
-            return s
+        class myext(etree.XSLTElementBase):
+            def __call__(self):
+                print tree, len(tree)
+                s = etree.Element("TEST")
+                s.append(tree.__copy__())
+                return s
 
         namespace = etree.XSLTElementNamespace('testextns')
         namespace['myext'] = myext
@@ -392,7 +394,7 @@
                           '<A><TEST><a><b>B</b><b>C</b></a></TEST></A>')
         etree.XSLTElementNamespace('testextns').clear()
 
-    def test_xslt_extension_elements_error(self):
+    def _test_xslt_extension_elements_error(self):
         tree = self.parse('<a><b>B</b></a>')
         style = self.parse('''\
 <xsl:stylesheet version="1.0"
@@ -408,8 +410,9 @@
         class LocalError(Exception):
             pass
 
-        def myext(_, tree):
-            raise LocalError
+        class myext(etree.XSLTElementBase):
+            def __call__(self):
+                raise LocalError
 
         namespace = etree.XSLTElementNamespace('testextns')
         namespace['myext'] = myext

Modified: lxml/branch/xsltext/src/lxml/xslt.pxd
==============================================================================
--- lxml/branch/xsltext/src/lxml/xslt.pxd	(original)
+++ lxml/branch/xsltext/src/lxml/xslt.pxd	Sat Apr  8 23:16:35 2006
@@ -13,6 +13,12 @@
 
     ctypedef struct xsltElemPreComp:
         pass
+
+    ctypedef struct xsltStackElem:
+        pass
+
+    ctypedef struct xsltTemplate:
+        pass
     
     cdef xsltStylesheet* xsltParseStylesheetDoc(xmlDoc* doc)
     cdef void xsltFreeStylesheet(xsltStylesheet* sheet)
@@ -59,3 +65,8 @@
                                         void* ctxt,
                                         void (*handler)(void* ctxt, char* msg, ...))
     
+cdef extern from "libxslt/pattern.h":
+    cdef xsltTemplate* xsltProcessOneNode(xsltTransformContext* ctxt,
+                                          xmlNode* node,
+                                          xsltStackElem* params)
+

Modified: lxml/branch/xsltext/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/xsltext/src/lxml/xslt.pxi	(original)
+++ lxml/branch/xsltext/src/lxml/xslt.pxi	Sat Apr  8 23:16:35 2006
@@ -828,58 +828,102 @@
         return
     xpath.valuePush(ctxt, obj)
 
+cdef class XSLTElementBase(_Element):
+    cdef xslt.xsltTransformContext* _ctxt
+    cdef _Document _result_doc
+    cdef xmlNode* _c_self
+    def __call__(self, element):
+        return None
+
+    def apply_templates(self, element=None):
+        if element is None:
+            # apply to all children
+            result = []
+            for child in self:
+                new_child = self._apply_templates_to_node(child)
+                result.append(new_child)
+        else:
+            return self._apply_templates_to_node(element)
+
+    cdef object _apply_templates_to_node(self, _NodeBase element):
+        cdef xmlNode* c_node
+        c_node = element._c_node
+        xslt.xsltProcessOneNode(self._ctxt, c_node, NULL)
+        return Element("test") # FIXME: FAKE!!
+
+cdef object _extensionNodeFactory(_Document doc, xmlNode* c_node):
+    if c_node is NULL:
+        return None
+    if c_node.type == tree.XML_ELEMENT_NODE:
+        return _elementFactory(doc, c_node)
+    elif c_node.type == tree.XML_TEXT_NODE:
+        return funicode(c_node.content)
+    elif c_node.type == tree.XML_ATTRIBUTE_NODE:
+        # FIXME: is this a) used and b) correct?
+        return (_namespacedName(c_node), funicode(c_node.content))
+    elif c_node.type == tree.XML_DOCUMENT_NODE or \
+             c_node.type == tree.XML_HTML_DOCUMENT_NODE:
+        return None
+    else:
+        assert 0, "Unknown node type: %s" % c_node.type
+
 cdef void _xsltCallback(xslt.xsltTransformContext* ctxt,
                         xmlNode* c_tree_node, xmlNode* instr,
                         xslt.xsltElemPreComp* comp):
     cdef xpath.xmlXPathContext* rctxt
     cdef xmlNode*  c_node
-    cdef _Document out_doc
-    cdef _Document tree_doc
-    cdef _NodeBase insert_element
+    cdef xmlDoc*   c_doc
+    cdef _Document instr_doc, out_doc, tree_doc
+    cdef _NodeBase insertion_point
     cdef _NodeBase context_tree
+    cdef XSLTElementBase instr_element
     cdef XSLTContext extensions
-    cdef int remove_proxy
+    cdef char* c_ns_href
 
     # get our context
     rctxt = ctxt.xpathCtxt
     extensions = <XSLTContext>rctxt.userData
 
-    # find name and namespace of called element
-    name = instr.name
-    if instr.ns is NULL or instr.ns.href is NULL:
-        uri = None
-    else:
-        uri = instr.ns.href
-
-##     # build child tree
-##     c_node = instr.children
-##     while c_node is not NULL:
-##         if xslt.IS_XSLT_ELEM(c_node):
-##             pass
-##         c_node = c_node.next
-
-    # lookup up the extension element in the context
-    f = extensions.find_extension_element(uri, name)
-
-    out_doc  = _documentFactory(ctxt.output)
-    tree_doc = _documentFactory(c_tree_node.doc)
-
-    insert_element = _elementFactory(out_doc, ctxt.insert)
-    context_tree   = _elementFactory(tree_doc, c_tree_node)
-
-    instr_attribs  = _attributeDict(instr)
-    instr_tag      = _namespacedName(instr)
-
     try:
-        result = f(instr_tag, instr_attribs, context_tree)
+        # create Python representations of insertion point and context node
+        out_doc = _documentFactory(ctxt.output)
+        insertion_point = _elementFactory(out_doc, ctxt.insert)
+
+##         tree_doc = _documentFactory(c_tree_node.doc)
+##         context_tree = _extensionNodeFactory(tree_doc, c_tree_node)
+
+        # deep-copy the XSLT element to create a Python proxy
+        c_doc = tree.xmlCopyDoc(instr.doc, 0) # non recursive doc copy
+        instr_doc = _documentFactory(c_doc)
+        c_node = tree.xmlDocCopyNode(instr, c_doc, 1) # recursive
+        tree.xmlDocSetRootElement(c_doc, c_node)
+
+        # create Python representation
+        #assert c_node is not NULL and c_node.type == tree.XML_ELEMENT_NODE
+        if c_node.ns == NULL:
+            c_ns_href = NULL
+        else:
+            c_ns_href = c_node.ns.href
+        element_class = _find_extension_element_class(c_ns_href, c_node.name)
+
+        #assert issubclass(element_class, XSLTElementBase)
+        instr_element = _newElement(instr_doc, c_node, element_class)
+        instr_element._ctxt = ctxt
+        instr_element._c_self = instr
+        instr_element._result_doc = out_doc
+
+        # call the extension
+        assert callable(instr_element), "Extension objects must be callable."
+        result = instr_element()
 
+        # handle the result (if any)
         if isinstance(result, _NodeBase):
-            insert_element.append(result)
+            insertion_point.append(result)
         elif python.PyString_Check(result) or PyUnicode_Check(result):
-            insert_element.text = result
+            insertion_point.text = result
         elif python.PySequence_Check(result):
             for node in result:
-                insert_element.append(node)
+                insertion_point.append(node)
         elif result is not None:
             raise TypeError, "Invalid return value from extension element."
     except Exception, e:
@@ -890,30 +934,11 @@
 
     # prevent garbage collection of document C structures
     out_doc._c_doc  = NULL
-    tree_doc._c_doc = NULL
-    #insert_element._c_node = NULL
+##    tree_doc._c_doc = NULL
 
-    # FIXME: insert_element? current_tree?
+    # FIXME: insertion_point? current_tree?
     # if result_tree_node is in the result document it can be GCed,
     # otherwise it won't be
-    #insert_element._c_node = NULL
+    #insertion_point._c_node = NULL
     #current_tree._c_node = NULL
-    #del insert_element, current_tree, doc
-
-cdef object _attributeDict(xmlNode* c_element):
-    cdef xmlNode* c_node
-    result = {}
-    c_node = <xmlNode*>(c_element.properties)
-    while c_node is not NULL:
-        if c_node.type == tree.XML_ATTRIBUTE_NODE:
-            python.PyDict_SetItem(
-                result, _namespacedName(c_node),
-                _attributeValue(c_element, c_node)
-                )
-        c_node = c_node.next
-    return result
-
-cdef xmlNode* _applyTemplate(xslt.xsltTransformContext* ctxt, xmlNode* c_node):
-    cdef xslt.xsltStylesheet* stylesheet
-    stylesheet = ctxt.style
-    
+    #del insertion_point, current_tree, doc
From scoder at codespeak.net  Mon Apr 10 09:31:14 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 10 09:31:16 2006
Subject: [Lxml-checkins] r25645 - lxml/branch/htmlparser/doc
Message-ID: <20060410073114.490A010182@code0.codespeak.net>

Author: scoder
Date: Mon Apr 10 09:31:12 2006
New Revision: 25645

Modified:
   lxml/branch/htmlparser/doc/api.txt
Log:
doctest for HTML parser in api.txt

Modified: lxml/branch/htmlparser/doc/api.txt
==============================================================================
--- lxml/branch/htmlparser/doc/api.txt	(original)
+++ lxml/branch/htmlparser/doc/api.txt	Mon Apr 10 09:31:12 2006
@@ -19,13 +19,14 @@
   >>> from StringIO import StringIO
 
 
-XMLParser
----------
+Parsers
+-------
 
-One of the differences is the parser.  It is based on libxml2 and therefore
-only supports options that are backed by the library.  Parsers take a number
-of keyword arguments.  The following is an example for namespace cleanup
-during parsing, first with the default parser, then with a parametrized one::
+One of the differences is the parser.  There is support for both XML and
+(broken) HTML.  Both are based on libxml2 and therefore only support options
+that are backed by the library.  Parsers take a number of keyword arguments.
+The following is an example for namespace cleanup during parsing, first with
+the default parser, then with a parametrized one::
 
   >>> xml = '<a xmlns="test"><b xmlns="test"/></a>'
 
@@ -38,6 +39,23 @@
   >>> print lxml.etree.tostring(et.getroot())
   <a xmlns="test"><b/></a>
 
+HTML parsing is similarly simple::
+
+  >>> broken_html = "<html><head><title>test<body><h1>page title</body></html>"
+
+  >>> parser = lxml.etree.HTMLParser()
+  >>> et     = lxml.etree.parse(StringIO(broken_html), parser)
+
+  >>> print lxml.etree.tostring(et.getroot())
+  <html><head><title>test</title></head><body><h1>page title</h1></body></html>
+
+Lxml has an HTML function, similar to the XML shortcut known from
+ElementTree::
+
+  >>> html = lxml.etree.HTML(broken_html)
+  >>> print lxml.etree.tostring(html)
+  <html><head><title>test</title></head><body><h1>page title</h1></body></html>
+
 
 Error handling on exceptions
 ----------------------------
From scoder at codespeak.net  Thu Apr 13 07:53:04 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 07:53:08 2006
Subject: [Lxml-checkins] r25761 - lxml/trunk
Message-ID: <20060413055304.A552710082@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 07:53:01 2006
New Revision: 25761

Modified:
   lxml/trunk/version.txt
Log:
reflect current version 0.9.1 in version.txt

Modified: lxml/trunk/version.txt
==============================================================================
--- lxml/trunk/version.txt	(original)
+++ lxml/trunk/version.txt	Thu Apr 13 07:53:01 2006
@@ -1 +1 @@
-0.9
+0.9.1
From scoder at codespeak.net  Thu Apr 13 07:54:09 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 07:54:11 2006
Subject: [Lxml-checkins] r25762 - lxml/trunk/src/lxml/tests
Message-ID: <20060413055409.7333710082@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 07:54:08 2006
New Revision: 25762

Modified:
   lxml/trunk/src/lxml/tests/test_xmlschema.py
Log:
new XMLSchema test case that crashes lxml if parsed schema document is not in the XML-Schema namespace

Modified: lxml/trunk/src/lxml/tests/test_xmlschema.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xmlschema.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xmlschema.py	Thu Apr 13 07:54:08 2006
@@ -26,7 +26,7 @@
         self.assert_(schema.validate(tree_valid))
         self.assert_(not schema.validate(tree_invalid))
 
-    def test_xmlschema_invalid_schema(self):
+    def test_xmlschema_invalid_schema1(self):
         schema = self.parse('''\
 <xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema">
   <element name="a" type="AType"/>
@@ -40,6 +40,11 @@
         self.assertRaises(etree.XMLSchemaParseError,
                           etree.XMLSchema, schema)
 
+    def test_xmlschema_invalid_schema2(self):
+        schema = self.parse('<test/>')
+        self.assertRaises(etree.XMLSchemaParseError,
+                          etree.XMLSchema, schema)
+
 ##     def test_xmlschema_include(self):
 ##         # this will only work if we access the file through path or
 ##         # file object..
From scoder at codespeak.net  Thu Apr 13 07:56:30 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 07:56:31 2006
Subject: [Lxml-checkins] r25763 - lxml/trunk/src/lxml
Message-ID: <20060413055630.8915610082@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 07:56:29 2006
New Revision: 25763

Modified:
   lxml/trunk/src/lxml/xmlschema.pxi
Log:
make new XML-Schema test case by checking the root node namespace of the document passed to XMLSchema() ; also allow passing Elements instead of ElementTrees

Modified: lxml/trunk/src/lxml/xmlschema.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlschema.pxi	(original)
+++ lxml/trunk/src/lxml/xmlschema.pxi	Thu Apr 13 07:56:29 2006
@@ -19,37 +19,59 @@
     cdef xmlschema.xmlSchema* _c_schema
     cdef _ErrorLog _error_log
     
-    def __init__(self, _ElementTree etree):
+    def __init__(self, etree):
         cdef _Document doc
+        cdef _NodeBase root_node
+        cdef xmlDoc* fake_c_doc
+        cdef xmlNode* c_node
         cdef xmlschema.xmlSchemaParserCtxt* parser_ctxt
-        doc = etree._doc
-        parser_ctxt = xmlschema.xmlSchemaNewDocParserCtxt(doc._c_doc)
+
+        doc = _documentOrRaise(etree)
+        root_node = _rootNodeOf(etree)
+
+        # work around for libxml2 bug if document is not XML schema at all
+        c_node = root_node._c_node
+        if c_node.ns is NULL or c_node.ns.href is NULL or \
+               tree.strcmp(c_node.ns.href, 'http://www.w3.org/2001/XMLSchema') != 0:
+            raise XMLSchemaParseError, "Document is not XML Schema"
+
+        fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
+        parser_ctxt = xmlschema.xmlSchemaNewDocParserCtxt(fake_c_doc)
         if parser_ctxt is NULL:
+            _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise XMLSchemaParseError, "Document is not parsable as XML Schema"
         self._c_schema = xmlschema.xmlSchemaParse(parser_ctxt)
+
+        xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
+        _destroyFakeDoc(doc._c_doc, fake_c_doc)
+
         if self._c_schema is NULL:
-            xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
             raise XMLSchemaParseError, "Document is not valid XML Schema"
-        xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
         self._error_log = _ErrorLog()
-        
+
     def __dealloc__(self):
         xmlschema.xmlSchemaFree(self._c_schema)
 
-    def validate(self, _ElementTree etree):
+    def validate(self, etree):
         """Validate doc using XML Schema.
 
         Returns true if document is valid, false if not.
         """
         cdef xmlschema.xmlSchemaValidCtxt* valid_ctxt
+        cdef _Document doc
+        cdef _NodeBase root_node
         cdef xmlDoc* c_doc
         cdef int ret
+
+        doc = _documentOrRaise(etree)
+        root_node = _rootNodeOf(etree)
+
         self._error_log.connect()
         valid_ctxt = xmlschema.xmlSchemaNewValidCtxt(self._c_schema)
 
-        c_doc = _fakeRootDoc(etree._doc._c_doc, etree._context_node._c_node)
+        c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
         ret = xmlschema.xmlSchemaValidateDoc(valid_ctxt, c_doc)
-        _destroyFakeDoc(etree._doc._c_doc, c_doc)
+        _destroyFakeDoc(doc._c_doc, c_doc)
 
         xmlschema.xmlSchemaFreeValidCtxt(valid_ctxt)
         self._error_log.disconnect()
From scoder at codespeak.net  Thu Apr 13 07:59:33 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 07:59:34 2006
Subject: [Lxml-checkins] r25764 - lxml/trunk/src/lxml
Message-ID: <20060413055933.55E6810082@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 07:59:32 2006
New Revision: 25764

Modified:
   lxml/trunk/src/lxml/relaxng.pxi
Log:
allow passing Elements to RelaxNG() and .validate(), not only ElementTrees

Modified: lxml/trunk/src/lxml/relaxng.pxi
==============================================================================
--- lxml/trunk/src/lxml/relaxng.pxi	(original)
+++ lxml/trunk/src/lxml/relaxng.pxi	Thu Apr 13 07:59:32 2006
@@ -20,11 +20,18 @@
     cdef relaxng.xmlRelaxNG* _c_schema
     cdef _ErrorLog _error_log
     
-    def __init__(self, _ElementTree etree=None, file=None):
+    def __init__(self, etree=None, file=None):
+        cdef _Document doc
+        cdef _NodeBase root_node
+        cdef xmlDoc* fake_c_doc
         cdef relaxng.xmlRelaxNGParserCtxt* parser_ctxt
 
+        fake_c_doc = NULL
         if etree is not None:
-            parser_ctxt = relaxng.xmlRelaxNGNewDocParserCtxt(etree._doc._c_doc)
+            doc = _documentOrRaise(etree)
+            root_node = _rootNodeOf(etree)
+            fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
+            parser_ctxt = relaxng.xmlRelaxNGNewDocParserCtxt(fake_c_doc)
         elif file is not None:
             filename = _getFilenameForFile(file)
             if filename is None:
@@ -33,31 +40,44 @@
             parser_ctxt = relaxng.xmlRelaxNGNewParserCtxt(filename)
         else:
             raise RelaxNGParseError, "No tree or file given"
+
         if parser_ctxt is NULL:
+            if fake_c_doc is not NULL:
+                _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise RelaxNGParseError, "Document is not parsable as Relax NG"
         self._c_schema = relaxng.xmlRelaxNGParse(parser_ctxt)
+
         if self._c_schema is NULL:
+            if fake_c_doc is not NULL:
+                _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise RelaxNGParseError, "Document is not valid Relax NG"
         relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
-
+        if fake_c_doc is not NULL:
+            _destroyFakeDoc(doc._c_doc, fake_c_doc)
         self._error_log = _ErrorLog()
         
     def __dealloc__(self):
         relaxng.xmlRelaxNGFree(self._c_schema)
         
-    def validate(self, _ElementTree etree):
+    def validate(self, etree):
         """Validate doc using Relax NG.
 
         Returns true if document is valid, false if not."""
+        cdef _Document doc
+        cdef _NodeBase root_node
         cdef xmlDoc* c_doc
         cdef relaxng.xmlRelaxNGValidCtxt* valid_ctxt
         cdef int ret
+
+        doc = _documentOrRaise(etree)
+        root_node = _rootNodeOf(etree)
+
         self._error_log.connect()
         valid_ctxt = relaxng.xmlRelaxNGNewValidCtxt(self._c_schema)
 
-        c_doc = _fakeRootDoc(etree._doc._c_doc, etree._context_node._c_node)
+        c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
         ret = relaxng.xmlRelaxNGValidateDoc(valid_ctxt, c_doc)
-        _destroyFakeDoc(etree._doc._c_doc, c_doc)
+        _destroyFakeDoc(doc._c_doc, c_doc)
 
         relaxng.xmlRelaxNGFreeValidCtxt(valid_ctxt)
         self._error_log.disconnect()
From scoder at codespeak.net  Thu Apr 13 08:06:51 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 08:06:54 2006
Subject: [Lxml-checkins] r25765 - lxml/trunk/src/lxml
Message-ID: <20060413060651.C3D3810078@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 08:06:50 2006
New Revision: 25765

Modified:
   lxml/trunk/src/lxml/relaxng.pxi
Log:
doc typo

Modified: lxml/trunk/src/lxml/relaxng.pxi
==============================================================================
--- lxml/trunk/src/lxml/relaxng.pxi	(original)
+++ lxml/trunk/src/lxml/relaxng.pxi	Thu Apr 13 08:06:50 2006
@@ -14,7 +14,7 @@
 # RelaxNG
 
 cdef class RelaxNG:
-    """Turn a document into an Relax NG validator.
+    """Turn a document into a Relax NG validator.
     Can also load from filesystem directly given file object or filename.
     """
     cdef relaxng.xmlRelaxNG* _c_schema
From scoder at codespeak.net  Thu Apr 13 10:17:34 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 10:17:36 2006
Subject: [Lxml-checkins] r25768 - lxml/trunk/src/lxml
Message-ID: <20060413081734.C27851007D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 10:17:33 2006
New Revision: 25768

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
restrict type test to _ElementTree in _rootNodeOf()

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 13 10:17:33 2006
@@ -1292,8 +1292,8 @@
 cdef _NodeBase _rootNodeOf(object input):
     # call this to get the root node of a
     # _Document, _ElementTree or _NodeBase object
-    if hasattr(input, 'getroot'): # ElementTree
-        return <_NodeBase>(input.getroot())
+    if isinstance(input, _ElementTree):
+        return (<_ElementTree>input)._context_node
     elif isinstance(input, _NodeBase):
         return <_NodeBase>input
     elif isinstance(input, _Document):
From scoder at codespeak.net  Thu Apr 13 10:17:53 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 10:17:55 2006
Subject: [Lxml-checkins] r25769 - lxml/trunk/src/lxml
Message-ID: <20060413081753.44D621007D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 10:17:52 2006
New Revision: 25769

Modified:
   lxml/trunk/src/lxml/relaxng.pxi
Log:
work arounds for libxml2 RelaxNG bug to make the RNG test cases valgrind clean without crashing

Modified: lxml/trunk/src/lxml/relaxng.pxi
==============================================================================
--- lxml/trunk/src/lxml/relaxng.pxi	(original)
+++ lxml/trunk/src/lxml/relaxng.pxi	Thu Apr 13 10:17:52 2006
@@ -23,6 +23,7 @@
     def __init__(self, etree=None, file=None):
         cdef _Document doc
         cdef _NodeBase root_node
+        cdef xmlNode* c_node
         cdef xmlDoc* fake_c_doc
         cdef relaxng.xmlRelaxNGParserCtxt* parser_ctxt
 
@@ -30,6 +31,12 @@
         if etree is not None:
             doc = _documentOrRaise(etree)
             root_node = _rootNodeOf(etree)
+            c_node = root_node._c_node
+            # work around for libxml2 bug if document is not RNG at all
+            if c_node.ns is NULL or c_node.ns.href is NULL or \
+                   tree.strcmp(c_node.ns.href,
+                               'http://relaxng.org/ns/structure/1.0') != 0:
+                raise RelaxNGParseError, "Document is not Relax NG"
             fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
             parser_ctxt = relaxng.xmlRelaxNGNewDocParserCtxt(fake_c_doc)
         elif file is not None:
@@ -47,8 +54,11 @@
             raise RelaxNGParseError, "Document is not parsable as Relax NG"
         self._c_schema = relaxng.xmlRelaxNGParse(parser_ctxt)
 
+        # XXX: freeing parser context will crash if document was not RNG!!
+        #relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
         if self._c_schema is NULL:
             if fake_c_doc is not NULL:
+                relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
                 _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise RelaxNGParseError, "Document is not valid Relax NG"
         relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
From scoder at codespeak.net  Thu Apr 13 10:25:35 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 10:25:37 2006
Subject: [Lxml-checkins] r25770 - in lxml/branch/lxml-0.9.x/src/lxml: . tests
Message-ID: <20060413082535.A97081007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 10:25:34 2006
New Revision: 25770

Modified:
   lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
   lxml/branch/lxml-0.9.x/src/lxml/relaxng.pxi
   lxml/branch/lxml-0.9.x/src/lxml/tests/test_xmlschema.py
   lxml/branch/lxml-0.9.x/src/lxml/xmlschema.pxi
Log:
merged in fixes for crashes and memory leaks in RelaxNG and XML Schema

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	Thu Apr 13 10:25:34 2006
@@ -1277,8 +1277,8 @@
 cdef _NodeBase _rootNodeOf(object input):
     # call this to get the root node of a
     # _Document, _ElementTree or _NodeBase object
-    if hasattr(input, 'getroot'): # ElementTree
-        return <_NodeBase>(input.getroot())
+    if isinstance(input, _ElementTree):
+        return (<_ElementTree>input)._context_node
     elif isinstance(input, _NodeBase):
         return <_NodeBase>input
     elif isinstance(input, _Document):

Modified: lxml/branch/lxml-0.9.x/src/lxml/relaxng.pxi
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/relaxng.pxi	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/relaxng.pxi	Thu Apr 13 10:25:34 2006
@@ -14,17 +14,31 @@
 # RelaxNG
 
 cdef class RelaxNG:
-    """Turn a document into an Relax NG validator.
+    """Turn a document into a Relax NG validator.
     Can also load from filesystem directly given file object or filename.
     """
     cdef relaxng.xmlRelaxNG* _c_schema
     cdef _ErrorLog _error_log
     
-    def __init__(self, _ElementTree etree=None, file=None):
+    def __init__(self, etree=None, file=None):
+        cdef _Document doc
+        cdef _NodeBase root_node
+        cdef xmlNode* c_node
+        cdef xmlDoc* fake_c_doc
         cdef relaxng.xmlRelaxNGParserCtxt* parser_ctxt
 
+        fake_c_doc = NULL
         if etree is not None:
-            parser_ctxt = relaxng.xmlRelaxNGNewDocParserCtxt(etree._doc._c_doc)
+            doc = _documentOrRaise(etree)
+            root_node = _rootNodeOf(etree)
+            c_node = root_node._c_node
+            # work around for libxml2 bug if document is not RNG at all
+            if c_node.ns is NULL or c_node.ns.href is NULL or \
+                   tree.strcmp(c_node.ns.href,
+                               'http://relaxng.org/ns/structure/1.0') != 0:
+                raise RelaxNGParseError, "Document is not Relax NG"
+            fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
+            parser_ctxt = relaxng.xmlRelaxNGNewDocParserCtxt(fake_c_doc)
         elif file is not None:
             filename = _getFilenameForFile(file)
             if filename is None:
@@ -33,31 +47,47 @@
             parser_ctxt = relaxng.xmlRelaxNGNewParserCtxt(filename)
         else:
             raise RelaxNGParseError, "No tree or file given"
+
         if parser_ctxt is NULL:
+            if fake_c_doc is not NULL:
+                _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise RelaxNGParseError, "Document is not parsable as Relax NG"
         self._c_schema = relaxng.xmlRelaxNGParse(parser_ctxt)
+
+        # XXX: freeing parser context will crash if document was not RNG!!
+        #relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
         if self._c_schema is NULL:
+            if fake_c_doc is not NULL:
+                relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
+                _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise RelaxNGParseError, "Document is not valid Relax NG"
         relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
-
+        if fake_c_doc is not NULL:
+            _destroyFakeDoc(doc._c_doc, fake_c_doc)
         self._error_log = _ErrorLog()
         
     def __dealloc__(self):
         relaxng.xmlRelaxNGFree(self._c_schema)
         
-    def validate(self, _ElementTree etree):
+    def validate(self, etree):
         """Validate doc using Relax NG.
 
         Returns true if document is valid, false if not."""
+        cdef _Document doc
+        cdef _NodeBase root_node
         cdef xmlDoc* c_doc
         cdef relaxng.xmlRelaxNGValidCtxt* valid_ctxt
         cdef int ret
+
+        doc = _documentOrRaise(etree)
+        root_node = _rootNodeOf(etree)
+
         self._error_log.connect()
         valid_ctxt = relaxng.xmlRelaxNGNewValidCtxt(self._c_schema)
 
-        c_doc = _fakeRootDoc(etree._doc._c_doc, etree._context_node._c_node)
+        c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
         ret = relaxng.xmlRelaxNGValidateDoc(valid_ctxt, c_doc)
-        _destroyFakeDoc(etree._doc._c_doc, c_doc)
+        _destroyFakeDoc(doc._c_doc, c_doc)
 
         relaxng.xmlRelaxNGFreeValidCtxt(valid_ctxt)
         self._error_log.disconnect()

Modified: lxml/branch/lxml-0.9.x/src/lxml/tests/test_xmlschema.py
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/tests/test_xmlschema.py	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/tests/test_xmlschema.py	Thu Apr 13 10:25:34 2006
@@ -26,7 +26,7 @@
         self.assert_(schema.validate(tree_valid))
         self.assert_(not schema.validate(tree_invalid))
 
-    def test_xmlschema_invalid_schema(self):
+    def test_xmlschema_invalid_schema1(self):
         schema = self.parse('''\
 <xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema">
   <element name="a" type="AType"/>
@@ -40,6 +40,11 @@
         self.assertRaises(etree.XMLSchemaParseError,
                           etree.XMLSchema, schema)
 
+    def test_xmlschema_invalid_schema2(self):
+        schema = self.parse('<test/>')
+        self.assertRaises(etree.XMLSchemaParseError,
+                          etree.XMLSchema, schema)
+
 ##     def test_xmlschema_include(self):
 ##         # this will only work if we access the file through path or
 ##         # file object..

Modified: lxml/branch/lxml-0.9.x/src/lxml/xmlschema.pxi
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/xmlschema.pxi	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/xmlschema.pxi	Thu Apr 13 10:25:34 2006
@@ -19,37 +19,59 @@
     cdef xmlschema.xmlSchema* _c_schema
     cdef _ErrorLog _error_log
     
-    def __init__(self, _ElementTree etree):
+    def __init__(self, etree):
         cdef _Document doc
+        cdef _NodeBase root_node
+        cdef xmlDoc* fake_c_doc
+        cdef xmlNode* c_node
         cdef xmlschema.xmlSchemaParserCtxt* parser_ctxt
-        doc = etree._doc
-        parser_ctxt = xmlschema.xmlSchemaNewDocParserCtxt(doc._c_doc)
+
+        doc = _documentOrRaise(etree)
+        root_node = _rootNodeOf(etree)
+
+        # work around for libxml2 bug if document is not XML schema at all
+        c_node = root_node._c_node
+        if c_node.ns is NULL or c_node.ns.href is NULL or \
+               tree.strcmp(c_node.ns.href, 'http://www.w3.org/2001/XMLSchema') != 0:
+            raise XMLSchemaParseError, "Document is not XML Schema"
+
+        fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
+        parser_ctxt = xmlschema.xmlSchemaNewDocParserCtxt(fake_c_doc)
         if parser_ctxt is NULL:
+            _destroyFakeDoc(doc._c_doc, fake_c_doc)
             raise XMLSchemaParseError, "Document is not parsable as XML Schema"
         self._c_schema = xmlschema.xmlSchemaParse(parser_ctxt)
+
+        xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
+        _destroyFakeDoc(doc._c_doc, fake_c_doc)
+
         if self._c_schema is NULL:
-            xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
             raise XMLSchemaParseError, "Document is not valid XML Schema"
-        xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
         self._error_log = _ErrorLog()
-        
+
     def __dealloc__(self):
         xmlschema.xmlSchemaFree(self._c_schema)
 
-    def validate(self, _ElementTree etree):
+    def validate(self, etree):
         """Validate doc using XML Schema.
 
         Returns true if document is valid, false if not.
         """
         cdef xmlschema.xmlSchemaValidCtxt* valid_ctxt
+        cdef _Document doc
+        cdef _NodeBase root_node
         cdef xmlDoc* c_doc
         cdef int ret
+
+        doc = _documentOrRaise(etree)
+        root_node = _rootNodeOf(etree)
+
         self._error_log.connect()
         valid_ctxt = xmlschema.xmlSchemaNewValidCtxt(self._c_schema)
 
-        c_doc = _fakeRootDoc(etree._doc._c_doc, etree._context_node._c_node)
+        c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
         ret = xmlschema.xmlSchemaValidateDoc(valid_ctxt, c_doc)
-        _destroyFakeDoc(etree._doc._c_doc, c_doc)
+        _destroyFakeDoc(doc._c_doc, c_doc)
 
         xmlschema.xmlSchemaFreeValidCtxt(valid_ctxt)
         self._error_log.disconnect()
From scoder at codespeak.net  Thu Apr 13 14:22:23 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:22:24 2006
Subject: [Lxml-checkins] r25781 - lxml/trunk/src/lxml
Message-ID: <20060413122223.A6C551007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:22:22 2006
New Revision: 25781

Modified:
   lxml/trunk/src/lxml/etree.h
   lxml/trunk/src/lxml/python.pxd
Log:
cleaned up declarations in python.pxd and etree.h, added str() macro

Modified: lxml/trunk/src/lxml/etree.h
==============================================================================
--- lxml/trunk/src/lxml/etree.h	(original)
+++ lxml/trunk/src/lxml/etree.h	Thu Apr 13 14:22:22 2006
@@ -1,9 +1,10 @@
 #ifndef HAS_ETREE_H
 #define HAS_ETREE_H
 
-#define isinstance(a,b) PyObject_IsInstance(a,b)
-#define hasattr(a,b)    PyObject_HasAttr(a,b)
-#define callable(a)     PyCallable_Check(a)
+#define isinstance(o,c) PyObject_IsInstance(o,c)
+#define hasattr(o,a)    PyObject_HasAttr(o,a)
+#define callable(o)     PyCallable_Check(o)
+#define str(o)          PyObject_Str(o)
 #define _cstr(s)        PyString_AS_STRING(s)
 #define _isElement(c_node) \
         ((c_node)->type == XML_ELEMENT_NODE || \

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Thu Apr 13 14:22:22 2006
@@ -33,12 +33,10 @@
     cdef int PyBool_Check(object instance)
     cdef int PySequence_Check(object instance)
     cdef int PyType_Check(object instance)
-    cdef int PyCallable_Check(object instance)
-    cdef int PyObject_IsInstance(object instance, object classes)
-    cdef int PyObject_HasAttr(object obj, object attr)
 
 cdef extern from "etree.h": # redefines some functions as macros
     cdef int isinstance(object instance, object classes)
     cdef int hasattr(object obj, object attr)
     cdef int callable(object obj)
+    cdef object str(object obj)
     cdef char* _cstr(object s)
From scoder at codespeak.net  Thu Apr 13 14:23:32 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:23:33 2006
Subject: [Lxml-checkins] r25782 - in lxml/trunk/src/lxml: . tests
Message-ID: <20060413122332.7D1E81007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:23:31 2006
New Revision: 25782

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/tests/test_elementtree.py
   lxml/trunk/src/lxml/tests/test_unicode.py
Log:
new QName class, allow QName objects to be passed into API functions like Element() etc.

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 13 14:23:31 2006
@@ -1,6 +1,6 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, hasattr, callable, _cstr
+from python cimport isinstance, hasattr, callable, str, _cstr
 cimport xpath
 cimport xslt
 cimport xmlerror
@@ -1203,6 +1203,20 @@
         python.PyDict_SetItem(dic, elem.get('id'), elem)
     return (root, dic)
 
+cdef class QName:
+    cdef readonly object text
+    def __init__(self, text_or_uri, tag=None):
+        if tag is not None:
+            text_or_uri = "{%s}%s" % (text_or_uri, tag)
+        elif not python.PyString_Check(text_or_uri) and \
+             not python.PyUnicode_Check(text_or_uri):
+            text_or_uri = str(text_or_uri)
+        self.text = text_or_uri
+    def __str__(self):
+        return self.text
+    def __hash__(self):
+        return self.text.__hash__()
+
 def iselement(element):
     return isinstance(element, _Element)
 
@@ -1559,6 +1573,8 @@
     cdef char* c_tag
     cdef char* c_pos
     cdef int nslen
+    if isinstance(tag, QName):
+        tag = (<QName>tag).text
     tag = _utf8(tag)
     c_tag = _cstr(tag)
     if c_tag[0] == c'{':

Modified: lxml/trunk/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_elementtree.py	Thu Apr 13 14:23:31 2006
@@ -1787,6 +1787,14 @@
         self.assertEquals(self._rootstring(t1), '<a><b/></a>')
         self.assertEquals(self._rootstring(t),  '<a><b/></a>')
 
+    def test_qname(self):
+        etree = self.etree
+        qname = etree.QName('myns', 'a')
+        a1 = etree.Element(qname)
+        a2 = etree.SubElement(a1, qname)
+        self.assertEquals(a1.tag, "{myns}a")
+        self.assertEquals(a2.tag, "{myns}a")
+
     def _writeElement(self, element, encoding='us-ascii'):
         """Write out element for comparison.
         """

Modified: lxml/trunk/src/lxml/tests/test_unicode.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_unicode.py	(original)
+++ lxml/trunk/src/lxml/tests/test_unicode.py	Thu Apr 13 14:23:31 2006
@@ -22,6 +22,13 @@
         el = etree.Element(tag)
         self.assertEquals(tag, el.tag)
 
+    def test_unicode_qname(self):
+        qname = etree.QName(uni, uni)
+        tag = u"{%s}%s" % (uni, uni)
+        self.assertEquals(qname.text, tag)
+        self.assertEquals(qname.__str__(), tag)
+        self.assertEquals(unicode(qname), tag)
+
     def test_unicode_attr(self):
         el = etree.Element('foo', {'bar': uni})
         self.assertEquals(uni, el.attrib['bar'])
From scoder at codespeak.net  Thu Apr 13 14:35:52 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:35:53 2006
Subject: [Lxml-checkins] r25783 - in lxml/branch/lxml-0.9.x/src/lxml: . tests
Message-ID: <20060413123552.194321007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:35:50 2006
New Revision: 25783

Modified:
   lxml/branch/lxml-0.9.x/src/lxml/etree.h
   lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
   lxml/branch/lxml-0.9.x/src/lxml/python.pxd
   lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py
   lxml/branch/lxml-0.9.x/src/lxml/tests/test_unicode.py
Log:
merged in QName class from trunk

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.h
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.h	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.h	Thu Apr 13 14:35:50 2006
@@ -1,9 +1,10 @@
 #ifndef HAS_ETREE_H
 #define HAS_ETREE_H
 
-#define isinstance(a,b) PyObject_IsInstance(a,b)
-#define hasattr(a,b)    PyObject_HasAttr(a,b)
-#define callable(a)     PyCallable_Check(a)
+#define isinstance(o,c) PyObject_IsInstance(o,c)
+#define hasattr(o,a)    PyObject_HasAttr(o,a)
+#define callable(o)     PyCallable_Check(o)
+#define str(o)          PyObject_Str(o)
 #define _isElement(c_node) \
         ((c_node)->type == XML_ELEMENT_NODE || \
 	 (c_node)->type == XML_COMMENT_NODE)

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	Thu Apr 13 14:35:50 2006
@@ -1,6 +1,6 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, hasattr
+from python cimport isinstance, hasattr, callable, str
 cimport xpath
 cimport xslt
 cimport xmlerror
@@ -1188,6 +1188,20 @@
         python.PyDict_SetItem(dic, elem.get('id'), elem)
     return (root, dic)
 
+cdef class QName:
+    cdef readonly object text
+    def __init__(self, text_or_uri, tag=None):
+        if tag is not None:
+            text_or_uri = "{%s}%s" % (text_or_uri, tag)
+        elif not python.PyString_Check(text_or_uri) and \
+             not python.PyUnicode_Check(text_or_uri):
+            text_or_uri = str(text_or_uri)
+        self.text = text_or_uri
+    def __str__(self):
+        return self.text
+    def __hash__(self):
+        return self.text.__hash__()
+
 def iselement(element):
     return isinstance(element, _Element)
 
@@ -1544,6 +1558,8 @@
     cdef char* c_tag
     cdef char* c_pos
     cdef int nslen
+    if isinstance(tag, QName):
+        tag = (<QName>tag).text
     tag = _utf8(tag)
     c_tag = tag
     if c_tag[0] == c'{':

Modified: lxml/branch/lxml-0.9.x/src/lxml/python.pxd
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/python.pxd	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/python.pxd	Thu Apr 13 14:35:50 2006
@@ -33,11 +33,9 @@
     cdef int PyBool_Check(object instance)
     cdef int PySequence_Check(object instance)
     cdef int PyType_Check(object instance)
-    cdef int PyCallable_Check(object instance)
-    cdef int PyObject_IsInstance(object instance, object classes)
-    cdef int PyObject_HasAttr(object obj, object attr)
 
 cdef extern from "etree.h": # redefines some functions as macros
     cdef int isinstance(object instance, object classes)
     cdef int hasattr(object obj, object attr)
     cdef int callable(object obj)
+    cdef object str(object obj)

Modified: lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py	Thu Apr 13 14:35:50 2006
@@ -1787,6 +1787,14 @@
         self.assertEquals(self._rootstring(t1), '<a><b/></a>')
         self.assertEquals(self._rootstring(t),  '<a><b/></a>')
 
+    def test_qname(self):
+        etree = self.etree
+        qname = etree.QName('myns', 'a')
+        a1 = etree.Element(qname)
+        a2 = etree.SubElement(a1, qname)
+        self.assertEquals(a1.tag, "{myns}a")
+        self.assertEquals(a2.tag, "{myns}a")
+
     def _writeElement(self, element, encoding='us-ascii'):
         """Write out element for comparison.
         """

Modified: lxml/branch/lxml-0.9.x/src/lxml/tests/test_unicode.py
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/tests/test_unicode.py	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/tests/test_unicode.py	Thu Apr 13 14:35:50 2006
@@ -22,6 +22,12 @@
         el = etree.Element(tag)
         self.assertEquals(tag, el.tag)
 
+    def test_unicode_qname(self):
+        qname = etree.QName(uni, uni)
+        tag = u"{%s}%s" % (uni, uni)
+        self.assertEquals(qname.text, tag)
+        self.assertEquals(unicode(qname), tag)
+
     def test_unicode_attr(self):
         el = etree.Element('foo', {'bar': uni})
         self.assertEquals(uni, el.attrib['bar'])
From scoder at codespeak.net  Thu Apr 13 14:40:43 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:40:44 2006
Subject: [Lxml-checkins] r25784 - lxml/trunk/src/lxml/tests
Message-ID: <20060413124043.513FF1007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:40:42 2006
New Revision: 25784

Modified:
   lxml/trunk/src/lxml/tests/test_unicode.py
Log:
clean up test case

Modified: lxml/trunk/src/lxml/tests/test_unicode.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_unicode.py	(original)
+++ lxml/trunk/src/lxml/tests/test_unicode.py	Thu Apr 13 14:40:42 2006
@@ -26,7 +26,6 @@
         qname = etree.QName(uni, uni)
         tag = u"{%s}%s" % (uni, uni)
         self.assertEquals(qname.text, tag)
-        self.assertEquals(qname.__str__(), tag)
         self.assertEquals(unicode(qname), tag)
 
     def test_unicode_attr(self):
From scoder at codespeak.net  Thu Apr 13 14:48:40 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:48:42 2006
Subject: [Lxml-checkins] r25785 - lxml/trunk/src/lxml
Message-ID: <20060413124840.CD8461007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:48:39 2006
New Revision: 25785

Modified:
   lxml/trunk/src/lxml/etree.h
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/nsclasses.pxi
   lxml/trunk/src/lxml/python.pxd
Log:
make issubclass a macro

Modified: lxml/trunk/src/lxml/etree.h
==============================================================================
--- lxml/trunk/src/lxml/etree.h	(original)
+++ lxml/trunk/src/lxml/etree.h	Thu Apr 13 14:48:39 2006
@@ -2,6 +2,7 @@
 #define HAS_ETREE_H
 
 #define isinstance(o,c) PyObject_IsInstance(o,c)
+#define issubclass(c,csuper) PyObject_IsSubclass(c,csuper)
 #define hasattr(o,a)    PyObject_HasAttr(o,a)
 #define callable(o)     PyCallable_Check(o)
 #define str(o)          PyObject_Str(o)

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 13 14:48:39 2006
@@ -1,6 +1,6 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, hasattr, callable, str, _cstr
+from python cimport isinstance, issubclass, hasattr, callable, str, _cstr
 cimport xpath
 cimport xslt
 cimport xmlerror

Modified: lxml/trunk/src/lxml/nsclasses.pxi
==============================================================================
--- lxml/trunk/src/lxml/nsclasses.pxi	(original)
+++ lxml/trunk/src/lxml/nsclasses.pxi	Thu Apr 13 14:48:39 2006
@@ -10,7 +10,7 @@
     persistent state of elements must be stored in the underlying XML."""
     pass
 
-class XSLTElement(object):
+cdef class XSLTElement:
     "NOT IMPLEMENTED YET!"
     pass
 

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Thu Apr 13 14:48:39 2006
@@ -36,6 +36,7 @@
 
 cdef extern from "etree.h": # redefines some functions as macros
     cdef int isinstance(object instance, object classes)
+    cdef int issubclass(object instance, object classes)
     cdef int hasattr(object obj, object attr)
     cdef int callable(object obj)
     cdef object str(object obj)
From scoder at codespeak.net  Thu Apr 13 14:56:51 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:56:52 2006
Subject: [Lxml-checkins] r25787 - lxml/trunk/src/lxml
Message-ID: <20060413125651.D9DFC1007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:56:51 2006
New Revision: 25787

Modified:
   lxml/trunk/src/lxml/python.pxd
Log:
clean up in python.pxd

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Thu Apr 13 14:56:51 2006
@@ -36,7 +36,7 @@
 
 cdef extern from "etree.h": # redefines some functions as macros
     cdef int isinstance(object instance, object classes)
-    cdef int issubclass(object instance, object classes)
+    cdef int issubclass(object derived,  object superclasses)
     cdef int hasattr(object obj, object attr)
     cdef int callable(object obj)
     cdef object str(object obj)
From scoder at codespeak.net  Thu Apr 13 14:57:49 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 14:57:50 2006
Subject: [Lxml-checkins] r25788 - lxml/branch/lxml-0.9.x/src/lxml
Message-ID: <20060413125749.7327D1007C@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 14:57:48 2006
New Revision: 25788

Modified:
   lxml/branch/lxml-0.9.x/src/lxml/etree.h
   lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
   lxml/branch/lxml-0.9.x/src/lxml/nsclasses.pxi
   lxml/branch/lxml-0.9.x/src/lxml/python.pxd
Log:
merged in issubclass macro from trunk

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.h
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.h	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.h	Thu Apr 13 14:57:48 2006
@@ -2,6 +2,7 @@
 #define HAS_ETREE_H
 
 #define isinstance(o,c) PyObject_IsInstance(o,c)
+#define issubclass(c,csuper) PyObject_IsSubclass(c,csuper)
 #define hasattr(o,a)    PyObject_HasAttr(o,a)
 #define callable(o)     PyCallable_Check(o)
 #define str(o)          PyObject_Str(o)

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	Thu Apr 13 14:57:48 2006
@@ -1,6 +1,6 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, hasattr, callable, str
+from python cimport isinstance, issubclass, hasattr, callable, str
 cimport xpath
 cimport xslt
 cimport xmlerror

Modified: lxml/branch/lxml-0.9.x/src/lxml/nsclasses.pxi
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/nsclasses.pxi	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/nsclasses.pxi	Thu Apr 13 14:57:48 2006
@@ -10,7 +10,7 @@
     persistent state of elements must be stored in the underlying XML."""
     pass
 
-class XSLTElement(object):
+cdef class XSLTElement:
     "NOT IMPLEMENTED YET!"
     pass
 

Modified: lxml/branch/lxml-0.9.x/src/lxml/python.pxd
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/python.pxd	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/python.pxd	Thu Apr 13 14:57:48 2006
@@ -36,6 +36,7 @@
 
 cdef extern from "etree.h": # redefines some functions as macros
     cdef int isinstance(object instance, object classes)
+    cdef int issubclass(object derived,  object superclasses)
     cdef int hasattr(object obj, object attr)
     cdef int callable(object obj)
     cdef object str(object obj)
From scoder at codespeak.net  Thu Apr 13 15:22:55 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 15:22:57 2006
Subject: [Lxml-checkins] r25790 - lxml/trunk
Message-ID: <20060413132255.F02711007B@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 15:22:55 2006
New Revision: 25790

Modified:
   lxml/trunk/CHANGES.txt
Log:
updated CHANGES.txt

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Thu Apr 13 15:22:55 2006
@@ -1,6 +1,23 @@
 lxml changelog
 ==============
 
+current
+=======
+
+Features added
+--------------
+
+* ElementTree compatible QName class
+
+* RelaxNG and XMLSchema now accept any Element, not only ElementTrees
+
+Bugs fixed
+----------
+
+* Crash in XMLSchema and RelaxNG when passing non-schema documents
+
+* Memory leak in RelaxNG() when RelaxNGParseError is raised
+
 0.9.1 (2006-03-30)
 ==================
 
From scoder at codespeak.net  Thu Apr 13 15:24:28 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 13 15:24:30 2006
Subject: [Lxml-checkins] r25791 - lxml/branch/lxml-0.9.x
Message-ID: <20060413132428.B03D11007B@code0.codespeak.net>

Author: scoder
Date: Thu Apr 13 15:24:27 2006
New Revision: 25791

Modified:
   lxml/branch/lxml-0.9.x/CHANGES.txt
Log:
updated CHANGES.txt

Modified: lxml/branch/lxml-0.9.x/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-0.9.x/CHANGES.txt	(original)
+++ lxml/branch/lxml-0.9.x/CHANGES.txt	Thu Apr 13 15:24:27 2006
@@ -1,6 +1,23 @@
 lxml changelog
 ==============
 
+current
+=======
+
+Features added
+--------------
+
+* ElementTree compatible QName class
+
+* RelaxNG and XMLSchema now accept any Element, not only ElementTrees
+
+Bugs fixed
+----------
+
+* Crash in XMLSchema and RelaxNG when passing non-schema documents
+
+* Memory leak in RelaxNG() when RelaxNGParseError is raised
+
 0.9.1 (2006-03-30)
 ==================
 
From scoder at codespeak.net  Sat Apr 15 16:48:19 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 15 16:48:20 2006
Subject: [Lxml-checkins] r25858 - lxml/trunk
Message-ID: <20060415144819.A8913100BC@code0.codespeak.net>

Author: scoder
Date: Sat Apr 15 16:48:16 2006
New Revision: 25858

Modified:
   lxml/trunk/bench.py
Log:
benchmarks for reversed() iteration over children

Modified: lxml/trunk/bench.py
==============================================================================
--- lxml/trunk/bench.py	(original)
+++ lxml/trunk/bench.py	Sat Apr 15 16:48:16 2006
@@ -240,6 +240,14 @@
 ############################################################
 
 class BenchMark(BenchMarkBase):
+    def bench_iter_children(self, root):
+        for child in root:
+            pass
+
+    def bench_iter_children_reversed(self, root):
+        for child in reversed(root):
+            pass
+
     def bench_append_from_document(self, root1, root2):
         # == "1,2 2,3 1,3 3,1 3,2 2,1" # trees 1 and 2, or 2 and 3, or ...
         for el in root2:
From scoder at codespeak.net  Sat Apr 15 17:12:37 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 15 17:12:39 2006
Subject: [Lxml-checkins] r25859 - lxml/trunk/src/lxml/tests
Message-ID: <20060415151237.7F3D4100BC@code0.codespeak.net>

Author: scoder
Date: Sat Apr 15 17:12:36 2006
New Revision: 25859

Modified:
   lxml/trunk/src/lxml/tests/test_elementtree.py
Log:
better names for element child iteration test cases, new test case for reversed(element) (used in Python >= 2.4 only)

Modified: lxml/trunk/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_elementtree.py	Sat Apr 15 17:12:36 2006
@@ -393,7 +393,7 @@
             result.append(el.tag)
         self.assertEquals(['one', 'two', 'three'], result)
 
-    def test_iteration2(self):
+    def test_iteration_empty(self):
         XML = self.etree.XML
         
         root = XML('<doc></doc>')
@@ -402,7 +402,7 @@
             result.append(el.tag)
         self.assertEquals([], result)
 
-    def test_iteration3(self):
+    def test_iteration_text_only(self):
         XML = self.etree.XML
         
         root = XML('<doc>Text</doc>')
@@ -418,7 +418,22 @@
         for elem in root:
             elem.tail = ''
 
-    def test_iteration4(self):
+    def test_iteration_reversed(self):
+        XML = self.etree.XML
+
+        try:
+            reversed(())
+        except NameError:
+            # before Python 2.4
+            return
+
+        root = XML('<doc><one/><two>Two</two>Hm<three/></doc>')
+        result = []
+        for el in reversed(root):
+            result.append(el.tag)
+        self.assertEquals(['three', 'two', 'one'], result)
+
+    def test_iteration_subelement(self):
         XML = self.etree.XML
 
         root = XML('<doc><one/><two>Two</two>Hm<three/></doc>')
@@ -431,7 +446,7 @@
                 add = False
         self.assertEquals(['one', 'two', 'three', 'four'], result)
 
-    def test_iteration5(self):
+    def test_iteration_del_child(self):
         XML = self.etree.XML
 
         root = XML('<doc><one/><two>Two</two>Hm<three/></doc>')
@@ -441,7 +456,7 @@
             del root[-1]
         self.assertEquals(['one', 'two'], result)
 
-    def test_iteration6(self):
+    def test_iteration_double(self):
         XML = self.etree.XML
 
         root = XML('<doc><one/><two/></doc>')
From scoder at codespeak.net  Sat Apr 15 17:14:23 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 15 17:14:25 2006
Subject: [Lxml-checkins] r25860 - lxml/trunk/src/lxml
Message-ID: <20060415151423.CB615100BC@code0.codespeak.net>

Author: scoder
Date: Sat Apr 15 17:14:22 2006
New Revision: 25860

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
new 'reversed' keyword for ElementChildIterator for efficient reversed iteration, used in new method Element.__reversed__ to remove quadratic behaviour

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Sat Apr 15 17:14:22 2006
@@ -628,6 +628,9 @@
     def __iter__(self):
         return ElementChildIterator(self)
 
+    def __reversed__(self):
+        return ElementChildIterator(self, reversed=True)
+
     def index(self, _Element x, start=None, stop=None):
         cdef int k
         cdef int l
@@ -1002,12 +1005,20 @@
     registerProxy(result, PROXY_ATTRIB)
     return result
 
+ctypedef xmlNode* (*_node_to_node_function)(xmlNode*)
+
 cdef class ElementChildIterator:
     # we keep Python references here to control GC
     cdef _NodeBase _node
-    def __init__(self, _NodeBase node): # Python ref!
+    cdef _node_to_node_function _next_element
+    def __init__(self, _NodeBase node, reversed=False): # Python ref!
         cdef xmlNode* c_node
-        c_node = _findChildForwards(node._c_node, 0)
+        if reversed:
+            c_node = _findChildBackwards(node._c_node, 0)
+            self._next_element = _previousElement
+        else:
+            c_node = _findChildForwards(node._c_node, 0)
+            self._next_element = _nextElement
         if c_node is NULL:
             self._node = None
         else:
@@ -1021,7 +1032,7 @@
         current_node = self._node
         if current_node is None:
             raise StopIteration
-        c_node = _nextElement(current_node._c_node)
+        c_node = self._next_element(current_node._c_node)
         if c_node is NULL:
             self._node = None
         else:
@@ -1501,6 +1512,16 @@
         c_node = c_node.next
     return NULL
 
+cdef xmlNode* _previousElement(xmlNode* c_node):
+    """Given a node, find the next sibling that is an element.
+    """
+    c_node = c_node.prev
+    while c_node is not NULL:
+        if _isElement(c_node):
+            return c_node
+        c_node = c_node.prev
+    return NULL
+
 cdef void _removeNode(xmlNode* c_node):
     """Unlink and free a node and subnodes if possible.
     """
From scoder at codespeak.net  Sat Apr 15 17:16:39 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 15 17:16:40 2006
Subject: [Lxml-checkins] r25861 - lxml/trunk
Message-ID: <20060415151639.5E484100BC@code0.codespeak.net>

Author: scoder
Date: Sat Apr 15 17:16:38 2006
New Revision: 25861

Modified:
   lxml/trunk/CHANGES.txt
Log:
updated CHANGES.txt

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Sat Apr 15 17:16:38 2006
@@ -7,6 +7,8 @@
 Features added
 --------------
 
+* Speedup for reversed() iteration over element children (Py2.4+ only)
+
 * ElementTree compatible QName class
 
 * RelaxNG and XMLSchema now accept any Element, not only ElementTrees
From scoder at codespeak.net  Sat Apr 15 17:25:19 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 15 17:25:20 2006
Subject: [Lxml-checkins] r25862 - in lxml/branch/lxml-0.9.x: . src/lxml
	src/lxml/tests
Message-ID: <20060415152519.4E288100BC@code0.codespeak.net>

Author: scoder
Date: Sat Apr 15 17:25:16 2006
New Revision: 25862

Modified:
   lxml/branch/lxml-0.9.x/CHANGES.txt
   lxml/branch/lxml-0.9.x/bench.py
   lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
   lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py
Log:
merged in updates for efficient reversed element child iteration from trunk

Modified: lxml/branch/lxml-0.9.x/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-0.9.x/CHANGES.txt	(original)
+++ lxml/branch/lxml-0.9.x/CHANGES.txt	Sat Apr 15 17:25:16 2006
@@ -7,6 +7,8 @@
 Features added
 --------------
 
+* Speedup for reversed() iteration over element children (Py2.4+ only)
+
 * ElementTree compatible QName class
 
 * RelaxNG and XMLSchema now accept any Element, not only ElementTrees

Modified: lxml/branch/lxml-0.9.x/bench.py
==============================================================================
--- lxml/branch/lxml-0.9.x/bench.py	(original)
+++ lxml/branch/lxml-0.9.x/bench.py	Sat Apr 15 17:25:16 2006
@@ -240,6 +240,14 @@
 ############################################################
 
 class BenchMark(BenchMarkBase):
+    def bench_iter_children(self, root):
+        for child in root:
+            pass
+
+    def bench_iter_children_reversed(self, root):
+        for child in reversed(root):
+            pass
+
     def bench_append_from_document(self, root1, root2):
         # == "1,2 2,3 1,3 3,1 3,2 2,1" # trees 1 and 2, or 2 and 3, or ...
         for el in root2:

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	Sat Apr 15 17:25:16 2006
@@ -627,6 +627,9 @@
     def __iter__(self):
         return ElementChildIterator(self)
 
+    def __reversed__(self):
+        return ElementChildIterator(self, reversed=True)
+
     def index(self, _Element x, start=None, stop=None):
         cdef int k
         cdef int l
@@ -987,12 +990,20 @@
     registerProxy(result, PROXY_ATTRIB)
     return result
 
+ctypedef xmlNode* (*_node_to_node_function)(xmlNode*)
+
 cdef class ElementChildIterator:
     # we keep Python references here to control GC
     cdef _NodeBase _node
-    def __init__(self, _NodeBase node): # Python ref!
+    cdef _node_to_node_function _next_element
+    def __init__(self, _NodeBase node, reversed=False): # Python ref!
         cdef xmlNode* c_node
-        c_node = _findChildForwards(node._c_node, 0)
+        if reversed:
+            c_node = _findChildBackwards(node._c_node, 0)
+            self._next_element = _previousElement
+        else:
+            c_node = _findChildForwards(node._c_node, 0)
+            self._next_element = _nextElement
         if c_node is NULL:
             self._node = None
         else:
@@ -1006,7 +1017,7 @@
         current_node = self._node
         if current_node is None:
             raise StopIteration
-        c_node = _nextElement(current_node._c_node)
+        c_node = self._next_element(current_node._c_node)
         if c_node is NULL:
             self._node = None
         else:
@@ -1486,6 +1497,16 @@
         c_node = c_node.next
     return NULL
 
+cdef xmlNode* _previousElement(xmlNode* c_node):
+    """Given a node, find the next sibling that is an element.
+    """
+    c_node = c_node.prev
+    while c_node is not NULL:
+        if _isElement(c_node):
+            return c_node
+        c_node = c_node.prev
+    return NULL
+
 cdef void _removeNode(xmlNode* c_node):
     """Unlink and free a node and subnodes if possible.
     """

Modified: lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py	Sat Apr 15 17:25:16 2006
@@ -393,7 +393,7 @@
             result.append(el.tag)
         self.assertEquals(['one', 'two', 'three'], result)
 
-    def test_iteration2(self):
+    def test_iteration_empty(self):
         XML = self.etree.XML
         
         root = XML('<doc></doc>')
@@ -402,7 +402,7 @@
             result.append(el.tag)
         self.assertEquals([], result)
 
-    def test_iteration3(self):
+    def test_iteration_text_only(self):
         XML = self.etree.XML
         
         root = XML('<doc>Text</doc>')
@@ -418,7 +418,22 @@
         for elem in root:
             elem.tail = ''
 
-    def test_iteration4(self):
+    def test_iteration_reversed(self):
+        XML = self.etree.XML
+
+        try:
+            reversed(())
+        except NameError:
+            # before Python 2.4
+            return
+
+        root = XML('<doc><one/><two>Two</two>Hm<three/></doc>')
+        result = []
+        for el in reversed(root):
+            result.append(el.tag)
+        self.assertEquals(['three', 'two', 'one'], result)
+
+    def test_iteration_subelement(self):
         XML = self.etree.XML
 
         root = XML('<doc><one/><two>Two</two>Hm<three/></doc>')
@@ -431,7 +446,7 @@
                 add = False
         self.assertEquals(['one', 'two', 'three', 'four'], result)
 
-    def test_iteration5(self):
+    def test_iteration_del_child(self):
         XML = self.etree.XML
 
         root = XML('<doc><one/><two>Two</two>Hm<three/></doc>')
@@ -441,7 +456,7 @@
             del root[-1]
         self.assertEquals(['one', 'two'], result)
 
-    def test_iteration6(self):
+    def test_iteration_double(self):
         XML = self.etree.XML
 
         root = XML('<doc><one/><two/></doc>')
From scoder at codespeak.net  Mon Apr 17 17:34:47 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 17 17:34:53 2006
Subject: [Lxml-checkins] r25906 - lxml/branch/htmlparser/src/lxml
Message-ID: <20060417153447.1ED33100B0@code0.codespeak.net>

Author: scoder
Date: Mon Apr 17 17:34:46 2006
New Revision: 25906

Modified:
   lxml/branch/htmlparser/src/lxml/parser.pxi
Log:
small refactoring to reduce code duplication between parsers

Modified: lxml/branch/htmlparser/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/htmlparser/src/lxml/parser.pxi	(original)
+++ lxml/branch/htmlparser/src/lxml/parser.pxi	Mon Apr 17 17:34:46 2006
@@ -63,14 +63,17 @@
 
 cdef class BaseParser:
     cdef _ErrorLog _error_log
-    def __init__(self):
+    cdef object _syntax_error_class
+    def __init__(self, syntax_error_class):
         self._error_log = _ErrorLog()
+        self._syntax_error_class = syntax_error_class
 
     property error_log:
         def __get__(self):
             return self._error_log.copy()
 
-    cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt, xmlDoc* result):
+    cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,
+                               xmlDoc* result) except NULL:
         if ctxt.wellFormed:
             __GLOBAL_PARSER_CONTEXT._initDocDict(result)
         elif result is not NULL:
@@ -78,6 +81,8 @@
             tree.xmlFreeDoc(result)
             result = NULL
         self._error_log.disconnect()
+        if result is NULL:
+            raise self._syntax_error_class
         return result
 
 
@@ -113,7 +118,7 @@
                  no_network=False, ns_clean=False):
         cdef int parse_options
         self._file_parser_ctxt = NULL
-        BaseParser.__init__(self)
+        BaseParser.__init__(self, XMLSyntaxError)
 
         parse_options = _XML_DEFAULT_PARSE_OPTIONS
         if dtd_validation:
@@ -158,10 +163,7 @@
         __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
         result = xmlparser.xmlCtxtReadDoc(
             pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
-        result = self._handleResult(pctxt, result)
-        if result is NULL:
-            raise XMLSyntaxError
-        return result
+        return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
         cdef xmlDoc* result
@@ -179,10 +181,7 @@
             if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
                 self._error_log.disconnect()
                 raise IOError, "Could not open file %s" % filename
-        result = self._handleResult(pctxt, result)
-        if result is NULL:
-            raise XMLSyntaxError
-        return result
+        return self._handleResult(pctxt, result)
 
 
 cdef XMLParser __DEFAULT_XML_PARSER
@@ -231,7 +230,7 @@
         cdef int parse_options
         self._memory_parser_ctxt = NULL
         self._file_parser_ctxt   = NULL
-        BaseParser.__init__(self)
+        BaseParser.__init__(self, HTMLSyntaxError)
 
         parse_options = _HTML_DEFAULT_PARSE_OPTIONS
         if not recover:
@@ -270,10 +269,7 @@
         __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
         result = htmlparser.htmlCtxtReadDoc(
             pctxt, c_text, NULL, NULL, self._parse_options)
-        result = self._handleResult(pctxt, result)
-        if result is NULL:
-            raise HTMLSyntaxError
-        return result
+        return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
         cdef xmlDoc* result
@@ -290,10 +286,7 @@
         __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
         result = htmlparser.htmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
-        result = self._handleResult(pctxt, result)
-        if result is NULL:
-            raise HTMLSyntaxError
-        return result
+        return self._handleResult(pctxt, result)
 
 cdef HTMLParser __DEFAULT_HTML_PARSER
 __DEFAULT_HTML_PARSER = HTMLParser()
From scoder at codespeak.net  Mon Apr 17 17:35:34 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 17 17:35:36 2006
Subject: [Lxml-checkins] r25907 - lxml/branch/htmlparser/src/lxml/tests
Message-ID: <20060417153534.05ADE100B0@code0.codespeak.net>

Author: scoder
Date: Mon Apr 17 17:35:33 2006
New Revision: 25907

Modified:
   lxml/branch/htmlparser/src/lxml/tests/test_elementtree.py
Log:
clean up

Modified: lxml/branch/htmlparser/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/branch/htmlparser/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/branch/htmlparser/src/lxml/tests/test_elementtree.py	Mon Apr 17 17:35:33 2006
@@ -9,11 +9,9 @@
 """
 
 import unittest, doctest
-
-from StringIO import StringIO
 import os, shutil, tempfile, copy
 
-from common_imports import etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
+from common_imports import StringIO, etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
 
 class ETreeTestCaseBase(unittest.TestCase):
     etree = None
From scoder at codespeak.net  Mon Apr 17 17:36:20 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 17 17:36:22 2006
Subject: [Lxml-checkins] r25908 - lxml/trunk/src/lxml/tests
Message-ID: <20060417153620.B2FCF100B0@code0.codespeak.net>

Author: scoder
Date: Mon Apr 17 17:36:19 2006
New Revision: 25908

Modified:
   lxml/trunk/src/lxml/tests/test_elementtree.py
Log:
merged in clean up from branch

Modified: lxml/trunk/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_elementtree.py	Mon Apr 17 17:36:19 2006
@@ -9,11 +9,9 @@
 """
 
 import unittest, doctest
-
-from StringIO import StringIO
 import os, shutil, tempfile, copy
 
-from common_imports import etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
+from common_imports import StringIO, etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
 
 class ETreeTestCaseBase(unittest.TestCase):
     etree = None
From scoder at codespeak.net  Mon Apr 17 17:52:33 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 17 17:52:34 2006
Subject: [Lxml-checkins] r25909 - lxml/branch/htmlparser/src/lxml
Message-ID: <20060417155233.0ECC3100B0@code0.codespeak.net>

Author: scoder
Date: Mon Apr 17 17:52:32 2006
New Revision: 25909

Modified:
   lxml/branch/htmlparser/src/lxml/parser.pxi
Log:
clean up parse option setup in HTMLParser.__init__

Modified: lxml/branch/htmlparser/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/htmlparser/src/lxml/parser.pxi	(original)
+++ lxml/branch/htmlparser/src/lxml/parser.pxi	Mon Apr 17 17:52:32 2006
@@ -211,9 +211,7 @@
 
 cdef int _HTML_DEFAULT_PARSE_OPTIONS
 _HTML_DEFAULT_PARSE_OPTIONS = (
-    htmlparser.HTML_PARSE_RECOVER |
     htmlparser.HTML_PARSE_NOWARNING |
-    htmlparser.HTML_PARSE_COMPACT |
     htmlparser.HTML_PARSE_NOERROR
     )
 
@@ -233,10 +231,10 @@
         BaseParser.__init__(self, HTMLSyntaxError)
 
         parse_options = _HTML_DEFAULT_PARSE_OPTIONS
-        if not recover:
-            parse_options = parse_options & ~htmlparser.HTML_PARSE_RECOVER
-        if not compact_text:
-            parse_options = parse_options & ~htmlparser.HTML_PARSE_COMPACT
+        if recover:
+            parse_options = parse_options | htmlparser.HTML_PARSE_RECOVER
+        if compact_text:
+            parse_options = parse_options | htmlparser.HTML_PARSE_COMPACT
         if no_network:
             parse_options = parse_options | htmlparser.HTML_PARSE_NONET
         if remove_blank_text:
From scoder at codespeak.net  Mon Apr 17 17:55:12 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 17 17:55:13 2006
Subject: [Lxml-checkins] r25910 - lxml/branch/htmlparser/src/lxml/tests
Message-ID: <20060417155512.9C675100B0@code0.codespeak.net>

Author: scoder
Date: Mon Apr 17 17:55:11 2006
New Revision: 25910

Modified:
   lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
Log:
new test case for HTMLSyntexError, extended test case for default parser

Modified: lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
==============================================================================
--- lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py	(original)
+++ lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py	Mon Apr 17 17:55:11 2006
@@ -26,6 +26,13 @@
         self.assertEqual(self.etree.tostring(element),
                          self.html_str)
 
+    def test_module_parse_html_error(self):
+        parser = self.etree.HTMLParser(recover=False)
+        parse = self.etree.parse
+        f = StringIO("<html></body>")
+        self.assertRaises(self.etree.HTMLSyntaxError,
+                          parse, f, parser)
+
     def test_module_HTML_broken(self):
         element = self.etree.HTML(self.broken_html_str)
         self.assertEqual(self.etree.tostring(element),
@@ -54,6 +61,10 @@
         self.assertEqual(self.etree.tostring(tree.getroot()),
                          self.html_str)
 
+        self.etree.set_default_parser()
+
+        self.assertRaises(self.etree.XMLSyntaxError,
+                          self.etree.parse, StringIO(self.broken_html_str))
 
 def test_suite():
     suite = unittest.TestSuite()
From scoder at codespeak.net  Mon Apr 17 18:14:59 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 17 18:15:00 2006
Subject: [Lxml-checkins] r25911 - lxml/branch/htmlparser/src/lxml
Message-ID: <20060417161459.75802100A4@code0.codespeak.net>

Author: scoder
Date: Mon Apr 17 18:14:58 2006
New Revision: 25911

Modified:
   lxml/branch/htmlparser/src/lxml/parser.pxi
Log:
doc note on avoiding default parser in threads

Modified: lxml/branch/htmlparser/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/htmlparser/src/lxml/parser.pxi	(original)
+++ lxml/branch/htmlparser/src/lxml/parser.pxi	Mon Apr 17 18:14:58 2006
@@ -195,6 +195,10 @@
     parser is supplied to the various parse functions of the lxml API.  If
     this function is called without a parser (or if it is None), the default
     parser is reset to the original configuration.
+
+    Note that the default parser is not thread-safe.  Avoid the default parser
+    in multi-threaded environments.  You can create a separate parser for each
+    thread explicitly or use a parser pool.
     """
     global __DEFAULT_PARSER
     if parser is None:
From scoder at codespeak.net  Tue Apr 18 12:29:26 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 18 12:29:27 2006
Subject: [Lxml-checkins] r25937 - lxml/trunk
Message-ID: <20060418102926.DC82110088@code0.codespeak.net>

Author: scoder
Date: Tue Apr 18 12:29:25 2006
New Revision: 25937

Modified:
   lxml/trunk/INSTALL.txt
Log:
much shorter INSTALL.txt for Pyrex 0.9.4.1

Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Tue Apr 18 12:29:25 2006
@@ -14,9 +14,17 @@
 * libxslt 1.1.12 (newer versions should work). It can be found here:
   http://xmlsoft.org/XSLT/downloads.html
 
-See below for instructions how to get these for Windows.  On MacOS-X 10.4, you
-can use the installed system libraries and the binary egg distribution of
-lxml.
+For Windows, there is a `binary distribution`_ of libxml2 and libxslt.  Note
+that you need both libxml2 and libxslt, as well as iconv and zlib.  You can
+then install the `binary egg distribution`_ of lxml (see below).
+
+.. _`binary distribution`: http://www.zlatkovic.com/libxml.en.html
+.. _`binary egg distribution`: http://cheeseshop.python.org/pypi/lxml
+
+On MacOS-X 10.4, you can use the installed system libraries and the binary egg
+distribution of lxml.  Note that the libxslt version on this system is older
+than the required version above.  While there were not any bug reports so far,
+you may still encounter certain differences in behaviour in rare cases.
 
 If you want to build lxml from SVN, you also need Pyrex_. If you are using a
 released version of lxml, it should come with the generated C file in the
@@ -24,8 +32,9 @@
 
 .. _Pyrex: http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/
 
-See also the notes on building with gcc 4.0 below if you are having
-trouble with Pyrex.
+Note that Pyrex up to version 0.9.4 has known problems when compiling lxml
+with gcc 4.0 or Python 2.4.  Do not use it.  If you want to build lxml from
+non-release sources, please install Pyrex version 0.9.4.1 or later.
 
 If you have read these instructions and still cannot manage to install lxml,
 you can check the archives of the `mailing list`_ to see if your problem is
@@ -44,8 +53,8 @@
 .. _easy_install: http://peak.telecommunity.com/DevCenter/EasyInstall
 
 This has been reported to work on Linux, MacOS-X 10.4 and Windows, as long as
-libxml2 and libxslt are installed.  To compile and install lxml without
-easy_install, download the source tar-ball, unpack it and type::
+libxml2 and libxslt are properly installed.  To compile and install lxml
+without easy_install, download the source tar-ball, unpack it and type::
 
   python setup.py install
 
@@ -62,78 +71,6 @@
 import lxml.etree and play with it.
 
 
-Installation on Windows
------------------------
-
-As always, installation on Windows is different. If you do not want to go
-through the hassle of compiling everything by hand, you can use the binary
-distribution of libxml2 and libxslt. It is available here:
-
-http://www.zlatkovic.com/libxml.en.html
-
-Note that you need both libxml2 and libxslt, as well as iconv and zlib. You
-can then download a binary version of lxml 0.9 for Python 2.4 from the
-following address:
-
-http://carcass.dhs.org/lxml-0.9.win32-py2.4.exe
-
-or the egg distribution from
-
-http://cheeseshop.python.org/pypi/lxml
-
-The egg can directly be installed using easy_install_. Both builds were kindly
-contributed by Steve Howe. If they do not work for you, feel free to report to
-the mailing list.
-
-
-Building lxml with gcc 4.0 or Python 2.4
-----------------------------------------
-
-Pyrex 0.9.3.1 generates C code that gcc 4.0 does not accept. Pending an
-official release of a version of Pyrex that does work with gcc 4.0, the lxml
-project currently provides an updated version of Pyrex in its Subversion
-repository:
-
-http://codespeak.net/svn/lxml/pyrex/
-
-To install it, you can just download one of the following files:
-
-http://codespeak.net/svn/lxml/pyrex/dist/Pyrex-0.9.3.1.tar.gz
-
-http://codespeak.net/svn/lxml/pyrex/dist/Pyrex-0.9.3.1-1.src.rpm
-
-It is based on Pyrex 0.9.3.1 and contains a number of patches that make lxml
-compile and appear to work with gcc 4.0. If you use this version, you can
-simply skip the rest of the section. In case you want to apply them yourself,
-the first one is:
-
-http://codespeak.net/lxml/Pyrex-0.9.3-gcc4.patch
-
-Some Linux distributions such as Fedora Core 4 and Ubuntu Linux may
-already have most of this applied. In that case, this smaller patch
-may be applicable to make lxml compile properly:
-
-http://codespeak.net/lxml/Pyrex-0.9.3-gcc4-small.patch
-
-It may however actually be that at the time you read this, this extra patch
-has been applied by the distributions as well. You may still encounter the
-following problem when building the extension on Python 2.4::
-
-  TypeError: swig_sources() takes exactly 2 arguments (3 given)
-
-To fix this, look for the following line in Pyrex/Distutils/build_ext.py
-(around line 35)::
-
-  def swig_sources (self, sources):
-
-and change it to::
-
-  def swig_sources (self, sources, *otherargs):
-
-The above install files have these changes applied. It should do no harm if
-you install them instead of the official Pyrex version.
-
-
 Running the tests and reporting errors
 --------------------------------------
 
@@ -146,7 +83,7 @@
 above), as it searches the "src" directory.  You can use the following
 one-step command to trigger an in-place build and test it::
 
-  make test
+  make clean test
 
 To run the ElementTree and cElementTree compatibility tests, make sure
 you have lxml on your PYTHONPATH first, then run::
From scoder at codespeak.net  Tue Apr 18 16:49:39 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 18 16:49:40 2006
Subject: [Lxml-checkins] r25952 - in lxml/branch/htmlparser/src/lxml: . tests
Message-ID: <20060418144939.0C3D610088@code0.codespeak.net>

Author: scoder
Date: Tue Apr 18 16:49:38 2006
New Revision: 25952

Modified:
   lxml/branch/htmlparser/src/lxml/parser.pxi
   lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
Log:
raise IOError on IO errors in HTMLParser (file not found, etc.)

Modified: lxml/branch/htmlparser/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/htmlparser/src/lxml/parser.pxi	(original)
+++ lxml/branch/htmlparser/src/lxml/parser.pxi	Tue Apr 18 16:49:38 2006
@@ -191,10 +191,10 @@
 __DEFAULT_PARSER = __DEFAULT_XML_PARSER
 
 def set_default_parser(parser=None):
-    """Set a default XMLParser.  This parser is used globally whenever no
-    parser is supplied to the various parse functions of the lxml API.  If
-    this function is called without a parser (or if it is None), the default
-    parser is reset to the original configuration.
+    """Set a default parser.  This parser is used globally whenever no parser
+    is supplied to the various parse functions of the lxml API.  If this
+    function is called without a parser (or if it is None), the default parser
+    is reset to the original configuration.
 
     Note that the default parser is not thread-safe.  Avoid the default parser
     in multi-threaded environments.  You can create a separate parser for each
@@ -283,6 +283,9 @@
             pctxt = htmlparser.htmlCreateFileParserCtxt(filename, NULL)
             if pctxt is NULL:
                 self._error_log.disconnect()
+                warnings = self._error_log.filter_from_warnings()
+                if warnings and warnings[-1].domain == xmlerror.XML_FROM_IO:
+                    raise IOError, "Could not open file %s" % filename
                 raise ParserError, "Failed to create parser context"
             self._file_parser_ctxt = pctxt
         __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)

Modified: lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
==============================================================================
--- lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py	(original)
+++ lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py	Tue Apr 18 16:49:38 2006
@@ -51,6 +51,13 @@
         tree = self.etree.parse(f, parser)
         self.assertEqual(self.etree.tostring(tree.getroot()), self.html_str)
 
+    def test_html_file_error(self):
+        parser = self.etree.HTMLParser()
+        parse = self.etree.parse
+        self.assertRaises(IOError,
+                          parse, "__some_hopefully_nonexisting_file__.html",
+                          parser)
+
     def test_default_parser(self):
         self.assertRaises(self.etree.XMLSyntaxError,
                           self.etree.parse, StringIO(self.broken_html_str))
From scoder at codespeak.net  Wed Apr 19 10:42:40 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 10:42:46 2006
Subject: [Lxml-checkins] r25973 - lxml/trunk/src/lxml
Message-ID: <20060419084240.D0A9C100B5@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 10:42:39 2006
New Revision: 25973

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
reimplemented _Element.makeelement by copying code from Element(), avoids instantiating an empty throw-away document for the new element

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Wed Apr 19 10:42:39 2006
@@ -755,8 +755,19 @@
         else:
             return ElementTagFilter(iterator, tag)
 
-    def makeelement(self, tag, attrib):
-        return Element(tag, attrib)
+    def makeelement(self, _tag, attrib=None, nsmap=None, **_extra):
+        "Creates a new element associated with the same document."
+        # a little code duplication, but less overhead through doc reuse
+        cdef xmlNode*  c_node
+        cdef xmlDoc*   c_doc
+        cdef _Document doc
+        ns_utf, name_utf = _getNsTag(_tag)
+        doc = self._doc
+        c_doc = doc._c_doc
+        c_node = _createElement(c_doc, name_utf, attrib, _extra)
+        # add namespaces to node if necessary
+        doc._setNodeNamespaces(c_node, ns_utf, nsmap)
+        return _elementFactory(doc, c_node)
 
     def find(self, path):
         return _elementpath.find(self, path)
From scoder at codespeak.net  Wed Apr 19 11:50:32 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 11:50:33 2006
Subject: [Lxml-checkins] r25974 - lxml/trunk
Message-ID: <20060419095032.B998C100B5@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 11:50:31 2006
New Revision: 25974

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/bench.py
Log:
new benchmark for makeelement, shows that it's about 50% faster now

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Wed Apr 19 11:50:31 2006
@@ -7,6 +7,9 @@
 Features added
 --------------
 
+* Speedup for Element.makeelement(): the new element now reuses the original
+  libxml2 document instead of creating a new empty one
+
 * Speedup for reversed() iteration over element children (Py2.4+ only)
 
 * ElementTree compatible QName class

Modified: lxml/trunk/bench.py
==============================================================================
--- lxml/trunk/bench.py	(original)
+++ lxml/trunk/bench.py	Wed Apr 19 11:50:31 2006
@@ -299,6 +299,12 @@
             el = Element('{test}test')
             child.append(el)
 
+    def bench_makeelement(self, root):
+        Element = self.etree.Element
+        empty_attrib = {}
+        for child in root:
+            child.makeelement('{test}test', empty_attrib)
+
     def bench_replace_children(self, root):
         Element = self.etree.Element
         for child in root:
From scoder at codespeak.net  Wed Apr 19 11:57:18 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 11:57:19 2006
Subject: [Lxml-checkins] r25975 - in lxml/branch/lxml-0.9.x: . src/lxml
	src/lxml/tests
Message-ID: <20060419095718.4E8B8100B5@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 11:57:15 2006
New Revision: 25975

Modified:
   lxml/branch/lxml-0.9.x/CHANGES.txt
   lxml/branch/lxml-0.9.x/bench.py
   lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
   lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py
Log:
merged in makeelement() updates from trunk

Modified: lxml/branch/lxml-0.9.x/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-0.9.x/CHANGES.txt	(original)
+++ lxml/branch/lxml-0.9.x/CHANGES.txt	Wed Apr 19 11:57:15 2006
@@ -7,6 +7,9 @@
 Features added
 --------------
 
+* Speedup for Element.makeelement(): the new element now reuses the original
+  libxml2 document instead of creating a new empty one
+
 * Speedup for reversed() iteration over element children (Py2.4+ only)
 
 * ElementTree compatible QName class

Modified: lxml/branch/lxml-0.9.x/bench.py
==============================================================================
--- lxml/branch/lxml-0.9.x/bench.py	(original)
+++ lxml/branch/lxml-0.9.x/bench.py	Wed Apr 19 11:57:15 2006
@@ -299,6 +299,12 @@
             el = Element('{test}test')
             child.append(el)
 
+    def bench_makeelement(self, root):
+        Element = self.etree.Element
+        empty_attrib = {}
+        for child in root:
+            child.makeelement('{test}test', empty_attrib)
+
     def bench_replace_children(self, root):
         Element = self.etree.Element
         for child in root:

Modified: lxml/branch/lxml-0.9.x/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/etree.pyx	Wed Apr 19 11:57:15 2006
@@ -752,8 +752,19 @@
         else:
             return ElementTagFilter(iterator, tag)
 
-    def makeelement(self, tag, attrib):
-        return Element(tag, attrib)
+    def makeelement(self, _tag, attrib=None, nsmap=None, **_extra):
+        "Creates a new element associated with the same document."
+        # a little code duplication, but less overhead through doc reuse
+        cdef xmlNode*  c_node
+        cdef xmlDoc*   c_doc
+        cdef _Document doc
+        ns_utf, name_utf = _getNsTag(_tag)
+        doc = self._doc
+        c_doc = doc._c_doc
+        c_node = _createElement(c_doc, name_utf, attrib, _extra)
+        # add namespaces to node if necessary
+        doc._setNodeNamespaces(c_node, ns_utf, nsmap)
+        return _elementFactory(doc, c_node)
 
     def find(self, path):
         return _elementpath.find(self, path)

Modified: lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/branch/lxml-0.9.x/src/lxml/tests/test_elementtree.py	Wed Apr 19 11:57:15 2006
@@ -9,11 +9,9 @@
 """
 
 import unittest, doctest
-
-from StringIO import StringIO
 import os, shutil, tempfile, copy
 
-from common_imports import etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
+from common_imports import StringIO, etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
 
 class ETreeTestCaseBase(unittest.TestCase):
     etree = None
From scoder at codespeak.net  Wed Apr 19 13:47:24 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 13:47:27 2006
Subject: [Lxml-checkins] r25981 - lxml/trunk/src/lxml
Message-ID: <20060419114724.8CA151009A@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 13:47:21 2006
New Revision: 25981

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/xslt.pxi
Log:
let Pyrex handle None arguments in API functions: raises TypeError as before, less code, more readable

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Wed Apr 19 13:47:21 2006
@@ -411,13 +411,13 @@
             c_node = _deleteSlice(c_node, start, stop)
         # if the insertion point is at the end, append there
         if c_node is NULL:
+            append = self.append
             for node in value:
-                self.append(node)
+                append(node)
             return
         # if the next element is in the list, insert before it
-        for node in value:
-            _raiseIfNone(node)
-            mynode = node
+        for mynode in value:
+            _raiseIfNone(mynode)
             foreign = self._doc is not mynode._doc
             # store possible text tail
             c_next = mynode._c_node.next
@@ -444,11 +444,10 @@
     def set(self, key, value):
         self.attrib[key] = value
         
-    def append(self, _Element element):
+    def append(self, _Element element not None):
         cdef xmlNode* c_next
         cdef xmlNode* c_node
         cdef int foreign
-        _raiseIfNone(element)
         foreign = self._doc is not element._doc
         c_node = element._c_node
         # store possible text node
@@ -487,11 +486,10 @@
                 _removeNode(c_node)
             c_node = c_node_next
     
-    def insert(self, index, _Element element):
+    def insert(self, index, _Element element not None):
         cdef xmlNode* c_node
         cdef xmlNode* c_next
         cdef int foreign
-        _raiseIfNone(element)
         c_node = _findChild(self._c_node, index)
         if c_node is NULL:
             self.append(element)
@@ -502,9 +500,8 @@
         _moveTail(c_next, element._c_node)
         changeDocumentBelow(element, self._doc, foreign)
 
-    def remove(self, _Element element):
+    def remove(self, _Element element not None):
         cdef xmlNode* c_node
-        _raiseIfNone(element)
         c_node = element._c_node
         if c_node.parent is not self._c_node:
             raise ValueError, "Element is not a child of this node."
@@ -631,14 +628,13 @@
     def __reversed__(self):
         return ElementChildIterator(self, reversed=True)
 
-    def index(self, _Element x, start=None, stop=None):
+    def index(self, _Element x not None, start=None, stop=None):
         cdef int k
         cdef int l
         cdef int c_stop
         cdef int c_start
         cdef xmlNode* c_child
         cdef xmlNode* c_start_node
-        _raiseIfNone(x)
         c_child = x._c_node
         if c_child.parent is not self._c_node:
             raise ValueError, "Element is not a child of this node."
@@ -1058,9 +1054,8 @@
     # keep next node to return and a stack of position state in the tree
     cdef object _stack
     cdef _NodeBase _next_node
-    def __init__(self, _NodeBase node):
+    def __init__(self, _NodeBase node not None):
         cdef xmlNode* c_node
-        _raiseIfNone(node)
         self._next_node = node
         self._stack = []
         self._findAndPushNextNode(node)
@@ -1176,10 +1171,10 @@
     tree.xmlAddChild(<xmlNode*>doc._c_doc, c_node)
     return _commentFactory(doc, c_node)
 
-def SubElement(_Element _parent, _tag, attrib=None, nsmap=None, **_extra):
+def SubElement(_Element _parent not None, _tag,
+               attrib=None, nsmap=None, **_extra):
     cdef xmlNode*  c_node
     cdef _Document doc
-    _raiseIfNone(_parent)
     ns_utf, name_utf = _getNsTag(_tag)
     doc = _parent._doc
     c_node = _createElement(doc._c_doc, name_utf, attrib, _extra)

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Wed Apr 19 13:47:21 2006
@@ -493,8 +493,10 @@
 
         if isinstance(etree, _Document):
             doc = <_Document>etree # for internal use only!
-        else:
+        elif isinstance(etree, _ElementTree):
             doc = (<_ElementTree>etree)._doc
+        else:
+            raise TypeError, "XPathDocumentEvaluator can only work on ElementTree objects"
         
         xpathCtxt = xpath.xmlXPathNewContext(doc._c_doc)
         if xpathCtxt is NULL:
@@ -555,7 +557,7 @@
     """
     cdef _Element _element
 
-    def __init__(self, _Element element, namespaces=None, extensions=None):
+    def __init__(self, _Element element not None, namespaces=None, extensions=None):
         XPathDocumentEvaluator.__init__(
             self, element._doc, namespaces, extensions)
         self._element = element
From scoder at codespeak.net  Wed Apr 19 14:50:37 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 14:50:39 2006
Subject: [Lxml-checkins] r25988 - lxml/trunk/src/lxml
Message-ID: <20060419125037.ABDCD100B9@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 14:50:36 2006
New Revision: 25988

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
use xmlDocCopy for Element.__deepcopy__ instead of xmlDocCopyNode to avoid creating a generic empty document and to copy namespaces etc. - slower, but safer

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Wed Apr 19 14:50:36 2006
@@ -435,11 +435,14 @@
     def __copy__(self):
         cdef xmlNode* c_node
         cdef xmlDoc* c_doc
-        c_doc = theParser.newDoc()
+        cdef xmlDoc* fake_c_doc
+        cdef _Document doc
+        doc = self._doc
+        fake_c_doc = _fakeRootDoc(doc._c_doc, self._c_node)
+        c_doc = tree.xmlCopyDoc(fake_c_doc, 1) # recursive copy
+        _destroyFakeDoc(doc._c_doc, fake_c_doc)
         doc = _documentFactory(c_doc)
-        c_node = tree.xmlDocCopyNode(self._c_node, c_doc, 1)
-        tree.xmlDocSetRootElement(c_doc, c_node)
-        return _elementFactory(doc, c_node)
+        return doc.getroot()
         
     def set(self, key, value):
         self.attrib[key] = value
From scoder at codespeak.net  Wed Apr 19 17:03:44 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 17:03:45 2006
Subject: [Lxml-checkins] r25991 - lxml/trunk/src/lxml/tests
Message-ID: <20060419150344.CAA9A100B5@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 17:03:43 2006
New Revision: 25991

Modified:
   lxml/trunk/src/lxml/tests/test_elementtree.py
Log:
fix the XMLID test case by adding DTD ID information

Modified: lxml/trunk/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_elementtree.py	Wed Apr 19 17:03:43 2006
@@ -336,6 +336,13 @@
         XMLID = self.etree.XMLID
         XML   = self.etree.XML
         xml_text = '''
+        <!DOCTYPE document [
+        <!ELEMENT document (h1,p)*>
+        <!ELEMENT h1 (#PCDATA)>
+        <!ATTLIST h1 id ID #REQUIRED>
+        <!ELEMENT p  (#PCDATA)>
+        <!ATTLIST p  id ID #REQUIRED>
+        ]>
         <document>
           <h1 id="chapter1">...</h1>
           <p id="note1" class="note">...</p>
From scoder at codespeak.net  Wed Apr 19 17:24:05 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 17:24:08 2006
Subject: [Lxml-checkins] r25992 - lxml/trunk/src/lxml
Message-ID: <20060419152405.ADD26100B5@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 17:24:02 2006
New Revision: 25992

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/tree.pxd
Log:
rewrite of XMLID to follow the xml:id spec as implemented by libxml2 => only ElementTree compatible if DTD is provided

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Wed Apr 19 17:24:02 2006
@@ -1212,15 +1212,30 @@
     return etree
 
 def XML(text):
+    "Parse the XML text and return the root node."
     return _parseMemoryDocument(text, None).getroot()
 
 fromstring = XML
 
 def XMLID(text):
+    """Parse the text and return a tuple (root node, ID dictionary).  The root
+    node is the same as returned by the XML() function.  The dictionary
+    contains string-element pairs.  The dictionary keys are the values of ID
+    attributes as specified by the XML DTD.  The elements referenced by the ID
+    are stored as dictionary values.
+    """
+    cdef _NodeBase root
     root = XML(text)
+    assert root is not None
     dic = {}
-    for elem in root.xpath('//*[string(@id)]'):
-        python.PyDict_SetItem(dic, elem.get('id'), elem)
+    if root._doc._c_doc.ids is not NULL:
+        context = (dic, root._doc)
+        tree.xmlHashScan(<tree.xmlHashTable*>root._doc._c_doc.ids,
+                         _collectIdHashItems, <python.PyObject*>context)
+    elif 0:
+        # the ElementTree compatible implementation
+        for elem in root.xpath('//*[string(@id)]'):
+            python.PyDict_SetItem(dic, elem.get('id'), elem)
     return (root, dic)
 
 cdef class QName:
@@ -1394,6 +1409,15 @@
                                   c_attrib_node.ns.href)
     return funicode(value)
 
+cdef void _collectIdHashItems(void* payload, void* context, char* name):
+    # collect elements from ID attribute hash table (used by XMLID)
+    cdef tree.xmlID* c_id
+    c_id = <tree.xmlID*>payload
+    if c_id is NULL or c_id.attr is NULL:
+        return
+    dic, doc = <object>context
+    element = _elementFactory(doc, c_id.attr.parent)
+    python.PyDict_SetItemString(dic, name, element)
 
 cdef _dumpToFile(f, xmlDoc* c_doc, xmlNode* c_node):
     cdef python.PyObject* o

Modified: lxml/trunk/src/lxml/tree.pxd
==============================================================================
--- lxml/trunk/src/lxml/tree.pxd	(original)
+++ lxml/trunk/src/lxml/tree.pxd	Wed Apr 19 17:24:02 2006
@@ -8,7 +8,12 @@
 cdef extern from "libxml/encoding.h":
     ctypedef struct xmlCharEncodingHandler
     cdef xmlCharEncodingHandler* xmlFindCharEncodingHandler(char* name)
-    
+
+cdef extern from "libxml/hash.h":
+    ctypedef struct xmlHashTable
+    ctypedef void xmlHashScanner(void* payload, void* data, char* name)
+    void xmlHashScan(xmlHashTable* table, xmlHashScanner f, void* data)
+
 cdef extern from "libxml/tree.h":
 
     # for some reason need to define this in this section;
@@ -69,6 +74,7 @@
         xmlNode* prev
         xmlDoc* doc
         xmlDict* dict
+        xmlHashTable* ids
         char* URL
         
     ctypedef struct xmlAttr:
@@ -93,6 +99,11 @@
         xmlNode* prev
         xmlDoc* doc
 
+    ctypedef struct xmlID:
+        char* value
+        xmlAttr* attr
+        xmlDoc* doc
+        
     ctypedef struct xmlBuffer
     
     ctypedef struct xmlOutputBuffer:
From scoder at codespeak.net  Wed Apr 19 21:29:51 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 21:29:52 2006
Subject: [Lxml-checkins] r26003 - lxml/trunk/src/lxml
Message-ID: <20060419192951.53B0910090@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 21:29:50 2006
New Revision: 26003

Modified:
   lxml/trunk/src/lxml/etree.h
   lxml/trunk/src/lxml/python.pxd
Log:
make iter() a C macro

Modified: lxml/trunk/src/lxml/etree.h
==============================================================================
--- lxml/trunk/src/lxml/etree.h	(original)
+++ lxml/trunk/src/lxml/etree.h	Wed Apr 19 21:29:50 2006
@@ -6,6 +6,7 @@
 #define hasattr(o,a)    PyObject_HasAttr(o,a)
 #define callable(o)     PyCallable_Check(o)
 #define str(o)          PyObject_Str(o)
+#define iter(o)         PyObject_GetIter(o)
 #define _cstr(s)        PyString_AS_STRING(s)
 #define _isElement(c_node) \
         ((c_node)->type == XML_ELEMENT_NODE || \

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Wed Apr 19 21:29:50 2006
@@ -40,4 +40,5 @@
     cdef int hasattr(object obj, object attr)
     cdef int callable(object obj)
     cdef object str(object obj)
+    cdef object iter(object obj)
     cdef char* _cstr(object s)
From scoder at codespeak.net  Wed Apr 19 22:25:37 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 19 22:25:38 2006
Subject: [Lxml-checkins] r26006 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20060419202537.2B7A9100BA@code0.codespeak.net>

Author: scoder
Date: Wed Apr 19 22:25:34 2006
New Revision: 26006

Added:
   lxml/trunk/src/lxml/xmlid.pxi
Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/python.pxd
   lxml/trunk/src/lxml/tests/test_elementtree.py
   lxml/trunk/src/lxml/tests/test_etree.py
   lxml/trunk/src/lxml/tree.pxd
   lxml/trunk/src/lxml/xmlerror.pxi
Log:
new include file xmlid.pxi, reverted XMLID() to old ElementTree compatible implementation, new XMLDTDID function that returns a tuple (root node, ID dict) with a custom dictionary representation backed by a libxml2 hashtable

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Wed Apr 19 22:25:34 2006
@@ -7,6 +7,9 @@
 Features added
 --------------
 
+* XMLDTDID function parses XML into tuple (root node, ID dict) based on xml:id
+  implementation of libxml2 (as opposed to ET compatible XMLID)
+
 * Speedup for Element.makeelement(): the new element now reuses the original
   libxml2 document instead of creating a new empty one
 
@@ -68,7 +71,7 @@
 
 * XPath class for compiled XPath expressions
 
-* XMLID module level function
+* XMLID module level function (ElementTree compatible)
 
 * XMLParser API for customized libxml2 parser configuration
 

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Wed Apr 19 22:25:34 2006
@@ -1,6 +1,6 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, issubclass, hasattr, callable, str, _cstr
+from python cimport isinstance, issubclass, hasattr, callable, iter, str, _cstr
 cimport xpath
 cimport xslt
 cimport xmlerror
@@ -1217,27 +1217,6 @@
 
 fromstring = XML
 
-def XMLID(text):
-    """Parse the text and return a tuple (root node, ID dictionary).  The root
-    node is the same as returned by the XML() function.  The dictionary
-    contains string-element pairs.  The dictionary keys are the values of ID
-    attributes as specified by the XML DTD.  The elements referenced by the ID
-    are stored as dictionary values.
-    """
-    cdef _NodeBase root
-    root = XML(text)
-    assert root is not None
-    dic = {}
-    if root._doc._c_doc.ids is not NULL:
-        context = (dic, root._doc)
-        tree.xmlHashScan(<tree.xmlHashTable*>root._doc._c_doc.ids,
-                         _collectIdHashItems, <python.PyObject*>context)
-    elif 0:
-        # the ElementTree compatible implementation
-        for elem in root.xpath('//*[string(@id)]'):
-            python.PyDict_SetItem(dic, elem.get('id'), elem)
-    return (root, dic)
-
 cdef class QName:
     cdef readonly object text
     def __init__(self, text_or_uri, tag=None):
@@ -1300,6 +1279,7 @@
 
 # include submodules
 include "xmlerror.pxi"  # error and log handling
+include "xmlid.pxi"     # XMLID and IDDict
 include "nsclasses.pxi" # Namespace implementation and registry
 include "xslt.pxi"      # XPath and XSLT
 include "relaxng.pxi"   # RelaxNG
@@ -1409,16 +1389,6 @@
                                   c_attrib_node.ns.href)
     return funicode(value)
 
-cdef void _collectIdHashItems(void* payload, void* context, char* name):
-    # collect elements from ID attribute hash table (used by XMLID)
-    cdef tree.xmlID* c_id
-    c_id = <tree.xmlID*>payload
-    if c_id is NULL or c_id.attr is NULL:
-        return
-    dic, doc = <object>context
-    element = _elementFactory(doc, c_id.attr.parent)
-    python.PyDict_SetItemString(dic, name, element)
-
 cdef _dumpToFile(f, xmlDoc* c_doc, xmlNode* c_node):
     cdef python.PyObject* o
     cdef tree.xmlOutputBuffer* c_buffer

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Wed Apr 19 22:25:34 2006
@@ -27,7 +27,8 @@
     cdef int PyDict_DelItem(object d, object key)
     cdef int PyDict_Clear(object d)
     cdef object PyList_AsTuple(object o)
-    cdef object PyObject_GetIter(object o)
+    cdef object PySequence_List(object o)
+    cdef object PySequence_Tuple(object o)
 
     cdef int PyNumber_Check(object instance)
     cdef int PyBool_Check(object instance)

Modified: lxml/trunk/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_elementtree.py	Wed Apr 19 22:25:34 2006
@@ -336,13 +336,6 @@
         XMLID = self.etree.XMLID
         XML   = self.etree.XML
         xml_text = '''
-        <!DOCTYPE document [
-        <!ELEMENT document (h1,p)*>
-        <!ELEMENT h1 (#PCDATA)>
-        <!ATTLIST h1 id ID #REQUIRED>
-        <!ELEMENT p  (#PCDATA)>
-        <!ATTLIST p  id ID #REQUIRED>
-        ]>
         <document>
           <h1 id="chapter1">...</h1>
           <p id="note1" class="note">...</p>

Modified: lxml/trunk/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_etree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_etree.py	Wed Apr 19 22:25:34 2006
@@ -140,6 +140,70 @@
             b,
             d.getparent())
 
+    def test_XMLDTDID(self):
+        XMLDTDID = etree.XMLDTDID
+        XML      = etree.XML
+        xml_text = '''
+        <!DOCTYPE document [
+        <!ELEMENT document (h1,p)*>
+        <!ELEMENT h1 (#PCDATA)>
+        <!ATTLIST h1 myid ID #REQUIRED>
+        <!ELEMENT p  (#PCDATA)>
+        <!ATTLIST p  someid ID #REQUIRED>
+        ]>
+        <document>
+          <h1 myid="chapter1">...</h1>
+          <p id="note1" class="note">...</p>
+          <p>Regular paragraph.</p>
+          <p someid="warn1" class="warning">...</p>
+        </document>
+        '''
+
+        root, dic = XMLDTDID(xml_text)
+        root2 = XML(xml_text)
+        self.assertEquals(self._writeElement(root),
+                          self._writeElement(root2))
+        expected = {
+            "chapter1" : root[0],
+            "warn1"    : root[3]
+            }
+
+        self.assertEquals(dic, expected)
+        self.assertEquals(sorted(dic.items()),
+                          sorted(expected.items()))
+        self.assertEquals(sorted(dic.iteritems()),
+                          sorted(expected.iteritems()))
+        self.assertEquals(sorted(dic.keys()),
+                          sorted(expected.keys()))
+        self.assert_("chapter1" in dic)
+        self.assert_("warn1" in dic)
+
+    def test_XMLDTDID_empty(self):
+        XMLDTDID = etree.XMLDTDID
+        XML      = etree.XML
+        xml_text = '''
+        <document>
+          <h1 myid="chapter1">...</h1>
+          <p id="note1" class="note">...</p>
+          <p>Regular paragraph.</p>
+          <p someid="warn1" class="warning">...</p>
+        </document>
+        '''
+
+        root, dic = XMLDTDID(xml_text)
+        root2 = XML(xml_text)
+        self.assertEquals(self._writeElement(root),
+                          self._writeElement(root2))
+        expected = {}
+
+        self.assertEquals(dic, expected)
+        self.assertEquals(sorted(dic.items()),
+                          sorted(expected.items()))
+        self.assertEquals(sorted(dic.iteritems()),
+                          sorted(expected.iteritems()))
+        self.assertEquals(sorted(dic.keys()),
+                          sorted(expected.keys()))
+
     def test_namespaces(self):
         etree = self.etree
 
@@ -225,7 +289,7 @@
         data = f.getvalue()
         return canonicalize(data)
 
-
+        
 class ETreeXIncludeTestCase(HelperTestCase):
     def test_xinclude(self):
         tree = etree.parse(fileInTestDir('test_xinclude.xml'))

Modified: lxml/trunk/src/lxml/tree.pxd
==============================================================================
--- lxml/trunk/src/lxml/tree.pxd	(original)
+++ lxml/trunk/src/lxml/tree.pxd	Wed Apr 19 22:25:34 2006
@@ -13,6 +13,7 @@
     ctypedef struct xmlHashTable
     ctypedef void xmlHashScanner(void* payload, void* data, char* name)
     void xmlHashScan(xmlHashTable* table, xmlHashScanner f, void* data)
+    void* xmlHashLookup(xmlHashTable* table, char* name)
 
 cdef extern from "libxml/tree.h":
 

Modified: lxml/trunk/src/lxml/xmlerror.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlerror.pxi	(original)
+++ lxml/trunk/src/lxml/xmlerror.pxi	Wed Apr 19 22:25:34 2006
@@ -67,7 +67,7 @@
         return _BaseErrorLog(self._entries)
 
     def __iter__(self):
-        return python.PyObject_GetIter(self._entries)
+        return iter(self._entries)
 
     def __repr__(self):
         return '\n'.join(map(repr, self._entries))
@@ -142,7 +142,7 @@
         return _BaseErrorLog(self._entries[:])
 
     def __iter__(self):
-        return python.PyObject_GetIter(self._entries[:])
+        return iter(self._entries[:])
 
     cdef void connect(self):
         del self._entries[:]

Added: lxml/trunk/src/lxml/xmlid.pxi
==============================================================================
--- (empty file)
+++ lxml/trunk/src/lxml/xmlid.pxi	Wed Apr 19 22:25:34 2006
@@ -0,0 +1,140 @@
+from UserDict import DictMixin
+
+def XMLID(text):
+    """Parse the text and return a tuple (root node, ID dictionary).  The root
+    node is the same as returned by the XML() function.  The dictionary
+    contains string-element pairs.  The dictionary keys are the values of 'id'
+    attributes.  The elements referenced by the ID are stored as dictionary
+    values.
+    """
+    root = XML(text)
+    # ElementTree compatible implementation: look for 'id' attributes
+    dic = {}
+    for elem in root.xpath('//*[string(@id)]'):
+        python.PyDict_SetItem(dic, elem.get('id'), elem)
+    return (root, dic)
+
+def XMLDTDID(text):
+    """Parse the text and return a tuple (root node, ID dictionary).  The root
+    node is the same as returned by the XML() function.  The dictionary
+    contains string-element pairs.  The dictionary keys are the values of ID
+    attributes as defined by the DTD.  The elements referenced by the ID are
+    stored as dictionary values.
+    """
+    cdef _NodeBase root
+    root = XML(text)
+    # xml:id spec compatible implementation: use DTD ID attributes from libxml2
+    if root._doc._c_doc.ids is NULL:
+        return (root, {})
+    else:
+        return (root, _IDDict(root))
+
+class _IDDict(DictMixin):
+    """A dictionary class that mapps ID attributes to elements.
+
+    The dictionary must be instantiated with the root element of a parsed XML
+    document, otherwise the behaviour is undefined.  Elements and XML trees
+    that were created or modified through the API are not supported.
+    """
+    def __init__(self, etree):
+        cdef _Document doc
+        doc = _documentOrRaise(etree)
+        if doc._c_doc.ids is NULL:
+            raise ValueError, "No ID dictionary available."
+        self.__doc = doc
+        self.__keys  = None
+        self.__items = None
+
+    def copy(self):
+        return IDDict(self._doc)
+
+    def __getitem__(self, id_name):
+        cdef tree.xmlHashTable* c_ids
+        cdef tree.xmlID* c_id
+        cdef xmlAttr* c_attr
+        cdef _Document doc
+        doc = self.__doc
+        c_ids = doc._c_doc.ids
+        id_utf = _utf8(id_name)
+        c_id = <tree.xmlID*>tree.xmlHashLookup(c_ids, _cstr(id_utf))
+        if c_id is NULL:
+            raise KeyError, "Key not found."
+        c_attr = c_id.attr
+        if c_attr is NULL or c_attr.parent is NULL:
+            raise KeyError, "ID attribute not found."
+        return _elementFactory(doc, c_attr.parent)
+
+    def __contains__(self, id_name):
+        cdef tree.xmlID* c_id
+        cdef _Document doc
+        doc = self.__doc
+        id_utf = _utf8(id_name)
+        c_id = <tree.xmlID*>tree.xmlHashLookup(doc._c_doc.ids, _cstr(id_utf))
+        return c_id is not NULL
+
+    def keys(self):
+        keys = self.__keys
+        if keys is not None:
+            return python.PySequence_List(keys)
+        keys = self.__build_keys()
+        self.__keys = python.PySequence_Tuple(keys)
+        return keys
+
+    def __build_keys(self):
+        cdef _Document doc
+        keys = []
+        doc = self.__doc
+        tree.xmlHashScan(<tree.xmlHashTable*>doc._c_doc.ids,
+                         _collectIdHashKeys, <python.PyObject*>keys)
+        return keys
+
+    def items(self):
+        items = self.__items
+        if items is not None:
+            return python.PySequence_List(items)
+        items = self.__build_items()
+        self.__items = python.PySequence_Tuple(items)
+        return items
+
+    def iteritems(self):
+        items = self.__items
+        if items is None:
+            items = self.items()
+        return iter(items)
+
+    def __build_items(self):
+        cdef _Document doc
+        items = []
+        doc = self.__doc
+        context = (items, doc)
+        tree.xmlHashScan(<tree.xmlHashTable*>doc._c_doc.ids,
+                         _collectIdHashItemList, <python.PyObject*>context)
+        return items
+        
+
+cdef void _collectIdHashItemDict(void* payload, void* context, char* name):
+    # collect elements from ID attribute hash table
+    cdef tree.xmlID* c_id
+    c_id = <tree.xmlID*>payload
+    if c_id is NULL or c_id.attr is NULL or c_id.attr.parent is NULL:
+        return
+    dic, doc = <object>context
+    element = _elementFactory(doc, c_id.attr.parent)
+    python.PyDict_SetItem(dic, funicode(name), element)
+
+cdef void _collectIdHashItemList(void* payload, void* context, char* name):
+    # collect elements from ID attribute hash table
+    cdef tree.xmlID* c_id
+    c_id = <tree.xmlID*>payload
+    if c_id is NULL or c_id.attr is NULL or c_id.attr.parent is NULL:
+        return
+    lst, doc = <object>context
+    element = _elementFactory(doc, c_id.attr.parent)
+    python.PyList_Append(lst, (funicode(name), element))
+
+cdef void _collectIdHashKeys(void* payload, void* collect_list, char* name):
+    cdef tree.xmlID* c_id
+    c_id = <tree.xmlID*>payload
+    if c_id is NULL or c_id.attr is NULL or c_id.attr.parent is NULL:
+        return
+    python.PyList_Append(<object>collect_list, funicode(name))
From scoder at codespeak.net  Thu Apr 20 08:08:35 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 08:08:38 2006
Subject: [Lxml-checkins] r26013 - lxml/branch/htmlparser/src/lxml
Message-ID: <20060420060835.31230100C6@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 08:08:31 2006
New Revision: 26013

Modified:
   lxml/branch/htmlparser/src/lxml/htmlparser.pxd
   lxml/branch/htmlparser/src/lxml/parser.pxi
   lxml/branch/htmlparser/src/lxml/xmlparser.pxd
Log:
replaced HTML_PARSE_RECOVER by XML_PARSE_RECOVER to make it compile with libxml2 2.6.16, removed compact_text (HTML_PARSE_COMPACT) option completely for the same reason

Modified: lxml/branch/htmlparser/src/lxml/htmlparser.pxd
==============================================================================
--- lxml/branch/htmlparser/src/lxml/htmlparser.pxd	(original)
+++ lxml/branch/htmlparser/src/lxml/htmlparser.pxd	Thu Apr 20 08:08:31 2006
@@ -4,13 +4,14 @@
 
 cdef extern from "libxml/HTMLparser.h":
     ctypedef enum htmlParserOption:
-        HTML_PARSE_RECOVER    # Relaxed parsing
         HTML_PARSE_NOERROR    # suppress error reports
         HTML_PARSE_NOWARNING  # suppress warning reports
         HTML_PARSE_PEDANTIC   # pedantic error reporting
         HTML_PARSE_NOBLANKS   # remove blank nodes
         HTML_PARSE_NONET      # Forbid network access
-        HTML_PARSE_COMPACT    # compact small text nodes
+# libxml2 2.6.21+ only:
+#        HTML_PARSE_RECOVER    # Relaxed parsing
+#        HTML_PARSE_COMPACT    # compact small text nodes
 
     xmlParserCtxt* htmlCreateMemoryParserCtxt(char* buffer, int size)
     xmlParserCtxt* htmlCreateFileParserCtxt(char* filename, char* encoding)

Modified: lxml/branch/htmlparser/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/htmlparser/src/lxml/parser.pxi	(original)
+++ lxml/branch/htmlparser/src/lxml/parser.pxi	Thu Apr 20 08:08:31 2006
@@ -220,15 +220,16 @@
     )
 
 cdef class HTMLParser(BaseParser):
-    """The HTML parser.  This parser allows reading broken HTML into XML.
+    """The HTML parser.  This parser allows reading HTML into a normal XML
+    tree.  On libxml2 2.6.21 and later, it can read broken (non well-formed)
+    HTML as well.
 
     Note that you must not share parsers between threads.
     """
     cdef int _parse_options
     cdef xmlParserCtxt* _memory_parser_ctxt
     cdef xmlParserCtxt* _file_parser_ctxt
-    def __init__(self, recover=True, no_network=False,
-                 compact_text=True, remove_blank_text=False):
+    def __init__(self, recover=True, no_network=False, remove_blank_text=False):
         cdef int parse_options
         self._memory_parser_ctxt = NULL
         self._file_parser_ctxt   = NULL
@@ -236,9 +237,9 @@
 
         parse_options = _HTML_DEFAULT_PARSE_OPTIONS
         if recover:
-            parse_options = parse_options | htmlparser.HTML_PARSE_RECOVER
-        if compact_text:
-            parse_options = parse_options | htmlparser.HTML_PARSE_COMPACT
+            # XXX: make it compile on libxml2 < 2.6.21
+            #parse_options = parse_options | htmlparser.HTML_PARSE_RECOVER
+            parse_options = parse_options | xmlparser.XML_PARSE_RECOVER
         if no_network:
             parse_options = parse_options | htmlparser.HTML_PARSE_NONET
         if remove_blank_text:

Modified: lxml/branch/htmlparser/src/lxml/xmlparser.pxd
==============================================================================
--- lxml/branch/htmlparser/src/lxml/xmlparser.pxd	(original)
+++ lxml/branch/htmlparser/src/lxml/xmlparser.pxd	Thu Apr 20 08:08:31 2006
@@ -30,6 +30,8 @@
         XML_PARSE_NSCLEAN = 8192 # remove redundant namespaces declarations
         XML_PARSE_NOCDATA = 16384 # merge CDATA as text nodes
         XML_PARSE_NOXINCNODE = 32768 # do not generate XINCLUDE START/END nodes
+# libxml2 2.6.21+ only:
+#        XML_PARSE_COMPACT = 65536 # compact small text nodes
        
     cdef void xmlInitParser()
     cdef xmlParserCtxt* xmlNewParserCtxt()
From scoder at codespeak.net  Thu Apr 20 08:39:57 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 08:39:59 2006
Subject: [Lxml-checkins] r26015 - lxml/trunk/src/lxml/tests
Message-ID: <20060420063957.DD86910092@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 08:39:54 2006
New Revision: 26015

Modified:
   lxml/trunk/src/lxml/tests/test_elementtree.py
   lxml/trunk/src/lxml/tests/test_etree.py
Log:
extended test cases for handling xml:id attributes in XMLID and XMLDTDID

Modified: lxml/trunk/src/lxml/tests/test_elementtree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_elementtree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_elementtree.py	Thu Apr 20 08:39:54 2006
@@ -340,6 +340,7 @@
           <h1 id="chapter1">...</h1>
           <p id="note1" class="note">...</p>
           <p>Regular paragraph.</p>
+          <p xml:id="xmlid">XML:ID paragraph.</p>
           <p id="warn1" class="warning">...</p>
         </document>
         '''
@@ -351,7 +352,7 @@
         expected = {
             "chapter1" : root[0],
             "note1"    : root[1],
-            "warn1"    : root[3]
+            "warn1"    : root[4]
             }
         self.assertEquals(dic, expected)
 

Modified: lxml/trunk/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_etree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_etree.py	Thu Apr 20 08:39:54 2006
@@ -155,6 +155,7 @@
           <h1 myid="chapter1">...</h1>
           <p id="note1" class="note">...</p>
           <p>Regular paragraph.</p>
+          <p xml:id="xmlid">XML:ID paragraph.</p>
           <p someid="warn1" class="warning">...</p>
         </document>
         '''
@@ -165,7 +166,8 @@
                           self._writeElement(root2))
         expected = {
             "chapter1" : root[0],
-            "warn1"    : root[3]
+            "xmlid"    : root[3],
+            "warn1"    : root[4]
             }
 
         self.assertEquals(dic, expected)
@@ -177,6 +179,7 @@
                           sorted(expected.keys()))
         self.assert_("chapter1" in dic)
         self.assert_("warn1" in dic)
+        self.assert_("xmlid" in dic)
 
     def test_XMLDTDID_empty(self):
         XMLDTDID = etree.XMLDTDID
From scoder at codespeak.net  Thu Apr 20 08:44:42 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 08:44:43 2006
Subject: [Lxml-checkins] r26016 - lxml/branch/htmlparser/src/lxml/tests
Message-ID: <20060420064442.57F6810092@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 08:44:41 2006
New Revision: 26016

Modified:
   lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
Log:
clean up in HTML parser tests

Modified: lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
==============================================================================
--- lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py	(original)
+++ lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py	Thu Apr 20 08:44:41 2006
@@ -16,7 +16,6 @@
 
     html_str = "<html><head><title>test</title></head><body><h1>page title</h1></body></html>"
     broken_html_str = "<html><head><title>test<body><h1>page title</body></html>"
-    blank_text_html_str = "<html><head><title>  </title></head><body><h1>  </h1></body></html>"
 
     def tearDown(self):
         self.etree.set_default_parser()
@@ -58,7 +57,7 @@
                           parse, "__some_hopefully_nonexisting_file__.html",
                           parser)
 
-    def test_default_parser(self):
+    def test_default_parser_HTML_broken(self):
         self.assertRaises(self.etree.XMLSyntaxError,
                           self.etree.parse, StringIO(self.broken_html_str))
 
From scoder at codespeak.net  Thu Apr 20 11:35:32 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 11:35:34 2006
Subject: [Lxml-checkins] r26022 - in lxml/trunk/src/lxml: . tests
Message-ID: <20060420093532.D215F1008B@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 11:35:31 2006
New Revision: 26022

Modified:
   lxml/trunk/src/lxml/python.pxd
   lxml/trunk/src/lxml/tests/test_etree.py
   lxml/trunk/src/lxml/xmlid.pxi
Log:
removed IDDict dependency on UserDict, extended test cases

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Thu Apr 20 11:35:31 2006
@@ -29,6 +29,7 @@
     cdef object PyList_AsTuple(object o)
     cdef object PySequence_List(object o)
     cdef object PySequence_Tuple(object o)
+    cdef object PyTuple_GET_ITEM(object o, int pos)
 
     cdef int PyNumber_Check(object instance)
     cdef int PyBool_Check(object instance)

Modified: lxml/trunk/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_etree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_etree.py	Thu Apr 20 11:35:31 2006
@@ -171,12 +171,20 @@
             }
 
         self.assertEquals(dic, expected)
+        self.assertEquals(len(dic),
+                          len(expected))
         self.assertEquals(sorted(dic.items()),
                           sorted(expected.items()))
         self.assertEquals(sorted(dic.iteritems()),
                           sorted(expected.iteritems()))
         self.assertEquals(sorted(dic.keys()),
                           sorted(expected.keys()))
+        self.assertEquals(sorted(dic.iterkeys()),
+                          sorted(expected.iterkeys()))
+        self.assertEquals(sorted(dic.values()),
+                          sorted(expected.values()))
+        self.assertEquals(sorted(dic.itervalues()),
+                          sorted(expected.itervalues()))
         self.assert_("chapter1" in dic)
         self.assert_("warn1" in dic)
         self.assert_("xmlid" in dic)
@@ -200,12 +208,12 @@
         expected = {}
 
         self.assertEquals(dic, expected)
-        self.assertEquals(sorted(dic.items()),
-                          sorted(expected.items()))
+        self.assertEquals(dic.items(),
+                          expected.items())
         self.assertEquals(sorted(dic.iteritems()),
                           sorted(expected.iteritems()))
-        self.assertEquals(sorted(dic.keys()),
-                          sorted(expected.keys()))
+        self.assertEquals(dic.keys(),
+                          expected.keys())
 
     def test_namespaces(self):
         etree = self.etree

Modified: lxml/trunk/src/lxml/xmlid.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlid.pxi	(original)
+++ lxml/trunk/src/lxml/xmlid.pxi	Thu Apr 20 11:35:31 2006
@@ -1,5 +1,3 @@
-from UserDict import DictMixin
-
 def XMLID(text):
     """Parse the text and return a tuple (root node, ID dictionary).  The root
     node is the same as returned by the XML() function.  The dictionary
@@ -29,21 +27,24 @@
     else:
         return (root, _IDDict(root))
 
-class _IDDict(DictMixin):
-    """A dictionary class that mapps ID attributes to elements.
+cdef class _IDDict:
+    """A dictionary-like proxy class that mapps ID attributes to elements.
 
     The dictionary must be instantiated with the root element of a parsed XML
     document, otherwise the behaviour is undefined.  Elements and XML trees
     that were created or modified through the API are not supported.
     """
+    cdef _Document _doc
+    cdef object _keys
+    cdef object _items
     def __init__(self, etree):
         cdef _Document doc
         doc = _documentOrRaise(etree)
         if doc._c_doc.ids is NULL:
             raise ValueError, "No ID dictionary available."
-        self.__doc = doc
-        self.__keys  = None
-        self.__items = None
+        self._doc = doc
+        self._keys  = None
+        self._items = None
 
     def copy(self):
         return IDDict(self._doc)
@@ -52,9 +53,7 @@
         cdef tree.xmlHashTable* c_ids
         cdef tree.xmlID* c_id
         cdef xmlAttr* c_attr
-        cdef _Document doc
-        doc = self.__doc
-        c_ids = doc._c_doc.ids
+        c_ids = self._doc._c_doc.ids
         id_utf = _utf8(id_name)
         c_id = <tree.xmlID*>tree.xmlHashLookup(c_ids, _cstr(id_utf))
         if c_id is NULL:
@@ -62,55 +61,104 @@
         c_attr = c_id.attr
         if c_attr is NULL or c_attr.parent is NULL:
             raise KeyError, "ID attribute not found."
-        return _elementFactory(doc, c_attr.parent)
+        return _elementFactory(self._doc, c_attr.parent)
+
+    def get(self, id_name):
+        return self[id_name]
 
     def __contains__(self, id_name):
         cdef tree.xmlID* c_id
-        cdef _Document doc
-        doc = self.__doc
         id_utf = _utf8(id_name)
-        c_id = <tree.xmlID*>tree.xmlHashLookup(doc._c_doc.ids, _cstr(id_utf))
+        c_id = <tree.xmlID*>tree.xmlHashLookup(
+            self._doc._c_doc.ids, _cstr(id_utf))
         return c_id is not NULL
 
+    def has_key(self, id_name):
+        return self.__contains__(id_name)
+
+    def __cmp__(self, other):
+        if other is None:
+            return 1
+        else:
+            return cmp(dict(self), other)
+
+    def __richcmp__(self, other, int op):
+        cdef int c_cmp
+        if other is None:
+            return op == 0 or op == 1 or op == 3
+        c_cmp = cmp(dict(self), other)
+        if c_cmp == 0: # equal
+            return op == 1 or op == 2 or op == 5
+        elif c_cmp < 0:
+            return op == 0 or op == 1 or op == 3
+        else:
+            return op == 4 or op == 5 or op == 3
+
+    def __repr__(self):
+        return repr(dict(self))
+
     def keys(self):
-        keys = self.__keys
+        keys = self._keys
         if keys is not None:
             return python.PySequence_List(keys)
-        keys = self.__build_keys()
-        self.__keys = python.PySequence_Tuple(keys)
+        keys = self._build_keys()
+        self._keys = python.PySequence_Tuple(keys)
         return keys
 
-    def __build_keys(self):
-        cdef _Document doc
+    def __iter__(self):
+        keys = self._keys
+        if keys is None:
+            keys = self.keys()
+        return iter(keys)
+
+    def iterkeys(self):
+        return self.__iter__()
+
+    def __len__(self):
+        keys = self._keys
+        if keys is None:
+            keys = self.keys()
+        return len(keys)
+
+    cdef object _build_keys(self):
         keys = []
-        doc = self.__doc
-        tree.xmlHashScan(<tree.xmlHashTable*>doc._c_doc.ids,
+        tree.xmlHashScan(<tree.xmlHashTable*>self._doc._c_doc.ids,
                          _collectIdHashKeys, <python.PyObject*>keys)
         return keys
 
     def items(self):
-        items = self.__items
+        items = self._items
         if items is not None:
             return python.PySequence_List(items)
-        items = self.__build_items()
-        self.__items = python.PySequence_Tuple(items)
+        items = self._build_items()
+        self._items = python.PySequence_Tuple(items)
         return items
 
     def iteritems(self):
-        items = self.__items
+        items = self._items
         if items is None:
             items = self.items()
         return iter(items)
 
-    def __build_items(self):
-        cdef _Document doc
+    cdef object _build_items(self):
         items = []
-        doc = self.__doc
-        context = (items, doc)
-        tree.xmlHashScan(<tree.xmlHashTable*>doc._c_doc.ids,
+        context = (items, self._doc)
+        tree.xmlHashScan(<tree.xmlHashTable*>self._doc._c_doc.ids,
                          _collectIdHashItemList, <python.PyObject*>context)
         return items
-        
+
+    def values(self):
+        items = self._items
+        if items is None:
+            items = self.items()
+        values = []
+        for item in items:
+            value = python.PyTuple_GET_ITEM(item, 1)
+            python.PyList_Append(values, value)
+        return values
+
+    def itervalues(self):
+        return iter(self.values())
 
 cdef void _collectIdHashItemDict(void* payload, void* context, char* name):
     # collect elements from ID attribute hash table
From scoder at codespeak.net  Thu Apr 20 11:37:09 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 11:37:09 2006
Subject: [Lxml-checkins] r26023 - lxml/trunk/src/lxml
Message-ID: <20060420093709.5366C1008B@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 11:37:07 2006
New Revision: 26023

Added:
   lxml/trunk/src/lxml/htmlparser.pxd
      - copied unchanged from r26022, lxml/branch/htmlparser/src/lxml/htmlparser.pxd
Log:
merged in new file from htmlparser branch

From scoder at codespeak.net  Thu Apr 20 11:47:42 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 11:47:44 2006
Subject: [Lxml-checkins] r26024 - lxml/trunk/src/lxml/tests
Message-ID: <20060420094742.B56CB1008B@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 11:47:41 2006
New Revision: 26024

Added:
   lxml/trunk/src/lxml/tests/test_htmlparser.py
      - copied unchanged from r26023, lxml/branch/htmlparser/src/lxml/tests/test_htmlparser.py
Log:
copied test file from htmlparser branch
From scoder at codespeak.net  Thu Apr 20 12:12:05 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 12:12:06 2006
Subject: [Lxml-checkins] r26027 - lxml/trunk/src/lxml
Message-ID: <20060420101205.94F1E1008D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 12:12:03 2006
New Revision: 26027

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/parser.pxi
   lxml/trunk/src/lxml/xmlparser.pxd
Log:
merge of htmlparser branch

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 20 12:12:03 2006
@@ -70,7 +70,7 @@
         # the document
         #print "freeing document:", <int>self._c_doc
         #displayNode(<xmlNode*>self._c_doc, 0)
-        #print self._c_doc.dict is theParser._c_dict
+        #print <int>self._c_doc, self._c_doc.dict is __GLOBAL_PARSER_CONTEXT._c_dict
         tree.xmlFreeDoc(self._c_doc)
 
     cdef getroot(self):
@@ -141,12 +141,8 @@
 
 cdef _Document _parseDocument(source, parser):
     cdef xmlDoc* c_doc
-    # XXX simplistic (c)StringIO support
-    if hasattr(source, 'getvalue'):
-        return _parseMemoryDocument(source.getvalue(), parser)
-
     filename = _getFilenameForFile(source)
-    # Support for unamed file-like object (eg urlgrabber.urlopen)
+    # Support for unamed file-like object (StringIO, urlgrabber.urlopen, ...)
     if not filename and hasattr(source, 'read'):
         return _parseMemoryDocument(source.read(), parser)
 
@@ -154,14 +150,14 @@
     if filename is None:
         filename = source
     # open filename
-    c_doc = theParser.parseDocFromFile(filename, parser)
+    c_doc = _parseDocFromFile(filename, parser)
     return _documentFactory(c_doc)
 
 cdef _Document _parseMemoryDocument(text, parser):
     cdef xmlDoc* c_doc
     if python.PyUnicode_Check(text):
         text = _stripDeclaration(_utf8(text))
-    c_doc = theParser.parseDoc(text, parser)
+    c_doc = _parseDoc(text, parser)
     return _documentFactory(c_doc)
 
 cdef _Document _documentFactory(xmlDoc* c_doc):
@@ -1154,7 +1150,7 @@
     cdef xmlDoc*   c_doc
     cdef _Document doc
     ns_utf, name_utf = _getNsTag(_tag)
-    c_doc = theParser.newDoc()
+    c_doc = _newDoc()
     c_node = _createElement(c_doc, name_utf, attrib, _extra)
     tree.xmlDocSetRootElement(c_doc, c_node)
     doc = _documentFactory(c_doc)
@@ -1169,7 +1165,7 @@
         text = '  '
     else:
         text = ' %s ' % _utf8(text)
-    doc = _documentFactory( theParser.newDoc() )
+    doc = _documentFactory( _newDoc() )
     c_node = _createComment(doc._c_doc, text)
     tree.xmlAddChild(<xmlNode*>doc._c_doc, c_node)
     return _commentFactory(doc, c_node)
@@ -1198,7 +1194,7 @@
     elif file is not None:
         doc = _parseDocument(file, parser)
     else:
-        doc = _documentFactory( theParser.newDoc() )
+        doc = _documentFactory( _newDoc() )
 
     etree = _elementTreeFactory(doc, element)
 
@@ -1211,9 +1207,15 @@
     
     return etree
 
+def HTML(text):
+    cdef _Document doc
+    doc = _parseMemoryDocument(text, __DEFAULT_HTML_PARSER)
+    return doc.getroot()
+
 def XML(text):
-    "Parse the XML text and return the root node."
-    return _parseMemoryDocument(text, None).getroot()
+    cdef _Document doc
+    doc = _parseMemoryDocument(text, __DEFAULT_XML_PARSER)
+    return doc.getroot()
 
 fromstring = XML
 
@@ -1288,11 +1290,6 @@
 include "proxy.pxi"     # Proxy handling (element backpointers/memory/etc.)
 
 
-# Instantiate globally shared XML parser to enable dictionary sharing
-cdef Parser theParser
-theParser = Parser()
-
-
 # Private helper functions
 cdef void _raiseIfNone(el):
     if el is None:

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Thu Apr 20 12:12:03 2006
@@ -1,24 +1,104 @@
 # XML parser that provides dictionary sharing
 
 cimport xmlparser
+cimport htmlparser
 from xmlparser cimport xmlParserCtxt, xmlDict
 
 class XMLSyntaxError(LxmlSyntaxError):
     pass
 
-cdef int _DEFAULT_PARSE_OPTIONS
-_DEFAULT_PARSE_OPTIONS = (
+class HTMLSyntaxError(LxmlSyntaxError):
+    pass
+
+class ParserError(LxmlError):
+    pass
+
+cdef class _ParserContext:
+    """Global parser context to share the string dictionary.
+    """
+    cdef xmlDict* _c_dict
+    cdef int _initialized
+
+    def __init__(self):
+        self._c_dict = NULL
+        self._initialized = 0
+
+    def __dealloc__(self):
+        if self._c_dict is not NULL:
+            xmlparser.xmlDictFree(self._c_dict)
+
+    cdef void _initParser(self):
+        if not self._initialized:
+            xmlparser.xmlInitParser()
+            self._initialized = 1
+
+    cdef void _initParserDict(self, xmlParserCtxt* pctxt):
+        "Assure we always use the same string dictionary."
+        if self._c_dict is NULL or self._c_dict is pctxt.dict:
+            return
+        if pctxt.dict is not NULL:
+            xmlparser.xmlDictFree(pctxt.dict)
+        pctxt.dict = self._c_dict
+        xmlparser.xmlDictReference(pctxt.dict)
+
+    cdef void _initDocDict(self, xmlDoc* result):
+        "Store dict of last object parsed if no shared dict yet"
+        if result is NULL:
+            return
+        if self._c_dict is NULL:
+            #print "storing shared dict"
+            if result.dict is NULL:
+                result.dict = xmlparser.xmlDictCreate()
+            self._c_dict = result.dict
+            xmlparser.xmlDictReference(result.dict)
+        elif result.dict != self._c_dict:
+            if result.dict is not NULL:
+                xmlparser.xmlDictFree(result.dict)
+            result.dict = self._c_dict
+            xmlparser.xmlDictReference(self._c_dict)
+
+cdef _ParserContext __GLOBAL_PARSER_CONTEXT
+__GLOBAL_PARSER_CONTEXT = _ParserContext()
+
+
+cdef class BaseParser:
+    cdef _ErrorLog _error_log
+    cdef object _syntax_error_class
+    def __init__(self, syntax_error_class):
+        self._error_log = _ErrorLog()
+        self._syntax_error_class = syntax_error_class
+
+    property error_log:
+        def __get__(self):
+            return self._error_log.copy()
+
+    cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,
+                               xmlDoc* result) except NULL:
+        if ctxt.wellFormed:
+            __GLOBAL_PARSER_CONTEXT._initDocDict(result)
+        elif result is not NULL:
+            # free broken document
+            tree.xmlFreeDoc(result)
+            result = NULL
+        self._error_log.disconnect()
+        if result is NULL:
+            raise self._syntax_error_class
+        return result
+
+
+############################################################
+## XML parser
+############################################################
+
+cdef int _XML_DEFAULT_PARSE_OPTIONS
+_XML_DEFAULT_PARSE_OPTIONS = (
     xmlparser.XML_PARSE_NOENT |
     xmlparser.XML_PARSE_NOCDATA |
     xmlparser.XML_PARSE_NOWARNING |
     xmlparser.XML_PARSE_NOERROR
     )
 
-cdef int _ORIG_DEFAULT_PARSE_OPTIONS
-_ORIG_DEFAULT_PARSE_OPTIONS = _DEFAULT_PARSE_OPTIONS
-
-
-cdef class XMLParser:
+cdef class XMLParser(BaseParser):
     """The XML parser.  Parsers can be supplied as additional argument to
     various parse functions of the lxml API.  A default parser is always
     available and can be replaced by a call to the global function
@@ -28,13 +108,19 @@
     The keyword arguments in the constructor are mainly based on the libxml2
     parser configuration.  A DTD will only be loaded if validation or
     attribute default values are requested.
+
+    Note that you must not share parsers between threads.
     """
     cdef int _parse_options
+    cdef xmlParserCtxt* _file_parser_ctxt
+    cdef xmlParserCtxt* _memory_parser_ctxt
     def __init__(self, attribute_defaults=False, dtd_validation=False,
                  no_network=False, ns_clean=False):
         cdef int parse_options
-        parse_options = _ORIG_DEFAULT_PARSE_OPTIONS
+        self._file_parser_ctxt = NULL
+        BaseParser.__init__(self, XMLSyntaxError)
 
+        parse_options = _XML_DEFAULT_PARSE_OPTIONS
         if dtd_validation:
             parse_options = parse_options | xmlparser.XML_PARSE_DTDLOAD | \
                             xmlparser.XML_PARSE_DTDVALID
@@ -52,156 +138,197 @@
         def __get__(self):
             return __build_error_log_tuple(self)
 
-##     def copy(self, attribute_defaults=None, dtd_validation=None,
-##              no_network=None, ns_clean=None):
-##         cdef int parse_options
-##         parse_options = self._parse_options
-##         if attribute_defaults is None:
-##             attribute_defaults = parse_options & xmlparser.XML_PARSE_DTDATTR
-##         if dtd_validation is None:
-##             dtd_validation = parse_options & xmlparser.XML_PARSE_DTDVALID
-##         if no_network is None:
-##             no_network = parse_options & xmlparser.XML_PARSE_NONET
-##         if ns_clean is None:
-##             ns_clean = parse_options & xmlparser.XML_PARSE_NSCLEAN
-
-##         return self.__class__(attribute_defaults=attribute_defaults,
-##                               dtd_validation=dtd_validation,
-##                               no_network=no_network, ns_clean=ns_clean)
-
-
-def set_default_parser(parser=None):
-    """Set a default XMLParser.  This parser is used globally whenever no
-    parser is supplied to the various parse functions of the lxml API.  If
-    this function is called without a parser (or if it is None), the default
-    parser is reset to the original configuration.
-    """
-    if parser is not None:
-        _DEFAULT_PARSE_OPTIONS = (<XMLParser>parser)._parse_options
-    else:
-        _DEFAULT_PARSE_OPTIONS = _ORIG_DEFAULT_PARSE_OPTIONS
-
+    def __dealloc__(self):
+        if self._file_parser_ctxt != NULL:
+            xmlparser.xmlFreeParserCtxt(self._file_parser_ctxt)
+        if self._memory_parser_ctxt != NULL:
+            xmlparser.xmlFreeParserCtxt(self._memory_parser_ctxt)
 
-cdef class Parser:
+    cdef xmlParserCtxt* _createContext(self) except NULL:
+        cdef xmlParserCtxt* pctxt
+        pctxt = xmlparser.xmlNewParserCtxt()
+        if pctxt is NULL:
+            self._error_log.disconnect()
+            raise ParserError, "Failed to create parser context"
+        return pctxt
 
-    cdef xmlDict* _c_dict
-    cdef int _parser_initialized
-    
-    def __init__(self):
-        self._c_dict = NULL
-        self._parser_initialized = 0
-        
-    def __dealloc__(self):
-        #print "cleanup parser"
-        if self._c_dict is not NULL:
-            #print "freeing dictionary (cleanup parser)"
-            xmlparser.xmlDictFree(self._c_dict)
-        
-    cdef xmlDoc* parseDoc(self, text, parser) except NULL:
+    cdef xmlDoc* _parseDoc(self, text_utf) except NULL:
         """Parse document, share dictionary if possible.
         """
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
         cdef int parse_error
-
-        if parser is not None:
-            parse_options = (<XMLParser>parser)._parse_options
-        else:
-            parse_options = _DEFAULT_PARSE_OPTIONS
-
-        self._initParse()
-        pctxt = xmlparser.xmlCreateDocParserCtxt(_cstr(text))
+        self._error_log.connect()
+        pctxt = self._memory_parser_ctxt
         if pctxt is NULL:
-            raise XMLSyntaxError
+            pctxt = self._createContext()
+            self._memory_parser_ctxt = pctxt
 
-        self._prepareParse(pctxt)
-        xmlparser.xmlCtxtUseOptions(
-            pctxt,
-            parse_options)
-        parse_error = xmlparser.xmlParseDocument(pctxt)
-        # in case of errors, clean up context plus any document
-        if parse_error != 0 or not pctxt.wellFormed:
-            if pctxt.myDoc is not NULL:
-                tree.xmlFreeDoc(pctxt.myDoc)
-                pctxt.myDoc = NULL
-            xmlparser.xmlFreeParserCtxt(pctxt)
-            raise XMLSyntaxError
-        result = pctxt.myDoc
-        self._finalizeParse(result)
-        xmlparser.xmlFreeParserCtxt(pctxt)
-        return result
+        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        result = xmlparser.xmlCtxtReadDoc(
+            pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
+        return self._handleResult(pctxt, result)
 
-    cdef xmlDoc* parseDocFromFile(self, char* filename, parser) except NULL:
-        cdef int parse_options
+    cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
+        self._error_log.connect()
+        pctxt = self._file_parser_ctxt
+        if pctxt is NULL:
+            pctxt = self._createContext()
+            self._file_parser_ctxt = pctxt
 
-        if parser is not None:
-            parse_options = (<XMLParser>parser)._parse_options
-        else:
-            parse_options = _DEFAULT_PARSE_OPTIONS
-
-        self._initParse()
-        pctxt = xmlparser.xmlNewParserCtxt()
-        self._prepareParse(pctxt)
-        # XXX set options twice? needed to shut up libxml2
-        xmlparser.xmlCtxtUseOptions(pctxt, parse_options)
-        result = xmlparser.xmlCtxtReadFile(pctxt, filename,
-                                           NULL, parse_options)
+        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        result = xmlparser.xmlCtxtReadFile(
+            pctxt, filename, NULL, self._parse_options)
         if result is NULL:
             if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
+                self._error_log.disconnect()
                 raise IOError, "Could not open file %s" % filename
-        # in case of errors, clean up context plus any document
-        # XXX other errors?
-        if not pctxt.wellFormed:
-            if pctxt.myDoc is not NULL:
-                tree.xmlFreeDoc(pctxt.myDoc)
-                pctxt.myDoc = NULL
-            xmlparser.xmlFreeParserCtxt(pctxt)
-            raise XMLSyntaxError
-        self._finalizeParse(result)
-        xmlparser.xmlFreeParserCtxt(pctxt)
-        return result
-    
-    cdef void _initParse(self):
-        if not self._parser_initialized:
-            xmlparser.xmlInitParser()
-            self._parser_initialized = 1
-            
-    cdef void _prepareParse(self, xmlParserCtxt* pctxt):
-        if self._c_dict is not NULL and pctxt.dict is not NULL:
-            #print "sharing dictionary (parseDoc)"
-            xmlparser.xmlDictFree(pctxt.dict)
-            pctxt.dict = self._c_dict
-            xmlparser.xmlDictReference(pctxt.dict)
+        return self._handleResult(pctxt, result)
 
-    cdef void _finalizeParse(self, xmlDoc* result):
-        # store dict of last object parsed if no shared dict yet
-        if self._c_dict is NULL:
-            #print "storing shared dict"
-            self._c_dict = result.dict
-        xmlparser.xmlDictReference(self._c_dict)
-    
-    cdef xmlDoc* newDoc(self):
+
+cdef XMLParser __DEFAULT_XML_PARSER
+__DEFAULT_XML_PARSER = XMLParser()
+
+cdef BaseParser __DEFAULT_PARSER
+__DEFAULT_PARSER = __DEFAULT_XML_PARSER
+
+def set_default_parser(parser=None):
+    """Set a default parser.  This parser is used globally whenever no parser
+    is supplied to the various parse functions of the lxml API.  If this
+    function is called without a parser (or if it is None), the default parser
+    is reset to the original configuration.
+
+    Note that the default parser is not thread-safe.  Avoid the default parser
+    in multi-threaded environments.  You can create a separate parser for each
+    thread explicitly or use a parser pool.
+    """
+    global __DEFAULT_PARSER
+    if parser is None:
+        __DEFAULT_PARSER = __DEFAULT_XML_PARSER
+    elif isinstance(parser, (HTMLParser, XMLParser)):
+        __DEFAULT_PARSER = parser
+    else:
+        raise TypeError, "Invalid parser"
+
+
+############################################################
+## HTML parser
+############################################################
+
+cdef int _HTML_DEFAULT_PARSE_OPTIONS
+_HTML_DEFAULT_PARSE_OPTIONS = (
+    htmlparser.HTML_PARSE_NOWARNING |
+    htmlparser.HTML_PARSE_NOERROR
+    )
+
+cdef class HTMLParser(BaseParser):
+    """The HTML parser.  This parser allows reading HTML into a normal XML
+    tree.  By default, it can read broken (non well-formed) HTML, depending on
+    the capabilities of libxml2.  Use the 'recover' option to switch this off.
+
+    Note that you must not share parsers between threads.
+    """
+    cdef int _parse_options
+    cdef xmlParserCtxt* _memory_parser_ctxt
+    cdef xmlParserCtxt* _file_parser_ctxt
+    def __init__(self, recover=True, no_network=False, remove_blank_text=False):
+        cdef int parse_options
+        self._memory_parser_ctxt = NULL
+        self._file_parser_ctxt   = NULL
+        BaseParser.__init__(self, HTMLSyntaxError)
+
+        parse_options = _HTML_DEFAULT_PARSE_OPTIONS
+        if recover:
+            # XXX: make it compile on libxml2 < 2.6.21
+            #parse_options = parse_options | htmlparser.HTML_PARSE_RECOVER
+            parse_options = parse_options | xmlparser.XML_PARSE_RECOVER
+        if no_network:
+            parse_options = parse_options | htmlparser.HTML_PARSE_NONET
+        if remove_blank_text:
+            parse_options = parse_options | htmlparser.HTML_PARSE_NOBLANKS
+
+        self._parse_options = parse_options
+
+    def __dealloc__(self):
+        if self._file_parser_ctxt != NULL:
+            htmlparser.htmlFreeParserCtxt(self._file_parser_ctxt)
+        if self._memory_parser_ctxt != NULL:
+            htmlparser.htmlFreeParserCtxt(self._memory_parser_ctxt)
+
+    cdef xmlDoc* _parseDoc(self, text_utf) except NULL:
+        """Parse HTML document, share dictionary if possible.
+        """
         cdef xmlDoc* result
-        cdef xmlDict* d
+        cdef xmlParserCtxt* pctxt
+        cdef char* c_text
+        cdef int c_len
+        self._error_log.connect()
+        c_text = _cstr(text_utf)
+        pctxt = self._memory_parser_ctxt
+        if pctxt is NULL:
+            pctxt = htmlparser.htmlCreateMemoryParserCtxt('dummy', 5)
+            if pctxt is NULL:
+                self._error_log.disconnect()
+                raise ParserError, "Failed to create parser context"
+            self._memory_parser_ctxt = pctxt
+        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        result = htmlparser.htmlCtxtReadDoc(
+            pctxt, c_text, NULL, NULL, self._parse_options)
+        return self._handleResult(pctxt, result)
 
-        result = tree.xmlNewDoc("1.0")
+    cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
+        cdef xmlDoc* result
+        cdef xmlParserCtxt* pctxt
+        cdef int parser_error
+        self._error_log.connect()
+        pctxt = self._file_parser_ctxt
+        if pctxt is NULL:
+            pctxt = htmlparser.htmlCreateFileParserCtxt(filename, NULL)
+            if pctxt is NULL:
+                self._error_log.disconnect()
+                warnings = self._error_log.filter_from_warnings()
+                if warnings and warnings[-1].domain == xmlerror.XML_FROM_IO:
+                    raise IOError, "Could not open file %s" % filename
+                raise ParserError, "Failed to create parser context"
+            self._file_parser_ctxt = pctxt
+        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        result = htmlparser.htmlCtxtReadFile(
+            pctxt, filename, NULL, self._parse_options)
+        return self._handleResult(pctxt, result)
+
+cdef HTMLParser __DEFAULT_HTML_PARSER
+__DEFAULT_HTML_PARSER = HTMLParser()
+
+############################################################
+## helper functions for document creation
+############################################################
+
+cdef xmlDoc* _parseDoc(text_utf, parser) except NULL:
+    if parser is None:
+        parser = __DEFAULT_PARSER
+    __GLOBAL_PARSER_CONTEXT._initParser()
+    if isinstance(parser, XMLParser):
+        return (<XMLParser>parser)._parseDoc(text_utf)
+    elif isinstance(parser, HTMLParser):
+        return (<HTMLParser>parser)._parseDoc(text_utf)
+    else:
+        raise TypeError, "invalid parser"
 
-        if self._c_dict is NULL:
-            # we need to get dict from the new document if it's there,
-            # otherwise make one
-            if result.dict is not NULL:
-                d = result.dict
-            else:
-                d = xmlparser.xmlDictCreate()
-                result.dict = d
-            self._c_dict = d
-            xmlparser.xmlDictReference(self._c_dict)
-        else:
-            # we need to reuse the central dict and get rid of the new one
-            if result.dict is not NULL:
-                xmlparser.xmlDictFree(result.dict)
-            result.dict = self._c_dict
-            xmlparser.xmlDictReference(result.dict)
-        return result
+cdef xmlDoc* _parseDocFromFile(filename, parser) except NULL:
+    if parser is None:
+        parser = __DEFAULT_PARSER
+    __GLOBAL_PARSER_CONTEXT._initParser()
+    if isinstance(parser, XMLParser):
+        return (<XMLParser>parser)._parseDocFromFile(_cstr(filename))
+    elif isinstance(parser, HTMLParser):
+        return (<HTMLParser>parser)._parseDocFromFile(_cstr(filename))
+    else:
+        raise TypeError, "invalid parser"
+
+cdef xmlDoc* _newDoc():
+    cdef xmlDoc* result
+    result = tree.xmlNewDoc("1.0")
+    __GLOBAL_PARSER_CONTEXT._initDocDict(result)
+    return result

Modified: lxml/trunk/src/lxml/xmlparser.pxd
==============================================================================
--- lxml/trunk/src/lxml/xmlparser.pxd	(original)
+++ lxml/trunk/src/lxml/xmlparser.pxd	Thu Apr 20 12:12:03 2006
@@ -30,15 +30,15 @@
         XML_PARSE_NSCLEAN = 8192 # remove redundant namespaces declarations
         XML_PARSE_NOCDATA = 16384 # merge CDATA as text nodes
         XML_PARSE_NOXINCNODE = 32768 # do not generate XINCLUDE START/END nodes
+# libxml2 2.6.21+ only:
+#        XML_PARSE_COMPACT = 65536 # compact small text nodes
        
     cdef void xmlInitParser()
-    cdef xmlParserCtxt* xmlCreateDocParserCtxt(char* cur)
     cdef xmlParserCtxt* xmlNewParserCtxt()
     cdef void xmlFreeParserCtxt(xmlParserCtxt* ctxt)
- 
-    cdef int xmlCtxtUseOptions(xmlParserCtxt* ctxt, int options)
-    cdef int xmlParseDocument(xmlParserCtxt* ctxt)
-    cdef xmlDoc* xmlParseDoc(char* cur)
+
+    cdef xmlDoc* xmlCtxtReadDoc(xmlParserCtxt* ctxt,
+                                char* cur, char* URL, char* encoding,
+                                int options)
     cdef xmlDoc* xmlCtxtReadFile(xmlParserCtxt* ctxt,
                                  char* filename, char* encoding, int options)
-    
From scoder at codespeak.net  Thu Apr 20 12:21:40 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 12:21:41 2006
Subject: [Lxml-checkins] r26029 - in lxml/trunk/src/lxml: . tests
Message-ID: <20060420102140.6A99D1008D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 12:21:39 2006
New Revision: 26029

Modified:
   lxml/trunk/src/lxml/parser.pxi
   lxml/trunk/src/lxml/tests/test_htmlparser.py
Log:
removed HTMLSyntaxError, replaced by XMLSyntaxError to simplify catching exception e.g. from the default parser

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Thu Apr 20 12:21:39 2006
@@ -7,9 +7,6 @@
 class XMLSyntaxError(LxmlSyntaxError):
     pass
 
-class HTMLSyntaxError(LxmlSyntaxError):
-    pass
-
 class ParserError(LxmlError):
     pass
 
@@ -63,10 +60,8 @@
 
 cdef class BaseParser:
     cdef _ErrorLog _error_log
-    cdef object _syntax_error_class
-    def __init__(self, syntax_error_class):
+    def __init__(self):
         self._error_log = _ErrorLog()
-        self._syntax_error_class = syntax_error_class
 
     property error_log:
         def __get__(self):
@@ -82,7 +77,7 @@
             result = NULL
         self._error_log.disconnect()
         if result is NULL:
-            raise self._syntax_error_class
+            raise XMLSyntaxError
         return result
 
 
@@ -118,7 +113,7 @@
                  no_network=False, ns_clean=False):
         cdef int parse_options
         self._file_parser_ctxt = NULL
-        BaseParser.__init__(self, XMLSyntaxError)
+        BaseParser.__init__(self)
 
         parse_options = _XML_DEFAULT_PARSE_OPTIONS
         if dtd_validation:
@@ -237,7 +232,7 @@
         cdef int parse_options
         self._memory_parser_ctxt = NULL
         self._file_parser_ctxt   = NULL
-        BaseParser.__init__(self, HTMLSyntaxError)
+        BaseParser.__init__(self)
 
         parse_options = _HTML_DEFAULT_PARSE_OPTIONS
         if recover:

Modified: lxml/trunk/src/lxml/tests/test_htmlparser.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_htmlparser.py	(original)
+++ lxml/trunk/src/lxml/tests/test_htmlparser.py	Thu Apr 20 12:21:39 2006
@@ -29,7 +29,7 @@
         parser = self.etree.HTMLParser(recover=False)
         parse = self.etree.parse
         f = StringIO("<html></body>")
-        self.assertRaises(self.etree.HTMLSyntaxError,
+        self.assertRaises(self.etree.XMLSyntaxError,
                           parse, f, parser)
 
     def test_module_HTML_broken(self):
From scoder at codespeak.net  Thu Apr 20 12:28:54 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 12:28:55 2006
Subject: [Lxml-checkins] r26030 - lxml/trunk/src/lxml
Message-ID: <20060420102854.EA37C10088@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 12:28:53 2006
New Revision: 26030

Modified:
   lxml/trunk/src/lxml/parser.pxi
Log:
clean up left over after merge

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Thu Apr 20 12:28:53 2006
@@ -129,10 +129,6 @@
 
         self._parse_options = parse_options
 
-    property error_log:
-        def __get__(self):
-            return __build_error_log_tuple(self)
-
     def __dealloc__(self):
         if self._file_parser_ctxt != NULL:
             xmlparser.xmlFreeParserCtxt(self._file_parser_ctxt)
From scoder at codespeak.net  Thu Apr 20 12:34:25 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 12:34:26 2006
Subject: [Lxml-checkins] r26031 - lxml/trunk
Message-ID: <20060420103425.0257A1008D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 12:34:25 2006
New Revision: 26031

Modified:
   lxml/trunk/CHANGES.txt
Log:
updated CHANGES.txt

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Thu Apr 20 12:34:25 2006
@@ -7,6 +7,8 @@
 Features added
 --------------
 
+* HTMLParser for parsing (broken) HTML
+
 * XMLDTDID function parses XML into tuple (root node, ID dict) based on xml:id
   implementation of libxml2 (as opposed to ET compatible XMLID)
 
From scoder at codespeak.net  Thu Apr 20 12:47:35 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 12:47:35 2006
Subject: [Lxml-checkins] r26032 - lxml/trunk
Message-ID: <20060420104735.444B910090@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 12:47:34 2006
New Revision: 26032

Modified:
   lxml/trunk/INSTALL.txt
Log:
state that newer-than-required versions of libxml2/xslt are 'recommended'

Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Thu Apr 20 12:47:34 2006
@@ -8,10 +8,10 @@
 
 You need libxml2 and libxslt, in particular:
 
-* libxml 2.6.16 (newer versions should work). It can be found here:
+* libxml 2.6.16 (newer versions are recommended). It can be found here:
   http://xmlsoft.org/downloads.html
 
-* libxslt 1.1.12 (newer versions should work). It can be found here:
+* libxslt 1.1.12 (newer versions are recommended). It can be found here:
   http://xmlsoft.org/XSLT/downloads.html
 
 For Windows, there is a `binary distribution`_ of libxml2 and libxslt.  Note
From scoder at codespeak.net  Thu Apr 20 12:57:16 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 12:57:18 2006
Subject: [Lxml-checkins] r26033 - lxml/trunk/src/lxml
Message-ID: <20060420105716.F08FB10090@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 12:57:16 2006
New Revision: 26033

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/xmlid.pxi
Log:
doc updates, make clear that trees that have ID dictionaries must not be modified

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 20 12:57:16 2006
@@ -1272,7 +1272,8 @@
     return result
 
 def parse(source, parser=None):
-    """Return an ElementTree object loaded with source elements
+    """Return an ElementTree object loaded with source elements.  If no parser
+    is provided as second argument, the default parser is used.
     """
     cdef _Document doc
     doc = _parseDocument(source, parser)

Modified: lxml/trunk/src/lxml/xmlid.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlid.pxi	(original)
+++ lxml/trunk/src/lxml/xmlid.pxi	Thu Apr 20 12:57:16 2006
@@ -18,6 +18,9 @@
     contains string-element pairs.  The dictionary keys are the values of ID
     attributes as defined by the DTD.  The elements referenced by the ID are
     stored as dictionary values.
+
+    Note that you must not modify the XML tree if you use the ID dictionary.
+    The results are undefined.
     """
     cdef _NodeBase root
     root = XML(text)
From scoder at codespeak.net  Thu Apr 20 13:05:45 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 13:05:47 2006
Subject: [Lxml-checkins] r26034 - in lxml/trunk/src/lxml: . tests
Message-ID: <20060420110545.AD65310088@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 13:05:44 2006
New Revision: 26034

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/tests/test_etree.py
Log:
new module level function parseid() as parse() equivalent to XMLDTDID()/XML(), cleanup in related test cases

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 20 13:05:44 2006
@@ -1279,6 +1279,18 @@
     doc = _parseDocument(source, parser)
     return ElementTree(doc.getroot())
 
+def parseid(source, parser=None):
+    """Parses the source into a tuple containing an ElementTree object and an
+    ID dictionary.  If no parser is provided as second argument, the default
+    parser is used.
+
+    Note that you must not modify the XML tree if you use the ID dictionary.
+    The results are undefined.
+    """
+    cdef _Document doc
+    doc = _parseDocument(source, parser)
+    return (ElementTree(doc.getroot()), _IDDict(doc))
+
 
 # include submodules
 include "xmlerror.pxi"  # error and log handling

Modified: lxml/trunk/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_etree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_etree.py	Thu Apr 20 13:05:44 2006
@@ -140,9 +140,44 @@
             b,
             d.getparent())
 
+    def test_parseid(self):
+        parseid = self.etree.parseid
+        XML     = self.etree.XML
+        xml_text = '''
+        <!DOCTYPE document [
+        <!ELEMENT document (h1,p)*>
+        <!ELEMENT h1 (#PCDATA)>
+        <!ATTLIST h1 myid ID #REQUIRED>
+        <!ELEMENT p  (#PCDATA)>
+        <!ATTLIST p  someid ID #REQUIRED>
+        ]>
+        <document>
+          <h1 myid="chapter1">...</h1>
+          <p id="note1" class="note">...</p>
+          <p>Regular paragraph.</p>
+          <p xml:id="xmlid">XML:ID paragraph.</p>
+          <p someid="warn1" class="warning">...</p>
+        </document>
+        '''
+
+        tree, dic = parseid(StringIO(xml_text))
+        root = tree.getroot()
+        root2 = XML(xml_text)
+        self.assertEquals(self._writeElement(root),
+                          self._writeElement(root2))
+        expected = {
+            "chapter1" : root[0],
+            "xmlid"    : root[3],
+            "warn1"    : root[4]
+            }
+        self.assert_("chapter1" in dic)
+        self.assert_("warn1" in dic)
+        self.assert_("xmlid" in dic)
+        self._checkIDDict(dic, expected)
+
     def test_XMLDTDID(self):
-        XMLDTDID = etree.XMLDTDID
-        XML      = etree.XML
+        XMLDTDID = self.etree.XMLDTDID
+        XML      = self.etree.XML
         xml_text = '''
         <!DOCTYPE document [
         <!ELEMENT document (h1,p)*>
@@ -169,29 +204,14 @@
             "xmlid"    : root[3],
             "warn1"    : root[4]
             }
-
-        self.assertEquals(dic, expected)
-        self.assertEquals(len(dic),
-                          len(expected))
-        self.assertEquals(sorted(dic.items()),
-                          sorted(expected.items()))
-        self.assertEquals(sorted(dic.iteritems()),
-                          sorted(expected.iteritems()))
-        self.assertEquals(sorted(dic.keys()),
-                          sorted(expected.keys()))
-        self.assertEquals(sorted(dic.iterkeys()),
-                          sorted(expected.iterkeys()))
-        self.assertEquals(sorted(dic.values()),
-                          sorted(expected.values()))
-        self.assertEquals(sorted(dic.itervalues()),
-                          sorted(expected.itervalues()))
         self.assert_("chapter1" in dic)
         self.assert_("warn1" in dic)
         self.assert_("xmlid" in dic)
+        self._checkIDDict(dic, expected)
 
     def test_XMLDTDID_empty(self):
-        XMLDTDID = etree.XMLDTDID
-        XML      = etree.XML
+        XMLDTDID = self.etree.XMLDTDID
+        XML      = self.etree.XML
         xml_text = '''
         <document>
           <h1 myid="chapter1">...</h1>
@@ -206,14 +226,24 @@
         self.assertEquals(self._writeElement(root),
                           self._writeElement(root2))
         expected = {}
+        self._checkIDDict(dic, expected)
 
+    def _checkIDDict(self, dic, expected):
         self.assertEquals(dic, expected)
-        self.assertEquals(dic.items(),
-                          expected.items())
+        self.assertEquals(len(dic),
+                          len(expected))
+        self.assertEquals(sorted(dic.items()),
+                          sorted(expected.items()))
         self.assertEquals(sorted(dic.iteritems()),
                           sorted(expected.iteritems()))
-        self.assertEquals(dic.keys(),
-                          expected.keys())
+        self.assertEquals(sorted(dic.keys()),
+                          sorted(expected.keys()))
+        self.assertEquals(sorted(dic.iterkeys()),
+                          sorted(expected.iterkeys()))
+        self.assertEquals(sorted(dic.values()),
+                          sorted(expected.values()))
+        self.assertEquals(sorted(dic.itervalues()),
+                          sorted(expected.itervalues()))
 
     def test_namespaces(self):
         etree = self.etree
From scoder at codespeak.net  Thu Apr 20 13:08:56 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 13:08:58 2006
Subject: [Lxml-checkins] r26035 - lxml/trunk/src/lxml
Message-ID: <20060420110856.EFBA510088@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 13:08:56 2006
New Revision: 26035

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
removed unused helper function _raiseIfNone

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 20 13:08:56 2006
@@ -413,7 +413,8 @@
             return
         # if the next element is in the list, insert before it
         for mynode in value:
-            _raiseIfNone(mynode)
+            if mynode is None:
+                raise TypeError, "Node must not be None."
             foreign = self._doc is not mynode._doc
             # store possible text tail
             c_next = mynode._c_node.next
@@ -1238,7 +1239,7 @@
 
 def dump(_NodeBase elem):
     assert elem is not None, "Must supply element."
-    # better, but not ET compatible : _raiseIfNone(elem)
+    # better, but not ET compatible : "_NodeBase elem not None"
     _dumpToFile(sys.stdout, elem._doc._c_doc, elem._c_node)
 
 def tostring(_NodeBase element, encoding='us-ascii'):
@@ -1248,7 +1249,7 @@
     cdef char* enc
 
     assert element is not None
-    # better, but not ET compatible : _raiseIfNone(element)
+    # better, but not ET compatible : "_NodeBase element not None"
     
     #if encoding is None:
     #    encoding = 'UTF-8'
@@ -1304,10 +1305,6 @@
 
 
 # Private helper functions
-cdef void _raiseIfNone(el):
-    if el is None:
-        raise TypeError, "Argument must not be None."
-
 cdef _Document _documentOrRaise(object input):
     cdef _Document doc
     doc = _documentOf(input)
From scoder at codespeak.net  Thu Apr 20 13:19:25 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 13:19:26 2006
Subject: [Lxml-checkins] r26036 - lxml/trunk/src/lxml
Message-ID: <20060420111925.84BC710088@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 13:19:24 2006
New Revision: 26036

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/xmlid.pxi
Log:
moved parseid() function into xmlid.pxi

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 20 13:19:24 2006
@@ -1280,18 +1280,6 @@
     doc = _parseDocument(source, parser)
     return ElementTree(doc.getroot())
 
-def parseid(source, parser=None):
-    """Parses the source into a tuple containing an ElementTree object and an
-    ID dictionary.  If no parser is provided as second argument, the default
-    parser is used.
-
-    Note that you must not modify the XML tree if you use the ID dictionary.
-    The results are undefined.
-    """
-    cdef _Document doc
-    doc = _parseDocument(source, parser)
-    return (ElementTree(doc.getroot()), _IDDict(doc))
-
 
 # include submodules
 include "xmlerror.pxi"  # error and log handling

Modified: lxml/trunk/src/lxml/xmlid.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlid.pxi	(original)
+++ lxml/trunk/src/lxml/xmlid.pxi	Thu Apr 20 13:19:24 2006
@@ -30,6 +30,18 @@
     else:
         return (root, _IDDict(root))
 
+def parseid(source, parser=None):
+    """Parses the source into a tuple containing an ElementTree object and an
+    ID dictionary.  If no parser is provided as second argument, the default
+    parser is used.
+
+    Note that you must not modify the XML tree if you use the ID dictionary.
+    The results are undefined.
+    """
+    cdef _Document doc
+    doc = _parseDocument(source, parser)
+    return (ElementTree(doc.getroot()), _IDDict(doc))
+
 cdef class _IDDict:
     """A dictionary-like proxy class that mapps ID attributes to elements.
 
From scoder at codespeak.net  Thu Apr 20 14:39:21 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 14:39:22 2006
Subject: [Lxml-checkins] r26039 - lxml/branch/resolver-new
Message-ID: <20060420123921.0D1F61008D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 14:39:19 2006
New Revision: 26039

Added:
   lxml/branch/resolver-new/
      - copied from r26038, lxml/trunk/
Log:
new branch for resolvers (2nd try)
From scoder at codespeak.net  Thu Apr 20 18:24:27 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 18:24:29 2006
Subject: [Lxml-checkins] r26055 - in lxml/branch/resolver-new/src/lxml: .
	tests
Message-ID: <20060420162427.7703D10088@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 18:24:25 2006
New Revision: 26055

Modified:
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/tests/test_etree.py
   lxml/branch/resolver-new/src/lxml/xmlparser.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
initial working implementation of parser entity resolvers

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Thu Apr 20 18:24:25 2006
@@ -1281,6 +1281,19 @@
     return ElementTree(doc.getroot())
 
 
+# class for temporary storage of Python references
+cdef class _TempStore:
+    cdef object _storage
+    def __init__(self):
+        self._storage = {}
+
+    cdef void add(self, obj):
+        python.PyDict_SetItem(self._storage, id(obj), obj)
+
+    cdef void clear(self):
+        python.PyDict_Clear(self._storage)
+
+
 # include submodules
 include "xmlerror.pxi"  # error and log handling
 include "xmlid.pxi"     # XMLID and IDDict

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Thu Apr 20 18:24:25 2006
@@ -58,15 +58,156 @@
 __GLOBAL_PARSER_CONTEXT = _ParserContext()
 
 
+############################################################
+## Custom resolver API
+############################################################
+
+cdef class _ResolverRegistry # forward declaration
+
+cdef class _ParserInput:
+    cdef xmlparser.xmlParserInput* _input
+    cdef object _pyref # to keep Python references
+
+cdef class _ResolverContext:
+    cdef xmlparser.xmlParserCtxt* _ctxt
+    cdef _ResolverRegistry _resolvers
+    cdef _TempStore _storage
+
+cdef class Resolver:
+    def resolve(self, system_url, public_id, _ResolverContext context not None):
+        cdef _ParserInput parser_input
+        cdef char* c_url
+        cdef char* c_id
+        if __DEFAULT_ENTITY_LOADER is NULL:
+            return None
+        if system_url is None:
+            c_url = NULL
+        else:
+            url_utf = _utf8(system_url)
+            c_url = _cstr(url_utf)
+        if public_id is None:
+            c_id = NULL
+        else:
+            id_utf = _utf8(public_id)
+            c_id = _cstr(id_utf)
+        parser_input = _ParserInput()
+        parser_input._input = __DEFAULT_ENTITY_LOADER(
+            c_url, c_id, context._ctxt)
+        return parser_input
+
+    def resolve_string(self, string, _ResolverContext context not None):
+        cdef _ParserInput parser_input
+        string_utf = _utf8(string)
+        parser_input = _ParserInput()
+        parser_input._input = xmlparser.xmlNewStringInputStream(
+            context._ctxt, _cstr(string_utf))
+        parser_input._pyref = string_utf
+        return parser_input
+
+    def resolve_filename(self, filename, _ResolverContext context not None):
+        cdef _ParserInput parser_input
+        filename_utf = _utf8(filename)
+        parser_input = _ParserInput()
+        parser_input._input = xmlparser.xmlNewInputFromFile(
+            context._ctxt, _cstr(filename_utf))
+        return parser_input
+
+cdef class _ResolverRegistry:
+    cdef object _resolvers
+    cdef Resolver _default_resolver
+    def __init__(self, Resolver default_resolver=None):
+        try:
+            self._resolvers = set()
+        except NameError:
+            from sets import Set
+            self._resolvers = Set()
+        if default_resolver is None:
+            self._default_resolver = Resolver()
+        else:
+            self._default_resolver = default_resolver
+
+    def add(self, Resolver resolver not None):
+        """Register a resolver.
+
+        For each requested entity, the 'resolve' method of the resolver will
+        be called and the result will be passed to the parser.  If this method
+        returns None, the request will be delegated to other resolvers or the
+        default resolver.  The resolvers will be tested in an arbitrary order
+        until the first match is found.
+        """
+        self._resolvers.add(resolver)
+
+    def remove(self, resolver):
+        self._resolvers.discard(resolver)
+
+    def resolve(self, system_url, public_id, _ResolverContext context not None):
+        for resolver in self._resolvers:
+            result = resolver.resolve(system_url, public_id, context)
+            if result is not None:
+                return result
+        return self._default_resolver.resolve(system_url, public_id, context)
+
+cdef xmlparser.xmlParserInput* _local_resolver(char* c_url, char* c_pubid,
+                                               xmlParserCtxt* c_context):
+    cdef _ResolverContext context
+    cdef _ParserInput     parser_input
+    if c_context._private is NULL:
+        return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
+
+    if c_url is NULL:
+        url = None
+    else:
+        url = funicode(c_url)
+    if c_pubid is NULL:
+        pubid = None
+    else:
+        pubid = funicode(c_pubid)
+
+    context = <_ResolverContext>c_context._private
+    try:
+        parser_input = context._resolvers.resolve(url, pubid, context)
+    except Exception, e:
+        print e
+    if parser_input is None:
+        return NULL
+    context._storage.add(parser_input)
+    return parser_input._input
+
+cdef xmlparser.xmlExternalEntityLoader __DEFAULT_ENTITY_LOADER
+__DEFAULT_ENTITY_LOADER = xmlparser.xmlGetExternalEntityLoader()
+
+xmlparser.xmlSetExternalEntityLoader(_local_resolver)
+
+############################################################
+## Parsers
+############################################################
+
 cdef class BaseParser:
     cdef _ErrorLog _error_log
+    cdef readonly object resolvers
+    cdef _ResolverContext _context
     def __init__(self):
         self._error_log = _ErrorLog()
+        self.resolvers = _ResolverRegistry()
+        self._context = None
 
     property error_log:
         def __get__(self):
             return self._error_log.copy()
 
+    cdef _initContext(self, xmlParserCtxt* c_ctxt):
+        cdef _ResolverContext context
+        __GLOBAL_PARSER_CONTEXT._initParserDict(c_ctxt)
+        context = _ResolverContext()
+        context._ctxt = c_ctxt
+        context._resolvers = self.resolvers
+        context._storage = _TempStore()
+        self._context = context
+        c_ctxt._private = <python.PyObject*>context
+
+    cdef _clearContext(self):
+        self._context = None
+
     cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,
                                xmlDoc* result) except NULL:
         if ctxt.wellFormed:
@@ -80,7 +221,6 @@
             raise XMLSyntaxError
         return result
 
-
 ############################################################
 ## XML parser
 ############################################################
@@ -154,10 +294,10 @@
         if pctxt is NULL:
             pctxt = self._createContext()
             self._memory_parser_ctxt = pctxt
-
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadDoc(
             pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
+        self._clearContext()
         return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
@@ -168,10 +308,10 @@
         if pctxt is NULL:
             pctxt = self._createContext()
             self._file_parser_ctxt = pctxt
-
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
+        self._clearContext()
         if result is NULL:
             if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
                 self._error_log.disconnect()
@@ -203,6 +343,8 @@
     else:
         raise TypeError, "Invalid parser"
 
+def get_default_parser():
+    return __DEFAULT_PARSER
 
 ############################################################
 ## HTML parser
@@ -264,9 +406,10 @@
                 self._error_log.disconnect()
                 raise ParserError, "Failed to create parser context"
             self._memory_parser_ctxt = pctxt
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadDoc(
             pctxt, c_text, NULL, NULL, self._parse_options)
+        self._clearContext()
         return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
@@ -284,9 +427,10 @@
                     raise IOError, "Could not open file %s" % filename
                 raise ParserError, "Failed to create parser context"
             self._file_parser_ctxt = pctxt
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
+        self._clearContext()
         return self._handleResult(pctxt, result)
 
 cdef HTMLParser __DEFAULT_HTML_PARSER

Modified: lxml/branch/resolver-new/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/resolver-new/src/lxml/tests/test_etree.py	Thu Apr 20 18:24:25 2006
@@ -49,7 +49,26 @@
         f = open(fileInTestDir('test_broken.xml'), 'r')
         self.assertRaises(SyntaxError, parse, f)
         f.close()
-        
+
+    def test_resolve_string_dtd(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(dtd_validation=True)
+        assertEqual = self.assertEqual
+        test_url = u"__nosuch.dtd"
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                assertEqual(url, test_url)
+                return self.resolve_string(
+                    u'<!ENTITY myentity "%s">' % url, context)
+
+        parser.resolvers.add(MyResolver())
+
+        xml = u'<!DOCTYPE doc SYSTEM "%s"><doc>&myentity;</doc>' % test_url
+        tree = parse(StringIO(xml), parser)
+        root = tree.getroot()
+        self.assertEquals(root.text, test_url)
+
     # TypeError in etree, AssertionError in ElementTree;
     def test_setitem_assert(self):
         Element = self.etree.Element

Modified: lxml/branch/resolver-new/src/lxml/xmlparser.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xmlparser.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xmlparser.pxd	Thu Apr 20 18:24:25 2006
@@ -1,6 +1,9 @@
 from tree cimport xmlDoc, xmlDict
 from xmlerror cimport xmlError
 
+cdef extern from "libxml/tree.h":
+    ctypedef struct xmlParserInput
+
 cdef extern from "libxml/parser.h":
 
     cdef xmlDict* xmlDictCreate()
@@ -10,6 +13,7 @@
     ctypedef struct xmlParserCtxt:
         xmlDoc* myDoc
         xmlDict* dict
+        void* _private
         int wellFormed
         xmlError lastError
         
@@ -42,3 +46,17 @@
                                 int options)
     cdef xmlDoc* xmlCtxtReadFile(xmlParserCtxt* ctxt,
                                  char* filename, char* encoding, int options)
+
+    # entity loaders
+
+    ctypedef xmlParserInput* (*xmlExternalEntityLoader)(char * URL,
+                                                        char * ID, 
+                                                        xmlParserCtxt* context)
+    cdef xmlExternalEntityLoader xmlGetExternalEntityLoader()
+    cdef void xmlSetExternalEntityLoader(xmlExternalEntityLoader f)
+
+cdef extern from "libxml/parserInternals.h":
+    cdef xmlParserInput* xmlNewStringInputStream(xmlParserCtxt* ctxt, 
+						 char* buffer)
+    cdef xmlParserInput* xmlNewInputFromFile(xmlParserCtxt* ctxt, 
+                                             char* filename)

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Thu Apr 20 18:24:25 2006
@@ -40,8 +40,8 @@
     cdef object _extension_functions
     cdef object _utf_refs
     # for exception handling and temporary reference keeping:
-    cdef object _temp_elements
-    cdef object _temp_docs
+    cdef _TempStore _temp_elements
+    cdef _TempStore _temp_docs
     cdef object _exc_info
 
     def __init__(self, namespaces, extensions):
@@ -68,8 +68,8 @@
         self._registered_namespaces = []
         self._registered_extensions = []
         self._extension_functions = {}
-        self._temp_elements = {}
-        self._temp_docs = {}
+        self._temp_elements = _TempStore()
+        self._temp_docs     = _TempStore()
 
     cdef object _to_utf(self, s):
         "Convert to UTF-8 and keep a reference to the encoded string"
@@ -174,8 +174,8 @@
 
     cdef _release_temp_refs(self):
         "Free temporarily referenced objects from this context."
-        python.PyDict_Clear(self._temp_elements)
-        python.PyDict_Clear(self._temp_docs)
+        self._temp_elements.clear()
+        self._temp_docs.clear()
         
     cdef _hold(self, obj):
         """A way to temporarily hold references to nodes in the evaluator.
@@ -193,9 +193,9 @@
             if isinstance(o, _NodeBase):
                 element = <_NodeBase>o
                 #print "Holding element:", <int>element._c_node
-                python.PyDict_SetItem(self._temp_elements, id(element), element)
+                self._temp_elements.add(element)
                 #print "Holding document:", <int>element._doc._c_doc
-                python.PyDict_SetItem(self._temp_docs, id(element._doc), element._doc)
+                self._temp_docs.add(element._doc)
 
 
 ################################################################################
From scoder at codespeak.net  Thu Apr 20 18:51:19 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 18:51:20 2006
Subject: [Lxml-checkins] r26060 - in lxml/branch/resolver-new/src/lxml: .
	tests
Message-ID: <20060420165119.A945710073@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 18:51:18 2006
New Revision: 26060

Modified:
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/tests/test_etree.py
Log:
make exception raising work from within resolvers

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Thu Apr 20 18:51:18 2006
@@ -72,6 +72,16 @@
     cdef xmlparser.xmlParserCtxt* _ctxt
     cdef _ResolverRegistry _resolvers
     cdef _TempStore _storage
+    cdef object _exc_info
+
+    cdef _store_exception(self, e):
+        self._exc_info = sys.exc_info()
+
+    cdef _check_exception(self):
+        _exc_info = self._exc_info
+        if _exc_info is not None:
+            type, value, traceback = _exc_info
+            raise type, value, traceback
 
 cdef class Resolver:
     def resolve(self, system_url, public_id, _ResolverContext context not None):
@@ -154,20 +164,22 @@
     if c_context._private is NULL:
         return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
 
-    if c_url is NULL:
-        url = None
-    else:
-        url = funicode(c_url)
-    if c_pubid is NULL:
-        pubid = None
-    else:
-        pubid = funicode(c_pubid)
-
-    context = <_ResolverContext>c_context._private
     try:
+        if c_url is NULL:
+            url = None
+        else:
+            url = funicode(c_url)
+        if c_pubid is NULL:
+            pubid = None
+        else:
+            pubid = funicode(c_pubid)
+
+        context = <_ResolverContext>c_context._private
+
         parser_input = context._resolvers.resolve(url, pubid, context)
     except Exception, e:
-        print e
+        context._store_exception(e)
+        return NULL
     if parser_input is None:
         return NULL
     context._storage.add(parser_input)
@@ -206,6 +218,8 @@
         c_ctxt._private = <python.PyObject*>context
 
     cdef _clearContext(self):
+        if self._context is not None:
+            self._context._check_exception()
         self._context = None
 
     cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,

Modified: lxml/branch/resolver-new/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/resolver-new/src/lxml/tests/test_etree.py	Thu Apr 20 18:51:18 2006
@@ -69,6 +69,23 @@
         root = tree.getroot()
         self.assertEquals(root.text, test_url)
 
+    def test_resolve_error(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(dtd_validation=True)
+        test_url = u"__nosuch.dtd"
+
+        class _LocalException(Exception):
+            pass
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                raise _LocalException
+
+        parser.resolvers.add(MyResolver())
+
+        xml = u'<!DOCTYPE doc SYSTEM "test"><doc>&myentity;</doc>'
+        self.assertRaises(_LocalException, parse, StringIO(xml), parser)
+
     # TypeError in etree, AssertionError in ElementTree;
     def test_setitem_assert(self):
         Element = self.etree.Element
From scoder at codespeak.net  Thu Apr 20 21:13:37 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 21:13:38 2006
Subject: [Lxml-checkins] r26063 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060420191337.452A910080@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 21:13:36 2006
New Revision: 26063

Modified:
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/xmlparser.pxd
Log:
some clean up, free input stream in case of exceptions

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Thu Apr 20 21:13:36 2006
@@ -179,6 +179,8 @@
         parser_input = context._resolvers.resolve(url, pubid, context)
     except Exception, e:
         context._store_exception(e)
+        if parser_input is not None and parser_input._input is not NULL:
+            xmlparser.xmlFreeInputStream(parser_input._input)
         return NULL
     if parser_input is None:
         return NULL
@@ -217,7 +219,7 @@
         self._context = context
         c_ctxt._private = <python.PyObject*>context
 
-    cdef _clearContext(self):
+    cdef _cleanupContext(self):
         if self._context is not None:
             self._context._check_exception()
         self._context = None
@@ -311,7 +313,7 @@
         self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadDoc(
             pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
-        self._clearContext()
+        self._cleanupContext()
         return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
@@ -325,7 +327,7 @@
         self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
-        self._clearContext()
+        self._cleanupContext()
         if result is NULL:
             if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
                 self._error_log.disconnect()
@@ -423,7 +425,7 @@
         self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadDoc(
             pctxt, c_text, NULL, NULL, self._parse_options)
-        self._clearContext()
+        self._cleanupContext()
         return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
@@ -444,7 +446,7 @@
         self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
-        self._clearContext()
+        self._cleanupContext()
         return self._handleResult(pctxt, result)
 
 cdef HTMLParser __DEFAULT_HTML_PARSER

Modified: lxml/branch/resolver-new/src/lxml/xmlparser.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xmlparser.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xmlparser.pxd	Thu Apr 20 21:13:36 2006
@@ -34,8 +34,8 @@
         XML_PARSE_NSCLEAN = 8192 # remove redundant namespaces declarations
         XML_PARSE_NOCDATA = 16384 # merge CDATA as text nodes
         XML_PARSE_NOXINCNODE = 32768 # do not generate XINCLUDE START/END nodes
-# libxml2 2.6.21+ only:
-#        XML_PARSE_COMPACT = 65536 # compact small text nodes
+        # libxml2 2.6.21+ only:
+        #XML_PARSE_COMPACT = 65536 # compact small text nodes
        
     cdef void xmlInitParser()
     cdef xmlParserCtxt* xmlNewParserCtxt()
@@ -47,7 +47,7 @@
     cdef xmlDoc* xmlCtxtReadFile(xmlParserCtxt* ctxt,
                                  char* filename, char* encoding, int options)
 
-    # entity loaders
+# entity loaders:
 
     ctypedef xmlParserInput* (*xmlExternalEntityLoader)(char * URL,
                                                         char * ID, 
@@ -60,3 +60,4 @@
 						 char* buffer)
     cdef xmlParserInput* xmlNewInputFromFile(xmlParserCtxt* ctxt, 
                                              char* filename)
+    cdef void xmlFreeInputStream(xmlParserInput* input)
From scoder at codespeak.net  Thu Apr 20 21:24:46 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 20 21:24:47 2006
Subject: [Lxml-checkins] r26064 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060420192446.6B08E10080@code0.codespeak.net>

Author: scoder
Date: Thu Apr 20 21:24:45 2006
New Revision: 26064

Modified:
   lxml/branch/resolver-new/src/lxml/parser.pxi
Log:
removed some race conditions related to exception handling

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Thu Apr 20 21:24:45 2006
@@ -219,7 +219,7 @@
         self._context = context
         c_ctxt._private = <python.PyObject*>context
 
-    cdef _cleanupContext(self):
+    cdef _cleanupContextAndRaisePending(self):
         if self._context is not None:
             self._context._check_exception()
         self._context = None
@@ -232,7 +232,6 @@
             # free broken document
             tree.xmlFreeDoc(result)
             result = NULL
-        self._error_log.disconnect()
         if result is NULL:
             raise XMLSyntaxError
         return result
@@ -313,7 +312,8 @@
         self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadDoc(
             pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
-        self._cleanupContext()
+        self._error_log.disconnect()
+        self._cleanupContextAndRaisePending()
         return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
@@ -327,10 +327,10 @@
         self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
-        self._cleanupContext()
+        self._error_log.disconnect()
+        self._cleanupContextAndRaisePending()
         if result is NULL:
             if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
-                self._error_log.disconnect()
                 raise IOError, "Could not open file %s" % filename
         return self._handleResult(pctxt, result)
 
@@ -425,7 +425,8 @@
         self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadDoc(
             pctxt, c_text, NULL, NULL, self._parse_options)
-        self._cleanupContext()
+        self._error_log.disconnect()
+        self._cleanupContextAndRaisePending()
         return self._handleResult(pctxt, result)
 
     cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
@@ -446,7 +447,8 @@
         self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadFile(
             pctxt, filename, NULL, self._parse_options)
-        self._cleanupContext()
+        self._error_log.disconnect()
+        self._cleanupContextAndRaisePending()
         return self._handleResult(pctxt, result)
 
 cdef HTMLParser __DEFAULT_HTML_PARSER
From scoder at codespeak.net  Fri Apr 21 11:07:44 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 11:07:46 2006
Subject: [Lxml-checkins] r26078 - lxml/trunk
Message-ID: <20060421090744.E5ED71007A@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 11:07:44 2006
New Revision: 26078

Modified:
   lxml/trunk/INSTALL.txt
Log:
slight clarification

Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Fri Apr 21 11:07:44 2006
@@ -32,9 +32,10 @@
 
 .. _Pyrex: http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/
 
-Note that Pyrex up to version 0.9.4 has known problems when compiling lxml
-with gcc 4.0 or Python 2.4.  Do not use it.  If you want to build lxml from
-non-release sources, please install Pyrex version 0.9.4.1 or later.
+Note that Pyrex up to and including version 0.9.4 has known problems when
+compiling lxml with gcc 4.0 or Python 2.4.  Do not use it.  If you want to
+build lxml from non-release sources, please install Pyrex version 0.9.4.1 or
+later.
 
 If you have read these instructions and still cannot manage to install lxml,
 you can check the archives of the `mailing list`_ to see if your problem is
From scoder at codespeak.net  Fri Apr 21 11:12:47 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 11:12:49 2006
Subject: [Lxml-checkins] r26079 - lxml/branch/lxml-0.9.x
Message-ID: <20060421091247.8C54F1007B@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 11:12:46 2006
New Revision: 26079

Modified:
   lxml/branch/lxml-0.9.x/INSTALL.txt
Log:
merged updated INSTALL.txt from trunk

Modified: lxml/branch/lxml-0.9.x/INSTALL.txt
==============================================================================
--- lxml/branch/lxml-0.9.x/INSTALL.txt	(original)
+++ lxml/branch/lxml-0.9.x/INSTALL.txt	Fri Apr 21 11:12:46 2006
@@ -8,15 +8,23 @@
 
 You need libxml2 and libxslt, in particular:
 
-* libxml 2.6.16 (newer versions should work). It can be found here:
+* libxml 2.6.16 (newer versions are recommended). It can be found here:
   http://xmlsoft.org/downloads.html
 
-* libxslt 1.1.12 (newer versions should work). It can be found here:
+* libxslt 1.1.12 (newer versions are recommended). It can be found here:
   http://xmlsoft.org/XSLT/downloads.html
 
-See below for instructions how to get these for Windows.  On MacOS-X 10.4, you
-can use the installed system libraries and the binary egg distribution of
-lxml.
+For Windows, there is a `binary distribution`_ of libxml2 and libxslt.  Note
+that you need both libxml2 and libxslt, as well as iconv and zlib.  You can
+then install the `binary egg distribution`_ of lxml (see below).
+
+.. _`binary distribution`: http://www.zlatkovic.com/libxml.en.html
+.. _`binary egg distribution`: http://cheeseshop.python.org/pypi/lxml
+
+On MacOS-X 10.4, you can use the installed system libraries and the binary egg
+distribution of lxml.  Note that the libxslt version on this system is older
+than the required version above.  While there were not any bug reports so far,
+you may still encounter certain differences in behaviour in rare cases.
 
 If you want to build lxml from SVN, you also need Pyrex_. If you are using a
 released version of lxml, it should come with the generated C file in the
@@ -24,8 +32,10 @@
 
 .. _Pyrex: http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/
 
-See also the notes on building with gcc 4.0 below if you are having
-trouble with Pyrex.
+Note that Pyrex up to and including version 0.9.4 has known problems when
+compiling lxml with gcc 4.0 or Python 2.4.  Do not use it.  If you want to
+build lxml from non-release sources, please install Pyrex version 0.9.4.1 or
+later.
 
 If you have read these instructions and still cannot manage to install lxml,
 you can check the archives of the `mailing list`_ to see if your problem is
@@ -44,8 +54,8 @@
 .. _easy_install: http://peak.telecommunity.com/DevCenter/EasyInstall
 
 This has been reported to work on Linux, MacOS-X 10.4 and Windows, as long as
-libxml2 and libxslt are installed.  To compile and install lxml without
-easy_install, download the source tar-ball, unpack it and type::
+libxml2 and libxslt are properly installed.  To compile and install lxml
+without easy_install, download the source tar-ball, unpack it and type::
 
   python setup.py install
 
@@ -62,78 +72,6 @@
 import lxml.etree and play with it.
 
 
-Installation on Windows
------------------------
-
-As always, installation on Windows is different. If you do not want to go
-through the hassle of compiling everything by hand, you can use the binary
-distribution of libxml2 and libxslt. It is available here:
-
-http://www.zlatkovic.com/libxml.en.html
-
-Note that you need both libxml2 and libxslt, as well as iconv and zlib. You
-can then download a binary version of lxml 0.9 for Python 2.4 from the
-following address:
-
-http://carcass.dhs.org/lxml-0.9.win32-py2.4.exe
-
-or the egg distribution from
-
-http://cheeseshop.python.org/pypi/lxml
-
-The egg can directly be installed using easy_install_. Both builds were kindly
-contributed by Steve Howe. If they do not work for you, feel free to report to
-the mailing list.
-
-
-Building lxml with gcc 4.0 or Python 2.4
-----------------------------------------
-
-Pyrex 0.9.3.1 generates C code that gcc 4.0 does not accept. Pending an
-official release of a version of Pyrex that does work with gcc 4.0, the lxml
-project currently provides an updated version of Pyrex in its Subversion
-repository:
-
-http://codespeak.net/svn/lxml/pyrex/
-
-To install it, you can just download one of the following files:
-
-http://codespeak.net/svn/lxml/pyrex/dist/Pyrex-0.9.3.1.tar.gz
-
-http://codespeak.net/svn/lxml/pyrex/dist/Pyrex-0.9.3.1-1.src.rpm
-
-It is based on Pyrex 0.9.3.1 and contains a number of patches that make lxml
-compile and appear to work with gcc 4.0. If you use this version, you can
-simply skip the rest of the section. In case you want to apply them yourself,
-the first one is:
-
-http://codespeak.net/lxml/Pyrex-0.9.3-gcc4.patch
-
-Some Linux distributions such as Fedora Core 4 and Ubuntu Linux may
-already have most of this applied. In that case, this smaller patch
-may be applicable to make lxml compile properly:
-
-http://codespeak.net/lxml/Pyrex-0.9.3-gcc4-small.patch
-
-It may however actually be that at the time you read this, this extra patch
-has been applied by the distributions as well. You may still encounter the
-following problem when building the extension on Python 2.4::
-
-  TypeError: swig_sources() takes exactly 2 arguments (3 given)
-
-To fix this, look for the following line in Pyrex/Distutils/build_ext.py
-(around line 35)::
-
-  def swig_sources (self, sources):
-
-and change it to::
-
-  def swig_sources (self, sources, *otherargs):
-
-The above install files have these changes applied. It should do no harm if
-you install them instead of the official Pyrex version.
-
-
 Running the tests and reporting errors
 --------------------------------------
 
@@ -146,7 +84,7 @@
 above), as it searches the "src" directory.  You can use the following
 one-step command to trigger an in-place build and test it::
 
-  make test
+  make clean test
 
 To run the ElementTree and cElementTree compatibility tests, make sure
 you have lxml on your PYTHONPATH first, then run::
From scoder at codespeak.net  Fri Apr 21 13:08:23 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 13:08:25 2006
Subject: [Lxml-checkins] r26083 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060421110823.05E581007B@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 13:08:22 2006
New Revision: 26083

Modified:
   lxml/branch/resolver-new/src/lxml/tree.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
initial implementation for replacing XSLT document loader, some debug output

Modified: lxml/branch/resolver-new/src/lxml/tree.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tree.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/tree.pxd	Fri Apr 21 13:08:22 2006
@@ -4,6 +4,7 @@
     ctypedef struct FILE
     cdef int strlen(char* s)
     cdef int strcmp(char* s1, char* s2)
+    cdef int strncmp(char* s1, char* s2, int n)
     
 cdef extern from "libxml/encoding.h":
     ctypedef struct xmlCharEncodingHandler

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxd	Fri Apr 21 13:08:22 2006
@@ -1,18 +1,16 @@
-from tree cimport xmlDoc
+from tree cimport xmlDoc, xmlDict
 from xpath cimport xmlXPathContext, xmlXPathFunction
 
 cdef extern from "libxslt/xsltInternals.h":
     ctypedef struct xsltStylesheet:
-        pass
+        xmlDoc* doc
 
     ctypedef struct xsltTransformContext:
         xmlXPathContext* xpathCtxt
+        void* _private
     
     cdef xsltStylesheet* xsltParseStylesheetDoc(xmlDoc* doc)
     cdef void xsltFreeStylesheet(xsltStylesheet* sheet)
-    
-#cdef extern from "libxslt/xslt.h":
-#    pass
 
 cdef extern from "libxslt/extensions.h":
     cdef int xsltRegisterExtFunction(xsltTransformContext* ctxt,
@@ -20,6 +18,19 @@
                                      char * URI,
                                      xmlXPathFunction function)
 
+cdef extern from "libxslt/documents.h":
+    ctypedef enum xsltLoadType:
+        XSLT_LOAD_START
+        XSLT_LOAD_STYLESHEET
+        XSLT_LOAD_DOCUMENT
+
+    ctypedef xmlDoc* (*xsltDocLoaderFunc)(char* URI, xmlDict* dict,
+                                          int options,
+                                          void* ctxt,
+                                          xsltLoadType type)
+    cdef xsltDocLoaderFunc xsltDocDefaultLoader
+    cdef void xsltSetLoaderFunc	(xsltDocLoaderFunc f)
+
 cdef extern from "libxslt/transform.h":
     cdef xmlDoc* xsltApplyStylesheet(xsltStylesheet* style, xmlDoc* doc,
                                      char** params)

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Fri Apr 21 13:08:22 2006
@@ -270,9 +270,10 @@
         # of libxml2)
         if doc._c_doc.URL is not NULL:
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
-            
+
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
+            tree.xmlFreeDoc(c_doc)
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
 
@@ -283,6 +284,9 @@
         # this cleans up copy of doc as well
         xslt.xsltFreeStylesheet(self._c_style)
 
+    cdef xmlDoc* _copyXslDoc(self):
+        return tree.xmlCopyDoc(self._c_style.doc, 1)
+
     def __call__(self, _input, **_kw):
         cdef _Document input_doc
         cdef _NodeBase root_node
@@ -303,6 +307,7 @@
         if transform_ctxt is NULL:
             _destroyFakeDoc(input_doc._c_doc, c_doc)
             raise XSLTApplyError, "Error preparing stylesheet run"
+        transform_ctxt._private = <python.PyObject*>self
 
         if _kw:
             # allocate space for parameters
@@ -375,6 +380,26 @@
 
 
 ################################################################################
+# XSLT document resolvers
+
+cdef xslt.xsltDocLoaderFunc __xsltDocDefaultLoader
+__xsltDocDefaultLoader = xslt.xsltDocDefaultLoader
+
+cdef xmlDoc* _doc_loader(char* uri, tree.xmlDict* dict, int options, void* ctxt,
+                         xslt.xsltLoadType type):
+    if tree.strncmp(uri, "__STRING__XSLT__", 16) != 0:
+        # use default loader for stylesheet
+        print uri, type, options
+        if type == xslt.XSLT_LOAD_DOCUMENT:
+            print uri
+        else:
+            print type, uri
+    return __xsltDocDefaultLoader(uri, dict, options, ctxt, type)
+
+xslt.xsltSetLoaderFunc(_doc_loader)
+
+
+################################################################################
 # XPath
 
 cdef class XPathContext(BaseContext):
From scoder at codespeak.net  Fri Apr 21 13:10:21 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 13:10:23 2006
Subject: [Lxml-checkins] r26084 - in lxml/branch/resolver-new/src/lxml: .
	tests
Message-ID: <20060421111021.BC35B1007D@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 13:10:20 2006
New Revision: 26084

Modified:
   lxml/branch/resolver-new/src/lxml/tests/test_xslt.py
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
fix document('') for XSLTs read from strings, includes test case

Modified: lxml/branch/resolver-new/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tests/test_xslt.py	(original)
+++ lxml/branch/resolver-new/src/lxml/tests/test_xslt.py	Fri Apr 21 13:10:20 2006
@@ -335,6 +335,25 @@
         self.assertEquals(self._rootstring(result),
                           '<A>X</A>')
 
+    def test_xslt_document_XML(self):
+        # make sure document('') works from loaded files
+        xslt = etree.XSLT(etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:template match="/">
+    <test><xsl:copy-of select="document('')//test"/></test>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        result = xslt(etree.XML('<a/>'))
+        root = result.getroot()
+        self.assertEquals(root.tag,
+                          'test')
+        self.assertEquals(root[0].tag,
+                          'test')
+        self.assertEquals(root[0][0].tag,
+                          '{http://www.w3.org/1999/XSL/Transform}copy-of')
+
     def test_xslt_document_parse(self):
         # make sure document('') works from loaded files
         xslt = etree.XSLT(etree.parse(fileInTestDir("test-document.xslt")))

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Fri Apr 21 13:10:20 2006
@@ -249,6 +249,7 @@
     """
     cdef XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
+    cdef object _doc_url_utf
     
     def __init__(self, xslt_input, extensions=None):
         # make a copy of the document as stylesheet needs to assume it
@@ -268,8 +269,14 @@
 
         # XXX work around bug in xmlCopyDoc (fix is upcoming in new release
         # of libxml2)
+        if c_doc.URL is not NULL and c_doc.URL != doc._c_doc.URL:
+            tree.xmlFree(c_doc.URL)
         if doc._c_doc.URL is not NULL:
+            self._doc_url_utf = doc._c_doc.URL
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
+        else:
+            self._doc_url_utf = "__STRING__XSLT__%s" % id(self)
+            c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
 
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
From scoder at codespeak.net  Fri Apr 21 13:17:53 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 13:17:55 2006
Subject: [Lxml-checkins] r26086 - in lxml/trunk/src/lxml: . tests
Message-ID: <20060421111753.DD79410082@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 13:17:52 2006
New Revision: 26086

Modified:
   lxml/trunk/src/lxml/tests/test_xslt.py
   lxml/trunk/src/lxml/xslt.pxi
Log:
merged document('') fixes from resolver-new branch

Modified: lxml/trunk/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xslt.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xslt.py	Fri Apr 21 13:17:52 2006
@@ -335,6 +335,25 @@
         self.assertEquals(self._rootstring(result),
                           '<A>X</A>')
 
+    def test_xslt_document_XML(self):
+        # make sure document('') works from loaded files
+        xslt = etree.XSLT(etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:template match="/">
+    <test><xsl:copy-of select="document('')//test"/></test>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        result = xslt(etree.XML('<a/>'))
+        root = result.getroot()
+        self.assertEquals(root.tag,
+                          'test')
+        self.assertEquals(root[0].tag,
+                          'test')
+        self.assertEquals(root[0][0].tag,
+                          '{http://www.w3.org/1999/XSL/Transform}copy-of')
+
     def test_xslt_document_parse(self):
         # make sure document('') works from loaded files
         xslt = etree.XSLT(etree.parse(fileInTestDir("test-document.xslt")))

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 21 13:17:52 2006
@@ -249,6 +249,7 @@
     """
     cdef XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
+    cdef object _doc_url_utf
     
     def __init__(self, xslt_input, extensions=None):
         # make a copy of the document as stylesheet needs to assume it
@@ -268,9 +269,15 @@
 
         # XXX work around bug in xmlCopyDoc (fix is upcoming in new release
         # of libxml2)
+        if c_doc.URL is not NULL and c_doc.URL != doc._c_doc.URL:
+            tree.xmlFree(c_doc.URL)
         if doc._c_doc.URL is not NULL:
+            self._doc_url_utf = doc._c_doc.URL
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
-            
+        else:
+            self._doc_url_utf = "__STRING__XSLT__%s" % id(self)
+            c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
+
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
             raise XSLTParseError, "Cannot parse style sheet"
From scoder at codespeak.net  Fri Apr 21 13:20:30 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 13:20:31 2006
Subject: [Lxml-checkins] r26088 - lxml/trunk
Message-ID: <20060421112030.578CE1007F@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 13:20:29 2006
New Revision: 26088

Modified:
   lxml/trunk/CHANGES.txt
Log:
updated CHANGES.txt to mark document('') bug fixed

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Fri Apr 21 13:20:29 2006
@@ -24,6 +24,8 @@
 Bugs fixed
 ----------
 
+* document('') now works in XSLT documents parsed from strings
+
 * Crash in XMLSchema and RelaxNG when passing non-schema documents
 
 * Memory leak in RelaxNG() when RelaxNGParseError is raised
From scoder at codespeak.net  Fri Apr 21 16:27:15 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 16:27:17 2006
Subject: [Lxml-checkins] r26103 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060421142715.58BF810086@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 16:27:13 2006
New Revision: 26103

Added:
   lxml/branch/resolver-new/src/lxml/docloader.pxi
Modified:
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/tree.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
moved document loader API into docloader.pxi, doc loader support in XSLT, merged with parser doc loader API, store parser reference in _Documents to be resused in related documents (XSLT etc.)

Added: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- (empty file)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Fri Apr 21 16:27:13 2006
@@ -0,0 +1,80 @@
+# Custom resolver API
+
+ctypedef enum _InputDocumentDataType:
+    PARSER_DATA_STRING
+    PARSER_DATA_FILENAME
+
+cdef class _InputDocument:
+    cdef _InputDocumentDataType _type
+    cdef object _data_utf
+
+cdef class Resolver:
+    "This is the base class of all resolvers."
+    def resolve(self, system_url, public_id, context):
+        return None
+
+    def resolve_string(self, string, context):
+        cdef _InputDocument doc_ref
+        doc_ref = _InputDocument()
+        doc_ref._type = PARSER_DATA_STRING
+        doc_ref._data_utf = _utf8(string)
+        return doc_ref
+
+    def resolve_filename(self, filename, context):
+        cdef _InputDocument doc_ref
+        doc_ref = _ParserInput()
+        doc_ref._type = PARSER_DATA_FILENAME
+        doc_ref._data_utf = _utf8(filename)
+        return doc_ref
+
+cdef class _ResolverRegistry:
+    cdef object _resolvers
+    cdef Resolver _default_resolver
+    def __init__(self, Resolver default_resolver=None):
+        try:
+            self._resolvers = set()
+        except NameError:
+            from sets import Set
+            self._resolvers = Set()
+        self._default_resolver = default_resolver
+
+    def add(self, Resolver resolver not None):
+        """Register a resolver.
+
+        For each requested entity, the 'resolve' method of the resolver will
+        be called and the result will be passed to the parser.  If this method
+        returns None, the request will be delegated to other resolvers or the
+        default resolver.  The resolvers will be tested in an arbitrary order
+        until the first match is found.
+        """
+        self._resolvers.add(resolver)
+
+    def remove(self, resolver):
+        self._resolvers.discard(resolver)
+
+    def resolve(self, system_url, public_id, context):
+        for resolver in self._resolvers:
+            result = resolver.resolve(system_url, public_id, context)
+            if result is not None:
+                return result
+        if self._default_resolver is None:
+            return None
+        return self._default_resolver.resolve(system_url, public_id, context)
+
+cdef class _ResolverContext:
+    cdef _ResolverRegistry _resolvers
+    cdef _TempStore _storage
+    cdef object _exc_info
+    def __init__(self, _ResolverRegistry resolvers not None):
+        self._resolvers = resolvers
+        self._storage = _TempStore()
+        self._exc_info = None
+
+    cdef _store_exception(self, e):
+        self._exc_info = sys.exc_info()
+
+    cdef _check_exception(self):
+        _exc_info = self._exc_info
+        if _exc_info is not None:
+            type, value, traceback = _exc_info
+            raise type, value, traceback

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Fri Apr 21 16:27:13 2006
@@ -63,6 +63,7 @@
     """
     cdef int _ns_counter
     cdef xmlDoc* _c_doc
+    cdef object _parser
     
     def __dealloc__(self):
         # if there are no more references to the document, it is safe
@@ -150,21 +151,22 @@
     if filename is None:
         filename = source
     # open filename
-    c_doc = _parseDocFromFile(filename, parser)
-    return _documentFactory(c_doc)
+    c_doc = _parseDocFromFile(_utf8(filename), parser)
+    return _documentFactory(c_doc, parser)
 
 cdef _Document _parseMemoryDocument(text, parser):
     cdef xmlDoc* c_doc
     if python.PyUnicode_Check(text):
         text = _stripDeclaration(_utf8(text))
     c_doc = _parseDoc(text, parser)
-    return _documentFactory(c_doc)
+    return _documentFactory(c_doc, parser)
 
-cdef _Document _documentFactory(xmlDoc* c_doc):
+cdef _Document _documentFactory(xmlDoc* c_doc, parser):
     cdef _Document result
     result = _Document()
     result._c_doc = c_doc
     result._ns_counter = 0
+    result._parser = parser
     return result
 
 # to help with debugging
@@ -438,7 +440,7 @@
         fake_c_doc = _fakeRootDoc(doc._c_doc, self._c_node)
         c_doc = tree.xmlCopyDoc(fake_c_doc, 1) # recursive copy
         _destroyFakeDoc(doc._c_doc, fake_c_doc)
-        doc = _documentFactory(c_doc)
+        doc = _documentFactory(c_doc, doc._parser)
         return doc.getroot()
         
     def set(self, key, value):
@@ -1154,7 +1156,7 @@
     c_doc = _newDoc()
     c_node = _createElement(c_doc, name_utf, attrib, _extra)
     tree.xmlDocSetRootElement(c_doc, c_node)
-    doc = _documentFactory(c_doc)
+    doc = _documentFactory(c_doc, None)
     # add namespaces to node if necessary
     doc._setNodeNamespaces(c_node, ns_utf, nsmap)
     return _elementFactory(doc, c_node)
@@ -1162,13 +1164,15 @@
 def Comment(text=None):
     cdef _Document doc
     cdef xmlNode*  c_node
+    cdef xmlDoc*   c_doc
     if text is None:
         text = '  '
     else:
         text = ' %s ' % _utf8(text)
-    doc = _documentFactory( _newDoc() )
-    c_node = _createComment(doc._c_doc, text)
-    tree.xmlAddChild(<xmlNode*>doc._c_doc, c_node)
+    c_doc = _newDoc()
+    doc = _documentFactory(c_doc, None)
+    c_node = _createComment(c_doc, text)
+    tree.xmlAddChild(<xmlNode*>c_doc, c_node)
     return _commentFactory(doc, c_node)
 
 def SubElement(_Element _parent not None, _tag,
@@ -1187,6 +1191,7 @@
     cdef xmlNode* c_next
     cdef xmlNode* c_node
     cdef xmlNode* c_node_copy
+    cdef xmlDoc*  c_doc
     cdef _ElementTree etree
     cdef _Document doc
 
@@ -1195,7 +1200,8 @@
     elif file is not None:
         doc = _parseDocument(file, parser)
     else:
-        doc = _documentFactory( _newDoc() )
+        c_doc = _newDoc()
+        doc = _documentFactory(c_doc, parser)
 
     etree = _elementTreeFactory(doc, element)
 
@@ -1298,6 +1304,7 @@
 include "xmlerror.pxi"  # error and log handling
 include "xmlid.pxi"     # XMLID and IDDict
 include "nsclasses.pxi" # Namespace implementation and registry
+include "docloader.pxi" # Support for custom document loaders
 include "xslt.pxi"      # XPath and XSLT
 include "relaxng.pxi"   # RelaxNG
 include "xmlschema.pxi" # XMLSchema

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Fri Apr 21 16:27:13 2006
@@ -59,109 +59,17 @@
 
 
 ############################################################
-## Custom resolver API
+## support for custom document loaders
 ############################################################
 
-cdef class _ResolverRegistry # forward declaration
-
-cdef class _ParserInput:
-    cdef xmlparser.xmlParserInput* _input
-    cdef object _pyref # to keep Python references
-
-cdef class _ResolverContext:
-    cdef xmlparser.xmlParserCtxt* _ctxt
-    cdef _ResolverRegistry _resolvers
-    cdef _TempStore _storage
-    cdef object _exc_info
-
-    cdef _store_exception(self, e):
-        self._exc_info = sys.exc_info()
-
-    cdef _check_exception(self):
-        _exc_info = self._exc_info
-        if _exc_info is not None:
-            type, value, traceback = _exc_info
-            raise type, value, traceback
-
-cdef class Resolver:
-    def resolve(self, system_url, public_id, _ResolverContext context not None):
-        cdef _ParserInput parser_input
-        cdef char* c_url
-        cdef char* c_id
-        if __DEFAULT_ENTITY_LOADER is NULL:
-            return None
-        if system_url is None:
-            c_url = NULL
-        else:
-            url_utf = _utf8(system_url)
-            c_url = _cstr(url_utf)
-        if public_id is None:
-            c_id = NULL
-        else:
-            id_utf = _utf8(public_id)
-            c_id = _cstr(id_utf)
-        parser_input = _ParserInput()
-        parser_input._input = __DEFAULT_ENTITY_LOADER(
-            c_url, c_id, context._ctxt)
-        return parser_input
-
-    def resolve_string(self, string, _ResolverContext context not None):
-        cdef _ParserInput parser_input
-        string_utf = _utf8(string)
-        parser_input = _ParserInput()
-        parser_input._input = xmlparser.xmlNewStringInputStream(
-            context._ctxt, _cstr(string_utf))
-        parser_input._pyref = string_utf
-        return parser_input
-
-    def resolve_filename(self, filename, _ResolverContext context not None):
-        cdef _ParserInput parser_input
-        filename_utf = _utf8(filename)
-        parser_input = _ParserInput()
-        parser_input._input = xmlparser.xmlNewInputFromFile(
-            context._ctxt, _cstr(filename_utf))
-        return parser_input
-
-cdef class _ResolverRegistry:
-    cdef object _resolvers
-    cdef Resolver _default_resolver
-    def __init__(self, Resolver default_resolver=None):
-        try:
-            self._resolvers = set()
-        except NameError:
-            from sets import Set
-            self._resolvers = Set()
-        if default_resolver is None:
-            self._default_resolver = Resolver()
-        else:
-            self._default_resolver = default_resolver
-
-    def add(self, Resolver resolver not None):
-        """Register a resolver.
-
-        For each requested entity, the 'resolve' method of the resolver will
-        be called and the result will be passed to the parser.  If this method
-        returns None, the request will be delegated to other resolvers or the
-        default resolver.  The resolvers will be tested in an arbitrary order
-        until the first match is found.
-        """
-        self._resolvers.add(resolver)
-
-    def remove(self, resolver):
-        self._resolvers.discard(resolver)
-
-    def resolve(self, system_url, public_id, _ResolverContext context not None):
-        for resolver in self._resolvers:
-            result = resolver.resolve(system_url, public_id, context)
-            if result is not None:
-                return result
-        return self._default_resolver.resolve(system_url, public_id, context)
-
 cdef xmlparser.xmlParserInput* _local_resolver(char* c_url, char* c_pubid,
                                                xmlParserCtxt* c_context):
     cdef _ResolverContext context
-    cdef _ParserInput     parser_input
+    cdef _InputDocument   doc_ref
+    cdef xmlparser.xmlParserInput* c_input
     if c_context._private is NULL:
+        if __DEFAULT_ENTITY_LOADER is NULL:
+            return NULL
         return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
 
     try:
@@ -175,17 +83,27 @@
             pubid = funicode(c_pubid)
 
         context = <_ResolverContext>c_context._private
-
-        parser_input = context._resolvers.resolve(url, pubid, context)
+        doc_ref = context._resolvers.resolve(url, pubid, context)
     except Exception, e:
         context._store_exception(e)
-        if parser_input is not None and parser_input._input is not NULL:
-            xmlparser.xmlFreeInputStream(parser_input._input)
         return NULL
-    if parser_input is None:
-        return NULL
-    context._storage.add(parser_input)
-    return parser_input._input
+
+    if doc_ref is None:
+        if __DEFAULT_ENTITY_LOADER is NULL:
+            return NULL
+        return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
+
+    c_input = NULL
+    if doc_ref._type == PARSER_DATA_STRING:
+        c_input = xmlparser.xmlNewStringInputStream(
+            c_context, _cstr(doc_ref._data_utf))
+    elif doc_ref._type == PARSER_DATA_FILENAME:
+        c_input = xmlparser.xmlNewInputFromFile(
+            c_context, _cstr(doc_ref._data_utf))
+
+    if c_input != NULL:
+        context._storage.add(doc_ref._data_utf)
+    return c_input
 
 cdef xmlparser.xmlExternalEntityLoader __DEFAULT_ENTITY_LOADER
 __DEFAULT_ENTITY_LOADER = xmlparser.xmlGetExternalEntityLoader()
@@ -201,28 +119,21 @@
     cdef readonly object resolvers
     cdef _ResolverContext _context
     def __init__(self):
+        cdef _ResolverContext context
         self._error_log = _ErrorLog()
         self.resolvers = _ResolverRegistry()
-        self._context = None
+        self._context = _ResolverContext(self.resolvers)
 
     property error_log:
         def __get__(self):
             return self._error_log.copy()
 
     cdef _initContext(self, xmlParserCtxt* c_ctxt):
-        cdef _ResolverContext context
         __GLOBAL_PARSER_CONTEXT._initParserDict(c_ctxt)
-        context = _ResolverContext()
-        context._ctxt = c_ctxt
-        context._resolvers = self.resolvers
-        context._storage = _TempStore()
-        self._context = context
-        c_ctxt._private = <python.PyObject*>context
+        c_ctxt._private = <python.PyObject*>self._context
 
     cdef _cleanupContextAndRaisePending(self):
-        if self._context is not None:
-            self._context._check_exception()
-        self._context = None
+        self._context._check_exception()
 
     cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,
                                xmlDoc* result) except NULL:

Modified: lxml/branch/resolver-new/src/lxml/tree.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tree.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/tree.pxd	Fri Apr 21 16:27:13 2006
@@ -78,6 +78,7 @@
         xmlDict* dict
         xmlHashTable* ids
         char* URL
+        void* _private
         
     ctypedef struct xmlAttr:
         void* _private

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxd	Fri Apr 21 16:27:13 2006
@@ -2,13 +2,16 @@
 from xpath cimport xmlXPathContext, xmlXPathFunction
 
 cdef extern from "libxslt/xsltInternals.h":
+    ctypedef struct xsltDocument:
+        xmlDoc* doc
+
     ctypedef struct xsltStylesheet:
         xmlDoc* doc
 
     ctypedef struct xsltTransformContext:
         xmlXPathContext* xpathCtxt
-        void* _private
-    
+        xsltDocument* document
+
     cdef xsltStylesheet* xsltParseStylesheetDoc(xmlDoc* doc)
     cdef void xsltFreeStylesheet(xsltStylesheet* sheet)
 

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Fri Apr 21 16:27:13 2006
@@ -250,6 +250,7 @@
     cdef XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
     cdef object _doc_url_utf
+    cdef object _parser
     
     def __init__(self, xslt_input, extensions=None):
         # make a copy of the document as stylesheet needs to assume it
@@ -275,14 +276,16 @@
             self._doc_url_utf = doc._c_doc.URL
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
         else:
-            self._doc_url_utf = "__STRING__XSLT__%s" % id(self)
+            self._doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
             c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
 
+        c_doc._private = <python.PyObject*>doc
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
             tree.xmlFreeDoc(c_doc)
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
+        self._parser = doc._parser
 
         self._context = XSLTContext(None, extensions)
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
@@ -309,12 +312,12 @@
         root_node = _rootNodeOf(_input)
 
         c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
+        c_doc._private = <python.PyObject*>input_doc
 
         transform_ctxt = xslt.xsltNewTransformContext(self._c_style, c_doc)
         if transform_ctxt is NULL:
             _destroyFakeDoc(input_doc._c_doc, c_doc)
             raise XSLTApplyError, "Error preparing stylesheet run"
-        transform_ctxt._private = <python.PyObject*>self
 
         if _kw:
             # allocate space for parameters
@@ -352,14 +355,14 @@
         if c_result is NULL:
             raise XSLTApplyError, "Error applying stylesheet"
 
-        result_doc = _documentFactory(c_result)
+        result_doc = _documentFactory(c_result, input_doc._parser)
         return _xsltResultTreeFactory(result_doc, self)
 
     def apply(self, _input, **_kw):
         return self(_input, **_kw)
 
     def tostring(self, _ElementTree result_tree):
-        """Save result doc to string using stylesheet as guidance.
+        """Save result doc to string based on stylesheet output method.
         """
         return str(result_tree)
 
@@ -387,21 +390,69 @@
 
 
 ################################################################################
-# XSLT document resolvers
+# XSLT document loaders
 
 cdef xslt.xsltDocLoaderFunc __xsltDocDefaultLoader
 __xsltDocDefaultLoader = xslt.xsltDocDefaultLoader
 
-cdef xmlDoc* _doc_loader(char* uri, tree.xmlDict* dict, int options, void* ctxt,
-                         xslt.xsltLoadType type):
-    if tree.strncmp(uri, "__STRING__XSLT__", 16) != 0:
-        # use default loader for stylesheet
-        print uri, type, options
-        if type == xslt.XSLT_LOAD_DOCUMENT:
-            print uri
-        else:
-            print type, uri
-    return __xsltDocDefaultLoader(uri, dict, options, ctxt, type)
+cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int c_options,
+                         void* c_ctxt, xslt.xsltLoadType c_type):
+    cdef xslt.xsltTransformContext* transform_ctxt
+    cdef xslt.xsltStylesheet* c_style
+    cdef _ResolverRegistry resolvers
+    cdef _InputDocument doc_ref
+    cdef xmlDoc* c_doc
+    cdef _Document reference_doc
+    # find reference _Document to retrieve resolvers
+    c_doc = NULL
+    reference_doc = None
+    if c_type == xslt.XSLT_LOAD_DOCUMENT:
+        transform_ctxt = <xslt.xsltTransformContext*>c_ctxt
+        if transform_ctxt.document is not NULL and \
+               transform_ctxt.document.doc is not NULL and \
+               transform_ctxt.document.doc._private is not NULL:
+            reference_doc = <_Document>transform_ctxt.document.doc._private
+    elif c_type == xslt.XSLT_LOAD_STYLESHEET:
+        c_style = <xslt.xsltStylesheet*>c_ctxt
+        if c_style.doc is not NULL and c_style.doc._private is not NULL:
+            reference_doc = <_Document>c_style.doc._private
+
+    # try default loader first to take advantage of libxslt's doc reuse
+    if tree.strncmp(c_uri, "py:", 3) != 0:
+        c_doc = __xsltDocDefaultLoader(c_uri, c_dict, c_options, c_ctxt, c_type)
+        if c_doc is not NULL:
+            if c_doc._private is NULL and reference_doc is not None:
+                c_doc._private = <python.PyObject*>reference_doc
+            return c_doc
+
+    # stylesheets always use default loader
+    if c_doc is NULL and tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 16) != 0:
+        # otherwise delegate to custom resolvers
+        if reference_doc is not None:
+            resolvers = reference_doc._parser.resolvers
+            # FIXME: we should pass a _ResolverContext here!
+            try:
+                uri = funicode(c_uri)
+                doc_ref = resolvers.resolve(uri, None, None)
+            except Exception, e:
+                # FIXME: where to put the exception?
+                print e
+                raise
+
+            if doc_ref is not None:
+                c_doc = NULL
+                if doc_ref._type == PARSER_DATA_STRING:
+                    c_doc = _parseDoc(doc_ref._data_utf,
+                                      reference_doc._parser)
+                elif doc_ref._type == PARSER_DATA_FILENAME:
+                    c_doc = _parseDocFromFile(doc_ref._data_utf,
+                                              reference_doc._parser)
+
+    if c_doc is NULL:
+        c_doc = __xsltDocDefaultLoader(c_uri, c_dict, c_options, c_ctxt, c_type)
+    if c_doc is not NULL and c_doc._private is NULL and reference_doc is not None:
+        c_doc._private = <python.PyObject*>reference_doc
+    return c_doc
 
 xslt.xsltSetLoaderFunc(_doc_loader)
 
From scoder at codespeak.net  Fri Apr 21 16:37:27 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 16:37:28 2006
Subject: [Lxml-checkins] r26105 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060421143727.5489A10086@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 16:37:26 2006
New Revision: 26105

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
clean up

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Fri Apr 21 16:37:26 2006
@@ -392,9 +392,6 @@
 ################################################################################
 # XSLT document loaders
 
-cdef xslt.xsltDocLoaderFunc __xsltDocDefaultLoader
-__xsltDocDefaultLoader = xslt.xsltDocDefaultLoader
-
 cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int c_options,
                          void* c_ctxt, xslt.xsltLoadType c_type):
     cdef xslt.xsltTransformContext* transform_ctxt
@@ -419,7 +416,8 @@
 
     # try default loader first to take advantage of libxslt's doc reuse
     if tree.strncmp(c_uri, "py:", 3) != 0:
-        c_doc = __xsltDocDefaultLoader(c_uri, c_dict, c_options, c_ctxt, c_type)
+        c_doc = XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, c_options, c_ctxt, c_type)
         if c_doc is not NULL:
             if c_doc._private is NULL and reference_doc is not None:
                 c_doc._private = <python.PyObject*>reference_doc
@@ -449,11 +447,15 @@
                                               reference_doc._parser)
 
     if c_doc is NULL:
-        c_doc = __xsltDocDefaultLoader(c_uri, c_dict, c_options, c_ctxt, c_type)
+        c_doc = XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, c_options, c_ctxt, c_type)
     if c_doc is not NULL and c_doc._private is NULL and reference_doc is not None:
         c_doc._private = <python.PyObject*>reference_doc
     return c_doc
 
+cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
+XSLT_DOC_DEFAULT_LOADER = xslt.xsltDocDefaultLoader
+
 xslt.xsltSetLoaderFunc(_doc_loader)
 
 
From scoder at codespeak.net  Fri Apr 21 16:38:30 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 16:38:32 2006
Subject: [Lxml-checkins] r26106 - lxml/branch/resolver-new/src/lxml/tests
Message-ID: <20060421143830.05CDA10086@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 16:38:30 2006
New Revision: 26106

Modified:
   lxml/branch/resolver-new/src/lxml/tests/test_xslt.py
Log:
extended test case test_xslt_document_XML to check for text in the stylesheet

Modified: lxml/branch/resolver-new/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tests/test_xslt.py	(original)
+++ lxml/branch/resolver-new/src/lxml/tests/test_xslt.py	Fri Apr 21 16:38:30 2006
@@ -336,12 +336,12 @@
                           '<A>X</A>')
 
     def test_xslt_document_XML(self):
-        # make sure document('') works from loaded files
+        # make sure document('') works from parsed strings
         xslt = etree.XSLT(etree.XML("""\
 <xsl:stylesheet version="1.0"
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
   <xsl:template match="/">
-    <test><xsl:copy-of select="document('')//test"/></test>
+    <test>TEXT<xsl:copy-of select="document('')//test"/></test>
   </xsl:template>
 </xsl:stylesheet>
 """))
@@ -351,6 +351,8 @@
                           'test')
         self.assertEquals(root[0].tag,
                           'test')
+        self.assertEquals(root[0].text,
+                          'TEXT')
         self.assertEquals(root[0][0].tag,
                           '{http://www.w3.org/1999/XSL/Transform}copy-of')
 
From scoder at codespeak.net  Fri Apr 21 22:24:55 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 22:24:55 2006
Subject: [Lxml-checkins] r26116 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060421202455.3A65D10086@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 22:24:54 2006
New Revision: 26116

Modified:
   lxml/branch/resolver-new/src/lxml/docloader.pxi
Log:
allow resolving to 'EMPTY'

Modified: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/docloader.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Fri Apr 21 22:24:54 2006
@@ -1,6 +1,7 @@
 # Custom resolver API
 
 ctypedef enum _InputDocumentDataType:
+    PARSER_DATA_EMPTY
     PARSER_DATA_STRING
     PARSER_DATA_FILENAME
 
@@ -13,6 +14,12 @@
     def resolve(self, system_url, public_id, context):
         return None
 
+    def resolve_empty(self, context):
+        cdef _InputDocument doc_ref
+        doc_ref = _InputDocument()
+        doc_ref._type = PARSER_DATA_EMPTY
+        return doc_ref
+
     def resolve_string(self, string, context):
         cdef _InputDocument doc_ref
         doc_ref = _InputDocument()
From scoder at codespeak.net  Fri Apr 21 22:30:15 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 21 22:30:17 2006
Subject: [Lxml-checkins] r26117 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060421203015.E12A010086@code0.codespeak.net>

Author: scoder
Date: Fri Apr 21 22:30:14 2006
New Revision: 26117

Modified:
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
use _XSLTResolverContext for XSLT document loaders, store exceptions to raise them after the transform

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Fri Apr 21 22:30:14 2006
@@ -55,6 +55,8 @@
 class C14NError(LxmlError):
     pass
 
+cdef class BaseParser # forward declaration
+
 cdef class _Document:
     """Internal base class to reference a libxml document.
 
@@ -63,7 +65,7 @@
     """
     cdef int _ns_counter
     cdef xmlDoc* _c_doc
-    cdef object _parser
+    cdef BaseParser _parser
     
     def __dealloc__(self):
         # if there are no more references to the document, it is safe
@@ -166,6 +168,8 @@
     result = _Document()
     result._c_doc = c_doc
     result._ns_counter = 0
+    if parser is None:
+        parser = __DEFAULT_PARSER
     result._parser = parser
     return result
 

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Fri Apr 21 22:30:14 2006
@@ -199,6 +199,81 @@
 
 
 ################################################################################
+# XSLT document loaders
+
+cdef class _XSLTResolverContext(_ResolverContext):
+    cdef BaseParser _parser
+    def __init__(self, BaseParser parser not None):
+        _ResolverContext.__init__(self, parser.resolvers)
+        self._parser = parser
+
+cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int c_options,
+                         void* c_ctxt, xslt.xsltLoadType c_type):
+    cdef xslt.xsltTransformContext* transform_ctxt
+    cdef _ResolverRegistry resolvers
+    cdef _InputDocument doc_ref
+    cdef xmlDoc* c_doc
+    cdef _XSLTResolverContext resolver_context
+    # find reference _Document to retrieve resolvers
+    c_doc = NULL
+    resolver_context = None
+    if c_type == xslt.XSLT_LOAD_DOCUMENT:
+        transform_ctxt = <xslt.xsltTransformContext*>c_ctxt
+        if transform_ctxt.document is not NULL:
+            c_doc = transform_ctxt.document.doc
+    elif c_type == xslt.XSLT_LOAD_STYLESHEET:
+        c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
+
+    if c_doc is NULL or c_doc._private is NULL:
+        # can't call Python without context, fall back to default loader
+        c_doc = XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, c_options, c_ctxt, c_type)
+        if c_doc is not NULL:
+            if c_doc._private is NULL and reference_doc is not None:
+                c_doc._private = <python.PyObject*>reference_doc
+            return c_doc
+
+    resolver_context = <_XSLTResolverContext>c_doc._private
+
+    # call the Python document loaders - except for stylesheets
+    c_doc = NULL
+    if tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 16) != 0:
+        resolvers = resolver_context._resolvers
+        try:
+            uri = funicode(c_uri)
+            # FIXME: we should pass a _ResolverContext here!
+            doc_ref = resolvers.resolve(uri, None, resolver_context)
+        except Exception, e:
+            resolver_context._store_exception(e)
+            return NULL
+
+        if doc_ref is not None:
+            c_doc = NULL
+            if doc_ref._type == PARSER_DATA_STRING:
+                c_doc = _parseDoc(doc_ref._data_utf,
+                                  resolver_context._parser)
+            elif doc_ref._type == PARSER_DATA_FILENAME:
+                c_doc = _parseDocFromFile(doc_ref._data_utf,
+                                          resolver_context._parser)
+            elif doc_ref._type == PARSER_DATA_EMPTY:
+                return NULL
+            if c_doc is not NULL and c_doc.URL is NULL:
+                c_doc.URL = tree.xmlStrdup(c_uri)
+
+    if c_doc is NULL:
+        c_doc = XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, c_options, c_ctxt, c_type)
+    if c_doc is not NULL and c_doc._private is NULL and resolver_context is not None:
+        c_doc._private = <python.PyObject*>resolver_context
+    return c_doc
+
+cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
+XSLT_DOC_DEFAULT_LOADER = xslt.xsltDocDefaultLoader
+
+xslt.xsltSetLoaderFunc(_doc_loader)
+
+
+################################################################################
 # XSLT
 
 cdef class XSLTContext(BaseContext):
@@ -250,8 +325,8 @@
     cdef XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
     cdef object _doc_url_utf
-    cdef object _parser
-    
+    cdef _XSLTResolverContext _xslt_resolver_context
+
     def __init__(self, xslt_input, extensions=None):
         # make a copy of the document as stylesheet needs to assume it
         # doesn't change
@@ -279,13 +354,13 @@
             self._doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
             c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
 
-        c_doc._private = <python.PyObject*>doc
+        self._xslt_resolver_context = _XSLTResolverContext(doc._parser)
+        c_doc._private = <python.PyObject*>self._xslt_resolver_context
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
             tree.xmlFreeDoc(c_doc)
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
-        self._parser = doc._parser
 
         self._context = XSLTContext(None, extensions)
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
@@ -301,6 +376,7 @@
         cdef _Document input_doc
         cdef _NodeBase root_node
         cdef _Document result_doc
+        cdef _XSLTResolverContext resolver_context
         cdef xslt.xsltTransformContext* transform_ctxt
         cdef xmlDoc* c_result
         cdef xmlDoc* c_doc
@@ -312,7 +388,9 @@
         root_node = _rootNodeOf(_input)
 
         c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
-        c_doc._private = <python.PyObject*>input_doc
+
+        resolver_context = _XSLTResolverContext(input_doc._parser)
+        c_doc._private = <python.PyObject*>resolver_context
 
         transform_ctxt = xslt.xsltNewTransformContext(self._c_style, c_doc)
         if transform_ctxt is NULL:
@@ -350,8 +428,10 @@
             cstd.free(params)
 
         self._context.free_context()
+        c_doc._private = <python.PyObject*>input_doc # restore just in case...
         _destroyFakeDoc(input_doc._c_doc, c_doc)
 
+        resolver_context._check_exception()
         if c_result is NULL:
             raise XSLTApplyError, "Error applying stylesheet"
 
@@ -390,76 +470,6 @@
 
 
 ################################################################################
-# XSLT document loaders
-
-cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int c_options,
-                         void* c_ctxt, xslt.xsltLoadType c_type):
-    cdef xslt.xsltTransformContext* transform_ctxt
-    cdef xslt.xsltStylesheet* c_style
-    cdef _ResolverRegistry resolvers
-    cdef _InputDocument doc_ref
-    cdef xmlDoc* c_doc
-    cdef _Document reference_doc
-    # find reference _Document to retrieve resolvers
-    c_doc = NULL
-    reference_doc = None
-    if c_type == xslt.XSLT_LOAD_DOCUMENT:
-        transform_ctxt = <xslt.xsltTransformContext*>c_ctxt
-        if transform_ctxt.document is not NULL and \
-               transform_ctxt.document.doc is not NULL and \
-               transform_ctxt.document.doc._private is not NULL:
-            reference_doc = <_Document>transform_ctxt.document.doc._private
-    elif c_type == xslt.XSLT_LOAD_STYLESHEET:
-        c_style = <xslt.xsltStylesheet*>c_ctxt
-        if c_style.doc is not NULL and c_style.doc._private is not NULL:
-            reference_doc = <_Document>c_style.doc._private
-
-    # try default loader first to take advantage of libxslt's doc reuse
-    if tree.strncmp(c_uri, "py:", 3) != 0:
-        c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, c_options, c_ctxt, c_type)
-        if c_doc is not NULL:
-            if c_doc._private is NULL and reference_doc is not None:
-                c_doc._private = <python.PyObject*>reference_doc
-            return c_doc
-
-    # stylesheets always use default loader
-    if c_doc is NULL and tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 16) != 0:
-        # otherwise delegate to custom resolvers
-        if reference_doc is not None:
-            resolvers = reference_doc._parser.resolvers
-            # FIXME: we should pass a _ResolverContext here!
-            try:
-                uri = funicode(c_uri)
-                doc_ref = resolvers.resolve(uri, None, None)
-            except Exception, e:
-                # FIXME: where to put the exception?
-                print e
-                raise
-
-            if doc_ref is not None:
-                c_doc = NULL
-                if doc_ref._type == PARSER_DATA_STRING:
-                    c_doc = _parseDoc(doc_ref._data_utf,
-                                      reference_doc._parser)
-                elif doc_ref._type == PARSER_DATA_FILENAME:
-                    c_doc = _parseDocFromFile(doc_ref._data_utf,
-                                              reference_doc._parser)
-
-    if c_doc is NULL:
-        c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, c_options, c_ctxt, c_type)
-    if c_doc is not NULL and c_doc._private is NULL and reference_doc is not None:
-        c_doc._private = <python.PyObject*>reference_doc
-    return c_doc
-
-cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
-XSLT_DOC_DEFAULT_LOADER = xslt.xsltDocDefaultLoader
-
-xslt.xsltSetLoaderFunc(_doc_loader)
-
-
-################################################################################
 # XPath
 
 cdef class XPathContext(BaseContext):
From scoder at codespeak.net  Sat Apr 22 10:12:47 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 10:12:50 2006
Subject: [Lxml-checkins] r26120 - lxml/branch/resolver-new/doc
Message-ID: <20060422081247.C0C531008D@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 10:12:46 2006
New Revision: 26120

Added:
   lxml/branch/resolver-new/doc/resolvers.txt
Log:
first shot on the doctest in doc/resolvers.txt

Added: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- (empty file)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Sat Apr 22 10:12:46 2006
@@ -0,0 +1,117 @@
+Document loading and URL resolving
+==================================
+
+Lxml has support for custom document loaders in both the parsers and XSL
+transformations.  These so-called resolvers are subclasses of the
+etree.Resolver class as in the following example::
+
+  >>> from lxml import etree
+
+  >>> class DTDResolver(etree.Resolver):
+  ...     def resolve(self, url, id, context):
+  ...         print "Resolving URL '%s'" % url
+  ...         return self.resolve_string(
+  ...             '<!ENTITY myentity "resolved text: %s">' % url, context)
+
+This defines a resolver that always returns a dynamically generated DTD
+fragment defining an entity.  The 'url' argument passes the system URL of the
+requested document, the 'id' argument is the public ID.  Note that any of
+these may be None.  The context object is not normally used by client code.
+
+Resolving is based on three methods of the Resolver object that build internal
+representations of the result document.  The method 'resolve_string' takes a
+document as parsable string, 'resolve_filename' takes a filename and
+'resolve_empty' resolves into an empty document.  The 'resolve' method may
+choose to return None, in which case the next registered resolver (or the
+default resolver) is consulted.
+
+Resolvers are registered local to a parser::
+
+  >>> parser = etree.XMLParser(dtd_validation=True)
+  >>> parser.resolvers.add( DTDResolver() )
+
+When we use this parser to parse a document that requires resolving a URL, it
+will call our custom resolver::
+
+  >>> xml = u'<!DOCTYPE doc SYSTEM "MissingDTD.dtd"><doc>&myentity;</doc>'
+  >>> from StringIO import StringIO
+  >>> tree = etree.parse(StringIO(xml), parser)
+  Resolving URL 'MissingDTD.dtd'
+  >>> root = tree.getroot()
+  >>> print root.text
+  resolved text: MissingDTD.dtd
+
+The entity was correctly resolved by the generated DTD fragment.
+
+
+Document loaders in context
+---------------------------
+
+XSLT document loading uses the same type of resolvers.  It distinguished
+between the context used when parsing the XSLT document (i.e. when resolving
+xsl:import and xsl:include elements) and the context at
+
+Let's defines a resolver that only responds to a specific prefix::
+
+  >>> class PrefixResolver(etree.Resolver):
+  ...     def __init__(self, prefix):
+  ...         self.prefix = prefix
+  ...         self.xml = "<test>%s-TEST</test>" % prefix
+  ...     def resolve(self, url, pubid, context):
+  ...         print "Resolving url %s as prefix %s ..." % (url, self.prefix),
+  ...         if url.startswith(self.prefix):
+  ...             print "done"
+  ...             return self.resolve_string(self.xml, context)
+  ...         print "failed"
+
+We use the following stylesheet as an example::
+
+  >>> xml_text = """\
+  ... <?xml version="1.0"?>
+  ... <xsl:stylesheet version="1.0"
+  ...    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  ...   <xsl:template match="/">
+  ...     <test>
+  ...       <xsl:copy-of select="document('uri:test')"/>
+  ...     </test>
+  ...   </xsl:template>
+  ... </xsl:stylesheet>
+  ... """
+
+If we now register different resolvers with two different parsers, we can
+parse our document twice in different resolver contexts::
+
+  >>> uri_parser = etree.XMLParser()
+  >>> uri_parser.resolvers.add( PrefixResolver("uri") )
+
+  >>> uro_parser = etree.XMLParser()
+  >>> uro_parser.resolvers.add( PrefixResolver("uro") )
+
+  >>> uri_doc = etree.parse(StringIO(xml_text), uri_parser)
+  >>> uro_doc = etree.parse(StringIO(xml_text), uro_parser)
+
+These contexts are important for the further behaviour of the documents.  They
+memorise their original parser so that the correct set of resolvers is used in
+subsequent lookups::
+
+  >>> transform = etree.XSLT(uro_doc)
+  >>> result = transform(uri_doc)
+  Resolving url uri:test as prefix uri ... done
+  >>> print str(result),
+  <?xml version="1.0"?>
+  <test><test>uri-TEST</test></test>
+
+We can see that the "uri" resolver was called to generate a document that was
+then inserted into the result document by the XSLT transformation.  Note that
+the "uri" resolver is attached to the transformed document, which defines the
+context for the transformation process.  If we reverse the contexts, the other
+resolver will get called.  Obviously, the "uro" resolver does not know how to
+resolve "uri" URLs, so the default resolver is called in this case::
+
+  >>> transform = etree.XSLT(uri_doc)
+  >>> result = transform(uro_doc)
+  Resolving url uri:test as prefix uro ... failed
+  >>> print str(result),
+  <?xml version="1.0"?>
+  <test/>
+
From scoder at codespeak.net  Sat Apr 22 13:21:04 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 13:21:05 2006
Subject: [Lxml-checkins] r26134 - lxml/branch/resolver-new/doc
Message-ID: <20060422112104.868CA1007E@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 13:21:03 2006
New Revision: 26134

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
Log:
restructured doctest example on document contexts in doc/resolvers.txt

Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Sat Apr 22 13:21:03 2006
@@ -11,7 +11,7 @@
   ...     def resolve(self, url, id, context):
   ...         print "Resolving URL '%s'" % url
   ...         return self.resolve_string(
-  ...             '<!ENTITY myentity "resolved text: %s">' % url, context)
+  ...             '<!ENTITY myentity "[resolved text: %s]">' % url, context)
 
 This defines a resolver that always returns a dynamically generated DTD
 fragment defining an entity.  The 'url' argument passes the system URL of the
@@ -27,11 +27,12 @@
 
 Resolvers are registered local to a parser::
 
-  >>> parser = etree.XMLParser(dtd_validation=True)
+  >>> parser = etree.XMLParser(load_dtd=True)
   >>> parser.resolvers.add( DTDResolver() )
 
-When we use this parser to parse a document that requires resolving a URL, it
-will call our custom resolver::
+Note that we instantiate a parser that loads the DTD.  This is not done by the
+default parser, which does no validation.  When we use this parser to parse a
+document that requires resolving a URL, it will call our custom resolver::
 
   >>> xml = u'<!DOCTYPE doc SYSTEM "MissingDTD.dtd"><doc>&myentity;</doc>'
   >>> from StringIO import StringIO
@@ -39,19 +40,26 @@
   Resolving URL 'MissingDTD.dtd'
   >>> root = tree.getroot()
   >>> print root.text
-  resolved text: MissingDTD.dtd
+  [resolved text: MissingDTD.dtd]
 
-The entity was correctly resolved by the generated DTD fragment.
+The entity in the document was correctly resolved by the generated DTD
+fragment.
 
 
 Document loaders in context
 ---------------------------
 
-XSLT document loading uses the same type of resolvers.  It distinguished
-between the context used when parsing the XSLT document (i.e. when resolving
-xsl:import and xsl:include elements) and the context at
+XML documents memorise their initial parser (and its resolvers) during their
+life-time.  This means that a lookup process related to a document will use
+the resolvers of the document's parser.
+
+This behaviour is most visible in XSLT, where two documents are used: the XSL
+document and the transformed XML document.  XSLT thus distinguishes between
+the context when parsing the XSLT document (i.e. when resolving xsl:import and
+xsl:include elements) and the context at transformation time (calls to the
+'document' function).
 
-Let's defines a resolver that only responds to a specific prefix::
+We start with a resolver that only responds to a specific prefix::
 
   >>> class PrefixResolver(etree.Resolver):
   ...     def __init__(self, prefix):
@@ -67,9 +75,9 @@
 We use the following stylesheet as an example::
 
   >>> xml_text = """\
-  ... <?xml version="1.0"?>
   ... <xsl:stylesheet version="1.0"
   ...    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  ...   <xsl:include href="uro:test"/>
   ...   <xsl:template match="/">
   ...     <test>
   ...       <xsl:copy-of select="document('uri:test')"/>
@@ -83,18 +91,21 @@
 
   >>> uri_parser = etree.XMLParser()
   >>> uri_parser.resolvers.add( PrefixResolver("uri") )
+  >>> uri_doc = etree.parse(StringIO(xml_text), uri_parser)
 
   >>> uro_parser = etree.XMLParser()
   >>> uro_parser.resolvers.add( PrefixResolver("uro") )
-
-  >>> uri_doc = etree.parse(StringIO(xml_text), uri_parser)
   >>> uro_doc = etree.parse(StringIO(xml_text), uro_parser)
 
 These contexts are important for the further behaviour of the documents.  They
 memorise their original parser so that the correct set of resolvers is used in
 subsequent lookups::
 
-  >>> transform = etree.XSLT(uro_doc)
+  >>> try:
+  ...     transform = etree.XSLT(uro_doc)
+  ... except Exception, e:
+  ...     print e, e.error_log
+  Resolving url uro:test as prefix uro ... done
   >>> result = transform(uri_doc)
   Resolving url uri:test as prefix uri ... done
   >>> print str(result),
@@ -114,4 +125,3 @@
   >>> print str(result),
   <?xml version="1.0"?>
   <test/>
-
From scoder at codespeak.net  Sat Apr 22 13:26:57 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 13:26:59 2006
Subject: [Lxml-checkins] r26135 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060422112657.D56FB1007E@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 13:26:56 2006
New Revision: 26135

Modified:
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
clean up in parsers, new internal parser functions used by XSLT to obey the libxslt-provided parse options and reuse resolver context

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Sat Apr 22 13:26:56 2006
@@ -1309,10 +1309,10 @@
 include "xmlid.pxi"     # XMLID and IDDict
 include "nsclasses.pxi" # Namespace implementation and registry
 include "docloader.pxi" # Support for custom document loaders
+include "parser.pxi"    # XML Parser
 include "xslt.pxi"      # XPath and XSLT
 include "relaxng.pxi"   # RelaxNG
 include "xmlschema.pxi" # XMLSchema
-include "parser.pxi"    # XML Parser
 include "proxy.pxi"     # Proxy handling (element backpointers/memory/etc.)
 
 

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Sat Apr 22 13:26:56 2006
@@ -132,20 +132,27 @@
         __GLOBAL_PARSER_CONTEXT._initParserDict(c_ctxt)
         c_ctxt._private = <python.PyObject*>self._context
 
-    cdef _cleanupContextAndRaisePending(self):
-        self._context._check_exception()
-
-    cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,
-                               xmlDoc* result) except NULL:
-        if ctxt.wellFormed:
-            __GLOBAL_PARSER_CONTEXT._initDocDict(result)
-        elif result is not NULL:
-            # free broken document
-            tree.xmlFreeDoc(result)
-            result = NULL
-        if result is NULL:
+cdef xmlDoc* _handleParseResult(xmlParserCtxt* ctxt, xmlDoc* result,
+                                char* c_filename) except NULL:
+    cdef _ResolverContext context
+    if ctxt.wellFormed:
+        __GLOBAL_PARSER_CONTEXT._initDocDict(result)
+    elif result is not NULL:
+        # free broken document
+        tree.xmlFreeDoc(result)
+        result = NULL
+
+    if ctxt._private is not NULL:
+        context = <_ResolverContext>ctxt._private
+        context._check_exception()
+
+    if result is NULL:
+        if c_filename is not NULL and \
+               ctxt.lastError.domain == xmlerror.XML_FROM_IO:
+            raise IOError, "Could not open file %s" % c_filename
+        else:
             raise XMLSyntaxError
-        return result
+    return result
 
 ############################################################
 ## XML parser
@@ -176,12 +183,14 @@
     cdef xmlParserCtxt* _file_parser_ctxt
     cdef xmlParserCtxt* _memory_parser_ctxt
     def __init__(self, attribute_defaults=False, dtd_validation=False,
-                 no_network=False, ns_clean=False):
+                 load_dtd=False, no_network=False, ns_clean=False):
         cdef int parse_options
         self._file_parser_ctxt = NULL
         BaseParser.__init__(self)
 
         parse_options = _XML_DEFAULT_PARSE_OPTIONS
+        if load_dtd:
+            parse_options = parse_options | xmlparser.XML_PARSE_DTDLOAD
         if dtd_validation:
             parse_options = parse_options | xmlparser.XML_PARSE_DTDLOAD | \
                             xmlparser.XML_PARSE_DTDVALID
@@ -209,7 +218,7 @@
             raise ParserError, "Failed to create parser context"
         return pctxt
 
-    cdef xmlDoc* _parseDoc(self, text_utf) except NULL:
+    cdef xmlDoc* _parseDoc(self, char* c_text) except NULL:
         """Parse document, share dictionary if possible.
         """
         cdef xmlDoc* result
@@ -222,12 +231,11 @@
             self._memory_parser_ctxt = pctxt
         self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadDoc(
-            pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
+            pctxt, c_text, NULL, NULL, self._parse_options)
         self._error_log.disconnect()
-        self._cleanupContextAndRaisePending()
-        return self._handleResult(pctxt, result)
+        return _handleParseResult(pctxt, result, NULL)
 
-    cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
+    cdef xmlDoc* _parseDocFromFile(self, char* c_filename) except NULL:
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
         self._error_log.connect()
@@ -237,13 +245,37 @@
             self._file_parser_ctxt = pctxt
         self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadFile(
-            pctxt, filename, NULL, self._parse_options)
+            pctxt, c_filename, NULL, self._parse_options)
         self._error_log.disconnect()
-        self._cleanupContextAndRaisePending()
-        if result is NULL:
-            if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
-                raise IOError, "Could not open file %s" % filename
-        return self._handleResult(pctxt, result)
+        return _handleParseResult(pctxt, result, c_filename)
+
+cdef xmlDoc* _internalParseDoc(char* c_text, int options,
+                               _ResolverContext context) except NULL:
+    # internal parser function for XSLT
+    cdef xmlParserCtxt* pctxt
+    cdef xmlDoc* c_doc
+    pctxt = xmlparser.xmlNewParserCtxt()
+    if pctxt is NULL:
+        return NULL
+    __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+    pctxt._private = <python.PyObject*>context
+    c_doc = xmlparser.xmlCtxtReadDoc(
+        pctxt, c_text, NULL, NULL, options)
+    return _handleParseResult(pctxt, c_doc, NULL)
+
+cdef xmlDoc* _internalParseDocFromFile(char* c_filename, int options,
+                                       _ResolverContext context) except NULL:
+    # internal parser function for XSLT
+    cdef xmlParserCtxt* pctxt
+    cdef xmlDoc* c_doc
+    pctxt = xmlparser.xmlNewParserCtxt()
+    if pctxt is NULL:
+        return NULL
+    __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+    pctxt._private = <python.PyObject*>context
+    c_doc = xmlparser.xmlCtxtReadFile(
+        pctxt, c_filename, NULL, options)
+    return _handleParseResult(pctxt, c_doc, c_filename)
 
 
 cdef XMLParser __DEFAULT_XML_PARSER
@@ -317,15 +349,13 @@
         if self._memory_parser_ctxt != NULL:
             htmlparser.htmlFreeParserCtxt(self._memory_parser_ctxt)
 
-    cdef xmlDoc* _parseDoc(self, text_utf) except NULL:
+    cdef xmlDoc* _parseDoc(self, char* c_text) except NULL:
         """Parse HTML document, share dictionary if possible.
         """
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
-        cdef char* c_text
         cdef int c_len
         self._error_log.connect()
-        c_text = _cstr(text_utf)
         pctxt = self._memory_parser_ctxt
         if pctxt is NULL:
             pctxt = htmlparser.htmlCreateMemoryParserCtxt('dummy', 5)
@@ -337,30 +367,28 @@
         result = htmlparser.htmlCtxtReadDoc(
             pctxt, c_text, NULL, NULL, self._parse_options)
         self._error_log.disconnect()
-        self._cleanupContextAndRaisePending()
-        return self._handleResult(pctxt, result)
+        return _handleParseResult(pctxt, result, NULL)
 
-    cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
+    cdef xmlDoc* _parseDocFromFile(self, char* c_filename) except NULL:
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
         cdef int parser_error
         self._error_log.connect()
         pctxt = self._file_parser_ctxt
         if pctxt is NULL:
-            pctxt = htmlparser.htmlCreateFileParserCtxt(filename, NULL)
+            pctxt = htmlparser.htmlCreateFileParserCtxt(c_filename, NULL)
             if pctxt is NULL:
                 self._error_log.disconnect()
                 warnings = self._error_log.filter_from_warnings()
                 if warnings and warnings[-1].domain == xmlerror.XML_FROM_IO:
-                    raise IOError, "Could not open file %s" % filename
+                    raise IOError, "Could not open file %s" % c_filename
                 raise ParserError, "Failed to create parser context"
             self._file_parser_ctxt = pctxt
         self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadFile(
-            pctxt, filename, NULL, self._parse_options)
+            pctxt, c_filename, NULL, self._parse_options)
         self._error_log.disconnect()
-        self._cleanupContextAndRaisePending()
-        return self._handleResult(pctxt, result)
+        return _handleParseResult(pctxt, result, c_filename)
 
 cdef HTMLParser __DEFAULT_HTML_PARSER
 __DEFAULT_HTML_PARSER = HTMLParser()
@@ -374,9 +402,9 @@
         parser = __DEFAULT_PARSER
     __GLOBAL_PARSER_CONTEXT._initParser()
     if isinstance(parser, XMLParser):
-        return (<XMLParser>parser)._parseDoc(text_utf)
+        return (<XMLParser>parser)._parseDoc(_cstr(text_utf))
     elif isinstance(parser, HTMLParser):
-        return (<HTMLParser>parser)._parseDoc(text_utf)
+        return (<HTMLParser>parser)._parseDoc(_cstr(text_utf))
     else:
         raise TypeError, "invalid parser"
 

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sat Apr 22 13:26:56 2006
@@ -207,16 +207,16 @@
         _ResolverContext.__init__(self, parser.resolvers)
         self._parser = parser
 
-cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int c_options,
+cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
                          void* c_ctxt, xslt.xsltLoadType c_type):
     cdef xslt.xsltTransformContext* transform_ctxt
+    cdef xmlDoc* c_doc
     cdef _ResolverRegistry resolvers
     cdef _InputDocument doc_ref
-    cdef xmlDoc* c_doc
     cdef _XSLTResolverContext resolver_context
+    cdef XMLParser parser
     # find reference _Document to retrieve resolvers
     c_doc = NULL
-    resolver_context = None
     if c_type == xslt.XSLT_LOAD_DOCUMENT:
         transform_ctxt = <xslt.xsltTransformContext*>c_ctxt
         if transform_ctxt.document is not NULL:
@@ -226,12 +226,8 @@
 
     if c_doc is NULL or c_doc._private is NULL:
         # can't call Python without context, fall back to default loader
-        c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, c_options, c_ctxt, c_type)
-        if c_doc is not NULL:
-            if c_doc._private is NULL and reference_doc is not None:
-                c_doc._private = <python.PyObject*>reference_doc
-            return c_doc
+        return XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, parse_options, c_ctxt, c_type)
 
     resolver_context = <_XSLTResolverContext>c_doc._private
 
@@ -248,22 +244,22 @@
             return NULL
 
         if doc_ref is not None:
+            if doc_ref._type == PARSER_DATA_EMPTY:
+                return NULL
             c_doc = NULL
             if doc_ref._type == PARSER_DATA_STRING:
-                c_doc = _parseDoc(doc_ref._data_utf,
-                                  resolver_context._parser)
+                c_doc = _internalParseDoc(
+                    doc_ref._data_utf, parse_options, resolver_context)
             elif doc_ref._type == PARSER_DATA_FILENAME:
-                c_doc = _parseDocFromFile(doc_ref._data_utf,
-                                          resolver_context._parser)
-            elif doc_ref._type == PARSER_DATA_EMPTY:
-                return NULL
+                c_doc = _internalParseDocFromFile(
+                    doc_ref._data_utf, parse_options, resolver_context)
             if c_doc is not NULL and c_doc.URL is NULL:
                 c_doc.URL = tree.xmlStrdup(c_uri)
 
     if c_doc is NULL:
         c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, c_options, c_ctxt, c_type)
-    if c_doc is not NULL and c_doc._private is NULL and resolver_context is not None:
+            c_uri, c_dict, parse_options, c_ctxt, c_type)
+    if c_doc is not NULL and c_doc._private is NULL:
         c_doc._private = <python.PyObject*>resolver_context
     return c_doc
 
From scoder at codespeak.net  Sat Apr 22 13:35:31 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 13:35:32 2006
Subject: [Lxml-checkins] r26137 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060422113531.A879510080@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 13:35:30 2006
New Revision: 26137

Modified:
   lxml/branch/resolver-new/src/lxml/parser.pxi
Log:
better error reporting after parsing: reuse libxml2 generated message

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Sat Apr 22 13:35:30 2006
@@ -149,7 +149,14 @@
     if result is NULL:
         if c_filename is not NULL and \
                ctxt.lastError.domain == xmlerror.XML_FROM_IO:
-            raise IOError, "Could not open file %s" % c_filename
+            if ctxt.lastError.message is not NULL:
+                message = "Error reading file %s: %s" % (
+                    funicode(c_filename), funicode(ctxt.lastError.message))
+            else:
+                message = "Error reading file %s" % funicode(c_filename)
+            raise IOError, message
+        elif ctxt.lastError.message is not NULL:
+            raise XMLSyntaxError, funicode(ctxt.lastError.message)
         else:
             raise XMLSyntaxError
     return result
From scoder at codespeak.net  Sat Apr 22 21:12:35 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 21:12:37 2006
Subject: [Lxml-checkins] r26146 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060422191235.CD7E110079@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 21:12:34 2006
New Revision: 26146

Modified:
   lxml/branch/resolver-new/src/lxml/docloader.pxi
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
clean up, fixed exception forwarding, removed some race conditions in memory handling (XSLT.__call__), fix document('') again: store second copy of stylesheet document and resolve its URL to a new copy on request

Modified: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/docloader.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Sat Apr 22 21:12:34 2006
@@ -68,6 +68,9 @@
             return None
         return self._default_resolver.resolve(system_url, public_id, context)
 
+    def __repr__(self):
+        return repr(self._resolvers)
+
 cdef class _ResolverContext:
     cdef _ResolverRegistry _resolvers
     cdef _TempStore _storage
@@ -77,11 +80,17 @@
         self._storage = _TempStore()
         self._exc_info = None
 
-    cdef _store_exception(self, e):
+    cdef void _store_raised(self):
         self._exc_info = sys.exc_info()
 
-    cdef _check_exception(self):
+    cdef void _store_exception(self, exception):
+        self._exc_info = (exception, None, None)
+
+    cdef _raise_if_stored(self):
         _exc_info = self._exc_info
         if _exc_info is not None:
             type, value, traceback = _exc_info
-            raise type, value, traceback
+            if traceback is None and value is None:
+                raise type
+            else:
+                raise type, value, traceback

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Sat Apr 22 21:12:34 2006
@@ -1,6 +1,7 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, issubclass, hasattr, callable, iter, str, _cstr
+from python cimport isinstance, issubclass, hasattr, callable
+from python cimport iter, str, _cstr
 cimport xpath
 cimport xslt
 cimport xmlerror

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Sat Apr 22 21:12:34 2006
@@ -84,8 +84,8 @@
 
         context = <_ResolverContext>c_context._private
         doc_ref = context._resolvers.resolve(url, pubid, context)
-    except Exception, e:
-        context._store_exception(e)
+    except Exception:
+        context._store_raised()
         return NULL
 
     if doc_ref is None:
@@ -144,7 +144,7 @@
 
     if ctxt._private is not NULL:
         context = <_ResolverContext>ctxt._private
-        context._check_exception()
+        context._raise_if_stored()
 
     if result is NULL:
         if c_filename is not NULL and \

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sat Apr 22 21:12:34 2006
@@ -202,6 +202,7 @@
 # XSLT document loaders
 
 cdef class _XSLTResolverContext(_ResolverContext):
+    cdef xmlDoc* _c_style_doc
     cdef BaseParser _parser
     def __init__(self, BaseParser parser not None):
         _ResolverContext.__init__(self, parser.resolvers)
@@ -224,41 +225,50 @@
     elif c_type == xslt.XSLT_LOAD_STYLESHEET:
         c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
 
-    if c_doc is NULL or c_doc._private is NULL:
+    if c_doc is NULL or c_doc._private is NULL or \
+           not isinstance(<object>c_doc._private, _XSLTResolverContext):
         # can't call Python without context, fall back to default loader
         return XSLT_DOC_DEFAULT_LOADER(
             c_uri, c_dict, parse_options, c_ctxt, c_type)
 
     resolver_context = <_XSLTResolverContext>c_doc._private
 
-    # call the Python document loaders - except for stylesheets
+    # quick check if we are looking for the stylesheet
+    if tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 21) == 0:
+        c_doc = resolver_context._c_style_doc
+        if c_doc is not NULL and c_doc.URL is not NULL:
+            if tree.strcmp(c_uri, c_doc.URL) == 0:
+                return tree.xmlCopyDoc(c_doc, 1)
+
+    # call the Python document loaders
     c_doc = NULL
-    if tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 16) != 0:
-        resolvers = resolver_context._resolvers
-        try:
-            uri = funicode(c_uri)
-            # FIXME: we should pass a _ResolverContext here!
-            doc_ref = resolvers.resolve(uri, None, resolver_context)
-        except Exception, e:
-            resolver_context._store_exception(e)
-            return NULL
+    resolvers = resolver_context._resolvers
+    try:
+        uri = funicode(c_uri)
+        doc_ref = resolvers.resolve(uri, None, resolver_context)
+    except Exception:
+        resolver_context._store_raised()
+        return NULL
 
-        if doc_ref is not None:
-            if doc_ref._type == PARSER_DATA_EMPTY:
-                return NULL
-            c_doc = NULL
-            if doc_ref._type == PARSER_DATA_STRING:
-                c_doc = _internalParseDoc(
-                    doc_ref._data_utf, parse_options, resolver_context)
-            elif doc_ref._type == PARSER_DATA_FILENAME:
-                c_doc = _internalParseDocFromFile(
-                    doc_ref._data_utf, parse_options, resolver_context)
-            if c_doc is not NULL and c_doc.URL is NULL:
-                c_doc.URL = tree.xmlStrdup(c_uri)
+    if doc_ref is not None:
+        if doc_ref._type == PARSER_DATA_EMPTY:
+            return NULL
+        if doc_ref._type == PARSER_DATA_STRING:
+            c_doc = _internalParseDoc(
+                doc_ref._data_utf, parse_options, resolver_context)
+        elif doc_ref._type == PARSER_DATA_FILENAME:
+            c_doc = _internalParseDocFromFile(
+                doc_ref._data_utf, parse_options, resolver_context)
+        if c_doc is not NULL and c_doc.URL is NULL:
+            c_doc.URL = tree.xmlStrdup(c_uri)
 
     if c_doc is NULL:
         c_doc = XSLT_DOC_DEFAULT_LOADER(
             c_uri, c_dict, parse_options, c_ctxt, c_type)
+        if c_doc is NULL:
+            exception = XSLTError("Cannot resolve URI %s" % funicode(c_uri))
+            resolver_context._store_exception(exception)
+            return NULL
     if c_doc is not NULL and c_doc._private is NULL:
         c_doc._private = <python.PyObject*>resolver_context
     return c_doc
@@ -341,7 +351,7 @@
 
         # XXX work around bug in xmlCopyDoc (fix is upcoming in new release
         # of libxml2)
-        if c_doc.URL is not NULL and c_doc.URL != doc._c_doc.URL:
+        if c_doc.URL is not NULL:
             tree.xmlFree(c_doc.URL)
         if doc._c_doc.URL is not NULL:
             self._doc_url_utf = doc._c_doc.URL
@@ -351,7 +361,9 @@
             c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
 
         self._xslt_resolver_context = _XSLTResolverContext(doc._parser)
+        self._xslt_resolver_context._c_style_doc = tree.xmlCopyDoc(c_doc, 1)
         c_doc._private = <python.PyObject*>self._xslt_resolver_context
+
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
             tree.xmlFreeDoc(c_doc)
@@ -365,9 +377,6 @@
         # this cleans up copy of doc as well
         xslt.xsltFreeStylesheet(self._c_style)
 
-    cdef xmlDoc* _copyXslDoc(self):
-        return tree.xmlCopyDoc(self._c_style.doc, 1)
-
     def __call__(self, _input, **_kw):
         cdef _Document input_doc
         cdef _NodeBase root_node
@@ -377,22 +386,26 @@
         cdef xmlDoc* c_result
         cdef xmlDoc* c_doc
         cdef char** params
+        cdef void* ptemp
         cdef int i
         cdef int j
 
         input_doc = _documentOrRaise(_input)
         root_node = _rootNodeOf(_input)
 
-        c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
-
         resolver_context = _XSLTResolverContext(input_doc._parser)
-        c_doc._private = <python.PyObject*>resolver_context
+        resolver_context._c_style_doc = self._xslt_resolver_context._c_style_doc
+
+        c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
 
         transform_ctxt = xslt.xsltNewTransformContext(self._c_style, c_doc)
         if transform_ctxt is NULL:
             _destroyFakeDoc(input_doc._c_doc, c_doc)
             raise XSLTApplyError, "Error preparing stylesheet run"
 
+        ptemp = c_doc._private
+        c_doc._private = <python.PyObject*>resolver_context
+
         if _kw:
             # allocate space for parameters
             # * 2 as we want an entry for both key and value,
@@ -424,10 +437,10 @@
             cstd.free(params)
 
         self._context.free_context()
-        c_doc._private = <python.PyObject*>input_doc # restore just in case...
+        c_doc._private = ptemp
         _destroyFakeDoc(input_doc._c_doc, c_doc)
 
-        resolver_context._check_exception()
+        resolver_context._raise_if_stored()
         if c_result is NULL:
             raise XSLTApplyError, "Error applying stylesheet"
 
From scoder at codespeak.net  Sat Apr 22 21:13:38 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 21:13:40 2006
Subject: [Lxml-checkins] r26147 - in lxml/branch/resolver-new: doc
	src/lxml/tests
Message-ID: <20060422191338.4CA1D10079@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 21:13:36 2006
New Revision: 26147

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
   lxml/branch/resolver-new/src/lxml/tests/test_etree.py
Log:
rewrite of doctest section on parser/resolver contexts, add resolver.txt doctest to test_etree.py

Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Sat Apr 22 21:13:36 2006
@@ -64,7 +64,11 @@
   >>> class PrefixResolver(etree.Resolver):
   ...     def __init__(self, prefix):
   ...         self.prefix = prefix
-  ...         self.xml = "<test>%s-TEST</test>" % prefix
+  ...         self.xml = '''\
+  ...              <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  ...                <test xmlns="testNS">%s-TEST</test>
+  ...              </xsl:stylesheet>
+  ...              ''' % prefix
   ...     def resolve(self, url, pubid, context):
   ...         print "Resolving url %s as prefix %s ..." % (url, self.prefix),
   ...         if url.startswith(self.prefix):
@@ -77,10 +81,10 @@
   >>> xml_text = """\
   ... <xsl:stylesheet version="1.0"
   ...    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
-  ...   <xsl:include href="uro:test"/>
+  ...   <xsl:include href="honk:test"/>
   ...   <xsl:template match="/">
   ...     <test>
-  ...       <xsl:copy-of select="document('uri:test')"/>
+  ...       <xsl:value-of select="document('hoi:test')/*/*/text()"/>
   ...     </test>
   ...   </xsl:template>
   ... </xsl:stylesheet>
@@ -89,39 +93,43 @@
 If we now register different resolvers with two different parsers, we can
 parse our document twice in different resolver contexts::
 
-  >>> uri_parser = etree.XMLParser()
-  >>> uri_parser.resolvers.add( PrefixResolver("uri") )
-  >>> uri_doc = etree.parse(StringIO(xml_text), uri_parser)
-
-  >>> uro_parser = etree.XMLParser()
-  >>> uro_parser.resolvers.add( PrefixResolver("uro") )
-  >>> uro_doc = etree.parse(StringIO(xml_text), uro_parser)
+  >>> hoi_parser = etree.XMLParser()
+  >>> hoi_parser.resolvers.add( PrefixResolver("hoi") )
+  >>> hoi_doc = etree.parse(StringIO(xml_text), hoi_parser)
+
+  >>> honk_parser = etree.XMLParser()
+  >>> honk_parser.resolvers.add( PrefixResolver("honk") )
+  >>> honk_doc = etree.parse(StringIO(xml_text), honk_parser)
 
 These contexts are important for the further behaviour of the documents.  They
 memorise their original parser so that the correct set of resolvers is used in
-subsequent lookups::
+subsequent lookups.  To compile the stylesheet, XSLT must resolve the
+honk:test URI in the xsl:include element.  The "hoi" resolver cannot do that::
 
-  >>> try:
-  ...     transform = etree.XSLT(uro_doc)
-  ... except Exception, e:
-  ...     print e, e.error_log
-  Resolving url uro:test as prefix uro ... done
-  >>> result = transform(uri_doc)
-  Resolving url uri:test as prefix uri ... done
-  >>> print str(result),
-  <?xml version="1.0"?>
-  <test><test>uri-TEST</test></test>
+  >>> transform = etree.XSLT(hoi_doc)
+  Traceback (most recent call last):
+    [...]
+  XSLTParseError: Cannot parse style sheet
+
+However, if we use the "honk" resolver associated with the second document,
+everything works fine::
 
-We can see that the "uri" resolver was called to generate a document that was
+  >>> transform = etree.XSLT(honk_doc)
+  Resolving url honk:test as prefix honk ... done
+
+We can see that the "hoi" resolver was called to generate a document that was
 then inserted into the result document by the XSLT transformation.  Note that
-the "uri" resolver is attached to the transformed document, which defines the
-context for the transformation process.  If we reverse the contexts, the other
-resolver will get called.  Obviously, the "uro" resolver does not know how to
-resolve "uri" URLs, so the default resolver is called in this case::
-
-  >>> transform = etree.XSLT(uri_doc)
-  >>> result = transform(uro_doc)
-  Resolving url uri:test as prefix uro ... failed
+the "hoi" resolver is attached to the transformed document, which defines the
+context for the transformation process.  
+
+  >>> result = transform(hoi_doc)
+  Resolving url hoi:test as prefix hoi ... done
   >>> print str(result),
   <?xml version="1.0"?>
-  <test/>
+  <test>hoi-TEST</test>
+
+If we reverse the contexts, the other
+resolver will get called.  Obviously, the "honk" resolver does not know how to
+resolve "hoi" URLs, so the default resolver is called in this case.  If the
+default resolver cannot find the requested document either, you will get an
+exception::

Modified: lxml/branch/resolver-new/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/resolver-new/src/lxml/tests/test_etree.py	Sat Apr 22 21:13:36 2006
@@ -393,6 +393,8 @@
     suite.addTests([unittest.makeSuite(ETreeC14NTestCase)])
     suite.addTests(
         [doctest.DocFileSuite('../../../doc/api.txt')])
+    suite.addTests(
+        [doctest.DocFileSuite('../../../doc/resolvers.txt')])
     return suite
 
 if __name__ == '__main__':
From scoder at codespeak.net  Sat Apr 22 21:26:22 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 21:26:24 2006
Subject: [Lxml-checkins] r26148 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060422192622.77CE51006E@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 21:26:21 2006
New Revision: 26148

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
give the default resolver the first shot on stylesheets

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sat Apr 22 21:26:21 2006
@@ -233,12 +233,17 @@
 
     resolver_context = <_XSLTResolverContext>c_doc._private
 
-    # quick check if we are looking for the stylesheet
+    # quick check if we are looking for a stylesheet
     if tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 21) == 0:
-        c_doc = resolver_context._c_style_doc
-        if c_doc is not NULL and c_doc.URL is not NULL:
-            if tree.strcmp(c_uri, c_doc.URL) == 0:
-                return tree.xmlCopyDoc(c_doc, 1)
+        # already stored by libxslt?
+        c_doc = XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, parse_options, c_ctxt, c_type)
+        if c_doc is NULL:
+            # otherwise, check if we are looking for the current stylesheet
+            c_doc = resolver_context._c_style_doc
+            if c_doc is not NULL and c_doc.URL is not NULL:
+                if tree.strcmp(c_uri, c_doc.URL) == 0:
+                    return tree.xmlCopyDoc(c_doc, 1)
 
     # call the Python document loaders
     c_doc = NULL
From scoder at codespeak.net  Sat Apr 22 21:43:13 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 21:43:14 2006
Subject: [Lxml-checkins] r26149 - lxml/branch/resolver-new/doc
Message-ID: <20060422194313.1DD721007A@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 21:43:12 2006
New Revision: 26149

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
Log:
doctest cleanup

Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Sat Apr 22 21:43:12 2006
@@ -64,8 +64,9 @@
   >>> class PrefixResolver(etree.Resolver):
   ...     def __init__(self, prefix):
   ...         self.prefix = prefix
-  ...         self.xml = '''\
-  ...              <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  ...         self.result_xml = '''\
+  ...              <xsl:stylesheet
+  ...                     xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
   ...                <test xmlns="testNS">%s-TEST</test>
   ...              </xsl:stylesheet>
   ...              ''' % prefix
@@ -73,7 +74,7 @@
   ...         print "Resolving url %s as prefix %s ..." % (url, self.prefix),
   ...         if url.startswith(self.prefix):
   ...             print "done"
-  ...             return self.resolve_string(self.xml, context)
+  ...             return self.resolve_string(self.result_xml, context)
   ...         print "failed"
 
 We use the following stylesheet as an example::
From scoder at codespeak.net  Sat Apr 22 21:48:40 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 21:48:41 2006
Subject: [Lxml-checkins] r26150 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060422194840.86D4C1007A@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 21:48:39 2006
New Revision: 26150

Modified:
   lxml/branch/resolver-new/src/lxml/parser.pxi
Log:
fix a potential race condition in parser.pxi:_local_resolver

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Sat Apr 22 21:48:39 2006
@@ -67,7 +67,8 @@
     cdef _ResolverContext context
     cdef _InputDocument   doc_ref
     cdef xmlparser.xmlParserInput* c_input
-    if c_context._private is NULL:
+    if c_context._private is NULL or \
+       not isinstance(<object>c_context._private, _ResolverContext):
         if __DEFAULT_ENTITY_LOADER is NULL:
             return NULL
         return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
From scoder at codespeak.net  Sat Apr 22 21:48:45 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sat Apr 22 21:48:47 2006
Subject: [Lxml-checkins] r26151 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060422194845.E35791007A@code0.codespeak.net>

Author: scoder
Date: Sat Apr 22 21:48:44 2006
New Revision: 26151

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
clean up

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sat Apr 22 21:48:44 2006
@@ -260,10 +260,10 @@
             return NULL
         if doc_ref._type == PARSER_DATA_STRING:
             c_doc = _internalParseDoc(
-                doc_ref._data_utf, parse_options, resolver_context)
+                _cstr(doc_ref._data_utf), parse_options, resolver_context)
         elif doc_ref._type == PARSER_DATA_FILENAME:
             c_doc = _internalParseDocFromFile(
-                doc_ref._data_utf, parse_options, resolver_context)
+                _cstr(doc_ref._data_utf), parse_options, resolver_context)
         if c_doc is not NULL and c_doc.URL is NULL:
             c_doc.URL = tree.xmlStrdup(c_uri)
 
From scoder at codespeak.net  Sun Apr 23 15:11:59 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 15:12:00 2006
Subject: [Lxml-checkins] r26182 - lxml/branch/resolver-new
Message-ID: <20060423131159.4C16210084@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 15:11:58 2006
New Revision: 26182

Modified:
   lxml/branch/resolver-new/bench.py
Log:
benchmark for XSLT and document(''), showed that deep copying is faster than parsing when building copies of small stylesheets

Modified: lxml/branch/resolver-new/bench.py
==============================================================================
--- lxml/branch/resolver-new/bench.py	(original)
+++ lxml/branch/resolver-new/bench.py	Sun Apr 23 15:11:58 2006
@@ -407,6 +407,25 @@
             xpath = self.etree.XPathElementEvaluator(child)
             xpath.evaluate("./*[0]")
 
+    @onlylib('lxe')
+    def bench_xslt_document(self, root):
+        transform = self.etree.XSLT(self.etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:l="test"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <l:data>TEST</l:data>
+  <xsl:template match="/">
+    <l:result>
+      <xsl:for-each select="*/*">
+        <l:test><xsl:copy-of select="document('')//l:data/text()"/></l:test>
+      </xsl:for-each>
+    </l:result>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        transform(root)
+
+
 ############################################################
 # Main program
 ############################################################
From scoder at codespeak.net  Sun Apr 23 15:26:50 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 15:26:51 2006
Subject: [Lxml-checkins] r26186 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060423132650.5B29210098@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 15:26:49 2006
New Revision: 26186

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
always special case the current stylesheet in XSLT document lookup

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 15:26:49 2006
@@ -203,10 +203,15 @@
 
 cdef class _XSLTResolverContext(_ResolverContext):
     cdef xmlDoc* _c_style_doc
+    cdef object _style_url_utf
+    cdef object _style_doc_utf
     cdef BaseParser _parser
     def __init__(self, BaseParser parser not None):
         _ResolverContext.__init__(self, parser.resolvers)
         self._parser = parser
+        self._c_style_doc = NULL
+        self._style_url_utf = None
+        self._style_doc_utf = None
 
 cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
                          void* c_ctxt, xslt.xsltLoadType c_type):
@@ -233,17 +238,11 @@
 
     resolver_context = <_XSLTResolverContext>c_doc._private
 
-    # quick check if we are looking for a stylesheet
-    if tree.strncmp(c_uri, "XSLT:__STRING__XSLT__", 21) == 0:
-        # already stored by libxslt?
-        c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, parse_options, c_ctxt, c_type)
-        if c_doc is NULL:
-            # otherwise, check if we are looking for the current stylesheet
-            c_doc = resolver_context._c_style_doc
-            if c_doc is not NULL and c_doc.URL is not NULL:
-                if tree.strcmp(c_uri, c_doc.URL) == 0:
-                    return tree.xmlCopyDoc(c_doc, 1)
+    # quick check if we are looking for the current stylesheet
+    c_doc = resolver_context._c_style_doc
+    if c_doc is not NULL and c_doc.URL is not NULL:
+        if tree.strcmp(c_uri, c_doc.URL) == 0:
+            return tree.xmlCopyDoc(c_doc, 1)
 
     # call the Python document loaders
     c_doc = NULL
@@ -335,12 +334,9 @@
     """
     cdef XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
-    cdef object _doc_url_utf
     cdef _XSLTResolverContext _xslt_resolver_context
 
     def __init__(self, xslt_input, extensions=None):
-        # make a copy of the document as stylesheet needs to assume it
-        # doesn't change
         cdef xslt.xsltStylesheet* c_style
         cdef xmlDoc* c_doc
         cdef xmlDoc* fake_c_doc
@@ -350,22 +346,22 @@
         doc = _documentOrRaise(xslt_input)
         root_node = _rootNodeOf(xslt_input)
 
+        # make a copy of the document as stylesheet parsing modifies it
         fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
         c_doc = tree.xmlCopyDoc(fake_c_doc, 1)
         _destroyFakeDoc(doc._c_doc, fake_c_doc)
 
-        # XXX work around bug in xmlCopyDoc (fix is upcoming in new release
-        # of libxml2)
+        # make sure we always have a stylesheet URL
         if c_doc.URL is not NULL:
             tree.xmlFree(c_doc.URL)
         if doc._c_doc.URL is not NULL:
-            self._doc_url_utf = doc._c_doc.URL
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
         else:
-            self._doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
-            c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
+            doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
+            c_doc.URL = tree.xmlStrdup(_cstr(doc_url_utf))
 
         self._xslt_resolver_context = _XSLTResolverContext(doc._parser)
+        # keep a copy in case we need to access the stylesheet via 'document()'
         self._xslt_resolver_context._c_style_doc = tree.xmlCopyDoc(c_doc, 1)
         c_doc._private = <python.PyObject*>self._xslt_resolver_context
 
From scoder at codespeak.net  Sun Apr 23 16:59:12 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 16:59:13 2006
Subject: [Lxml-checkins] r26190 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060423145912.184291009D@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 16:59:10 2006
New Revision: 26190

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
free document copy on XSLT.__dealloc__

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 16:59:10 2006
@@ -375,6 +375,9 @@
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
         
     def __dealloc__(self):
+        if self._xslt_resolver_context is not None and \
+               self._xslt_resolver_context._c_style_doc is not NULL:
+            tree.xmlFreeDoc(self._xslt_resolver_context._c_style_doc)
         # this cleans up copy of doc as well
         xslt.xsltFreeStylesheet(self._c_style)
 
From scoder at codespeak.net  Sun Apr 23 17:15:51 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 17:15:52 2006
Subject: [Lxml-checkins] r26191 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060423151551.BBC751009F@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 17:15:50 2006
New Revision: 26191

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
some C-ification

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 17:15:50 2006
@@ -825,6 +825,7 @@
     cdef xmlNode* c_node
     cdef char* s
     cdef _NodeBase element
+    cdef int i
     result = []
     if xpathObj.nodesetval is NULL:
         return result
@@ -850,6 +851,7 @@
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
     cdef BaseContext extensions
+    cdef int i
 
     rctxt = ctxt.context
 
@@ -869,7 +871,7 @@
     args = []
     doc = extensions._doc
     for i from 0 <= i < nargs:
-        args.append(_unwrapXPathObject(xpath.valuePop(ctxt), doc))
+        python.PyList_Append(args, _unwrapXPathObject(xpath.valuePop(ctxt), doc))
     args.reverse()
 
     try:
From scoder at codespeak.net  Sun Apr 23 18:44:30 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 18:44:31 2006
Subject: [Lxml-checkins] r26196 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060423164430.7BC7B1009D@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 18:44:29 2006
New Revision: 26196

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
prefix internal classnames by '_'

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 18:44:29 2006
@@ -30,7 +30,7 @@
 ################################################################################
 # support for extension functions in XPath/XSLT
 
-cdef class BaseContext:
+cdef class _BaseContext:
     cdef xpath.xmlXPathContext* _xpathCtxt
     cdef _Document _doc
     cdef object _extensions
@@ -221,7 +221,7 @@
     cdef _InputDocument doc_ref
     cdef _XSLTResolverContext resolver_context
     cdef XMLParser parser
-    # find reference _Document to retrieve resolvers
+    # find current xmlDoc to retrieve resolvers
     c_doc = NULL
     if c_type == xslt.XSLT_LOAD_DOCUMENT:
         transform_ctxt = <xslt.xsltTransformContext*>c_ctxt
@@ -286,11 +286,11 @@
 ################################################################################
 # XSLT
 
-cdef class XSLTContext(BaseContext):
+cdef class _XSLTContext(_BaseContext):
     cdef xslt.xsltTransformContext* _xsltCtxt
     def __init__(self, namespaces, extensions):
         self._xsltCtxt = NULL
-        BaseContext.__init__(self, namespaces, extensions)
+        _BaseContext.__init__(self, namespaces, extensions)
 
     cdef register_context(self, xslt.xsltTransformContext* xsltCtxt, _Document doc):
         self._xsltCtxt = xsltCtxt
@@ -332,7 +332,7 @@
 cdef class XSLT:
     """Turn a document into an XSLT object.
     """
-    cdef XSLTContext _context
+    cdef _XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
     cdef _XSLTResolverContext _xslt_resolver_context
 
@@ -371,7 +371,7 @@
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
 
-        self._context = XSLTContext(None, extensions)
+        self._context = _XSLTContext(None, extensions)
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
         
     def __dealloc__(self):
@@ -481,15 +481,14 @@
     result._xslt = xslt
     return result
 
-
 ################################################################################
 # XPath
 
-cdef class XPathContext(BaseContext):
+cdef class _XPathContext(_BaseContext):
     cdef object _variables
     cdef object _registered_variables
     def __init__(self, namespaces, extensions, variables):
-        BaseContext.__init__(self, namespaces, extensions)
+        _BaseContext.__init__(self, namespaces, extensions)
         self._variables  = variables
         self._registered_variables  = []
         
@@ -567,10 +566,10 @@
 
 
 cdef class XPathEvaluatorBase:
-    cdef XPathContext _context
+    cdef _XPathContext _context
 
     def __init__(self, namespaces, extensions, variables=None):
-        self._context = XPathContext(namespaces, extensions, variables)
+        self._context = _XPathContext(namespaces, extensions, variables)
 
     cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
         _exc_info = self._context._exc_info
@@ -706,7 +705,7 @@
         cdef xpath.xmlXPathObject*  xpathObj
         cdef _Document document
         cdef _NodeBase element
-        cdef XPathContext context
+        cdef _XPathContext context
 
         document = _documentOrRaise(_etree_or_element)
         element  = _rootNodeOf(_etree_or_element)
@@ -850,7 +849,7 @@
     cdef xpath.xmlXPathContext* rctxt
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
-    cdef BaseContext extensions
+    cdef _BaseContext extensions
     cdef int i
 
     rctxt = ctxt.context
@@ -863,7 +862,7 @@
         uri = None
 
     # get our evaluator
-    extensions = <BaseContext>(rctxt.userData)
+    extensions = <_BaseContext>(rctxt.userData)
 
     # lookup up the extension function in the context
     f = extensions.find_extension(uri, name)
From scoder at codespeak.net  Sun Apr 23 18:53:41 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 18:53:42 2006
Subject: [Lxml-checkins] r26197 - in lxml/branch/resolver-new: . src/lxml
	src/lxml/tests
Message-ID: <20060423165341.5B4551009D@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 18:53:39 2006
New Revision: 26197

Modified:
   lxml/branch/resolver-new/setup.py
   lxml/branch/resolver-new/src/lxml/tests/test_xslt.py
   lxml/branch/resolver-new/src/lxml/xslt.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
compile with EXSLT support

Modified: lxml/branch/resolver-new/setup.py
==============================================================================
--- lxml/branch/resolver-new/setup.py	(original)
+++ lxml/branch/resolver-new/setup.py	Sun Apr 23 18:53:39 2006
@@ -11,6 +11,7 @@
 try:
     from setuptools import setup
     from setuptools.extension import Extension
+    # prevent setuptools from making local etree.so copies:
     setup_args['zip_safe'] = False
 except ImportError:
     from distutils.core import setup
@@ -47,6 +48,16 @@
 
     changelog.close()
 
+# compile also against libexslt!
+xslt_libs = flags('xslt-config --libs')
+xslt_libs.append('-lexslt')
+for i, libname in (): # enumerate(xslt_libs):
+    if 'exslt' in libname:
+        break
+    if 'xslt' in libname:
+        xslt_libs.insert(i, libname.replace('xslt', 'exslt'))
+        break
+
 setup(
     name = "lxml",
     version = version,
@@ -85,7 +96,7 @@
         "lxml.etree", 
         sources = sources,
         extra_compile_args = ['-w'] + flags('xslt-config --cflags'),
-        extra_link_args = flags('xslt-config --libs')
+        extra_link_args = xslt_libs
     )],
     **setup_args
 )

Modified: lxml/branch/resolver-new/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/branch/resolver-new/src/lxml/tests/test_xslt.py	(original)
+++ lxml/branch/resolver-new/src/lxml/tests/test_xslt.py	Sun Apr 23 18:53:39 2006
@@ -29,6 +29,36 @@
 <foo>B</foo>
 ''',
                           st.tostring(res))
+
+    def test_exslt(self):
+        tree = self.parse('<a><b>B</b><c>C</c></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:math="http://exslt.org/math"
+    xmlns:str="http://exslt.org/strings"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    exclude-result-prefixes="math str xsl">
+  <xsl:template match="text()">
+    <xsl:value-of select="str:align(string(.), '---', 'center')" />
+  </xsl:template>
+  <xsl:template match="*">
+    <xsl:copy>
+      <xsl:attribute name="pi">
+        <xsl:value-of select="math:constant('PI', count(*)+2)"/>
+      </xsl:attribute>
+      <xsl:apply-templates/>
+    </xsl:copy>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        st = etree.XSLT(style)
+        res = st.apply(tree)
+        self.assertEquals('''\
+<?xml version="1.0"?>
+<a pi="3.14"><b pi="3">-B-</b><c pi="3">-C-</c></a>
+''',
+                          st.tostring(res))
+
     def test_xslt_input(self):
         tree = self.parse('<a><b>B</b><c>C</c></a>')
         style = self.parse('''\

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxd	Sun Apr 23 18:53:39 2006
@@ -56,4 +56,8 @@
     cdef void xsltSetTransformErrorFunc(xsltTransformContext*,
                                         void* ctxt,
                                         void (*handler)(void* ctxt, char* msg, ...))
-    
+
+    cdef void xsltRegisterAllExtras()
+
+cdef extern from "libexslt/exslt.h":
+    cdef void exsltRegisterAll()

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 18:53:39 2006
@@ -481,6 +481,10 @@
     result._xslt = xslt
     return result
 
+# enable EXSLT and Saxon support for XSLT
+xslt.xsltRegisterAllExtras()
+xslt.exsltRegisterAll()
+
 ################################################################################
 # XPath
 
From scoder at codespeak.net  Sun Apr 23 20:24:50 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 20:24:51 2006
Subject: [Lxml-checkins] r26200 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060423182450.5EC0F100B7@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 20:24:49 2006
New Revision: 26200

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
make self call in XSLT.apply() more explicit

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 20:24:49 2006
@@ -452,7 +452,7 @@
         return _xsltResultTreeFactory(result_doc, self)
 
     def apply(self, _input, **_kw):
-        return self(_input, **_kw)
+        return self.__call__(_input, **_kw)
 
     def tostring(self, _ElementTree result_tree):
         """Save result doc to string based on stylesheet output method.
From scoder at codespeak.net  Sun Apr 23 21:31:19 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 21:31:20 2006
Subject: [Lxml-checkins] r26202 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060423193119.389571008A@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 21:31:18 2006
New Revision: 26202

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
doc

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Sun Apr 23 21:31:18 2006
@@ -353,6 +353,7 @@
 
         # make sure we always have a stylesheet URL
         if c_doc.URL is not NULL:
+            # handle a bug in older libxml2 versions
             tree.xmlFree(c_doc.URL)
         if doc._c_doc.URL is not NULL:
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
From scoder at codespeak.net  Sun Apr 23 21:58:46 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 23 21:58:47 2006
Subject: [Lxml-checkins] r26203 - lxml/trunk/src/lxml
Message-ID: <20060423195846.36BF110087@code0.codespeak.net>

Author: scoder
Date: Sun Apr 23 21:58:45 2006
New Revision: 26203

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
trivial implementations for Attrib.iter*()

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Sun Apr 23 21:58:45 2006
@@ -931,9 +931,6 @@
         except KeyError:
             return default
 
-    def __iter__(self):
-        return iter(self.keys())
-    
     def keys(self):
         result = []
         cdef xmlNode* c_node
@@ -944,6 +941,12 @@
             c_node = c_node.next
         return result
 
+    def __iter__(self):
+        return iter(self.keys())
+    
+    def iterkeys(self):
+        return iter(self.keys())
+
     def values(self):
         cdef xmlNode* c_node
         result = []
@@ -954,7 +957,10 @@
                     result, _attributeValue(self._c_node, c_node))
             c_node = c_node.next
         return result
-    
+
+    def itervalues(self):
+        return iter(self.values())
+
     def items(self):
         result = []
         cdef xmlNode* c_node
@@ -968,6 +974,9 @@
             c_node = c_node.next
         return result
 
+    def iteritems(self):
+        return iter(self.items())
+
     def has_key(self, key):
         cdef xmlNs* c_ns
         cdef char* result
From scoder at codespeak.net  Mon Apr 24 14:51:01 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 24 14:51:03 2006
Subject: [Lxml-checkins] r26263 - lxml/trunk/src/lxml
Message-ID: <20060424125101.C82501007C@code0.codespeak.net>

Author: scoder
Date: Mon Apr 24 14:51:00 2006
New Revision: 26263

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
merged in name change of internal classes from resolver-new branch

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Mon Apr 24 14:51:00 2006
@@ -30,7 +30,7 @@
 ################################################################################
 # support for extension functions in XPath/XSLT
 
-cdef class BaseContext:
+cdef class _BaseContext:
     cdef xpath.xmlXPathContext* _xpathCtxt
     cdef _Document _doc
     cdef object _extensions
@@ -201,11 +201,11 @@
 ################################################################################
 # XSLT
 
-cdef class XSLTContext(BaseContext):
+cdef class _XSLTContext(_BaseContext):
     cdef xslt.xsltTransformContext* _xsltCtxt
     def __init__(self, namespaces, extensions):
         self._xsltCtxt = NULL
-        BaseContext.__init__(self, namespaces, extensions)
+        _BaseContext.__init__(self, namespaces, extensions)
 
     cdef register_context(self, xslt.xsltTransformContext* xsltCtxt, _Document doc):
         self._xsltCtxt = xsltCtxt
@@ -247,7 +247,7 @@
 cdef class XSLT:
     """Turn a document into an XSLT object.
     """
-    cdef XSLTContext _context
+    cdef _XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
     cdef object _doc_url_utf
     
@@ -283,7 +283,7 @@
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
 
-        self._context = XSLTContext(None, extensions)
+        self._context = _XSLTContext(None, extensions)
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
         
     def __dealloc__(self):
@@ -380,15 +380,14 @@
     result._xslt = xslt
     return result
 
-
 ################################################################################
 # XPath
 
-cdef class XPathContext(BaseContext):
+cdef class _XPathContext(_BaseContext):
     cdef object _variables
     cdef object _registered_variables
     def __init__(self, namespaces, extensions, variables):
-        BaseContext.__init__(self, namespaces, extensions)
+        _BaseContext.__init__(self, namespaces, extensions)
         self._variables  = variables
         self._registered_variables  = []
         
@@ -466,10 +465,10 @@
 
 
 cdef class XPathEvaluatorBase:
-    cdef XPathContext _context
+    cdef _XPathContext _context
 
     def __init__(self, namespaces, extensions, variables=None):
-        self._context = XPathContext(namespaces, extensions, variables)
+        self._context = _XPathContext(namespaces, extensions, variables)
 
     cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
         _exc_info = self._context._exc_info
@@ -605,7 +604,7 @@
         cdef xpath.xmlXPathObject*  xpathObj
         cdef _Document document
         cdef _NodeBase element
-        cdef XPathContext context
+        cdef _XPathContext context
 
         document = _documentOrRaise(_etree_or_element)
         element  = _rootNodeOf(_etree_or_element)
@@ -748,7 +747,7 @@
     cdef xpath.xmlXPathContext* rctxt
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
-    cdef BaseContext extensions
+    cdef _BaseContext extensions
 
     rctxt = ctxt.context
 
@@ -760,7 +759,7 @@
         uri = None
 
     # get our evaluator
-    extensions = <BaseContext>(rctxt.userData)
+    extensions = <_BaseContext>(rctxt.userData)
 
     # lookup up the extension function in the context
     f = extensions.find_extension(uri, name)
From scoder at codespeak.net  Mon Apr 24 16:42:58 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 24 16:42:59 2006
Subject: [Lxml-checkins] r26268 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20060424144258.64B6C1008D@code0.codespeak.net>

Author: scoder
Date: Mon Apr 24 16:42:56 2006
New Revision: 26268

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/tests/test_xslt.py
   lxml/trunk/src/lxml/xslt.pxi
Log:
Exslt:regexp implementation for XSLT based on Python 're' module

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Mon Apr 24 16:42:56 2006
@@ -7,6 +7,8 @@
 Features added
 --------------
 
+* Exslt:regexp implementation for XSLT based on the Python 're' module
+
 * HTMLParser for parsing (broken) HTML
 
 * XMLDTDID function parses XML into tuple (root node, ID dict) based on xml:id

Modified: lxml/trunk/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xslt.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xslt.py	Mon Apr 24 16:42:56 2006
@@ -374,6 +374,84 @@
         self.assertEquals(root[0].tag,
                           '{http://www.w3.org/1999/XSL/Transform}stylesheet')
 
+    def test_exslt_regexp_test(self):
+        xslt = etree.XSLT(etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:regexp="http://exslt.org/regular-expressions"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:template match="*">
+    <test><xsl:copy-of select="*[regexp:test(string(.), '8.')]"/></test>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        result = xslt(etree.XML('<a><b>123</b><b>098</b><b>987</b></a>'))
+        root = result.getroot()
+        self.assertEquals(root.tag,
+                          'test')
+        self.assertEquals(len(root), 1)
+        self.assertEquals(root[0].tag,
+                          'b')
+        self.assertEquals(root[0].text,
+                          '987')
+
+    def test_exslt_regexp_replace(self):
+        xslt = etree.XSLT(etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:regexp="http://exslt.org/regular-expressions"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:template match="*">
+    <test>
+      <xsl:copy-of select="regexp:replace(string(.), 'd.', '',   'XX')"/>
+      <xsl:text>-</xsl:text>
+      <xsl:copy-of select="regexp:replace(string(.), 'd.', 'gi', 'XX')"/>
+    </test>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        result = xslt(etree.XML('<a>abdCdEeDed</a>'))
+        root = result.getroot()
+        self.assertEquals(root.tag,
+                          'test')
+        self.assertEquals(len(root), 0)
+        self.assertEquals(root.text, 'abXXdEeDed-abXXXXeXXd')
+
+    def test_exslt_regexp_match(self):
+        xslt = etree.XSLT(etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:regexp="http://exslt.org/regular-expressions"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:template match="*">
+    <test>
+      <test1><xsl:copy-of select="regexp:match(string(.), 'd.')"/></test1>
+      <test2><xsl:copy-of select="regexp:match(string(.), 'd.', 'g')"/></test2>
+      <test2i><xsl:copy-of select="regexp:match(string(.), 'd.', 'gi')"/></test2i>
+    </test>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        result = xslt(etree.XML('<a>abdCdEeDed</a>'))
+        root = result.getroot()
+        self.assertEquals(root.tag,  'test')
+        self.assertEquals(len(root), 3)
+
+        self.assertEquals(len(root[0]), 1)
+        self.assertEquals(root[0][0].tag, 'match')
+        self.assertEquals(root[0][0].text, 'dC')
+
+        self.assertEquals(len(root[1]), 2)
+        self.assertEquals(root[1][0].tag, 'match')
+        self.assertEquals(root[1][0].text, 'dC')
+        self.assertEquals(root[1][1].tag, 'match')
+        self.assertEquals(root[1][1].text, 'dE')
+
+        self.assertEquals(len(root[2]), 3)
+        self.assertEquals(root[2][0].tag, 'match')
+        self.assertEquals(root[2][0].text, 'dC')
+        self.assertEquals(root[2][1].tag, 'match')
+        self.assertEquals(root[2][1].text, 'dE')
+        self.assertEquals(root[2][2].tag, 'match')
+        self.assertEquals(root[2][2].text, 'De')
+
 def test_suite():
     suite = unittest.TestSuite()
     suite.addTests([unittest.makeSuite(ETreeXSLTTestCase)])

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Mon Apr 24 16:42:56 2006
@@ -197,6 +197,72 @@
                 #print "Holding document:", <int>element._doc._c_doc
                 python.PyDict_SetItem(self._temp_docs, id(element._doc), element._doc)
 
+################################################################################
+# EXSLT regexp implementation
+
+cdef object RE_COMPILE
+RE_COMPILE = re.compile
+
+cdef class _ExsltRegExp:
+    cdef object _compile_map
+    def __init__(self):
+        self._compile_map = {}
+
+    cdef _compile(self, rexp, ignore_case):
+        cdef python.PyObject* c_result
+        key = (rexp, ignore_case)
+        c_result = python.PyDict_GetItem(self._compile_map, key)
+        if c_result is not NULL:
+            return <object>c_result
+        if ignore_case:
+            py_flags = re.IGNORECASE
+        else:
+            py_flags = 0
+        rexp_compiled = RE_COMPILE(rexp, py_flags)
+        python.PyDict_SetItem(self._compile_map, key, rexp_compiled)
+        return rexp_compiled
+
+    def test(self, ctxt, s, rexp, flags=''):
+        rexpc = self._compile(rexp, 'i' in flags)
+        if rexpc.search(s) is None:
+            return False
+        else:
+            return True
+
+    def match(self, ctxt, s, rexp, flags=''):
+        rexpc = self._compile(rexp, 'i' in flags)
+        if 'g' in flags:
+            results = rexpc.findall(s)
+            if not results:
+                return ()
+            result_list = []
+            root = Element('matches')
+            for s_match in results:
+                elem = SubElement(root, 'match')
+                elem.text = s_match
+                python.PyList_Append(result_list, elem)
+            return result_list
+        else:
+            result = rexpc.search(s)
+            if result is None:
+                return ()
+            root = Element('match')
+            root.text = result.group()
+            return (root,)
+
+    def replace(self, ctxt, s, rexp, flags, replacement):
+        rexpc = self._compile(rexp, 'i' in flags)
+        if 'g' in flags:
+            count = 0
+        else:
+            count = 1
+        return rexpc.sub(replacement, s, count)
+
+    cdef void _register_exslt_regexp(self, _BaseContext context):
+        ns = "http://exslt.org/regular-expressions"
+        context._registerExtensionFunction(ns, "test",    self.test)
+        context._registerExtensionFunction(ns, "match",   self.match)
+        context._registerExtensionFunction(ns, "replace", self.replace)
 
 ################################################################################
 # XSLT
@@ -249,6 +315,7 @@
     """
     cdef _XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
+    cdef _ExsltRegExp _regexp
     cdef object _doc_url_utf
     
     def __init__(self, xslt_input, extensions=None):
@@ -284,6 +351,7 @@
         self._c_style = c_style
 
         self._context = _XSLTContext(None, extensions)
+        self._regexp  = _ExsltRegExp()
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
         
     def __dealloc__(self):
@@ -333,6 +401,7 @@
 
         self._context._release_temp_refs()
         self._context.register_context(transform_ctxt, input_doc)
+        self._regexp._register_exslt_regexp(self._context)
 
         c_result = xslt.xsltApplyStylesheetUser(self._c_style, c_doc, params,
                                                 NULL, NULL, transform_ctxt)
From scoder at codespeak.net  Mon Apr 24 17:08:46 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Mon Apr 24 17:08:47 2006
Subject: [Lxml-checkins] r26271 - in lxml/trunk/src/lxml: . tests
Message-ID: <20060424150846.5F27F10093@code0.codespeak.net>

Author: scoder
Date: Mon Apr 24 17:08:45 2006
New Revision: 26271

Modified:
   lxml/trunk/src/lxml/tests/test_xslt.py
   lxml/trunk/src/lxml/xslt.pxi
Log:
clean up, check type of input arguments in exslt:regexp extension functions

Modified: lxml/trunk/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xslt.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xslt.py	Mon Apr 24 17:08:45 2006
@@ -422,8 +422,8 @@
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
   <xsl:template match="*">
     <test>
-      <test1><xsl:copy-of select="regexp:match(string(.), 'd.')"/></test1>
-      <test2><xsl:copy-of select="regexp:match(string(.), 'd.', 'g')"/></test2>
+      <test1><xsl:copy-of  select="regexp:match(string(.), 'd.')"/></test1>
+      <test2><xsl:copy-of  select="regexp:match(string(.), 'd.', 'g')"/></test2>
       <test2i><xsl:copy-of select="regexp:match(string(.), 'd.', 'gi')"/></test2i>
     </test>
   </xsl:template>

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Mon Apr 24 17:08:45 2006
@@ -208,8 +208,15 @@
     def __init__(self):
         self._compile_map = {}
 
+    cdef _make_string(self, value):
+        if python.PyString_Check(value) or python.PyUnicode_Check(value):
+            return value
+        else:
+            raise TypeError, "Invalid argument type %s" % type(value)
+
     cdef _compile(self, rexp, ignore_case):
         cdef python.PyObject* c_result
+        rexp = self._make_string(rexp)
         key = (rexp, ignore_case)
         c_result = python.PyDict_GetItem(self._compile_map, key)
         if c_result is not NULL:
@@ -223,6 +230,8 @@
         return rexp_compiled
 
     def test(self, ctxt, s, rexp, flags=''):
+        flags = self._make_string(flags)
+        s = self._make_string(s)
         rexpc = self._compile(rexp, 'i' in flags)
         if rexpc.search(s) is None:
             return False
@@ -230,6 +239,8 @@
             return True
 
     def match(self, ctxt, s, rexp, flags=''):
+        flags = self._make_string(flags)
+        s = self._make_string(s)
         rexpc = self._compile(rexp, 'i' in flags)
         if 'g' in flags:
             results = rexpc.findall(s)
@@ -251,6 +262,9 @@
             return (root,)
 
     def replace(self, ctxt, s, rexp, flags, replacement):
+        replacement = self._make_string(replacement)
+        flags = self._make_string(flags)
+        s = self._make_string(s)
         rexpc = self._compile(rexp, 'i' in flags)
         if 'g' in flags:
             count = 0
From scoder at codespeak.net  Tue Apr 25 08:05:58 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 08:06:00 2006
Subject: [Lxml-checkins] r26288 - in lxml/trunk: . src/lxml
Message-ID: <20060425060558.30E5F1007E@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 08:05:56 2006
New Revision: 26288

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/xslt.pxi
Log:
some cleanup in XPath/XSLT impl, reuse XPath contexts in XPath(): substantial speedup

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Tue Apr 25 08:05:56 2006
@@ -7,6 +7,8 @@
 Features added
 --------------
 
+* Substantial speedup in XPath.evaluate()
+
 * Exslt:regexp implementation for XSLT based on the Python 're' module
 
 * HTMLParser for parsing (broken) HTML
@@ -73,6 +75,8 @@
 * Support for registering extension functions through new FunctionNamespace
   class (see doc/extensions.txt)
 
+* ETXPath class for XPath expressions in ElementTree notation ('//{ns}tag')
+
 * Support for variables in XPath expressions (also in XPath class)
 
 * XPath class for compiled XPath expressions

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Tue Apr 25 08:05:56 2006
@@ -110,8 +110,8 @@
         self._free_context()
 
     cdef _free_context(self):
-        self._registered_namespaces = []
-        self._registered_extensions = []
+        del self._registered_namespaces[:]
+        del self._registered_extensions[:]
         python.PyDict_Clear(self._utf_refs)
         self._doc = None
         if self._xpathCtxt is not NULL:
@@ -122,9 +122,8 @@
 
     def addNamespace(self, prefix, uri):
         if self._namespaces is None:
-            self._namespaces = {prefix : uri}
-        else:
-            self._namespaces[prefix] = uri
+            self._namespaces = {}
+        python.PyDict_SetItem(self._namespaces, prefix, uri)
 
     def registerNamespaces(self, namespaces):
         for prefix, uri in namespaces.items():
@@ -134,7 +133,7 @@
         prefix_utf = self._to_utf(prefix)
         ns_uri_utf = self._to_utf(ns_uri)
         xpath.xmlXPathRegisterNs(self._xpathCtxt, prefix_utf, ns_uri_utf)
-        self._registered_namespaces.append(prefix_utf)
+        python.PyList_Append(self._registered_namespaces, prefix_utf)
 
     cdef _unregisterNamespaces(self):
         cdef xpath.xmlXPathContext* xpathCtxt
@@ -147,7 +146,8 @@
     def registerExtensionFunctions(self, extensions):
         for ns_uri, extension in extensions.items():
             for name, function in extension.items():
-                self.registerExtensionFunction(ns_uri, name, function)
+                self._registerExtensionFunction(
+                    self._to_utf(ns_uri), self._to_utf(name), function)
 
     def registerExtensionFunction(self, ns_uri, name, function):
         self._registerExtensionFunction(
@@ -293,14 +293,6 @@
         self._register_context(doc, 0)
         xsltCtxt.xpathCtxt.userData = <void*>self
 
-    cdef unregister_context(self):
-        cdef xslt.xsltTransformContext* xsltCtxt
-        xsltCtxt = self._xsltCtxt
-        if xsltCtxt is NULL:
-            return
-        self._unregister_context()
-        self._xsltCtxt = NULL
-
     cdef free_context(self):
         cdef xslt.xsltTransformContext* xsltCtxt
         xsltCtxt = self._xsltCtxt
@@ -471,7 +463,7 @@
     cdef object _registered_variables
     def __init__(self, namespaces, extensions, variables):
         _BaseContext.__init__(self, namespaces, extensions)
-        self._variables  = variables
+        self._variables = variables
         self._registered_variables  = []
         
     cdef register_context(self, xpath.xmlXPathContext* xpathCtxt, _Document doc):
@@ -488,46 +480,37 @@
         xpathCtxt = self._xpathCtxt
         if xpathCtxt is NULL:
             return
-        xpathCtxt.userData = NULL
-        self._unregister_context()
         self._unregisterVariables()
         self._registered_variables  = []
-        self._xpathCtxt = NULL
+        self._unregister_context()
 
-    cdef free_context(self):
+    cdef void _unregisterVariables(self):
         cdef xpath.xmlXPathContext* xpathCtxt
+        cdef xpath.xmlXPathObject* xpathVarValue
+        cdef char* c_name
         xpathCtxt = self._xpathCtxt
-        if xpathCtxt is NULL:
-            return
-        self._free_context()
-        self._registered_variables  = []
-        xpath.xmlXPathFreeContext(xpathCtxt)
+        for name_utf in self._registered_variables:
+            c_name = _cstr(name_utf)
+            xpathVarValue = xpath.xmlXPathVariableLookup(xpathCtxt, c_name)
+            if xpathVarValue is not NULL:
+                xpath.xmlXPathRegisterVariable(xpathCtxt, c_name, NULL)
+                xpath.xmlXPathFreeObject(xpathVarValue)
 
     def registerVariables(self, variable_dict):
         for name, value in variable_dict.items():
-            self.registerVariable(name, value)
-
-    cdef void _unregisterVariables(self):
-        for name in self._registered_variables:
-            self._unregisterVariable(name)
+            name_utf = self._to_utf(name)
+            self._registerVariable(name_utf, value)
+            python.PyList_Append(self._registered_variables, name_utf)
 
     def registerVariable(self, name, value):
-        self._registerVariable(self._to_utf(name), value)
-        self._registered_variables.append(name)
+        name_utf = self._to_utf(name)
+        self._registerVariable(name_utf, value)
+        python.PyList_Append(self._registered_variables, name_utf)
 
     cdef void _registerVariable(self, name_utf, value):
         xpath.xmlXPathRegisterVariable(
             self._xpathCtxt, _cstr(name_utf), _wrapXPathObject(value))
 
-    cdef void _unregisterVariable(self, name_utf):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject* xpathVarValue
-        xpathCtxt = self._xpathCtxt
-        xpathVarValue = xpath.xmlXPathVariableLookup(xpathCtxt, _cstr(name_utf))
-        if xpathVarValue is not NULL:
-            xpath.xmlXPathRegisterVariable(xpathCtxt, _cstr(name_utf), NULL)
-            xpath.xmlXPathFreeObject(xpathVarValue)
-
     def _contextRegisterExtensionFunction(self, ns_uri_utf, name_utf):
         if ns_uri_utf is not None:
             xpath.xmlXPathRegisterFuncNS(
@@ -598,7 +581,8 @@
         XPathEvaluatorBase.__init__(self, namespaces, extensions)
 
     def __dealloc__(self):
-        xpath.xmlXPathFreeContext(self._c_ctxt)
+        if self._c_ctxt is not NULL:
+            xpath.xmlXPathFreeContext(self._c_ctxt)
     
     def registerNamespace(self, prefix, uri):
         """Register a namespace with the XPath context.
@@ -670,6 +654,7 @@
     return {ns_uri : functions}
 
 cdef class XPath(XPathEvaluatorBase):
+    cdef xpath.xmlXPathContext* _xpathCtxt
     cdef xpath.xmlXPathCompExpr* _xpath
     cdef object _prefix_map
     cdef readonly object path
@@ -681,6 +666,7 @@
         self._xpath = xpath.xmlXPathCompile(_cstr(path))
         if self._xpath is NULL:
             raise XPathSyntaxError, "Error in xpath expression."
+        self._xpathCtxt = xpath.xmlXPathNewContext(NULL)
 
     def __call__(self, _etree_or_element, **_variables):
         cdef xpath.xmlXPathContext* xpathCtxt
@@ -692,7 +678,8 @@
         document = _documentOrRaise(_etree_or_element)
         element  = _rootNodeOf(_etree_or_element)
 
-        xpathCtxt = xpath.xmlXPathNewContext(document._c_doc)
+        xpathCtxt = self._xpathCtxt
+        xpathCtxt.doc = document._c_doc
         xpathCtxt.node = element._c_node
 
         context = self._context
@@ -701,17 +688,15 @@
         context.registerVariables(_variables)
 
         xpathObj = xpath.xmlXPathCompiledEval(self._xpath, xpathCtxt)
-
         context.unregister_context()
-
-        xpath.xmlXPathFreeContext(xpathCtxt)
-
         return self._handle_result(xpathObj, document)
 
     def evaluate(self, _tree, **_variables):
         return self(_tree, **_variables)
 
     def __dealloc__(self):
+        if self._xpathCtxt is not NULL:
+            xpath.xmlXPathFreeContext(self._xpathCtxt)
         if self._xpath is not NULL:
             xpath.xmlXPathFreeCompExpr(self._xpath)
 
From scoder at codespeak.net  Tue Apr 25 10:17:51 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 10:17:53 2006
Subject: [Lxml-checkins] r26306 - lxml/branch/py_ssize_t
Message-ID: <20060425081751.255C01007D@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 10:17:49 2006
New Revision: 26306

Added:
   lxml/branch/py_ssize_t/
      - copied from r26305, lxml/trunk/
Log:
new branch for Py_ssize_t migration
From scoder at codespeak.net  Tue Apr 25 13:08:13 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 13:08:14 2006
Subject: [Lxml-checkins] r26323 - lxml/trunk
Message-ID: <20060425110813.45CB110088@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 13:08:12 2006
New Revision: 26323

Modified:
   lxml/trunk/Makefile
Log:
allow overriding python executable in Makefile by passing PYTHON variable

Modified: lxml/trunk/Makefile
==============================================================================
--- lxml/trunk/Makefile	(original)
+++ lxml/trunk/Makefile	Tue Apr 25 13:08:12 2006
@@ -1,4 +1,4 @@
-PYTHON=python
+PYTHON?=python
 TESTFLAGS=-p -v
 TESTOPTS=
 SETUPFLAGS=
From scoder at codespeak.net  Tue Apr 25 14:06:04 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 14:06:05 2006
Subject: [Lxml-checkins] r26329 - lxml/branch/py_ssize_t/src/lxml
Message-ID: <20060425120604.78EBA1007E@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 14:06:02 2006
New Revision: 26329

Modified:
   lxml/branch/py_ssize_t/src/lxml/etree.pyx
   lxml/branch/py_ssize_t/src/lxml/parser.pxi
   lxml/branch/py_ssize_t/src/lxml/python.pxd
   lxml/branch/py_ssize_t/src/lxml/xslt.pxi
Log:
Py_ssize_t fixes for Python 2.5, needs patched Pyrex

Modified: lxml/branch/py_ssize_t/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/py_ssize_t/src/lxml/etree.pyx	(original)
+++ lxml/branch/py_ssize_t/src/lxml/etree.pyx	Tue Apr 25 14:06:02 2006
@@ -368,7 +368,7 @@
 
     # MANIPULATORS
 
-    def __setitem__(self, index, _NodeBase element):
+    def __setitem__(self, Py_ssize_t index, _NodeBase element):
         cdef xmlNode* c_node
         cdef xmlNode* c_next
         cdef int foreign
@@ -382,7 +382,7 @@
         _moveTail(c_next, element._c_node)
         changeDocumentBelow(element, self._doc, foreign)
         
-    def __delitem__(self, index):
+    def __delitem__(self, Py_ssize_t index):
         cdef xmlNode* c_node
         c_node = _findChild(self._c_node, index)
         if c_node is NULL:
@@ -390,12 +390,12 @@
         _removeText(c_node.next)
         _removeNode(c_node)
 
-    def __delslice__(self, start, stop):
+    def __delslice__(self, Py_ssize_t start, Py_ssize_t stop):
         cdef xmlNode* c_node
         c_node = _findChild(self._c_node, start)
         _deleteSlice(c_node, start, stop)
         
-    def __setslice__(self, start, stop, value):
+    def __setslice__(self, Py_ssize_t start, Py_ssize_t stop, value):
         cdef xmlNode* c_node
         cdef xmlNode* c_next
         cdef _Element mynode
@@ -576,17 +576,17 @@
     def __repr__(self):
         return "<Element %s at %x>" % (self.tag, id(self))
     
-    def __getitem__(self, index):
+    def __getitem__(self, Py_ssize_t index):
         cdef xmlNode* c_node
         c_node = _findChild(self._c_node, index)
         if c_node is NULL:
             raise IndexError, "list index out of range"
         return _elementFactory(self._doc, c_node)
 
-    def __getslice__(self, start, stop):
+    def __getslice__(self, Py_ssize_t start, Py_ssize_t stop):
         cdef xmlNode* c_node
         cdef _Document doc
-        cdef int c, c_stop
+        cdef Py_ssize_t c
         # this does not work for negative start, stop, however,
         # python seems to convert these to positive start, stop before
         # calling, so this all works perfectly (at the cost of a len() call)
@@ -594,10 +594,9 @@
         if c_node is NULL:
             return []
         c = start
-        c_stop = stop
         result = []
         doc = self._doc
-        while c_node is not NULL and c < c_stop:
+        while c_node is not NULL and c < stop:
             if _isElement(c_node):
                 ret = python.PyList_Append(result, _elementFactory(doc, c_node))
                 if ret:
@@ -607,7 +606,7 @@
         return result
             
     def __len__(self):
-        cdef int c
+        cdef Py_ssize_t c
         cdef xmlNode* c_node
         c = 0
         c_node = self._c_node.children
@@ -629,10 +628,8 @@
         return ElementChildIterator(self, reversed=True)
 
     def index(self, _Element x not None, start=None, stop=None):
-        cdef int k
-        cdef int l
-        cdef int c_stop
-        cdef int c_start
+        cdef Py_ssize_t k, l
+        cdef Py_ssize_t c_start, c_stop
         cdef xmlNode* c_child
         cdef xmlNode* c_start_node
         c_child = x._c_node
@@ -693,7 +690,7 @@
                     return k
             else:
                 return k
-        if c_start or c_stop:
+        if c_start != 0 or c_stop != 0:
             raise ValueError, "list.index(x): x not in slice"
         else:
             raise ValueError, "list.index(x): x not in list"
@@ -915,7 +912,7 @@
         return result
 
     def __len__(self):
-        cdef int c
+        cdef Py_ssize_t c
         cdef xmlNode* c_node
         c = 0
         c_node = <xmlNode*>(self._c_node.properties)
@@ -1431,7 +1428,7 @@
     
     If there was no text to collect, return None
     """
-    cdef int scount
+    cdef Py_ssize_t scount
     cdef char* text
     cdef xmlNode* c_node_cur
     # check for multiple text nodes
@@ -1470,17 +1467,17 @@
         tree.xmlFreeNode(c_node)
         c_node = c_next
 
-cdef xmlNode* _findChild(xmlNode* c_node, int index):
+cdef xmlNode* _findChild(xmlNode* c_node, Py_ssize_t index):
     if index < 0:
         return _findChildBackwards(c_node, -index - 1)
     else:
         return _findChildForwards(c_node, index)
     
-cdef xmlNode* _findChildForwards(xmlNode* c_node, int index):
+cdef xmlNode* _findChildForwards(xmlNode* c_node, Py_ssize_t index):
     """Return child element of c_node with index, or return NULL if not found.
     """
     cdef xmlNode* c_child
-    cdef int c
+    cdef Py_ssize_t c
     c_child = c_node.children
     c = 0
     while c_child is not NULL:
@@ -1492,12 +1489,12 @@
     else:
         return NULL
 
-cdef xmlNode* _findChildBackwards(xmlNode* c_node, int index):
+cdef xmlNode* _findChildBackwards(xmlNode* c_node, Py_ssize_t index):
     """Return child element of c_node with index, or return NULL if not found.
     Search from the end.
     """
     cdef xmlNode* c_child
-    cdef int c
+    cdef Py_ssize_t c
     c_child = c_node.last
     c = 0
     while c_child is not NULL:
@@ -1551,11 +1548,11 @@
 ##     return (c_node.type == tree.XML_ELEMENT_NODE or
 ##             c_node.type == tree.XML_COMMENT_NODE)
 
-cdef xmlNode* _deleteSlice(xmlNode* c_node, int start, int stop):
+cdef xmlNode* _deleteSlice(xmlNode* c_node, Py_ssize_t start, Py_ssize_t stop):
     """Delete slice, starting with c_node, start counting at start, end at stop.
     """
     cdef xmlNode* c_next
-    cdef int c
+    cdef Py_ssize_t c
     if c_node is NULL:
         return NULL
     # now start deleting nodes
@@ -1571,7 +1568,7 @@
     return c_node
 
 cdef int isutf8(char* string):
-    cdef int i
+    cdef Py_ssize_t i
     i = 0
     while 1:
         if string[i] == c'\0':
@@ -1600,7 +1597,7 @@
     """
     cdef char* c_tag
     cdef char* c_pos
-    cdef int nslen
+    cdef Py_ssize_t nslen
     if isinstance(tag, QName):
         tag = (<QName>tag).text
     tag = _utf8(tag)

Modified: lxml/branch/py_ssize_t/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/py_ssize_t/src/lxml/parser.pxi	(original)
+++ lxml/branch/py_ssize_t/src/lxml/parser.pxi	Tue Apr 25 14:06:02 2006
@@ -254,7 +254,6 @@
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
         cdef char* c_text
-        cdef int c_len
         self._error_log.connect()
         c_text = _cstr(text_utf)
         pctxt = self._memory_parser_ctxt

Modified: lxml/branch/py_ssize_t/src/lxml/python.pxd
==============================================================================
--- lxml/branch/py_ssize_t/src/lxml/python.pxd	(original)
+++ lxml/branch/py_ssize_t/src/lxml/python.pxd	Tue Apr 25 14:06:02 2006
@@ -12,13 +12,13 @@
 
     cdef object PyUnicode_FromEncodedObject(object s, char* encoding,
                                             char* errors)
-    cdef object PyUnicode_DecodeUTF8(char* s, int size, char* errors)
+    cdef object PyUnicode_DecodeUTF8(char* s, Py_ssize_t size, char* errors)
     cdef object PyUnicode_AsUTF8String(object ustring)
-    cdef object PyString_FromStringAndSize(char* s, int size)
+    cdef object PyString_FromStringAndSize(char* s, Py_ssize_t size)
     cdef object PyString_FromString(char* s)
     cdef object PyString_FromFormat(char* format, ...)
 
-    cdef int PyList_GET_SIZE(object l)
+    cdef Py_ssize_t PyList_GET_SIZE(object l)
     cdef int PyList_Append(object l, object obj)
     cdef int PyDict_SetItemString(object d, char* key, object value)
     cdef int PyDict_SetItem(object d, object key, object value)
@@ -29,7 +29,7 @@
     cdef object PyList_AsTuple(object o)
     cdef object PySequence_List(object o)
     cdef object PySequence_Tuple(object o)
-    cdef object PyTuple_GET_ITEM(object o, int pos)
+    cdef object PyTuple_GET_ITEM(object o, Py_ssize_t pos)
 
     cdef int PyNumber_Check(object instance)
     cdef int PyBool_Check(object instance)

Modified: lxml/branch/py_ssize_t/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/py_ssize_t/src/lxml/xslt.pxi	(original)
+++ lxml/branch/py_ssize_t/src/lxml/xslt.pxi	Tue Apr 25 14:06:02 2006
@@ -372,8 +372,7 @@
         cdef xmlDoc* c_result
         cdef xmlDoc* c_doc
         cdef char** params
-        cdef int i
-        cdef int j
+        cdef Py_ssize_t i
 
         input_doc = _documentOrRaise(_input)
         root_node = _rootNodeOf(_input)
@@ -816,6 +815,7 @@
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
     cdef _BaseContext extensions
+    cdef int i
 
     rctxt = ctxt.context
 
From scoder at codespeak.net  Tue Apr 25 17:27:35 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 17:27:37 2006
Subject: [Lxml-checkins] r26339 - lxml/trunk/src/lxml
Message-ID: <20060425152735.BD0C81008B@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 17:27:34 2006
New Revision: 26339

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
always use unicode regexps (it's XML after all)

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Tue Apr 25 17:27:34 2006
@@ -221,10 +221,9 @@
         c_result = python.PyDict_GetItem(self._compile_map, key)
         if c_result is not NULL:
             return <object>c_result
+        py_flags = re.UNICODE
         if ignore_case:
-            py_flags = re.IGNORECASE
-        else:
-            py_flags = 0
+            py_flags = py_flags | re.IGNORECASE
         rexp_compiled = RE_COMPILE(rexp, py_flags)
         python.PyDict_SetItem(self._compile_map, key, rexp_compiled)
         return rexp_compiled
From scoder at codespeak.net  Tue Apr 25 17:29:59 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 17:30:03 2006
Subject: [Lxml-checkins] r26340 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060425152959.839101008B@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 17:29:55 2006
New Revision: 26340

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
do not register all extra functions, just 'node-set()' to prevent potential security risks through write/output/debug extra elements

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxd	Tue Apr 25 17:29:55 2006
@@ -18,8 +18,11 @@
 cdef extern from "libxslt/extensions.h":
     cdef int xsltRegisterExtFunction(xsltTransformContext* ctxt,
                                      char* name,
-                                     char * URI,
+                                     char* URI,
                                      xmlXPathFunction function)
+    cdef int xsltRegisterExtModuleFunction(char* name, char* URI,
+                                           xmlXPathFunction function)
+    cdef int xsltUnregisterExtModuleFunction(char* name, char* URI)
 
 cdef extern from "libxslt/documents.h":
     ctypedef enum xsltLoadType:
@@ -32,7 +35,7 @@
                                           void* ctxt,
                                           xsltLoadType type)
     cdef xsltDocLoaderFunc xsltDocDefaultLoader
-    cdef void xsltSetLoaderFunc	(xsltDocLoaderFunc f)
+    cdef void xsltSetLoaderFunc(xsltDocLoaderFunc f)
 
 cdef extern from "libxslt/transform.h":
     cdef xmlDoc* xsltApplyStylesheet(xsltStylesheet* style, xmlDoc* doc,
@@ -57,6 +60,13 @@
                                         void* ctxt,
                                         void (*handler)(void* ctxt, char* msg, ...))
 
+cdef extern from "libxslt/extra.h":
+    cdef char* XSLT_LIBXSLT_NAMESPACE
+    cdef char* XSLT_XALAN_NAMESPACE
+    cdef char* XSLT_SAXON_NAMESPACE
+    cdef char* XSLT_XT_NAMESPACE
+
+    cdef xmlXPathFunction xsltFunctionNodeSet
     cdef void xsltRegisterAllExtras()
 
 cdef extern from "libexslt/exslt.h":

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Tue Apr 25 17:29:55 2006
@@ -482,10 +482,23 @@
     result._xslt = xslt
     return result
 
-# enable EXSLT and Saxon support for XSLT
-xslt.xsltRegisterAllExtras()
+# enable EXSLT support for XSLT
 xslt.exsltRegisterAll()
 
+# do not register all libxslt extra function, provide only "node-set"
+# functions like "output" and "write" are a potential security risk
+#xslt.xsltRegisterAllExtras()
+xslt.xsltRegisterExtModuleFunction("node-set",
+                                   xslt.XSLT_LIBXSLT_NAMESPACE,
+                                   xslt.xsltFunctionNodeSet);
+xslt.xsltRegisterExtModuleFunction("node-set",
+                                   xslt.XSLT_SAXON_NAMESPACE,
+                                   xslt.xsltFunctionNodeSet);
+xslt.xsltRegisterExtModuleFunction("node-set",
+                                   xslt.XSLT_XT_NAMESPACE,
+                                   xslt.xsltFunctionNodeSet);
+
+
 ################################################################################
 # XPath
 
From scoder at codespeak.net  Tue Apr 25 18:07:30 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Tue Apr 25 18:07:33 2006
Subject: [Lxml-checkins] r26341 - lxml/trunk/src/lxml
Message-ID: <20060425160730.D30581009F@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 18:07:28 2006
New Revision: 26341

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
clean up in XPath context/extension function registration

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Tue Apr 25 18:07:28 2006
@@ -36,7 +36,6 @@
     cdef object _extensions
     cdef object _namespaces
     cdef object _registered_namespaces
-    cdef object _registered_extensions
     cdef object _extension_functions
     cdef object _utf_refs
     # for exception handling and temporary reference keeping:
@@ -66,7 +65,6 @@
         self._extensions = extensions
         self._namespaces = namespaces
         self._registered_namespaces = []
-        self._registered_extensions = []
         self._extension_functions = {}
         self._temp_elements = {}
         self._temp_docs = {}
@@ -111,7 +109,7 @@
 
     cdef _free_context(self):
         del self._registered_namespaces[:]
-        del self._registered_extensions[:]
+        python.PyDict_Clear(self._extension_functions)
         python.PyDict_Clear(self._utf_refs)
         self._doc = None
         if self._xpathCtxt is not NULL:
@@ -161,10 +159,9 @@
     cdef _registerExtensionFunction(self, ns_uri_utf, name_utf, function):
         self._contextRegisterExtensionFunction(ns_uri_utf, name_utf)
         self._extension_functions[(ns_uri_utf, name_utf)] = function
-        self._registered_extensions.append((ns_uri_utf, name_utf))
 
     cdef _unregisterExtensionFunctions(self):
-        for ns_uri_utf, name_utf in self._registered_extensions:
+        for ns_uri_utf, name_utf in self._extension_functions:
             self._contextUnregisterExtensionFunction(ns_uri_utf, name_utf)
 
     def find_extension(self, ns_uri_utf, name_utf):
From scoder at codespeak.net  Wed Apr 26 12:38:25 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 26 12:38:26 2006
Subject: [Lxml-checkins] r26357 - lxml/trunk/src/lxml
Message-ID: <20060426103825.1874810084@code0.codespeak.net>

Author: scoder
Date: Wed Apr 26 12:38:22 2006
New Revision: 26357

Modified:
   lxml/trunk/src/lxml/nsclasses.pxi
   lxml/trunk/src/lxml/xslt.pxd
   lxml/trunk/src/lxml/xslt.pxi
Log:
rewrite of internal extension function registration
* register XSLT functions at module level directly in FunctionNamespace with xsltRegisterExtModuleFunction
* lookup XPath functions at XPath run-time rather than registering everything at call-time
* new 'regexp' keyword for XSLT() to switch off regexp support


Modified: lxml/trunk/src/lxml/nsclasses.pxi
==============================================================================
--- lxml/trunk/src/lxml/nsclasses.pxi	(original)
+++ lxml/trunk/src/lxml/nsclasses.pxi	Wed Apr 26 12:38:22 2006
@@ -53,11 +53,19 @@
 cdef class _NamespaceRegistry:
     "Dictionary-like registry for namespace implementations"
     cdef object _ns_uri
+    cdef object _ns_uri_utf
     cdef object _classes
     cdef object _extensions
     cdef object _xslt_elements
+    cdef char* _c_ns_uri_utf
     def __init__(self, ns_uri):
         self._ns_uri = ns_uri
+        if ns_uri is None:
+            self._ns_uri_utf = None
+            self._c_ns_uri_utf = NULL
+        else:
+            self._ns_uri_utf = _utf8(ns_uri)
+            self._c_ns_uri_utf = _cstr(self._ns_uri_utf)
         self._classes = {}
         self._extensions = {}
         self._xslt_elements = {}
@@ -95,14 +103,11 @@
         d[name_utf] = item
 
     def __getitem__(self, name):
-        name_utf = _utf8(name)
-        return self._get(name_utf)
-
-    cdef object _get(self, object name):
         cdef python.PyObject* dict_result
-        dict_result = python.PyDict_GetItem(self._classes, name)
+        name_utf = _utf8(name)
+        dict_result = python.PyDict_GetItem(self._classes, name_utf)
         if dict_result is NULL:
-            dict_result = python.PyDict_GetItem(self._extensions, name)
+            dict_result = python.PyDict_GetItem(self._extensions, name_utf)
         if dict_result is NULL:
             raise KeyError, "Name not registered."
         return <object>dict_result
@@ -130,34 +135,34 @@
             self._prefix_utf = _utf8(prefix)
             self._prefix = prefix
 
-    def __setitem__(self, name, item):
-        if not callable(item):
+    def __setitem__(self, name, function):
+        if not callable(function):
             raise NamespaceRegistryError, "Registered function must be callable."
         if name is None:
             name_utf = None
         else:
             name_utf = _utf8(name)
-        self._extensions[name_utf] = item
+        self._extensions[name_utf] = function
+        _register_global_xslt_function(self._c_ns_uri_utf, _cstr(name_utf))
 
-    cdef object _get(self, object name):
+    def __getitem__(self, name):
         cdef python.PyObject* dict_result
-        dict_result = python.PyDict_GetItem(self._extensions, name)
+        name_utf = _utf8(name)
+        dict_result = python.PyDict_GetItem(self._extensions, name_utf)
         if dict_result is NULL:
             raise KeyError, "Name not registered."
         return <object>dict_result
 
+    def clear(self):
+        cdef char* c_uri_utf
+        c_uri_utf = self._c_ns_uri_utf
+        for name_utf in self._extensions:
+            _unregister_global_xslt_function(c_uri_utf, _cstr(name_utf))
+        _NamespaceRegistry.clear(self)
+
     def __repr__(self):
         return "FunctionNamespace(%r)" % self._ns_uri
 
-cdef object _find_all_extensions():
-    "Internal lookup function to find all extension functions for XSLT/XPath."
-    cdef _NamespaceRegistry registry
-    ns_extensions = {}
-    for (ns_utf, registry) in __FUNCTION_NAMESPACE_REGISTRIES.iteritems():
-        if registry._extensions:
-            ns_extensions[ns_utf] = registry._extensions
-    return ns_extensions
-
 cdef object _find_all_extension_prefixes():
     "Internal lookup function to find all function prefixes for XSLT/XPath."
     cdef _FunctionNamespaceRegistry registry
@@ -167,25 +172,18 @@
             ns_prefixes[registry._prefix_utf] = ns_utf
     return ns_prefixes
 
-cdef _find_extensions(namespaces):
-    """Returns a dictionary that maps each namespace in the provided list to a
-    dictionary of name-function mappings defined under that namespace."""
+cdef object _find_extension(ns_uri_utf, name_utf):
     cdef python.PyObject* dict_result
-    cdef char* c_ns_utf
-    extension_dict = {}
-    for ns_uri in namespaces:
-        if ns_uri is None:
-            ns_utf = None
-        else:
-            ns_utf = _utf8(ns_uri)
-        dict_result = python.PyDict_GetItem(
-            __FUNCTION_NAMESPACE_REGISTRIES, ns_utf)
-        if dict_result is NULL:
-            continue
-        extensions = (<_NamespaceRegistry>dict_result)._extensions
-        if extensions:
-            python.PyDict_SetItem(extension_dict, ns_utf, extensions)
-    return extension_dict
+    dict_result = python.PyDict_GetItem(
+        __FUNCTION_NAMESPACE_REGISTRIES, ns_uri_utf)
+    if dict_result is NULL:
+        return None
+    extensions = (<_NamespaceRegistry>dict_result)._extensions
+    dict_result = python.PyDict_GetItem(extensions, name_utf)
+    if dict_result is NULL:
+        return None
+    else:
+        return <object>dict_result
 
 cdef object _find_element_class(char* c_namespace_utf,
                                 char* c_element_name_utf):

Modified: lxml/trunk/src/lxml/xslt.pxd
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxd	(original)
+++ lxml/trunk/src/lxml/xslt.pxd	Wed Apr 26 12:38:22 2006
@@ -19,6 +19,9 @@
                                      char* name,
                                      char * URI,
                                      xmlXPathFunction function)
+    cdef int xsltRegisterExtModuleFunction(char* name, char* URI,
+                                           xmlXPathFunction function)
+    cdef int xsltUnregisterExtModuleFunction(char* name, char* URI)
 
 cdef extern from "libxslt/transform.h":
     cdef xmlDoc* xsltApplyStylesheet(xsltStylesheet* style, xmlDoc* doc,

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Wed Apr 26 12:38:22 2006
@@ -36,8 +36,8 @@
     cdef object _extensions
     cdef object _namespaces
     cdef object _registered_namespaces
-    cdef object _extension_functions
     cdef object _utf_refs
+    cdef object _temp_last_function
     # for exception handling and temporary reference keeping:
     cdef object _temp_elements
     cdef object _temp_docs
@@ -46,8 +46,9 @@
     def __init__(self, namespaces, extensions):
         self._xpathCtxt = NULL
         self._utf_refs = {}
+        self._temp_last_function = (None, None, None)
 
-        # fix old format extensions
+        # convert old format extensions to UTF-8
         if isinstance(extensions, (list, tuple)):
             new_extensions = {}
             for extension in extensions:
@@ -65,7 +66,6 @@
         self._extensions = extensions
         self._namespaces = namespaces
         self._registered_namespaces = []
-        self._extension_functions = {}
         self._temp_elements = {}
         self._temp_docs = {}
 
@@ -88,28 +88,19 @@
     cdef _register_context(self, _Document doc, int allow_none_namespace):
         self._doc      = doc
         self._exc_info = None
+        self._temp_last_function = (None, None, None)
         namespaces = self._namespaces
         if namespaces is not None:
             self.registerNamespaces(namespaces)
-            extensions = _find_extensions(namespaces.values())
-        else:
-            extensions = _find_all_extensions()
-        if self._extensions is not None:
-            # add user provided extensions
-            extensions.update(self._extensions)
-        if extensions:
-            if not allow_none_namespace:
-                python.PyDict_DelItem(extensions, None)
-            self._registerExtensionFunctions(extensions)
+        xpath.xmlXPathRegisterFuncLookup(self._xpathCtxt, _function_check,
+                                         <python.PyObject*>self)
 
     cdef _unregister_context(self):
-        self._unregisterExtensionFunctions()
         self._unregisterNamespaces()
         self._free_context()
 
     cdef _free_context(self):
         del self._registered_namespaces[:]
-        python.PyDict_Clear(self._extension_functions)
         python.PyDict_Clear(self._utf_refs)
         self._doc = None
         if self._xpathCtxt is not NULL:
@@ -139,33 +130,25 @@
         for prefix_utf in self._registered_namespaces:
             xpath.xmlXPathRegisterNs(xpathCtxt, prefix_utf, NULL)
     
-    # extension functions (internal UTF-8 methods with leading '_')
-
-    def registerExtensionFunctions(self, extensions):
-        for ns_uri, extension in extensions.items():
-            for name, function in extension.items():
-                self._registerExtensionFunction(
-                    self._to_utf(ns_uri), self._to_utf(name), function)
-
-    def registerExtensionFunction(self, ns_uri, name, function):
-        self._registerExtensionFunction(
-            self._to_utf(ns_uri), self._to_utf(name), function)
+    # extension functions
 
-    cdef _registerExtensionFunctions(self, extensions_utf):
-        for ns_uri_utf, extension in extensions_utf.items():
-            for name_utf, function in extension.items():
-                self._registerExtensionFunction(ns_uri_utf, name_utf, function)
+    cdef _lookup_extension(self, ns_uri_utf, name_utf):
+        cdef python.PyObject* dict_result
+        if self._temp_last_function[0] == ns_uri_utf and \
+           self._temp_last_function[1] == name_utf:
+            return self._temp_last_function[2]
 
-    cdef _registerExtensionFunction(self, ns_uri_utf, name_utf, function):
-        self._contextRegisterExtensionFunction(ns_uri_utf, name_utf)
-        self._extension_functions[(ns_uri_utf, name_utf)] = function
-
-    cdef _unregisterExtensionFunctions(self):
-        for ns_uri_utf, name_utf in self._extension_functions:
-            self._contextUnregisterExtensionFunction(ns_uri_utf, name_utf)
+        dict_result = python.PyDict_GetItem(self._extensions, ns_uri_utf)
+        if dict_result is not NULL:
+            dict_result = python.PyDict_GetItem(<object>dict_result, name_utf)
+        if dict_result is not NULL:
+            function = <object>dict_result
+        else:
+            function = _find_extension(ns_uri_utf, name_utf)
 
-    def find_extension(self, ns_uri_utf, name_utf):
-        return self._extension_functions[(ns_uri_utf, name_utf)]
+        # store temporarily as it will be looked up again in the next call
+        self._temp_last_function = (ns_uri_utf, name_utf, function)
+        return function
 
     # Python reference keeping during XPath function evaluation
 
@@ -194,86 +177,27 @@
                 #print "Holding document:", <int>element._doc._c_doc
                 python.PyDict_SetItem(self._temp_docs, id(element._doc), element._doc)
 
-################################################################################
-# EXSLT regexp implementation
-
-cdef object RE_COMPILE
-RE_COMPILE = re.compile
-
-cdef class _ExsltRegExp:
-    cdef object _compile_map
-    def __init__(self):
-        self._compile_map = {}
-
-    cdef _make_string(self, value):
-        if python.PyString_Check(value) or python.PyUnicode_Check(value):
-            return value
-        else:
-            raise TypeError, "Invalid argument type %s" % type(value)
-
-    cdef _compile(self, rexp, ignore_case):
-        cdef python.PyObject* c_result
-        rexp = self._make_string(rexp)
-        key = (rexp, ignore_case)
-        c_result = python.PyDict_GetItem(self._compile_map, key)
-        if c_result is not NULL:
-            return <object>c_result
-        py_flags = re.UNICODE
-        if ignore_case:
-            py_flags = py_flags | re.IGNORECASE
-        rexp_compiled = RE_COMPILE(rexp, py_flags)
-        python.PyDict_SetItem(self._compile_map, key, rexp_compiled)
-        return rexp_compiled
-
-    def test(self, ctxt, s, rexp, flags=''):
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if rexpc.search(s) is None:
-            return False
-        else:
-            return True
-
-    def match(self, ctxt, s, rexp, flags=''):
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if 'g' in flags:
-            results = rexpc.findall(s)
-            if not results:
-                return ()
-            result_list = []
-            root = Element('matches')
-            for s_match in results:
-                elem = SubElement(root, 'match')
-                elem.text = s_match
-                python.PyList_Append(result_list, elem)
-            return result_list
-        else:
-            result = rexpc.search(s)
-            if result is None:
-                return ()
-            root = Element('match')
-            root.text = result.group()
-            return (root,)
+cdef xpath.xmlXPathFunction _function_check(void* ctxt, char* c_name, char* c_ns_uri):
+    cdef _BaseContext context
+    if c_name is NULL:
+        return NULL
+    if c_ns_uri is NULL:
+        ns_uri = None
+    else:
+        ns_uri = c_ns_uri
+    context = <_BaseContext>ctxt
+    if context._lookup_extension(ns_uri, c_name) is None:
+        return NULL
+    else:
+        return _xpath_function_call
 
-    def replace(self, ctxt, s, rexp, flags, replacement):
-        replacement = self._make_string(replacement)
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if 'g' in flags:
-            count = 0
-        else:
-            count = 1
-        return rexpc.sub(replacement, s, count)
+cdef void _register_global_xslt_function(char* ns_uri, char* name):
+    xslt.xsltRegisterExtModuleFunction(ns_uri, name, _xpath_function_call)
 
-    cdef void _register_exslt_regexp(self, _BaseContext context):
-        ns = "http://exslt.org/regular-expressions"
-        context._registerExtensionFunction(ns, "test",    self.test)
-        context._registerExtensionFunction(ns, "match",   self.match)
-        context._registerExtensionFunction(ns, "replace", self.replace)
+cdef void _unregister_global_xslt_function(char* ns_uri, char* name):
+    xslt.xsltUnRegisterExtModuleFunction(ns_uri, name)
 
+ 
 ################################################################################
 # XSLT
 
@@ -281,13 +205,17 @@
     cdef xslt.xsltTransformContext* _xsltCtxt
     def __init__(self, namespaces, extensions):
         self._xsltCtxt = NULL
+        if extensions and None in extensions:
+            raise XSLTExtensionError, "extensions must have non-empty namespaces"
         _BaseContext.__init__(self, namespaces, extensions)
 
-    cdef register_context(self, xslt.xsltTransformContext* xsltCtxt, _Document doc):
+    cdef register_context(self, xslt.xsltTransformContext* xsltCtxt,
+                               _Document doc):
         self._xsltCtxt = xsltCtxt
         self._set_xpath_context(xsltCtxt.xpathCtxt)
         self._register_context(doc, 0)
         xsltCtxt.xpathCtxt.userData = <void*>self
+        self._registerLocalExtensionFunctions()
 
     cdef free_context(self):
         cdef xslt.xsltTransformContext* xsltCtxt
@@ -298,19 +226,31 @@
         self._xsltCtxt = NULL
         xslt.xsltFreeTransformContext(xsltCtxt)
 
-    def _contextRegisterExtensionFunction(self, ns_uri_utf, name_utf):
-        if ns_uri_utf is None:
-            raise XSLTExtensionError, "extensions must have non-empty namespaces"
+    cdef _registerLocalExtensionFunction(self, ns_utf, name_utf, function):
+        extensions = self._extensions
+        if self._extensions is None:
+            self._extensions = {ns_utf:{name_utf:function}}
+        else:
+            if ns_utf in self._extensions:
+                self._extensions[ns_utf][name_utf] = function
+            else:
+                self._extensions[ns_utf] = ns_extensions = {name_utf:function}
         xslt.xsltRegisterExtFunction(
-            self._xsltCtxt, _cstr(name_utf), _cstr(ns_uri_utf),
-            _xpathCallback)
+            self._xsltCtxt, _cstr(name_utf), _cstr(ns_utf),
+            _xpath_function_call)
 
-    def _contextUnregisterExtensionFunction(self, ns_uri_utf, name_utf):
-        if ns_uri_utf is not None:
-            xslt.xsltRegisterExtFunction(
-                self._xsltCtxt, _cstr(name_utf), _cstr(ns_uri_utf),
-                _xpathCallback)
+    cdef _registerLocalExtensionFunctions(self):
+        cdef xslt.xsltTransformContext* xsltCtxt
+        if self._extensions is None:
+            return
+        xsltCtxt = self._xsltCtxt
+        for ns_uri_utf, extension in self._extensions.items():
+            for name_utf, function in extension.items():
+                xslt.xsltRegisterExtFunction(
+                    xsltCtxt, _cstr(name_utf), _cstr(ns_uri_utf),
+                    _xpath_function_call)
 
+cdef class _ExsltRegExp # forward declaration
 
 cdef class XSLT:
     """Turn a document into an XSLT object.
@@ -320,7 +260,7 @@
     cdef _ExsltRegExp _regexp
     cdef object _doc_url_utf
     
-    def __init__(self, xslt_input, extensions=None):
+    def __init__(self, xslt_input, extensions=None, regexp=True):
         # make a copy of the document as stylesheet needs to assume it
         # doesn't change
         cdef xslt.xsltStylesheet* c_style
@@ -353,7 +293,10 @@
         self._c_style = c_style
 
         self._context = _XSLTContext(None, extensions)
-        self._regexp  = _ExsltRegExp()
+        if regexp:
+            self._regexp  = _ExsltRegExp()
+        else:
+            self._regexp  = None
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
         
     def __dealloc__(self):
@@ -403,7 +346,8 @@
 
         self._context._release_temp_refs()
         self._context.register_context(transform_ctxt, input_doc)
-        self._regexp._register_exslt_regexp(self._context)
+        if self._regexp is not None:
+            self._regexp._register_in_context(self._context)
 
         c_result = xslt.xsltApplyStylesheetUser(self._c_style, c_doc, params,
                                                 NULL, NULL, transform_ctxt)
@@ -452,6 +396,86 @@
     return result
 
 ################################################################################
+# EXSLT regexp implementation
+
+cdef object RE_COMPILE
+RE_COMPILE = re.compile
+
+cdef class _ExsltRegExp:
+    cdef object _compile_map
+    def __init__(self):
+        self._compile_map = {}
+
+    cdef _make_string(self, value):
+        if python.PyString_Check(value) or python.PyUnicode_Check(value):
+            return value
+        else:
+            raise TypeError, "Invalid argument type %s" % type(value)
+
+    cdef _compile(self, rexp, ignore_case):
+        cdef python.PyObject* c_result
+        rexp = self._make_string(rexp)
+        key = (rexp, ignore_case)
+        c_result = python.PyDict_GetItem(self._compile_map, key)
+        if c_result is not NULL:
+            return <object>c_result
+        py_flags = re.UNICODE
+        if ignore_case:
+            py_flags = py_flags | re.IGNORECASE
+        rexp_compiled = RE_COMPILE(rexp, py_flags)
+        python.PyDict_SetItem(self._compile_map, key, rexp_compiled)
+        return rexp_compiled
+
+    def test(self, ctxt, s, rexp, flags=''):
+        flags = self._make_string(flags)
+        s = self._make_string(s)
+        rexpc = self._compile(rexp, 'i' in flags)
+        if rexpc.search(s) is None:
+            return False
+        else:
+            return True
+
+    def match(self, ctxt, s, rexp, flags=''):
+        flags = self._make_string(flags)
+        s = self._make_string(s)
+        rexpc = self._compile(rexp, 'i' in flags)
+        if 'g' in flags:
+            results = rexpc.findall(s)
+            if not results:
+                return ()
+            result_list = []
+            root = Element('matches')
+            for s_match in results:
+                elem = SubElement(root, 'match')
+                elem.text = s_match
+                python.PyList_Append(result_list, elem)
+            return result_list
+        else:
+            result = rexpc.search(s)
+            if result is None:
+                return ()
+            root = Element('match')
+            root.text = result.group()
+            return (root,)
+
+    def replace(self, ctxt, s, rexp, flags, replacement):
+        replacement = self._make_string(replacement)
+        flags = self._make_string(flags)
+        s = self._make_string(s)
+        rexpc = self._compile(rexp, 'i' in flags)
+        if 'g' in flags:
+            count = 0
+        else:
+            count = 1
+        return rexpc.sub(replacement, s, count)
+
+    cdef _register_in_context(self, _XSLTContext context):
+        ns = "http://exslt.org/regular-expressions"
+        context._registerLocalExtensionFunction(ns, "test",    self.test)
+        context._registerLocalExtensionFunction(ns, "match",   self.match)
+        context._registerLocalExtensionFunction(ns, "replace", self.replace)
+
+################################################################################
 # XPath
 
 cdef class _XPathContext(_BaseContext):
@@ -507,24 +531,6 @@
         xpath.xmlXPathRegisterVariable(
             self._xpathCtxt, _cstr(name_utf), _wrapXPathObject(value))
 
-    def _contextRegisterExtensionFunction(self, ns_uri_utf, name_utf):
-        if ns_uri_utf is not None:
-            xpath.xmlXPathRegisterFuncNS(
-                self._xpathCtxt, _cstr(name_utf), _cstr(ns_uri_utf),
-                _xpathCallback)
-        else:
-            xpath.xmlXPathRegisterFunc(
-                self._xpathCtxt, _cstr(name_utf),
-                _xpathCallback)
-
-    def _contextUnregisterExtensionFunction(self, ns_uri_utf, name_utf):
-        if ns_uri_utf is not None:
-            xpath.xmlXPathRegisterFuncNS(
-                self._xpathCtxt, _cstr(name_utf), _cstr(ns_uri_utf), NULL)
-        else:
-            xpath.xmlXPathRegisterFunc(
-                self._xpathCtxt, _cstr(name_utf), NULL)
-
 
 cdef class XPathEvaluatorBase:
     cdef _XPathContext _context
@@ -807,7 +813,7 @@
             raise NotImplementedError
     return result
 
-cdef void _xpathCallback(xpath.xmlXPathParserContext* ctxt, int nargs):
+cdef void _xpath_function_call(xpath.xmlXPathParserContext* ctxt, int nargs):
     cdef xpath.xmlXPathContext* rctxt
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
@@ -826,7 +832,7 @@
     extensions = <_BaseContext>(rctxt.userData)
 
     # lookup up the extension function in the context
-    f = extensions.find_extension(uri, name)
+    f = extensions._lookup_extension(uri, name)
 
     args = []
     doc = extensions._doc
From scoder at codespeak.net  Wed Apr 26 12:42:27 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 26 12:42:28 2006
Subject: [Lxml-checkins] r26358 - in lxml/trunk: . src/lxml
Message-ID: <20060426104227.4454910084@code0.codespeak.net>

Author: scoder
Date: Wed Apr 26 12:42:26 2006
New Revision: 26358

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/xslt.pxi
Log:
clean up, updated CHANGES.txt

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Wed Apr 26 12:42:26 2006
@@ -10,6 +10,7 @@
 * Substantial speedup in XPath.evaluate()
 
 * Exslt:regexp implementation for XSLT based on the Python 're' module
+  on by default, can be switched off with 'regexp=False' keyword argument
 
 * HTMLParser for parsing (broken) HTML
 

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Wed Apr 26 12:42:26 2006
@@ -191,12 +191,6 @@
     else:
         return _xpath_function_call
 
-cdef void _register_global_xslt_function(char* ns_uri, char* name):
-    xslt.xsltRegisterExtModuleFunction(ns_uri, name, _xpath_function_call)
-
-cdef void _unregister_global_xslt_function(char* ns_uri, char* name):
-    xslt.xsltUnRegisterExtModuleFunction(ns_uri, name)
-
  
 ################################################################################
 # XSLT
@@ -395,6 +389,13 @@
     result._xslt = xslt
     return result
 
+# used by FunctionNamespace():
+cdef void _register_global_xslt_function(char* ns_uri, char* name):
+    xslt.xsltRegisterExtModuleFunction(ns_uri, name, _xpath_function_call)
+
+cdef void _unregister_global_xslt_function(char* ns_uri, char* name):
+    xslt.xsltUnRegisterExtModuleFunction(ns_uri, name)
+
 ################################################################################
 # EXSLT regexp implementation
 
From scoder at codespeak.net  Wed Apr 26 13:39:57 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 26 13:39:59 2006
Subject: [Lxml-checkins] r26360 - lxml/trunk
Message-ID: <20060426113957.9609810084@code0.codespeak.net>

Author: scoder
Date: Wed Apr 26 13:39:56 2006
New Revision: 26360

Modified:
   lxml/trunk/bench.py
Log:
benchmarks for XPath/XSLT extension calls

Modified: lxml/trunk/bench.py
==============================================================================
--- lxml/trunk/bench.py	(original)
+++ lxml/trunk/bench.py	Wed Apr 26 13:39:56 2006
@@ -407,6 +407,44 @@
             xpath = self.etree.XPathElementEvaluator(child)
             xpath.evaluate("./*[0]")
 
+    @onlylib('lxe')
+    def bench_xpath_extensions_old(self, root):
+        def return_child(_, element):
+            if element:
+                return element[0]
+            else:
+                return ()
+        extensions = {None : {'child' : return_child}}
+        xpath = self.etree.XPath("child(.)", extensions=extensions)
+        for child in root:
+            xpath(child)
+
+    @onlylib('lxe')
+    def bench_xslt_extensions_old(self, root):
+        tree = self.etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:l="test"
+   xmlns:testns="testns"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <l:data>TEST</l:data>
+  <xsl:template match="/">
+    <l:result>
+      <xsl:for-each select="*/*">
+        <xsl:copy-of select="testns:child(.)"/>
+      </xsl:for-each>
+    </l:result>
+  </xsl:template>
+</xsl:stylesheet>
+""")
+        def return_child(_, elements):
+            return elements[0][0]
+
+        extensions = {'testns' : {'child' : return_child}}
+
+        transform = self.etree.XSLT(tree, extensions)
+        for i in range(10):
+            transform(root)
+
 ############################################################
 # Main program
 ############################################################
From scoder at codespeak.net  Wed Apr 26 13:48:41 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 26 13:48:42 2006
Subject: [Lxml-checkins] r26362 - lxml/trunk/src/lxml
Message-ID: <20060426114841.A977610084@code0.codespeak.net>

Author: scoder
Date: Wed Apr 26 13:48:40 2006
New Revision: 26362

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
rewrote fast path in _lookup_extension to use dictionary

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Wed Apr 26 13:48:40 2006
@@ -37,7 +37,7 @@
     cdef object _namespaces
     cdef object _registered_namespaces
     cdef object _utf_refs
-    cdef object _temp_last_function
+    cdef object _temp_functions
     # for exception handling and temporary reference keeping:
     cdef object _temp_elements
     cdef object _temp_docs
@@ -46,7 +46,7 @@
     def __init__(self, namespaces, extensions):
         self._xpathCtxt = NULL
         self._utf_refs = {}
-        self._temp_last_function = (None, None, None)
+        self._temp_functions = {}
 
         # convert old format extensions to UTF-8
         if isinstance(extensions, (list, tuple)):
@@ -88,7 +88,7 @@
     cdef _register_context(self, _Document doc, int allow_none_namespace):
         self._doc      = doc
         self._exc_info = None
-        self._temp_last_function = (None, None, None)
+        self._temp_functions.clear()
         namespaces = self._namespaces
         if namespaces is not None:
             self.registerNamespaces(namespaces)
@@ -134,9 +134,10 @@
 
     cdef _lookup_extension(self, ns_uri_utf, name_utf):
         cdef python.PyObject* dict_result
-        if self._temp_last_function[0] == ns_uri_utf and \
-           self._temp_last_function[1] == name_utf:
-            return self._temp_last_function[2]
+        key = (ns_uri_utf, name_utf)
+        dict_result = python.PyDict_GetItem(self._temp_functions, key)
+        if dict_result is not NULL:
+            return <object>dict_result
 
         dict_result = python.PyDict_GetItem(self._extensions, ns_uri_utf)
         if dict_result is not NULL:
@@ -146,8 +147,7 @@
         else:
             function = _find_extension(ns_uri_utf, name_utf)
 
-        # store temporarily as it will be looked up again in the next call
-        self._temp_last_function = (ns_uri_utf, name_utf, function)
+        python.PyDict_SetItem(self._temp_functions, key, function)
         return function
 
     # Python reference keeping during XPath function evaluation
@@ -182,6 +182,7 @@
     if c_name is NULL:
         return NULL
     if c_ns_uri is NULL:
+        c_ns_uri = ''
         ns_uri = None
     else:
         ns_uri = c_ns_uri
From scoder at codespeak.net  Wed Apr 26 22:05:20 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Wed Apr 26 22:05:25 2006
Subject: [Lxml-checkins] r26384 - lxml/trunk/src/lxml
Message-ID: <20060426200520.EF48210087@code0.codespeak.net>

Author: scoder
Date: Wed Apr 26 22:05:19 2006
New Revision: 26384

Modified:
   lxml/trunk/src/lxml/etree.pyx
Log:
about 30% faster isutf8()

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Wed Apr 26 22:05:19 2006
@@ -1570,15 +1570,15 @@
         c_node = c_next
     return c_node
 
-cdef int isutf8(char* string):
-    cdef int i
-    i = 0
-    while 1:
-        if string[i] == c'\0':
-            return 0
-        if string[i] & 0x80:
+cdef int isutf8(char* s):
+    cdef char c
+    c = s[0]
+    while c != c'\0':
+        if c & 0x80:
             return 1
-        i = i + 1
+        s = s + 1
+        c = s[0]
+    return 0
 
 cdef object funicode(char* s):
     if isutf8(s):
From scoder at codespeak.net  Thu Apr 27 07:24:40 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 07:24:42 2006
Subject: [Lxml-checkins] r26398 - in lxml/branch/resolver-new: doc src/lxml
Message-ID: <20060427052440.958A510095@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 07:24:38 2006
New Revision: 26398

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
   lxml/branch/resolver-new/src/lxml/docloader.pxi
   lxml/branch/resolver-new/src/lxml/parser.pxi
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
doc updates, new resolver method 'resolve_file', fix semantics of 'resolve_empty' to be treated as empty document instead of 'not found'

Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Thu Apr 27 07:24:38 2006
@@ -19,11 +19,17 @@
 these may be None.  The context object is not normally used by client code.
 
 Resolving is based on three methods of the Resolver object that build internal
-representations of the result document.  The method 'resolve_string' takes a
-document as parsable string, 'resolve_filename' takes a filename and
-'resolve_empty' resolves into an empty document.  The 'resolve' method may
-choose to return None, in which case the next registered resolver (or the
-default resolver) is consulted.
+representations of the result document.  The following methods exist:
+
+* 'resolve_string' takes a parsable string as result document
+* 'resolve_filename' takes a filename
+* 'resolve_file' takes an open file-like object that has at least a read() method
+* 'resolve_empty' resolves into an empty document
+
+The 'resolve' method may choose to return None, in which case the next
+registered resolver (or the default resolver) is consulted.  It is never
+called if the resolver returns the result of any of the above 'resolve_*'
+methods.
 
 Resolvers are registered local to a parser::
 

Modified: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/docloader.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Thu Apr 27 07:24:38 2006
@@ -4,10 +4,12 @@
     PARSER_DATA_EMPTY
     PARSER_DATA_STRING
     PARSER_DATA_FILENAME
+    PARSER_DATA_FILE
 
 cdef class _InputDocument:
     cdef _InputDocumentDataType _type
     cdef object _data_utf
+    cdef object _file
 
 cdef class Resolver:
     "This is the base class of all resolvers."
@@ -15,12 +17,14 @@
         return None
 
     def resolve_empty(self, context):
+        "Return an empty input document."
         cdef _InputDocument doc_ref
         doc_ref = _InputDocument()
         doc_ref._type = PARSER_DATA_EMPTY
         return doc_ref
 
     def resolve_string(self, string, context):
+        "Return a parsable string as input document."
         cdef _InputDocument doc_ref
         doc_ref = _InputDocument()
         doc_ref._type = PARSER_DATA_STRING
@@ -28,12 +32,23 @@
         return doc_ref
 
     def resolve_filename(self, filename, context):
+        "Return the name of a parsable file as input document."
         cdef _InputDocument doc_ref
         doc_ref = _ParserInput()
         doc_ref._type = PARSER_DATA_FILENAME
         doc_ref._data_utf = _utf8(filename)
         return doc_ref
 
+    def resolve_file(self, f, context):
+        "Return an open file-like object as input document."
+        cdef _InputDocument doc_ref
+        if not hasattr(f, 'read'):
+            raise TypeError, "Argument is not a file-like object"
+        doc_ref = _ParserInput()
+        doc_ref._type = PARSER_DATA_FILE
+        doc_ref._file = f
+        return doc_ref
+
 cdef class _ResolverRegistry:
     cdef object _resolvers
     cdef Resolver _default_resolver

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Thu Apr 27 07:24:38 2006
@@ -95,15 +95,21 @@
         return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
 
     c_input = NULL
+    data = None
     if doc_ref._type == PARSER_DATA_STRING:
+        data = doc_ref._data_utf
         c_input = xmlparser.xmlNewStringInputStream(
             c_context, _cstr(doc_ref._data_utf))
     elif doc_ref._type == PARSER_DATA_FILENAME:
         c_input = xmlparser.xmlNewInputFromFile(
             c_context, _cstr(doc_ref._data_utf))
+    elif doc_ref._type == PARSER_DATA_FILE:
+        data = doc_ref._file.read()
+        c_input = xmlparser.xmlNewStringInputStream(
+            c_context, _cstr(data))
 
-    if c_input != NULL:
-        context._storage.add(doc_ref._data_utf)
+    if data is not None:
+        context._storage.add(data)
     return c_input
 
 cdef xmlparser.xmlExternalEntityLoader __DEFAULT_ENTITY_LOADER
@@ -269,7 +275,11 @@
     pctxt._private = <python.PyObject*>context
     c_doc = xmlparser.xmlCtxtReadDoc(
         pctxt, c_text, NULL, NULL, options)
-    return _handleParseResult(pctxt, c_doc, NULL)
+    try:
+        c_doc = _handleParseResult(pctxt, c_doc, NULL)
+    finally:
+        xmlparser.xmlFreeParserCtxt(pctxt)
+    return c_doc
 
 cdef xmlDoc* _internalParseDocFromFile(char* c_filename, int options,
                                        _ResolverContext context) except NULL:
@@ -283,7 +293,11 @@
     pctxt._private = <python.PyObject*>context
     c_doc = xmlparser.xmlCtxtReadFile(
         pctxt, c_filename, NULL, options)
-    return _handleParseResult(pctxt, c_doc, c_filename)
+    try:
+        c_doc = _handleParseResult(pctxt, c_doc, c_filename)
+    finally:
+        xmlparser.xmlFreeParserCtxt(pctxt)
+    return c_doc
 
 
 cdef XMLParser __DEFAULT_XML_PARSER

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Thu Apr 27 07:24:38 2006
@@ -256,10 +256,14 @@
 
     if doc_ref is not None:
         if doc_ref._type == PARSER_DATA_EMPTY:
-            return NULL
+            c_doc = _newDoc()
         if doc_ref._type == PARSER_DATA_STRING:
             c_doc = _internalParseDoc(
                 _cstr(doc_ref._data_utf), parse_options, resolver_context)
+        elif doc_ref._type == PARSER_DATA_FILE:
+            data = doc_ref._file.read()
+            c_doc = _internalParseDoc(
+                _cstr(data), parse_options, resolver_context)
         elif doc_ref._type == PARSER_DATA_FILENAME:
             c_doc = _internalParseDocFromFile(
                 _cstr(doc_ref._data_utf), parse_options, resolver_context)
From scoder at codespeak.net  Thu Apr 27 07:33:59 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 07:34:00 2006
Subject: [Lxml-checkins] r26399 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060427053359.14E131008F@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 07:33:58 2006
New Revision: 26399

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
handle resolver exceptions that were raised in the parser functions

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Thu Apr 27 07:33:58 2006
@@ -250,26 +250,27 @@
     try:
         uri = funicode(c_uri)
         doc_ref = resolvers.resolve(uri, None, resolver_context)
+
+        if doc_ref is not None:
+            if doc_ref._type == PARSER_DATA_EMPTY:
+                c_doc = _newDoc()
+            if doc_ref._type == PARSER_DATA_STRING:
+                c_doc = _internalParseDoc(
+                    _cstr(doc_ref._data_utf), parse_options, resolver_context)
+            elif doc_ref._type == PARSER_DATA_FILE:
+                data = doc_ref._file.read()
+                c_doc = _internalParseDoc(
+                    _cstr(data), parse_options, resolver_context)
+            elif doc_ref._type == PARSER_DATA_FILENAME:
+                c_doc = _internalParseDocFromFile(
+                    _cstr(doc_ref._data_utf), parse_options, resolver_context)
+            if c_doc is not NULL and c_doc.URL is NULL:
+                c_doc.URL = tree.xmlStrdup(c_uri)
+
     except Exception:
         resolver_context._store_raised()
         return NULL
 
-    if doc_ref is not None:
-        if doc_ref._type == PARSER_DATA_EMPTY:
-            c_doc = _newDoc()
-        if doc_ref._type == PARSER_DATA_STRING:
-            c_doc = _internalParseDoc(
-                _cstr(doc_ref._data_utf), parse_options, resolver_context)
-        elif doc_ref._type == PARSER_DATA_FILE:
-            data = doc_ref._file.read()
-            c_doc = _internalParseDoc(
-                _cstr(data), parse_options, resolver_context)
-        elif doc_ref._type == PARSER_DATA_FILENAME:
-            c_doc = _internalParseDocFromFile(
-                _cstr(doc_ref._data_utf), parse_options, resolver_context)
-        if c_doc is not NULL and c_doc.URL is NULL:
-            c_doc.URL = tree.xmlStrdup(c_uri)
-
     if c_doc is NULL:
         c_doc = XSLT_DOC_DEFAULT_LOADER(
             c_uri, c_dict, parse_options, c_ctxt, c_type)
From scoder at codespeak.net  Thu Apr 27 09:03:10 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 09:03:11 2006
Subject: [Lxml-checkins] r26407 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060427070310.1071F1009D@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 09:03:08 2006
New Revision: 26407

Modified:
   lxml/branch/resolver-new/src/lxml/docloader.pxi
Log:
fix exception forwarding: clear after re-raising

Modified: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/docloader.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Thu Apr 27 09:03:08 2006
@@ -104,6 +104,7 @@
     cdef _raise_if_stored(self):
         _exc_info = self._exc_info
         if _exc_info is not None:
+            self._exc_info = None
             type, value, traceback = _exc_info
             if traceback is None and value is None:
                 raise type
From scoder at codespeak.net  Thu Apr 27 09:27:09 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 09:27:11 2006
Subject: [Lxml-checkins] r26410 - in lxml/branch/resolver-new: doc src/lxml
Message-ID: <20060427072709.4CDF1100A8@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 09:27:07 2006
New Revision: 26410

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
   lxml/branch/resolver-new/src/lxml/xslt.pxd
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
fix XSLT document lookup for xsl:document() function
* always use resolver context of XSL doc rather than transformed doc to get it right in most cases

FIXME: need to determine special cases where the transformed doc is actually meant
* infrastructure is there: xslt_resolver_context and doc_resolver_context are determined in _doc_loader function
* which one to take depends on what context document() is called for! Does libxslt tell us?


Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Thu Apr 27 09:27:07 2006
@@ -57,15 +57,8 @@
 
 XML documents memorise their initial parser (and its resolvers) during their
 life-time.  This means that a lookup process related to a document will use
-the resolvers of the document's parser.
-
-This behaviour is most visible in XSLT, where two documents are used: the XSL
-document and the transformed XML document.  XSLT thus distinguishes between
-the context when parsing the XSLT document (i.e. when resolving xsl:import and
-xsl:include elements) and the context at transformation time (calls to the
-'document' function).
-
-We start with a resolver that only responds to a specific prefix::
+the resolvers of the document's parser.  We can demonstrate this with a
+resolver that only responds to a specific prefix::
 
   >>> class PrefixResolver(etree.Resolver):
   ...     def __init__(self, prefix):
@@ -83,7 +76,7 @@
   ...             return self.resolve_string(self.result_xml, context)
   ...         print "failed"
 
-We use the following stylesheet as an example::
+We demonstrate this in XSLT and use the following stylesheet as an example::
 
   >>> xml_text = """\
   ... <xsl:stylesheet version="1.0"
@@ -97,8 +90,11 @@
   ... </xsl:stylesheet>
   ... """
 
-If we now register different resolvers with two different parsers, we can
-parse our document twice in different resolver contexts::
+Note that it needs to resolve two URIs: 'honk:test' when compiling the XSLT
+document (i.e. when resolving xsl:import and xsl:include elements) and
+'hoi:test' at transformation time, when calls to the 'document' function are
+resolved.  If we now register different resolvers with two different parsers,
+we can parse our document twice in different resolver contexts::
 
   >>> hoi_parser = etree.XMLParser()
   >>> hoi_parser.resolvers.add( PrefixResolver("hoi") )
@@ -116,7 +112,7 @@
   >>> transform = etree.XSLT(hoi_doc)
   Traceback (most recent call last):
     [...]
-  XSLTParseError: Cannot parse style sheet
+  XSLTParseError: Cannot resolve URI honk:test
 
 However, if we use the "honk" resolver associated with the second document,
 everything works fine::
@@ -124,19 +120,33 @@
   >>> transform = etree.XSLT(honk_doc)
   Resolving url honk:test as prefix honk ... done
 
-We can see that the "hoi" resolver was called to generate a document that was
-then inserted into the result document by the XSLT transformation.  Note that
-the "hoi" resolver is attached to the transformed document, which defines the
-context for the transformation process.  
+Running the transform accesses the same parser context again, but since it now
+needs to resolve the "hoi" URI in the call to the document function, its
+"honk" resolver will fail to do so::
+
+  >>> result = transform(hoi_doc)
+  Traceback (most recent call last):
+    [...]
+  XSLTApplyError: Cannot resolve URI hoi:test
 
+This can only be solved by adding a "hoi" resolver to the parser::
+
+  >>> honk_parser.resolvers.add( PrefixResolver("hoi") )
   >>> result = transform(hoi_doc)
+  Resolving url hoi:test as prefix honk ... failed
   Resolving url hoi:test as prefix hoi ... done
   >>> print str(result),
   <?xml version="1.0"?>
   <test>hoi-TEST</test>
 
-If we reverse the contexts, the other
-resolver will get called.  Obviously, the "honk" resolver does not know how to
-resolve "hoi" URLs, so the default resolver is called in this case.  If the
-default resolver cannot find the requested document either, you will get an
-exception::
+We can see that the "hoi" resolver was called to generate a document that was
+then inserted into the result document by the XSLT transformation.  Note that
+this is completely independent of the XML file you transform, as the URI is
+resolved from within the stylesheet context::
+
+  >>> result = transform(honk_doc)
+  Resolving url hoi:test as prefix honk ... failed
+  Resolving url hoi:test as prefix hoi ... done
+  >>> print str(result),
+  <?xml version="1.0"?>
+  <test>hoi-TEST</test>

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxd
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxd	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxd	Thu Apr 27 09:27:07 2006
@@ -9,6 +9,7 @@
         xmlDoc* doc
 
     ctypedef struct xsltTransformContext:
+        xsltStylesheet* style
         xmlXPathContext* xpathCtxt
         xsltDocument* document
 

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Thu Apr 27 09:27:07 2006
@@ -215,18 +215,17 @@
 
 cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
                          void* c_ctxt, xslt.xsltLoadType c_type):
-    cdef xslt.xsltTransformContext* transform_ctxt
     cdef xmlDoc* c_doc
     cdef _ResolverRegistry resolvers
     cdef _InputDocument doc_ref
+    cdef _XSLTResolverContext xslt_resolver_context
+    cdef _XSLTResolverContext doc_resolver_context
     cdef _XSLTResolverContext resolver_context
     cdef XMLParser parser
-    # find current xmlDoc to retrieve resolvers
+    # find stylesheet xmlDoc to retrieve resolvers
     c_doc = NULL
     if c_type == xslt.XSLT_LOAD_DOCUMENT:
-        transform_ctxt = <xslt.xsltTransformContext*>c_ctxt
-        if transform_ctxt.document is not NULL:
-            c_doc = transform_ctxt.document.doc
+        c_doc = (<xslt.xsltTransformContext*>c_ctxt).style.doc
     elif c_type == xslt.XSLT_LOAD_STYLESHEET:
         c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
 
@@ -236,16 +235,25 @@
         return XSLT_DOC_DEFAULT_LOADER(
             c_uri, c_dict, parse_options, c_ctxt, c_type)
 
-    resolver_context = <_XSLTResolverContext>c_doc._private
+    xslt_resolver_context = <_XSLTResolverContext>c_doc._private
+
+    # find resolver context for transformed document
+    doc_resolver_context = None
+    if c_type == xslt.XSLT_LOAD_DOCUMENT:
+        c_doc = (<xslt.xsltTransformContext*>c_ctxt).document.doc
+        if c_doc is not NULL and c_doc._private is not NULL and \
+               isinstance(<object>c_doc._private, _XSLTResolverContext):
+            doc_resolver_context = <_XSLTResolverContext>c_doc._private
 
     # quick check if we are looking for the current stylesheet
-    c_doc = resolver_context._c_style_doc
+    c_doc = xslt_resolver_context._c_style_doc
     if c_doc is not NULL and c_doc.URL is not NULL:
         if tree.strcmp(c_uri, c_doc.URL) == 0:
             return tree.xmlCopyDoc(c_doc, 1)
 
     # call the Python document loaders
     c_doc = NULL
+    resolver_context = xslt_resolver_context # currently use only XSLT resolvers
     resolvers = resolver_context._resolvers
     try:
         uri = funicode(c_uri)
@@ -256,30 +264,37 @@
                 c_doc = _newDoc()
             if doc_ref._type == PARSER_DATA_STRING:
                 c_doc = _internalParseDoc(
-                    _cstr(doc_ref._data_utf), parse_options, resolver_context)
+                    _cstr(doc_ref._data_utf), parse_options,
+                    resolver_context)
             elif doc_ref._type == PARSER_DATA_FILE:
                 data = doc_ref._file.read()
                 c_doc = _internalParseDoc(
-                    _cstr(data), parse_options, resolver_context)
+                    _cstr(data), parse_options,
+                    resolver_context)
             elif doc_ref._type == PARSER_DATA_FILENAME:
                 c_doc = _internalParseDocFromFile(
-                    _cstr(doc_ref._data_utf), parse_options, resolver_context)
+                    _cstr(doc_ref._data_utf), parse_options,
+                    resolver_context)
             if c_doc is not NULL and c_doc.URL is NULL:
                 c_doc.URL = tree.xmlStrdup(c_uri)
 
-    except Exception:
-        resolver_context._store_raised()
+    except Exception, e:
+        xslt_resolver_context._store_raised()
         return NULL
 
     if c_doc is NULL:
         c_doc = XSLT_DOC_DEFAULT_LOADER(
             c_uri, c_dict, parse_options, c_ctxt, c_type)
         if c_doc is NULL:
-            exception = XSLTError("Cannot resolve URI %s" % funicode(c_uri))
-            resolver_context._store_exception(exception)
+            message = "Cannot resolve URI %s" % funicode(c_uri)
+            if c_type == xslt.XSLT_LOAD_DOCUMENT:
+                exception = XSLTApplyError(message)
+            else:
+                exception = XSLTParseError(message)
+            xslt_resolver_context._store_exception(exception)
             return NULL
     if c_doc is not NULL and c_doc._private is NULL:
-        c_doc._private = <python.PyObject*>resolver_context
+        c_doc._private = <python.PyObject*>xslt_resolver_context
     return c_doc
 
 cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
@@ -374,6 +389,7 @@
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
             tree.xmlFreeDoc(c_doc)
+            self._xslt_resolver_context._raise_if_stored()
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
 
@@ -447,10 +463,10 @@
             cstd.free(params)
 
         self._context.free_context()
-        c_doc._private = ptemp
+        c_doc._private = ptemp # restore _private before _destroyFakeDoc!
         _destroyFakeDoc(input_doc._c_doc, c_doc)
 
-        resolver_context._raise_if_stored()
+        self._xslt_resolver_context._raise_if_stored()
         if c_result is NULL:
             raise XSLTApplyError, "Error applying stylesheet"
 
From scoder at codespeak.net  Thu Apr 27 09:47:19 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 09:47:20 2006
Subject: [Lxml-checkins] r26411 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060427074719.6398B100A8@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 09:47:18 2006
New Revision: 26411

Modified:
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
some cleanup in _doc_loader

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Thu Apr 27 09:47:18 2006
@@ -222,9 +222,14 @@
     cdef _XSLTResolverContext doc_resolver_context
     cdef _XSLTResolverContext resolver_context
     cdef XMLParser parser
-    # find stylesheet xmlDoc to retrieve resolvers
+    # find resolver contexts of stylesheet and transformed doc
     c_doc = NULL
+    doc_resolver_context = None
     if c_type == xslt.XSLT_LOAD_DOCUMENT:
+        c_doc = (<xslt.xsltTransformContext*>c_ctxt).document.doc
+        if c_doc is not NULL and c_doc._private is not NULL:
+            if isinstance(<object>c_doc._private, _XSLTResolverContext):
+                doc_resolver_context = <_XSLTResolverContext>c_doc._private
         c_doc = (<xslt.xsltTransformContext*>c_ctxt).style.doc
     elif c_type == xslt.XSLT_LOAD_STYLESHEET:
         c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
@@ -237,14 +242,6 @@
 
     xslt_resolver_context = <_XSLTResolverContext>c_doc._private
 
-    # find resolver context for transformed document
-    doc_resolver_context = None
-    if c_type == xslt.XSLT_LOAD_DOCUMENT:
-        c_doc = (<xslt.xsltTransformContext*>c_ctxt).document.doc
-        if c_doc is not NULL and c_doc._private is not NULL and \
-               isinstance(<object>c_doc._private, _XSLTResolverContext):
-            doc_resolver_context = <_XSLTResolverContext>c_doc._private
-
     # quick check if we are looking for the current stylesheet
     c_doc = xslt_resolver_context._c_style_doc
     if c_doc is not NULL and c_doc.URL is not NULL:
From scoder at codespeak.net  Thu Apr 27 10:16:12 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 10:16:14 2006
Subject: [Lxml-checkins] r26416 - in lxml/branch/resolver-new: doc src/lxml
Message-ID: <20060427081612.5116A10089@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 10:16:10 2006
New Revision: 26416

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
   lxml/branch/resolver-new/src/lxml/docloader.pxi
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/parser.pxi
Log:
parser.copy() method to allow copying parsers (including their registered resolvers), store a copy of the parser in each document instead of a reference to memorise the resolver state at parse time

Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Thu Apr 27 10:16:10 2006
@@ -97,6 +97,8 @@
 we can parse our document twice in different resolver contexts::
 
   >>> hoi_parser = etree.XMLParser()
+  >>> normal_doc = etree.parse(StringIO(xml_text), hoi_parser)
+
   >>> hoi_parser.resolvers.add( PrefixResolver("hoi") )
   >>> hoi_doc = etree.parse(StringIO(xml_text), hoi_parser)
 
@@ -109,6 +111,11 @@
 subsequent lookups.  To compile the stylesheet, XSLT must resolve the
 honk:test URI in the xsl:include element.  The "hoi" resolver cannot do that::
 
+  >>> transform = etree.XSLT(normal_doc)
+  Traceback (most recent call last):
+    [...]
+  XSLTParseError: Cannot resolve URI honk:test
+
   >>> transform = etree.XSLT(hoi_doc)
   Traceback (most recent call last):
     [...]
@@ -124,15 +131,38 @@
 needs to resolve the "hoi" URI in the call to the document function, its
 "honk" resolver will fail to do so::
 
+  >>> result = transform(normal_doc)
+  Traceback (most recent call last):
+    [...]
+  XSLTApplyError: Cannot resolve URI hoi:test
+
   >>> result = transform(hoi_doc)
   Traceback (most recent call last):
     [...]
   XSLTApplyError: Cannot resolve URI hoi:test
 
-This can only be solved by adding a "hoi" resolver to the parser::
+  >>> result = transform(honk_doc)
+  Traceback (most recent call last):
+    [...]
+  XSLTApplyError: Cannot resolve URI hoi:test
+
+This can only be solved by adding a "hoi" resolver to the parser.  Note that
+adding it after parsing the XSL document will not work as parsed documents
+remember the state of the parser at the time of their creation::
 
   >>> honk_parser.resolvers.add( PrefixResolver("hoi") )
-  >>> result = transform(hoi_doc)
+  >>> result = transform(honk_doc)
+  Traceback (most recent call last):
+    [...]
+  XSLTApplyError: Cannot resolve URI hoi:test
+
+You have to parse the document again with this updated parser to copy the
+resolver state to the stylesheet::
+
+  >>> honk_doc = etree.parse(StringIO(xml_text), honk_parser)
+  >>> transform = etree.XSLT(honk_doc)
+  Resolving url honk:test as prefix honk ... done
+  >>> result = transform(honk_doc)
   Resolving url hoi:test as prefix honk ... failed
   Resolving url hoi:test as prefix hoi ... done
   >>> print str(result),
@@ -144,7 +174,7 @@
 this is completely independent of the XML file you transform, as the URI is
 resolved from within the stylesheet context::
 
-  >>> result = transform(honk_doc)
+  >>> result = transform(normal_doc)
   Resolving url hoi:test as prefix honk ... failed
   Resolving url hoi:test as prefix hoi ... done
   >>> print str(result),

Modified: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/docloader.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Thu Apr 27 10:16:10 2006
@@ -74,6 +74,12 @@
     def remove(self, resolver):
         self._resolvers.discard(resolver)
 
+    def copy(self):
+        cdef _ResolverRegistry registry
+        registry = _ResolverRegistry(self._default_resolver)
+        registry._resolvers = self._resolvers.copy()
+        return registry
+
     def resolve(self, system_url, public_id, context):
         for resolver in self._resolvers:
             result = resolver.resolve(system_url, public_id, context)

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Thu Apr 27 10:16:10 2006
@@ -171,7 +171,7 @@
     result._ns_counter = 0
     if parser is None:
         parser = __DEFAULT_PARSER
-    result._parser = parser
+    result._parser = parser.copy()
     return result
 
 # to help with debugging

Modified: lxml/branch/resolver-new/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/parser.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/parser.pxi	Thu Apr 27 10:16:10 2006
@@ -135,6 +135,13 @@
         def __get__(self):
             return self._error_log.copy()
 
+    cdef _copy(self):
+        cdef BaseParser parser
+        parser = self.__class__()
+        parser.resolvers = self.resolvers.copy()
+        parser._context = _ResolverContext(parser.resolvers)
+        return parser
+
     cdef _initContext(self, xmlParserCtxt* c_ctxt):
         __GLOBAL_PARSER_CONTEXT._initParserDict(c_ctxt)
         c_ctxt._private = <python.PyObject*>self._context
@@ -224,6 +231,12 @@
         if self._memory_parser_ctxt != NULL:
             xmlparser.xmlFreeParserCtxt(self._memory_parser_ctxt)
 
+    def copy(self):
+        cdef XMLParser parser
+        parser = self._copy()
+        parser._parse_options = self._parse_options
+        return parser
+
     cdef xmlParserCtxt* _createContext(self) except NULL:
         cdef xmlParserCtxt* pctxt
         pctxt = xmlparser.xmlNewParserCtxt()
@@ -371,6 +384,12 @@
         if self._memory_parser_ctxt != NULL:
             htmlparser.htmlFreeParserCtxt(self._memory_parser_ctxt)
 
+    def copy(self):
+        cdef HTMLParser parser
+        parser = self._copy()
+        parser._parse_options = self._parse_options
+        return parser
+
     cdef xmlDoc* _parseDoc(self, char* c_text) except NULL:
         """Parse HTML document, share dictionary if possible.
         """
From scoder at codespeak.net  Thu Apr 27 10:31:08 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 10:31:11 2006
Subject: [Lxml-checkins] r26418 - lxml/branch/resolver-new/doc
Message-ID: <20060427083108.DD274100B3@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 10:31:06 2006
New Revision: 26418

Modified:
   lxml/branch/resolver-new/doc/resolvers.txt
Log:
state that XSLT generated documents inherit the parsers of their input document

Modified: lxml/branch/resolver-new/doc/resolvers.txt
==============================================================================
--- lxml/branch/resolver-new/doc/resolvers.txt	(original)
+++ lxml/branch/resolver-new/doc/resolvers.txt	Thu Apr 27 10:31:06 2006
@@ -121,8 +121,8 @@
     [...]
   XSLTParseError: Cannot resolve URI honk:test
 
-However, if we use the "honk" resolver associated with the second document,
-everything works fine::
+However, if we use the "honk" resolver associated with the respective
+document, everything works fine::
 
   >>> transform = etree.XSLT(honk_doc)
   Resolving url honk:test as prefix honk ... done
@@ -180,3 +180,8 @@
   >>> print str(result),
   <?xml version="1.0"?>
   <test>hoi-TEST</test>
+
+It may be seen as a matter of taste what resolvers the generated document
+inherits.  For XSLT, the output document inherits the resolvers of the input
+document and not those of the stylesheet.  Therefore, the last result does not
+inherit any resolvers at all.
From scoder at codespeak.net  Thu Apr 27 11:56:39 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 11:56:40 2006
Subject: [Lxml-checkins] r26428 - lxml/branch/resolver-new/src/lxml
Message-ID: <20060427095639.120CF100D2@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 11:56:36 2006
New Revision: 26428

Modified:
   lxml/branch/resolver-new/src/lxml/docloader.pxi
   lxml/branch/resolver-new/src/lxml/etree.pyx
   lxml/branch/resolver-new/src/lxml/xslt.pxi
Log:
refactoring of exception storage: new _ExceptionContext class, used by BaseContext (XSLT), inherited by _ResolverContext

Modified: lxml/branch/resolver-new/src/lxml/docloader.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/docloader.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/docloader.pxi	Thu Apr 27 11:56:36 2006
@@ -92,27 +92,10 @@
     def __repr__(self):
         return repr(self._resolvers)
 
-cdef class _ResolverContext:
+cdef class _ResolverContext(_ExceptionContext):
     cdef _ResolverRegistry _resolvers
     cdef _TempStore _storage
-    cdef object _exc_info
     def __init__(self, _ResolverRegistry resolvers not None):
+        _ExceptionContext.__init__(self)
         self._resolvers = resolvers
         self._storage = _TempStore()
-        self._exc_info = None
-
-    cdef void _store_raised(self):
-        self._exc_info = sys.exc_info()
-
-    cdef void _store_exception(self, exception):
-        self._exc_info = (exception, None, None)
-
-    cdef _raise_if_stored(self):
-        _exc_info = self._exc_info
-        if _exc_info is not None:
-            self._exc_info = None
-            type, value, traceback = _exc_info
-            if traceback is None and value is None:
-                raise type
-            else:
-                raise type, value, traceback

Modified: lxml/branch/resolver-new/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/resolver-new/src/lxml/etree.pyx	(original)
+++ lxml/branch/resolver-new/src/lxml/etree.pyx	Thu Apr 27 11:56:36 2006
@@ -56,6 +56,51 @@
 class C14NError(LxmlError):
     pass
 
+
+# class for temporary storage of Python references
+cdef class _TempStore:
+    cdef object _storage
+    def __init__(self):
+        self._storage = {}
+
+    cdef void add(self, obj):
+        python.PyDict_SetItem(self._storage, id(obj), obj)
+
+    cdef void clear(self):
+        python.PyDict_Clear(self._storage)
+
+    cdef object dictcopy(self):
+        return self._storage.copy()
+
+# class for temporarily storing exceptions raised in extensions
+cdef class _ExceptionContext:
+    cdef object _exc_info
+    def __init__(self):
+        self._exc_info = None
+
+    cdef void clear(self):
+        self._exc_info = None
+
+    cdef void _store_raised(self):
+        self._exc_info = sys.exc_info()
+
+    cdef void _store_exception(self, exception):
+        self._exc_info = (exception, None, None)
+
+    cdef _has_raised(self):
+        return self._exc_info is not None
+
+    cdef _raise_if_stored(self):
+        _exc_info = self._exc_info
+        if _exc_info is not None:
+            self._exc_info = None
+            type, value, traceback = _exc_info
+            if traceback is None and value is None:
+                raise type
+            else:
+                raise type, value, traceback
+
+
 cdef class BaseParser # forward declaration
 
 cdef class _Document:
@@ -1292,19 +1337,6 @@
     return ElementTree(doc.getroot())
 
 
-# class for temporary storage of Python references
-cdef class _TempStore:
-    cdef object _storage
-    def __init__(self):
-        self._storage = {}
-
-    cdef void add(self, obj):
-        python.PyDict_SetItem(self._storage, id(obj), obj)
-
-    cdef void clear(self):
-        python.PyDict_Clear(self._storage)
-
-
 # include submodules
 include "xmlerror.pxi"  # error and log handling
 include "xmlid.pxi"     # XMLID and IDDict

Modified: lxml/branch/resolver-new/src/lxml/xslt.pxi
==============================================================================
--- lxml/branch/resolver-new/src/lxml/xslt.pxi	(original)
+++ lxml/branch/resolver-new/src/lxml/xslt.pxi	Thu Apr 27 11:56:36 2006
@@ -40,9 +40,8 @@
     cdef object _extension_functions
     cdef object _utf_refs
     # for exception handling and temporary reference keeping:
-    cdef _TempStore _temp_elements
-    cdef _TempStore _temp_docs
-    cdef object _exc_info
+    cdef _TempStore _temp_refs
+    cdef _ExceptionContext _exc
 
     def __init__(self, namespaces, extensions):
         self._xpathCtxt = NULL
@@ -62,14 +61,13 @@
             extensions = new_extensions or None
 
         self._doc        = None
-        self._exc_info   = None
+        self._exc        = _ExceptionContext()
         self._extensions = extensions
         self._namespaces = namespaces
         self._registered_namespaces = []
         self._registered_extensions = []
         self._extension_functions = {}
-        self._temp_elements = _TempStore()
-        self._temp_docs     = _TempStore()
+        self._temp_refs = _TempStore()
 
     cdef object _to_utf(self, s):
         "Convert to UTF-8 and keep a reference to the encoded string"
@@ -88,8 +86,8 @@
         xpathCtxt.userData = <void*>self
 
     cdef _register_context(self, _Document doc, int allow_none_namespace):
-        self._doc      = doc
-        self._exc_info = None
+        self._doc = doc
+        self._exc.clear()
         namespaces = self._namespaces
         if namespaces is not None:
             self.registerNamespaces(namespaces)
@@ -167,15 +165,14 @@
         for ns_uri_utf, name_utf in self._registered_extensions:
             self._contextUnregisterExtensionFunction(ns_uri_utf, name_utf)
 
-    def find_extension(self, ns_uri_utf, name_utf):
+    cdef _find_extension(self, ns_uri_utf, name_utf):
         return self._extension_functions[(ns_uri_utf, name_utf)]
 
     # Python reference keeping during XPath function evaluation
 
     cdef _release_temp_refs(self):
         "Free temporarily referenced objects from this context."
-        self._temp_elements.clear()
-        self._temp_docs.clear()
+        self._temp_refs.clear()
         
     cdef _hold(self, obj):
         """A way to temporarily hold references to nodes in the evaluator.
@@ -193,9 +190,9 @@
             if isinstance(o, _NodeBase):
                 element = <_NodeBase>o
                 #print "Holding element:", <int>element._c_node
-                self._temp_elements.add(element)
+                self._temp_refs.add(element)
                 #print "Holding document:", <int>element._doc._c_doc
-                self._temp_docs.add(element._doc)
+                self._temp_refs.add(element._doc)
 
 
 ################################################################################
@@ -463,7 +460,11 @@
         c_doc._private = ptemp # restore _private before _destroyFakeDoc!
         _destroyFakeDoc(input_doc._c_doc, c_doc)
 
-        self._xslt_resolver_context._raise_if_stored()
+        if self._xslt_resolver_context._has_raised():
+            if c_result is not NULL:
+                tree.xmlFreeDoc(c_result)
+            self._xslt_resolver_context._raise_if_stored()
+
         if c_result is NULL:
             raise XSLTApplyError, "Error applying stylesheet"
 
@@ -608,17 +609,21 @@
         self._context = _XPathContext(namespaces, extensions, variables)
 
     cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
-        _exc_info = self._context._exc_info
-        if _exc_info is not None:
-            type, value, traceback = _exc_info
-            raise type, value, traceback
+        if self._context._exc._has_raised():
+            if xpathObj is not NULL:
+                xpath.xmlXPathFreeObject(xpathObj)
+                xpathObj = NULL
+            self._context._exc._raise_if_stored()
+
         if xpathObj is NULL:
             raise XPathSyntaxError, "Error in xpath expression."
+
         try:
             result = _unwrapXPathObject(xpathObj, doc)
         except XPathResultError:
             xpath.xmlXPathFreeObject(xpathObj)
             raise
+
         xpath.xmlXPathFreeObject(xpathObj)
         return result
 
@@ -885,7 +890,7 @@
     cdef xpath.xmlXPathContext* rctxt
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
-    cdef _BaseContext extensions
+    cdef _BaseContext context
     cdef int i
 
     rctxt = ctxt.context
@@ -898,13 +903,13 @@
         uri = None
 
     # get our evaluator
-    extensions = <_BaseContext>(rctxt.userData)
+    context = <_BaseContext>(rctxt.userData)
 
     # lookup up the extension function in the context
-    f = extensions.find_extension(uri, name)
+    f = context._find_extension(uri, name)
 
     args = []
-    doc = extensions._doc
+    doc = context._doc
     for i from 0 <= i < nargs:
         python.PyList_Append(args, _unwrapXPathObject(xpath.valuePop(ctxt), doc))
     args.reverse()
@@ -914,13 +919,13 @@
         res = f(None, *args)
         # hold python objects temporarily so that they won't get deallocated
         # during processing
-        extensions._hold(res)
+        context._hold(res)
         # now wrap for XPath consumption
         obj = _wrapXPathObject(res)
     except:
         xpath.xmlXPathErr(
             ctxt,
             xmlerror.XML_XPATH_EXPR_ERROR - xmlerror.XML_XPATH_EXPRESSION_OK)
-        extensions._exc_info = sys.exc_info()
+        context._exc._store_raised()
         return
     xpath.valuePush(ctxt, obj)
From scoder at codespeak.net  Thu Apr 27 22:51:47 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Thu Apr 27 22:51:49 2006
Subject: [Lxml-checkins] r26466 - in lxml/trunk: . doc src/lxml
	src/lxml/tests
Message-ID: <20060427205147.87558100B8@code0.codespeak.net>

Author: scoder
Date: Thu Apr 27 22:51:43 2006
New Revision: 26466

Added:
   lxml/trunk/doc/resolvers.txt
      - copied unchanged from r26455, lxml/branch/resolver-new/doc/resolvers.txt
   lxml/trunk/src/lxml/docloader.pxi
      - copied unchanged from r26455, lxml/branch/resolver-new/src/lxml/docloader.pxi
Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/bench.py
   lxml/trunk/setup.py
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/parser.pxi
   lxml/trunk/src/lxml/tests/test_etree.py
   lxml/trunk/src/lxml/tests/test_xslt.py
   lxml/trunk/src/lxml/tree.pxd
   lxml/trunk/src/lxml/xmlparser.pxd
   lxml/trunk/src/lxml/xslt.pxd
   lxml/trunk/src/lxml/xslt.pxi
Log:
merge of resolver-new branch: support for custom URI resolvers in parsers and XSLT

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Thu Apr 27 22:51:43 2006
@@ -7,11 +7,16 @@
 Features added
 --------------
 
-* Substantial speedup in XPath.evaluate()
+* Support for custom document loaders (URI resolvers) in parsers and XSLT,
+  resolvers are registered at parser level
 
 * Exslt:regexp implementation for XSLT based on the Python 're' module
   on by default, can be switched off with 'regexp=False' keyword argument
 
+* Support for exslt extensions (libexslt) and node-set function
+
+* Substantial speedup in XPath.evaluate()
+
 * HTMLParser for parsing (broken) HTML
 
 * XMLDTDID function parses XML into tuple (root node, ID dict) based on xml:id
@@ -29,7 +34,7 @@
 Bugs fixed
 ----------
 
-* document('') now works in XSLT documents parsed from strings
+* document('') now works in all XSLT documents
 
 * Crash in XMLSchema and RelaxNG when passing non-schema documents
 

Modified: lxml/trunk/bench.py
==============================================================================
--- lxml/trunk/bench.py	(original)
+++ lxml/trunk/bench.py	Thu Apr 27 22:51:43 2006
@@ -445,6 +445,25 @@
         for i in range(10):
             transform(root)
 
+    @onlylib('lxe')
+    def bench_xslt_document(self, root):
+        transform = self.etree.XSLT(self.etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:l="test"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <l:data>TEST</l:data>
+  <xsl:template match="/">
+    <l:result>
+      <xsl:for-each select="*/*">
+        <l:test><xsl:copy-of select="document('')//l:data/text()"/></l:test>
+      </xsl:for-each>
+    </l:result>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        transform(root)
+
+
 ############################################################
 # Main program
 ############################################################

Modified: lxml/trunk/setup.py
==============================================================================
--- lxml/trunk/setup.py	(original)
+++ lxml/trunk/setup.py	Thu Apr 27 22:51:43 2006
@@ -11,6 +11,7 @@
 try:
     from setuptools import setup
     from setuptools.extension import Extension
+    # prevent setuptools from making local etree.so copies:
     setup_args['zip_safe'] = False
 except ImportError:
     from distutils.core import setup
@@ -47,6 +48,16 @@
 
     changelog.close()
 
+# compile also against libexslt!
+xslt_libs = flags('xslt-config --libs')
+xslt_libs.append('-lexslt')
+for i, libname in (): # enumerate(xslt_libs):
+    if 'exslt' in libname:
+        break
+    if 'xslt' in libname:
+        xslt_libs.insert(i, libname.replace('xslt', 'exslt'))
+        break
+
 setup(
     name = "lxml",
     version = version,
@@ -85,7 +96,7 @@
         "lxml.etree", 
         sources = sources,
         extra_compile_args = ['-w'] + flags('xslt-config --cflags'),
-        extra_link_args = flags('xslt-config --libs')
+        extra_link_args = xslt_libs
     )],
     **setup_args
 )

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Thu Apr 27 22:51:43 2006
@@ -1,6 +1,7 @@
 cimport tree, python
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, issubclass, hasattr, callable, iter, str, _cstr
+from python cimport isinstance, issubclass, hasattr, callable
+from python cimport iter, str, _cstr
 cimport xpath
 cimport xslt
 cimport xmlerror
@@ -55,6 +56,53 @@
 class C14NError(LxmlError):
     pass
 
+
+# class for temporary storage of Python references
+cdef class _TempStore:
+    cdef object _storage
+    def __init__(self):
+        self._storage = {}
+
+    cdef void add(self, obj):
+        python.PyDict_SetItem(self._storage, id(obj), obj)
+
+    cdef void clear(self):
+        python.PyDict_Clear(self._storage)
+
+    cdef object dictcopy(self):
+        return self._storage.copy()
+
+# class for temporarily storing exceptions raised in extensions
+cdef class _ExceptionContext:
+    cdef object _exc_info
+    def __init__(self):
+        self._exc_info = None
+
+    cdef void clear(self):
+        self._exc_info = None
+
+    cdef void _store_raised(self):
+        self._exc_info = sys.exc_info()
+
+    cdef void _store_exception(self, exception):
+        self._exc_info = (exception, None, None)
+
+    cdef _has_raised(self):
+        return self._exc_info is not None
+
+    cdef _raise_if_stored(self):
+        _exc_info = self._exc_info
+        if _exc_info is not None:
+            self._exc_info = None
+            type, value, traceback = _exc_info
+            if traceback is None and value is None:
+                raise type
+            else:
+                raise type, value, traceback
+
+
+cdef class BaseParser # forward declaration
+
 cdef class _Document:
     """Internal base class to reference a libxml document.
 
@@ -63,6 +111,7 @@
     """
     cdef int _ns_counter
     cdef xmlDoc* _c_doc
+    cdef BaseParser _parser
     
     def __dealloc__(self):
         # if there are no more references to the document, it is safe
@@ -150,21 +199,24 @@
     if filename is None:
         filename = source
     # open filename
-    c_doc = _parseDocFromFile(filename, parser)
-    return _documentFactory(c_doc)
+    c_doc = _parseDocFromFile(_utf8(filename), parser)
+    return _documentFactory(c_doc, parser)
 
 cdef _Document _parseMemoryDocument(text, parser):
     cdef xmlDoc* c_doc
     if python.PyUnicode_Check(text):
         text = _stripDeclaration(_utf8(text))
     c_doc = _parseDoc(text, parser)
-    return _documentFactory(c_doc)
+    return _documentFactory(c_doc, parser)
 
-cdef _Document _documentFactory(xmlDoc* c_doc):
+cdef _Document _documentFactory(xmlDoc* c_doc, parser):
     cdef _Document result
     result = _Document()
     result._c_doc = c_doc
     result._ns_counter = 0
+    if parser is None:
+        parser = __DEFAULT_PARSER
+    result._parser = parser.copy()
     return result
 
 # to help with debugging
@@ -438,7 +490,7 @@
         fake_c_doc = _fakeRootDoc(doc._c_doc, self._c_node)
         c_doc = tree.xmlCopyDoc(fake_c_doc, 1) # recursive copy
         _destroyFakeDoc(doc._c_doc, fake_c_doc)
-        doc = _documentFactory(c_doc)
+        doc = _documentFactory(c_doc, doc._parser)
         return doc.getroot()
         
     def set(self, key, value):
@@ -1163,7 +1215,7 @@
     c_doc = _newDoc()
     c_node = _createElement(c_doc, name_utf, attrib, _extra)
     tree.xmlDocSetRootElement(c_doc, c_node)
-    doc = _documentFactory(c_doc)
+    doc = _documentFactory(c_doc, None)
     # add namespaces to node if necessary
     doc._setNodeNamespaces(c_node, ns_utf, nsmap)
     return _elementFactory(doc, c_node)
@@ -1171,13 +1223,15 @@
 def Comment(text=None):
     cdef _Document doc
     cdef xmlNode*  c_node
+    cdef xmlDoc*   c_doc
     if text is None:
         text = '  '
     else:
         text = ' %s ' % _utf8(text)
-    doc = _documentFactory( _newDoc() )
-    c_node = _createComment(doc._c_doc, text)
-    tree.xmlAddChild(<xmlNode*>doc._c_doc, c_node)
+    c_doc = _newDoc()
+    doc = _documentFactory(c_doc, None)
+    c_node = _createComment(c_doc, text)
+    tree.xmlAddChild(<xmlNode*>c_doc, c_node)
     return _commentFactory(doc, c_node)
 
 def SubElement(_Element _parent not None, _tag,
@@ -1196,6 +1250,7 @@
     cdef xmlNode* c_next
     cdef xmlNode* c_node
     cdef xmlNode* c_node_copy
+    cdef xmlDoc*  c_doc
     cdef _ElementTree etree
     cdef _Document doc
 
@@ -1204,7 +1259,8 @@
     elif file is not None:
         doc = _parseDocument(file, parser)
     else:
-        doc = _documentFactory( _newDoc() )
+        c_doc = _newDoc()
+        doc = _documentFactory(c_doc, parser)
 
     etree = _elementTreeFactory(doc, element)
 
@@ -1294,10 +1350,11 @@
 include "xmlerror.pxi"  # error and log handling
 include "xmlid.pxi"     # XMLID and IDDict
 include "nsclasses.pxi" # Namespace implementation and registry
+include "docloader.pxi" # Support for custom document loaders
+include "parser.pxi"    # XML Parser
 include "xslt.pxi"      # XPath and XSLT
 include "relaxng.pxi"   # RelaxNG
 include "xmlschema.pxi" # XMLSchema
-include "parser.pxi"    # XML Parser
 include "proxy.pxi"     # Proxy handling (element backpointers/memory/etc.)
 
 

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Thu Apr 27 22:51:43 2006
@@ -58,28 +58,122 @@
 __GLOBAL_PARSER_CONTEXT = _ParserContext()
 
 
+############################################################
+## support for custom document loaders
+############################################################
+
+cdef xmlparser.xmlParserInput* _local_resolver(char* c_url, char* c_pubid,
+                                               xmlParserCtxt* c_context):
+    cdef _ResolverContext context
+    cdef _InputDocument   doc_ref
+    cdef xmlparser.xmlParserInput* c_input
+    if c_context._private is NULL or \
+       not isinstance(<object>c_context._private, _ResolverContext):
+        if __DEFAULT_ENTITY_LOADER is NULL:
+            return NULL
+        return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
+
+    try:
+        if c_url is NULL:
+            url = None
+        else:
+            url = funicode(c_url)
+        if c_pubid is NULL:
+            pubid = None
+        else:
+            pubid = funicode(c_pubid)
+
+        context = <_ResolverContext>c_context._private
+        doc_ref = context._resolvers.resolve(url, pubid, context)
+    except Exception:
+        context._store_raised()
+        return NULL
+
+    if doc_ref is None:
+        if __DEFAULT_ENTITY_LOADER is NULL:
+            return NULL
+        return __DEFAULT_ENTITY_LOADER(c_url, c_pubid, c_context)
+
+    c_input = NULL
+    data = None
+    if doc_ref._type == PARSER_DATA_STRING:
+        data = doc_ref._data_utf
+        c_input = xmlparser.xmlNewStringInputStream(
+            c_context, _cstr(doc_ref._data_utf))
+    elif doc_ref._type == PARSER_DATA_FILENAME:
+        c_input = xmlparser.xmlNewInputFromFile(
+            c_context, _cstr(doc_ref._data_utf))
+    elif doc_ref._type == PARSER_DATA_FILE:
+        data = doc_ref._file.read()
+        c_input = xmlparser.xmlNewStringInputStream(
+            c_context, _cstr(data))
+
+    if data is not None:
+        context._storage.add(data)
+    return c_input
+
+cdef xmlparser.xmlExternalEntityLoader __DEFAULT_ENTITY_LOADER
+__DEFAULT_ENTITY_LOADER = xmlparser.xmlGetExternalEntityLoader()
+
+xmlparser.xmlSetExternalEntityLoader(_local_resolver)
+
+############################################################
+## Parsers
+############################################################
+
 cdef class BaseParser:
     cdef _ErrorLog _error_log
+    cdef readonly object resolvers
+    cdef _ResolverContext _context
     def __init__(self):
+        cdef _ResolverContext context
         self._error_log = _ErrorLog()
+        self.resolvers = _ResolverRegistry()
+        self._context = _ResolverContext(self.resolvers)
 
     property error_log:
         def __get__(self):
             return self._error_log.copy()
 
-    cdef xmlDoc* _handleResult(self, xmlParserCtxt* ctxt,
-                               xmlDoc* result) except NULL:
-        if ctxt.wellFormed:
-            __GLOBAL_PARSER_CONTEXT._initDocDict(result)
-        elif result is not NULL:
-            # free broken document
-            tree.xmlFreeDoc(result)
-            result = NULL
-        self._error_log.disconnect()
-        if result is NULL:
+    cdef _copy(self):
+        cdef BaseParser parser
+        parser = self.__class__()
+        parser.resolvers = self.resolvers.copy()
+        parser._context = _ResolverContext(parser.resolvers)
+        return parser
+
+    cdef _initContext(self, xmlParserCtxt* c_ctxt):
+        __GLOBAL_PARSER_CONTEXT._initParserDict(c_ctxt)
+        c_ctxt._private = <python.PyObject*>self._context
+
+cdef xmlDoc* _handleParseResult(xmlParserCtxt* ctxt, xmlDoc* result,
+                                char* c_filename) except NULL:
+    cdef _ResolverContext context
+    if ctxt.wellFormed:
+        __GLOBAL_PARSER_CONTEXT._initDocDict(result)
+    elif result is not NULL:
+        # free broken document
+        tree.xmlFreeDoc(result)
+        result = NULL
+
+    if ctxt._private is not NULL:
+        context = <_ResolverContext>ctxt._private
+        context._raise_if_stored()
+
+    if result is NULL:
+        if c_filename is not NULL and \
+               ctxt.lastError.domain == xmlerror.XML_FROM_IO:
+            if ctxt.lastError.message is not NULL:
+                message = "Error reading file %s: %s" % (
+                    funicode(c_filename), funicode(ctxt.lastError.message))
+            else:
+                message = "Error reading file %s" % funicode(c_filename)
+            raise IOError, message
+        elif ctxt.lastError.message is not NULL:
+            raise XMLSyntaxError, funicode(ctxt.lastError.message)
+        else:
             raise XMLSyntaxError
-        return result
-
+    return result
 
 ############################################################
 ## XML parser
@@ -110,12 +204,14 @@
     cdef xmlParserCtxt* _file_parser_ctxt
     cdef xmlParserCtxt* _memory_parser_ctxt
     def __init__(self, attribute_defaults=False, dtd_validation=False,
-                 no_network=False, ns_clean=False):
+                 load_dtd=False, no_network=False, ns_clean=False):
         cdef int parse_options
         self._file_parser_ctxt = NULL
         BaseParser.__init__(self)
 
         parse_options = _XML_DEFAULT_PARSE_OPTIONS
+        if load_dtd:
+            parse_options = parse_options | xmlparser.XML_PARSE_DTDLOAD
         if dtd_validation:
             parse_options = parse_options | xmlparser.XML_PARSE_DTDLOAD | \
                             xmlparser.XML_PARSE_DTDVALID
@@ -135,6 +231,12 @@
         if self._memory_parser_ctxt != NULL:
             xmlparser.xmlFreeParserCtxt(self._memory_parser_ctxt)
 
+    def copy(self):
+        cdef XMLParser parser
+        parser = self._copy()
+        parser._parse_options = self._parse_options
+        return parser
+
     cdef xmlParserCtxt* _createContext(self) except NULL:
         cdef xmlParserCtxt* pctxt
         pctxt = xmlparser.xmlNewParserCtxt()
@@ -143,7 +245,7 @@
             raise ParserError, "Failed to create parser context"
         return pctxt
 
-    cdef xmlDoc* _parseDoc(self, text_utf) except NULL:
+    cdef xmlDoc* _parseDoc(self, char* c_text) except NULL:
         """Parse document, share dictionary if possible.
         """
         cdef xmlDoc* result
@@ -154,13 +256,13 @@
         if pctxt is NULL:
             pctxt = self._createContext()
             self._memory_parser_ctxt = pctxt
-
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadDoc(
-            pctxt, _cstr(text_utf), NULL, NULL, self._parse_options)
-        return self._handleResult(pctxt, result)
+            pctxt, c_text, NULL, NULL, self._parse_options)
+        self._error_log.disconnect()
+        return _handleParseResult(pctxt, result, NULL)
 
-    cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
+    cdef xmlDoc* _parseDocFromFile(self, char* c_filename) except NULL:
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
         self._error_log.connect()
@@ -168,15 +270,47 @@
         if pctxt is NULL:
             pctxt = self._createContext()
             self._file_parser_ctxt = pctxt
-
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = xmlparser.xmlCtxtReadFile(
-            pctxt, filename, NULL, self._parse_options)
-        if result is NULL:
-            if pctxt.lastError.domain == xmlerror.XML_FROM_IO:
-                self._error_log.disconnect()
-                raise IOError, "Could not open file %s" % filename
-        return self._handleResult(pctxt, result)
+            pctxt, c_filename, NULL, self._parse_options)
+        self._error_log.disconnect()
+        return _handleParseResult(pctxt, result, c_filename)
+
+cdef xmlDoc* _internalParseDoc(char* c_text, int options,
+                               _ResolverContext context) except NULL:
+    # internal parser function for XSLT
+    cdef xmlParserCtxt* pctxt
+    cdef xmlDoc* c_doc
+    pctxt = xmlparser.xmlNewParserCtxt()
+    if pctxt is NULL:
+        return NULL
+    __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+    pctxt._private = <python.PyObject*>context
+    c_doc = xmlparser.xmlCtxtReadDoc(
+        pctxt, c_text, NULL, NULL, options)
+    try:
+        c_doc = _handleParseResult(pctxt, c_doc, NULL)
+    finally:
+        xmlparser.xmlFreeParserCtxt(pctxt)
+    return c_doc
+
+cdef xmlDoc* _internalParseDocFromFile(char* c_filename, int options,
+                                       _ResolverContext context) except NULL:
+    # internal parser function for XSLT
+    cdef xmlParserCtxt* pctxt
+    cdef xmlDoc* c_doc
+    pctxt = xmlparser.xmlNewParserCtxt()
+    if pctxt is NULL:
+        return NULL
+    __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+    pctxt._private = <python.PyObject*>context
+    c_doc = xmlparser.xmlCtxtReadFile(
+        pctxt, c_filename, NULL, options)
+    try:
+        c_doc = _handleParseResult(pctxt, c_doc, c_filename)
+    finally:
+        xmlparser.xmlFreeParserCtxt(pctxt)
+    return c_doc
 
 
 cdef XMLParser __DEFAULT_XML_PARSER
@@ -203,6 +337,8 @@
     else:
         raise TypeError, "Invalid parser"
 
+def get_default_parser():
+    return __DEFAULT_PARSER
 
 ############################################################
 ## HTML parser
@@ -248,15 +384,19 @@
         if self._memory_parser_ctxt != NULL:
             htmlparser.htmlFreeParserCtxt(self._memory_parser_ctxt)
 
-    cdef xmlDoc* _parseDoc(self, text_utf) except NULL:
+    def copy(self):
+        cdef HTMLParser parser
+        parser = self._copy()
+        parser._parse_options = self._parse_options
+        return parser
+
+    cdef xmlDoc* _parseDoc(self, char* c_text) except NULL:
         """Parse HTML document, share dictionary if possible.
         """
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
-        cdef char* c_text
         cdef int c_len
         self._error_log.connect()
-        c_text = _cstr(text_utf)
         pctxt = self._memory_parser_ctxt
         if pctxt is NULL:
             pctxt = htmlparser.htmlCreateMemoryParserCtxt('dummy', 5)
@@ -264,30 +404,32 @@
                 self._error_log.disconnect()
                 raise ParserError, "Failed to create parser context"
             self._memory_parser_ctxt = pctxt
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadDoc(
             pctxt, c_text, NULL, NULL, self._parse_options)
-        return self._handleResult(pctxt, result)
+        self._error_log.disconnect()
+        return _handleParseResult(pctxt, result, NULL)
 
-    cdef xmlDoc* _parseDocFromFile(self, char* filename) except NULL:
+    cdef xmlDoc* _parseDocFromFile(self, char* c_filename) except NULL:
         cdef xmlDoc* result
         cdef xmlParserCtxt* pctxt
         cdef int parser_error
         self._error_log.connect()
         pctxt = self._file_parser_ctxt
         if pctxt is NULL:
-            pctxt = htmlparser.htmlCreateFileParserCtxt(filename, NULL)
+            pctxt = htmlparser.htmlCreateFileParserCtxt(c_filename, NULL)
             if pctxt is NULL:
                 self._error_log.disconnect()
                 warnings = self._error_log.filter_from_warnings()
                 if warnings and warnings[-1].domain == xmlerror.XML_FROM_IO:
-                    raise IOError, "Could not open file %s" % filename
+                    raise IOError, "Could not open file %s" % c_filename
                 raise ParserError, "Failed to create parser context"
             self._file_parser_ctxt = pctxt
-        __GLOBAL_PARSER_CONTEXT._initParserDict(pctxt)
+        self._initContext(pctxt)
         result = htmlparser.htmlCtxtReadFile(
-            pctxt, filename, NULL, self._parse_options)
-        return self._handleResult(pctxt, result)
+            pctxt, c_filename, NULL, self._parse_options)
+        self._error_log.disconnect()
+        return _handleParseResult(pctxt, result, c_filename)
 
 cdef HTMLParser __DEFAULT_HTML_PARSER
 __DEFAULT_HTML_PARSER = HTMLParser()
@@ -301,9 +443,9 @@
         parser = __DEFAULT_PARSER
     __GLOBAL_PARSER_CONTEXT._initParser()
     if isinstance(parser, XMLParser):
-        return (<XMLParser>parser)._parseDoc(text_utf)
+        return (<XMLParser>parser)._parseDoc(_cstr(text_utf))
     elif isinstance(parser, HTMLParser):
-        return (<HTMLParser>parser)._parseDoc(text_utf)
+        return (<HTMLParser>parser)._parseDoc(_cstr(text_utf))
     else:
         raise TypeError, "invalid parser"
 

Modified: lxml/trunk/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_etree.py	(original)
+++ lxml/trunk/src/lxml/tests/test_etree.py	Thu Apr 27 22:51:43 2006
@@ -49,7 +49,67 @@
         f = open(fileInTestDir('test_broken.xml'), 'r')
         self.assertRaises(SyntaxError, parse, f)
         f.close()
-        
+
+    def test_resolve_string_dtd(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(dtd_validation=True)
+        assertEqual = self.assertEqual
+        test_url = u"__nosuch.dtd"
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                assertEqual(url, test_url)
+                return self.resolve_string(
+                    u'<!ENTITY myentity "%s">' % url, context)
+
+        parser.resolvers.add(MyResolver())
+
+        xml = u'<!DOCTYPE doc SYSTEM "%s"><doc>&myentity;</doc>' % test_url
+        tree = parse(StringIO(xml), parser)
+        root = tree.getroot()
+        self.assertEquals(root.text, test_url)
+
+    def test_resolve_empty(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(dtd_validation=True)
+        assertEqual = self.assertEqual
+        test_url = u"__nosuch.dtd"
+
+        class check(object):
+            resolved = False
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                assertEqual(url, test_url)
+                check.resolved = True
+                return self.resolve_empty(context)
+
+        parser.resolvers.add(MyResolver())
+
+        xml = u'<!DOCTYPE doc SYSTEM "%s"><doc>&myentity;</doc>' % test_url
+        tree = parse(StringIO(xml), parser)
+        self.assert_(check.resolved)
+
+        root = tree.getroot()
+        self.assertEquals(root.text, None)
+
+    def test_resolve_error(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(dtd_validation=True)
+        test_url = u"__nosuch.dtd"
+
+        class _LocalException(Exception):
+            pass
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                raise _LocalException
+
+        parser.resolvers.add(MyResolver())
+
+        xml = u'<!DOCTYPE doc SYSTEM "test"><doc>&myentity;</doc>'
+        self.assertRaises(_LocalException, parse, StringIO(xml), parser)
+
     # TypeError in etree, AssertionError in ElementTree;
     def test_setitem_assert(self):
         Element = self.etree.Element
@@ -357,6 +417,8 @@
     suite.addTests([unittest.makeSuite(ETreeC14NTestCase)])
     suite.addTests(
         [doctest.DocFileSuite('../../../doc/api.txt')])
+    suite.addTests(
+        [doctest.DocFileSuite('../../../doc/resolvers.txt')])
     return suite
 
 if __name__ == '__main__':

Modified: lxml/trunk/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xslt.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xslt.py	Thu Apr 27 22:51:43 2006
@@ -29,6 +29,36 @@
 <foo>B</foo>
 ''',
                           st.tostring(res))
+
+    def test_exslt(self):
+        tree = self.parse('<a><b>B</b><c>C</c></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:math="http://exslt.org/math"
+    xmlns:str="http://exslt.org/strings"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    exclude-result-prefixes="math str xsl">
+  <xsl:template match="text()">
+    <xsl:value-of select="str:align(string(.), '---', 'center')" />
+  </xsl:template>
+  <xsl:template match="*">
+    <xsl:copy>
+      <xsl:attribute name="pi">
+        <xsl:value-of select="math:constant('PI', count(*)+2)"/>
+      </xsl:attribute>
+      <xsl:apply-templates/>
+    </xsl:copy>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        st = etree.XSLT(style)
+        res = st(tree)
+        self.assertEquals('''\
+<?xml version="1.0"?>
+<a pi="3.14"><b pi="3">-B-</b><c pi="3">-C-</c></a>
+''',
+                          st.tostring(res))
+
     def test_xslt_input(self):
         tree = self.parse('<a><b>B</b><c>C</c></a>')
         style = self.parse('''\
@@ -336,12 +366,12 @@
                           '<A>X</A>')
 
     def test_xslt_document_XML(self):
-        # make sure document('') works from loaded files
+        # make sure document('') works from parsed strings
         xslt = etree.XSLT(etree.XML("""\
 <xsl:stylesheet version="1.0"
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
   <xsl:template match="/">
-    <test><xsl:copy-of select="document('')//test"/></test>
+    <test>TEXT<xsl:copy-of select="document('')//test"/></test>
   </xsl:template>
 </xsl:stylesheet>
 """))
@@ -351,6 +381,8 @@
                           'test')
         self.assertEquals(root[0].tag,
                           'test')
+        self.assertEquals(root[0].text,
+                          'TEXT')
         self.assertEquals(root[0][0].tag,
                           '{http://www.w3.org/1999/XSL/Transform}copy-of')
 

Modified: lxml/trunk/src/lxml/tree.pxd
==============================================================================
--- lxml/trunk/src/lxml/tree.pxd	(original)
+++ lxml/trunk/src/lxml/tree.pxd	Thu Apr 27 22:51:43 2006
@@ -77,6 +77,7 @@
         xmlDict* dict
         xmlHashTable* ids
         char* URL
+        void* _private
         
     ctypedef struct xmlAttr:
         void* _private

Modified: lxml/trunk/src/lxml/xmlparser.pxd
==============================================================================
--- lxml/trunk/src/lxml/xmlparser.pxd	(original)
+++ lxml/trunk/src/lxml/xmlparser.pxd	Thu Apr 27 22:51:43 2006
@@ -1,6 +1,9 @@
 from tree cimport xmlDoc, xmlDict
 from xmlerror cimport xmlError
 
+cdef extern from "libxml/tree.h":
+    ctypedef struct xmlParserInput
+
 cdef extern from "libxml/parser.h":
 
     cdef xmlDict* xmlDictCreate()
@@ -10,6 +13,7 @@
     ctypedef struct xmlParserCtxt:
         xmlDoc* myDoc
         xmlDict* dict
+        void* _private
         int wellFormed
         xmlError lastError
         
@@ -30,8 +34,8 @@
         XML_PARSE_NSCLEAN = 8192 # remove redundant namespaces declarations
         XML_PARSE_NOCDATA = 16384 # merge CDATA as text nodes
         XML_PARSE_NOXINCNODE = 32768 # do not generate XINCLUDE START/END nodes
-# libxml2 2.6.21+ only:
-#        XML_PARSE_COMPACT = 65536 # compact small text nodes
+        # libxml2 2.6.21+ only:
+        #XML_PARSE_COMPACT = 65536 # compact small text nodes
        
     cdef void xmlInitParser()
     cdef xmlParserCtxt* xmlNewParserCtxt()
@@ -42,3 +46,18 @@
                                 int options)
     cdef xmlDoc* xmlCtxtReadFile(xmlParserCtxt* ctxt,
                                  char* filename, char* encoding, int options)
+
+# entity loaders:
+
+    ctypedef xmlParserInput* (*xmlExternalEntityLoader)(char * URL,
+                                                        char * ID, 
+                                                        xmlParserCtxt* context)
+    cdef xmlExternalEntityLoader xmlGetExternalEntityLoader()
+    cdef void xmlSetExternalEntityLoader(xmlExternalEntityLoader f)
+
+cdef extern from "libxml/parserInternals.h":
+    cdef xmlParserInput* xmlNewStringInputStream(xmlParserCtxt* ctxt, 
+						 char* buffer)
+    cdef xmlParserInput* xmlNewInputFromFile(xmlParserCtxt* ctxt, 
+                                             char* filename)
+    cdef void xmlFreeInputStream(xmlParserInput* input)

Modified: lxml/trunk/src/lxml/xslt.pxd
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxd	(original)
+++ lxml/trunk/src/lxml/xslt.pxd	Thu Apr 27 22:51:43 2006
@@ -1,27 +1,43 @@
-from tree cimport xmlDoc
+from tree cimport xmlDoc, xmlDict
 from xpath cimport xmlXPathContext, xmlXPathFunction
 
 cdef extern from "libxslt/xsltInternals.h":
+    ctypedef struct xsltDocument:
+        xmlDoc* doc
+
     ctypedef struct xsltStylesheet:
-        pass
+        xmlDoc* doc
 
     ctypedef struct xsltTransformContext:
+        xsltStylesheet* style
         xmlXPathContext* xpathCtxt
-    
+        xsltDocument* document
+
     cdef xsltStylesheet* xsltParseStylesheetDoc(xmlDoc* doc)
     cdef void xsltFreeStylesheet(xsltStylesheet* sheet)
-    
-#cdef extern from "libxslt/xslt.h":
-#    pass
 
 cdef extern from "libxslt/extensions.h":
     cdef int xsltRegisterExtFunction(xsltTransformContext* ctxt,
                                      char* name,
-                                     char * URI,
+                                     char* URI,
                                      xmlXPathFunction function)
     cdef int xsltRegisterExtModuleFunction(char* name, char* URI,
                                            xmlXPathFunction function)
     cdef int xsltUnregisterExtModuleFunction(char* name, char* URI)
+    cdef xmlXPathFunction xsltExtModuleFunctionLookup(char* name, char* URI)
+
+cdef extern from "libxslt/documents.h":
+    ctypedef enum xsltLoadType:
+        XSLT_LOAD_START
+        XSLT_LOAD_STYLESHEET
+        XSLT_LOAD_DOCUMENT
+
+    ctypedef xmlDoc* (*xsltDocLoaderFunc)(char* URI, xmlDict* dict,
+                                          int options,
+                                          void* ctxt,
+                                          xsltLoadType type)
+    cdef xsltDocLoaderFunc xsltDocDefaultLoader
+    cdef void xsltSetLoaderFunc(xsltDocLoaderFunc f)
 
 cdef extern from "libxslt/transform.h":
     cdef xmlDoc* xsltApplyStylesheet(xsltStylesheet* style, xmlDoc* doc,
@@ -45,4 +61,15 @@
     cdef void xsltSetTransformErrorFunc(xsltTransformContext*,
                                         void* ctxt,
                                         void (*handler)(void* ctxt, char* msg, ...))
-    
+
+cdef extern from "libxslt/extra.h":
+    cdef char* XSLT_LIBXSLT_NAMESPACE
+    cdef char* XSLT_XALAN_NAMESPACE
+    cdef char* XSLT_SAXON_NAMESPACE
+    cdef char* XSLT_XT_NAMESPACE
+
+    cdef xmlXPathFunction xsltFunctionNodeSet
+    cdef void xsltRegisterAllExtras()
+
+cdef extern from "libexslt/exslt.h":
+    cdef void exsltRegisterAll()

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Thu Apr 27 22:51:43 2006
@@ -39,9 +39,8 @@
     cdef object _utf_refs
     cdef object _temp_functions
     # for exception handling and temporary reference keeping:
-    cdef object _temp_elements
-    cdef object _temp_docs
-    cdef object _exc_info
+    cdef _TempStore _temp_refs
+    cdef _ExceptionContext _exc
 
     def __init__(self, namespaces, extensions):
         self._xpathCtxt = NULL
@@ -62,12 +61,11 @@
             extensions = new_extensions or None
 
         self._doc        = None
-        self._exc_info   = None
+        self._exc        = _ExceptionContext()
         self._extensions = extensions
         self._namespaces = namespaces
         self._registered_namespaces = []
-        self._temp_elements = {}
-        self._temp_docs = {}
+        self._temp_refs = _TempStore()
 
     cdef object _to_utf(self, s):
         "Convert to UTF-8 and keep a reference to the encoded string"
@@ -86,9 +84,9 @@
         xpathCtxt.userData = <void*>self
 
     cdef _register_context(self, _Document doc, int allow_none_namespace):
-        self._doc      = doc
-        self._exc_info = None
-        self._temp_functions.clear()
+        self._doc = doc
+        self._exc.clear()
+        python.PyDict_Clear(self._temp_functions)
         namespaces = self._namespaces
         if namespaces is not None:
             self.registerNamespaces(namespaces)
@@ -154,8 +152,7 @@
 
     cdef _release_temp_refs(self):
         "Free temporarily referenced objects from this context."
-        python.PyDict_Clear(self._temp_elements)
-        python.PyDict_Clear(self._temp_docs)
+        self._temp_refs.clear()
         
     cdef _hold(self, obj):
         """A way to temporarily hold references to nodes in the evaluator.
@@ -173,27 +170,128 @@
             if isinstance(o, _NodeBase):
                 element = <_NodeBase>o
                 #print "Holding element:", <int>element._c_node
-                python.PyDict_SetItem(self._temp_elements, id(element), element)
+                self._temp_refs.add(element)
                 #print "Holding document:", <int>element._doc._c_doc
-                python.PyDict_SetItem(self._temp_docs, id(element._doc), element._doc)
+                self._temp_refs.add(element._doc)
 
 cdef xpath.xmlXPathFunction _function_check(void* ctxt, char* c_name, char* c_ns_uri):
     cdef _BaseContext context
     if c_name is NULL:
         return NULL
     if c_ns_uri is NULL:
-        c_ns_uri = ''
         ns_uri = None
     else:
         ns_uri = c_ns_uri
     context = <_BaseContext>ctxt
-    if context._lookup_extension(ns_uri, c_name) is None:
-        return NULL
-    else:
+    if context._lookup_extension(ns_uri, c_name) is not None:
         return _xpath_function_call
+    else:
+        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
 
  
 ################################################################################
+# XSLT document loaders
+
+cdef class _XSLTResolverContext(_ResolverContext):
+    cdef xmlDoc* _c_style_doc
+    cdef object _style_url_utf
+    cdef object _style_doc_utf
+    cdef BaseParser _parser
+    def __init__(self, BaseParser parser not None):
+        _ResolverContext.__init__(self, parser.resolvers)
+        self._parser = parser
+        self._c_style_doc = NULL
+        self._style_url_utf = None
+        self._style_doc_utf = None
+
+cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
+                         void* c_ctxt, xslt.xsltLoadType c_type):
+    cdef xmlDoc* c_doc
+    cdef _ResolverRegistry resolvers
+    cdef _InputDocument doc_ref
+    cdef _XSLTResolverContext xslt_resolver_context
+    cdef _XSLTResolverContext doc_resolver_context
+    cdef _XSLTResolverContext resolver_context
+    cdef XMLParser parser
+    # find resolver contexts of stylesheet and transformed doc
+    c_doc = NULL
+    doc_resolver_context = None
+    if c_type == xslt.XSLT_LOAD_DOCUMENT:
+        c_doc = (<xslt.xsltTransformContext*>c_ctxt).document.doc
+        if c_doc is not NULL and c_doc._private is not NULL:
+            if isinstance(<object>c_doc._private, _XSLTResolverContext):
+                doc_resolver_context = <_XSLTResolverContext>c_doc._private
+        c_doc = (<xslt.xsltTransformContext*>c_ctxt).style.doc
+    elif c_type == xslt.XSLT_LOAD_STYLESHEET:
+        c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
+
+    if c_doc is NULL or c_doc._private is NULL or \
+           not isinstance(<object>c_doc._private, _XSLTResolverContext):
+        # can't call Python without context, fall back to default loader
+        return XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, parse_options, c_ctxt, c_type)
+
+    xslt_resolver_context = <_XSLTResolverContext>c_doc._private
+
+    # quick check if we are looking for the current stylesheet
+    c_doc = xslt_resolver_context._c_style_doc
+    if c_doc is not NULL and c_doc.URL is not NULL:
+        if tree.strcmp(c_uri, c_doc.URL) == 0:
+            return tree.xmlCopyDoc(c_doc, 1)
+
+    # call the Python document loaders
+    c_doc = NULL
+    resolver_context = xslt_resolver_context # currently use only XSLT resolvers
+    resolvers = resolver_context._resolvers
+    try:
+        uri = funicode(c_uri)
+        doc_ref = resolvers.resolve(uri, None, resolver_context)
+
+        if doc_ref is not None:
+            if doc_ref._type == PARSER_DATA_EMPTY:
+                c_doc = _newDoc()
+            if doc_ref._type == PARSER_DATA_STRING:
+                c_doc = _internalParseDoc(
+                    _cstr(doc_ref._data_utf), parse_options,
+                    resolver_context)
+            elif doc_ref._type == PARSER_DATA_FILE:
+                data = doc_ref._file.read()
+                c_doc = _internalParseDoc(
+                    _cstr(data), parse_options,
+                    resolver_context)
+            elif doc_ref._type == PARSER_DATA_FILENAME:
+                c_doc = _internalParseDocFromFile(
+                    _cstr(doc_ref._data_utf), parse_options,
+                    resolver_context)
+            if c_doc is not NULL and c_doc.URL is NULL:
+                c_doc.URL = tree.xmlStrdup(c_uri)
+
+    except Exception, e:
+        xslt_resolver_context._store_raised()
+        return NULL
+
+    if c_doc is NULL:
+        c_doc = XSLT_DOC_DEFAULT_LOADER(
+            c_uri, c_dict, parse_options, c_ctxt, c_type)
+        if c_doc is NULL:
+            message = "Cannot resolve URI %s" % funicode(c_uri)
+            if c_type == xslt.XSLT_LOAD_DOCUMENT:
+                exception = XSLTApplyError(message)
+            else:
+                exception = XSLTParseError(message)
+            xslt_resolver_context._store_exception(exception)
+            return NULL
+    if c_doc is not NULL and c_doc._private is NULL:
+        c_doc._private = <python.PyObject*>xslt_resolver_context
+    return c_doc
+
+cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
+XSLT_DOC_DEFAULT_LOADER = xslt.xsltDocDefaultLoader
+
+xslt.xsltSetLoaderFunc(_doc_loader)
+
+
+################################################################################
 # XSLT
 
 cdef class _XSLTContext(_BaseContext):
@@ -223,13 +321,14 @@
 
     cdef _registerLocalExtensionFunction(self, ns_utf, name_utf, function):
         extensions = self._extensions
-        if self._extensions is None:
+        if extensions is None:
             self._extensions = {ns_utf:{name_utf:function}}
         else:
-            if ns_utf in self._extensions:
-                self._extensions[ns_utf][name_utf] = function
+            if ns_utf in extensions:
+                ns_extensions = extensions[ns_utf]
             else:
-                self._extensions[ns_utf] = ns_extensions = {name_utf:function}
+                ns_extensions = extensions[ns_utf] = {}
+            python.PyDict_SetItem(ns_extensions, name_utf, function)
         xslt.xsltRegisterExtFunction(
             self._xsltCtxt, _cstr(name_utf), _cstr(ns_utf),
             _xpath_function_call)
@@ -252,12 +351,10 @@
     """
     cdef _XSLTContext _context
     cdef xslt.xsltStylesheet* _c_style
+    cdef _XSLTResolverContext _xslt_resolver_context
     cdef _ExsltRegExp _regexp
-    cdef object _doc_url_utf
-    
+
     def __init__(self, xslt_input, extensions=None, regexp=True):
-        # make a copy of the document as stylesheet needs to assume it
-        # doesn't change
         cdef xslt.xsltStylesheet* c_style
         cdef xmlDoc* c_doc
         cdef xmlDoc* fake_c_doc
@@ -267,23 +364,30 @@
         doc = _documentOrRaise(xslt_input)
         root_node = _rootNodeOf(xslt_input)
 
+        # make a copy of the document as stylesheet parsing modifies it
         fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
         c_doc = tree.xmlCopyDoc(fake_c_doc, 1)
         _destroyFakeDoc(doc._c_doc, fake_c_doc)
 
-        # XXX work around bug in xmlCopyDoc (fix is upcoming in new release
-        # of libxml2)
-        if c_doc.URL is not NULL and c_doc.URL != doc._c_doc.URL:
+        # make sure we always have a stylesheet URL
+        if c_doc.URL is not NULL:
+            # handle a bug in older libxml2 versions
             tree.xmlFree(c_doc.URL)
         if doc._c_doc.URL is not NULL:
-            self._doc_url_utf = doc._c_doc.URL
             c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
         else:
-            self._doc_url_utf = "__STRING__XSLT__%s" % id(self)
-            c_doc.URL = tree.xmlStrdup(_cstr(self._doc_url_utf))
+            doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
+            c_doc.URL = tree.xmlStrdup(_cstr(doc_url_utf))
+
+        self._xslt_resolver_context = _XSLTResolverContext(doc._parser)
+        # keep a copy in case we need to access the stylesheet via 'document()'
+        self._xslt_resolver_context._c_style_doc = tree.xmlCopyDoc(c_doc, 1)
+        c_doc._private = <python.PyObject*>self._xslt_resolver_context
 
         c_style = xslt.xsltParseStylesheetDoc(c_doc)
         if c_style is NULL:
+            tree.xmlFreeDoc(c_doc)
+            self._xslt_resolver_context._raise_if_stored()
             raise XSLTParseError, "Cannot parse style sheet"
         self._c_style = c_style
 
@@ -295,6 +399,9 @@
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
         
     def __dealloc__(self):
+        if self._xslt_resolver_context is not None and \
+               self._xslt_resolver_context._c_style_doc is not NULL:
+            tree.xmlFreeDoc(self._xslt_resolver_context._c_style_doc)
         # this cleans up copy of doc as well
         xslt.xsltFreeStylesheet(self._c_style)
 
@@ -302,16 +409,20 @@
         cdef _Document input_doc
         cdef _NodeBase root_node
         cdef _Document result_doc
+        cdef _XSLTResolverContext resolver_context
         cdef xslt.xsltTransformContext* transform_ctxt
         cdef xmlDoc* c_result
         cdef xmlDoc* c_doc
         cdef char** params
+        cdef void* ptemp
         cdef int i
-        cdef int j
 
         input_doc = _documentOrRaise(_input)
         root_node = _rootNodeOf(_input)
 
+        resolver_context = _XSLTResolverContext(input_doc._parser)
+        resolver_context._c_style_doc = self._xslt_resolver_context._c_style_doc
+
         c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
 
         transform_ctxt = xslt.xsltNewTransformContext(self._c_style, c_doc)
@@ -319,6 +430,9 @@
             _destroyFakeDoc(input_doc._c_doc, c_doc)
             raise XSLTApplyError, "Error preparing stylesheet run"
 
+        ptemp = c_doc._private
+        c_doc._private = <python.PyObject*>resolver_context
+
         if _kw:
             # allocate space for parameters
             # * 2 as we want an entry for both key and value,
@@ -352,19 +466,25 @@
             cstd.free(params)
 
         self._context.free_context()
+        c_doc._private = ptemp # restore _private before _destroyFakeDoc!
         _destroyFakeDoc(input_doc._c_doc, c_doc)
 
+        if self._xslt_resolver_context._has_raised():
+            if c_result is not NULL:
+                tree.xmlFreeDoc(c_result)
+            self._xslt_resolver_context._raise_if_stored()
+
         if c_result is NULL:
             raise XSLTApplyError, "Error applying stylesheet"
 
-        result_doc = _documentFactory(c_result)
+        result_doc = _documentFactory(c_result, input_doc._parser)
         return _xsltResultTreeFactory(result_doc, self)
 
     def apply(self, _input, **_kw):
-        return self(_input, **_kw)
+        return self.__call__(_input, **_kw)
 
     def tostring(self, _ElementTree result_tree):
-        """Save result doc to string using stylesheet as guidance.
+        """Save result doc to string based on stylesheet output method.
         """
         return str(result_tree)
 
@@ -397,6 +517,23 @@
 cdef void _unregister_global_xslt_function(char* ns_uri, char* name):
     xslt.xsltUnRegisterExtModuleFunction(ns_uri, name)
 
+# do not register all libxslt extra function, provide only "node-set"
+# functions like "output" and "write" are a potential security risk
+#xslt.xsltRegisterAllExtras()
+xslt.xsltRegisterExtModuleFunction("node-set",
+                                   xslt.XSLT_LIBXSLT_NAMESPACE,
+                                   xslt.xsltFunctionNodeSet)
+xslt.xsltRegisterExtModuleFunction("node-set",
+                                   xslt.XSLT_SAXON_NAMESPACE,
+                                   xslt.xsltFunctionNodeSet)
+xslt.xsltRegisterExtModuleFunction("node-set",
+                                   xslt.XSLT_XT_NAMESPACE,
+                                   xslt.xsltFunctionNodeSet)
+
+# enable EXSLT support for XSLT
+xslt.exsltRegisterAll()
+
+
 ################################################################################
 # EXSLT regexp implementation
 
@@ -541,17 +678,21 @@
         self._context = _XPathContext(namespaces, extensions, variables)
 
     cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
-        _exc_info = self._context._exc_info
-        if _exc_info is not None:
-            type, value, traceback = _exc_info
-            raise type, value, traceback
+        if self._context._exc._has_raised():
+            if xpathObj is not NULL:
+                xpath.xmlXPathFreeObject(xpathObj)
+                xpathObj = NULL
+            self._context._exc._raise_if_stored()
+
         if xpathObj is NULL:
             raise XPathSyntaxError, "Error in xpath expression."
+
         try:
             result = _unwrapXPathObject(xpathObj, doc)
         except XPathResultError:
             xpath.xmlXPathFreeObject(xpathObj)
             raise
+
         xpath.xmlXPathFreeObject(xpathObj)
         return result
 
@@ -795,6 +936,7 @@
     cdef xmlNode* c_node
     cdef char* s
     cdef _NodeBase element
+    cdef int i
     result = []
     if xpathObj.nodesetval is NULL:
         return result
@@ -819,7 +961,8 @@
     cdef xpath.xmlXPathContext* rctxt
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
-    cdef _BaseContext extensions
+    cdef _BaseContext context
+    cdef int i
 
     rctxt = ctxt.context
 
@@ -831,15 +974,15 @@
         uri = None
 
     # get our evaluator
-    extensions = <_BaseContext>(rctxt.userData)
+    context = <_BaseContext>(rctxt.userData)
 
     # lookup up the extension function in the context
-    f = extensions._lookup_extension(uri, name)
+    f = context._lookup_extension(uri, name)
 
     args = []
-    doc = extensions._doc
+    doc = context._doc
     for i from 0 <= i < nargs:
-        args.append(_unwrapXPathObject(xpath.valuePop(ctxt), doc))
+        python.PyList_Append(args, _unwrapXPathObject(xpath.valuePop(ctxt), doc))
     args.reverse()
 
     try:
@@ -847,13 +990,13 @@
         res = f(None, *args)
         # hold python objects temporarily so that they won't get deallocated
         # during processing
-        extensions._hold(res)
+        context._hold(res)
         # now wrap for XPath consumption
         obj = _wrapXPathObject(res)
     except:
         xpath.xmlXPathErr(
             ctxt,
             xmlerror.XML_XPATH_EXPR_ERROR - xmlerror.XML_XPATH_EXPRESSION_OK)
-        extensions._exc_info = sys.exc_info()
+        context._exc._store_raised()
         return
     xpath.valuePush(ctxt, obj)
From scoder at codespeak.net  Fri Apr 28 08:14:41 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 08:14:42 2006
Subject: [Lxml-checkins] r26492 - lxml/trunk/src/lxml
Message-ID: <20060428061441.BC2E5100A7@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 08:14:40 2006
New Revision: 26492

Modified:
   lxml/trunk/src/lxml/nsclasses.pxi
   lxml/trunk/src/lxml/xslt.pxi
Log:
large cleanup in XSLT function calling: store function for next call after lookup, do not xsltExtModuleRegister XSLT functions (looked up by Python anyway)

Modified: lxml/trunk/src/lxml/nsclasses.pxi
==============================================================================
--- lxml/trunk/src/lxml/nsclasses.pxi	(original)
+++ lxml/trunk/src/lxml/nsclasses.pxi	Fri Apr 28 08:14:40 2006
@@ -143,7 +143,6 @@
         else:
             name_utf = _utf8(name)
         self._extensions[name_utf] = function
-        _register_global_xslt_function(self._c_ns_uri_utf, _cstr(name_utf))
 
     def __getitem__(self, name):
         cdef python.PyObject* dict_result
@@ -153,13 +152,6 @@
             raise KeyError, "Name not registered."
         return <object>dict_result
 
-    def clear(self):
-        cdef char* c_uri_utf
-        c_uri_utf = self._c_ns_uri_utf
-        for name_utf in self._extensions:
-            _unregister_global_xslt_function(c_uri_utf, _cstr(name_utf))
-        _NamespaceRegistry.clear(self)
-
     def __repr__(self):
         return "FunctionNamespace(%r)" % self._ns_uri
 

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 08:14:40 2006
@@ -37,7 +37,8 @@
     cdef object _namespaces
     cdef object _registered_namespaces
     cdef object _utf_refs
-    cdef object _temp_functions
+    cdef object _function_cache
+    cdef object _called_function
     # for exception handling and temporary reference keeping:
     cdef _TempStore _temp_refs
     cdef _ExceptionContext _exc
@@ -45,7 +46,8 @@
     def __init__(self, namespaces, extensions):
         self._xpathCtxt = NULL
         self._utf_refs = {}
-        self._temp_functions = {}
+        self._function_cache = {}
+        self._called_function = None
 
         # convert old format extensions to UTF-8
         if isinstance(extensions, (list, tuple)):
@@ -86,7 +88,7 @@
     cdef _register_context(self, _Document doc, int allow_none_namespace):
         self._doc = doc
         self._exc.clear()
-        python.PyDict_Clear(self._temp_functions)
+        python.PyDict_Clear(self._function_cache)
         namespaces = self._namespaces
         if namespaces is not None:
             self.registerNamespaces(namespaces)
@@ -130,12 +132,14 @@
     
     # extension functions
 
-    cdef _lookup_extension(self, ns_uri_utf, name_utf):
+    cdef _prepare_function_call(self, ns_uri_utf, name_utf):
         cdef python.PyObject* dict_result
         key = (ns_uri_utf, name_utf)
-        dict_result = python.PyDict_GetItem(self._temp_functions, key)
+        dict_result = python.PyDict_GetItem(self._function_cache, key)
         if dict_result is not NULL:
-            return <object>dict_result
+            function = <object>dict_result
+            self._called_function = function
+            return function
 
         dict_result = python.PyDict_GetItem(self._extensions, ns_uri_utf)
         if dict_result is not NULL:
@@ -145,7 +149,8 @@
         else:
             function = _find_extension(ns_uri_utf, name_utf)
 
-        python.PyDict_SetItem(self._temp_functions, key, function)
+        python.PyDict_SetItem(self._function_cache, key, function)
+        self._called_function = function
         return function
 
     # Python reference keeping during XPath function evaluation
@@ -183,8 +188,9 @@
     else:
         ns_uri = c_ns_uri
     context = <_BaseContext>ctxt
-    if context._lookup_extension(ns_uri, c_name) is not None:
-        return _xpath_function_call
+    function = context._prepare_function_call(ns_uri, c_name)
+    if function is not None:
+        return _call_prepared_function
     else:
         return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
 
@@ -194,15 +200,11 @@
 
 cdef class _XSLTResolverContext(_ResolverContext):
     cdef xmlDoc* _c_style_doc
-    cdef object _style_url_utf
-    cdef object _style_doc_utf
     cdef BaseParser _parser
     def __init__(self, BaseParser parser not None):
         _ResolverContext.__init__(self, parser.resolvers)
         self._parser = parser
         self._c_style_doc = NULL
-        self._style_url_utf = None
-        self._style_doc_utf = None
 
 cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
                          void* c_ctxt, xslt.xsltLoadType c_type):
@@ -308,7 +310,6 @@
         self._set_xpath_context(xsltCtxt.xpathCtxt)
         self._register_context(doc, 0)
         xsltCtxt.xpathCtxt.userData = <void*>self
-        self._registerLocalExtensionFunctions()
 
     cdef free_context(self):
         cdef xslt.xsltTransformContext* xsltCtxt
@@ -318,6 +319,7 @@
         self._free_context()
         self._xsltCtxt = NULL
         xslt.xsltFreeTransformContext(xsltCtxt)
+        self._release_temp_refs()
 
     cdef _registerLocalExtensionFunction(self, ns_utf, name_utf, function):
         extensions = self._extensions
@@ -329,20 +331,6 @@
             else:
                 ns_extensions = extensions[ns_utf] = {}
             python.PyDict_SetItem(ns_extensions, name_utf, function)
-        xslt.xsltRegisterExtFunction(
-            self._xsltCtxt, _cstr(name_utf), _cstr(ns_utf),
-            _xpath_function_call)
-
-    cdef _registerLocalExtensionFunctions(self):
-        cdef xslt.xsltTransformContext* xsltCtxt
-        if self._extensions is None:
-            return
-        xsltCtxt = self._xsltCtxt
-        for ns_uri_utf, extension in self._extensions.items():
-            for name_utf, function in extension.items():
-                xslt.xsltRegisterExtFunction(
-                    xsltCtxt, _cstr(name_utf), _cstr(ns_uri_utf),
-                    _xpath_function_call)
 
 cdef class _ExsltRegExp # forward declaration
 
@@ -453,7 +441,6 @@
         else:
             params = NULL
 
-        self._context._release_temp_refs()
         self._context.register_context(transform_ctxt, input_doc)
         if self._regexp is not None:
             self._regexp._register_in_context(self._context)
@@ -510,13 +497,6 @@
     result._xslt = xslt
     return result
 
-# used by FunctionNamespace():
-cdef void _register_global_xslt_function(char* ns_uri, char* name):
-    xslt.xsltRegisterExtModuleFunction(ns_uri, name, _xpath_function_call)
-
-cdef void _unregister_global_xslt_function(char* ns_uri, char* name):
-    xslt.xsltUnRegisterExtModuleFunction(ns_uri, name)
-
 # do not register all libxslt extra function, provide only "node-set"
 # functions like "output" and "write" are a potential security risk
 #xslt.xsltRegisterAllExtras()
@@ -774,7 +754,6 @@
     """Create an XPath evaluator for an element.
     """
     cdef _Element _element
-
     def __init__(self, _Element element not None, namespaces=None, extensions=None):
         XPathDocumentEvaluator.__init__(
             self, element._doc, namespaces, extensions)
@@ -957,28 +936,32 @@
             raise NotImplementedError
     return result
 
+
 cdef void _xpath_function_call(xpath.xmlXPathParserContext* ctxt, int nargs):
     cdef xpath.xmlXPathContext* rctxt
-    cdef _Document doc
-    cdef xpath.xmlXPathObject* obj
     cdef _BaseContext context
-    cdef int i
-
     rctxt = ctxt.context
-
-    # get information on what function is called
+    context = <_BaseContext>(rctxt.userData)
     name = rctxt.function
     if rctxt.functionURI is not NULL:
         uri = rctxt.functionURI
     else:
         uri = None
+    context._prepare_function_call(uri, name)
+    _extension_function_call(context, ctxt, nargs)
 
-    # get our evaluator
+cdef void _call_prepared_function(xpath.xmlXPathParserContext* ctxt, int nargs):
+    cdef xpath.xmlXPathContext* rctxt
+    cdef _BaseContext context
+    rctxt = ctxt.context
     context = <_BaseContext>(rctxt.userData)
+    _extension_function_call(context, ctxt, nargs)
 
-    # lookup up the extension function in the context
-    f = context._lookup_extension(uri, name)
-
+cdef void _extension_function_call(_BaseContext context,
+                                   xpath.xmlXPathParserContext* ctxt, int nargs):
+    cdef _Document doc
+    cdef xpath.xmlXPathObject* obj
+    cdef int i
     args = []
     doc = context._doc
     for i from 0 <= i < nargs:
@@ -987,16 +970,16 @@
 
     try:
         # call the function
-        res = f(None, *args)
+        res = context._called_function(None, *args)
         # hold python objects temporarily so that they won't get deallocated
         # during processing
         context._hold(res)
         # now wrap for XPath consumption
         obj = _wrapXPathObject(res)
+        xpath.valuePush(ctxt, obj)
     except:
         xpath.xmlXPathErr(
             ctxt,
             xmlerror.XML_XPATH_EXPR_ERROR - xmlerror.XML_XPATH_EXPRESSION_OK)
         context._exc._store_raised()
         return
-    xpath.valuePush(ctxt, obj)
From scoder at codespeak.net  Fri Apr 28 08:54:10 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 08:54:12 2006
Subject: [Lxml-checkins] r26494 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20060428065410.7D3A0100B0@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 08:54:07 2006
New Revision: 26494

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/tests/test_xpathevaluator.py
   lxml/trunk/src/lxml/xslt.pxi
Log:
refactoring of XPath(Document|Element)Evaluator: fix document evaluator by using the _ElementTree context node

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Fri Apr 28 08:54:07 2006
@@ -34,7 +34,10 @@
 Bugs fixed
 ----------
 
-* document('') now works in all XSLT documents
+* ElementTree.xpath() and XPathDocumentEvaluator were not using the
+  ElementTree root node as reference point
+
+* Calling document('') in XSLT failed to return the stylesheet
 
 * Crash in XMLSchema and RelaxNG when passing non-schema documents
 

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Fri Apr 28 08:54:07 2006
@@ -320,7 +320,8 @@
         against the same document, it is more efficient to use
         XPathEvaluator directly.
         """
-        return XPathDocumentEvaluator(self._doc, namespaces).evaluate(_path, **_variables)
+        evaluator = XPathElementEvaluator(self._context_node, namespaces)
+        return evaluator.evaluate(_path, **_variables)
 
     def xslt(self, _xslt, extensions=None, **_kw):
         """Transform this document using other document.
@@ -827,7 +828,8 @@
         return _elementpath.findall(self, path)
 
     def xpath(self, _path, namespaces=None, **_variables):
-        return XPathElementEvaluator(self, namespaces).evaluate(_path, **_variables)
+        evaluator = XPathElementEvaluator(self, namespaces)
+        return evaluator.evaluate(_path, **_variables)
 
 cdef _Element _elementFactory(_Document doc, xmlNode* c_node):
     cdef _Element result

Modified: lxml/trunk/src/lxml/tests/test_xpathevaluator.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xpathevaluator.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xpathevaluator.py	Fri Apr 28 08:54:07 2006
@@ -105,6 +105,18 @@
             [root],
             e.evaluate('//a'))
 
+    def test_xpath_evaluator_tree(self):
+        tree = self.parse('<a><b><c></c></b></a>')
+        child_tree = etree.ElementTree(tree.getroot()[0])
+        e = etree.XPathEvaluator(child_tree)
+        self.assertEquals(
+            [],
+            e.evaluate('.//a'))
+        root = child_tree.getroot()
+        self.assertEquals(
+            [root[0]],
+            e.evaluate('.//c'))
+
     def test_xpath_evaluator_element(self):
         tree = self.parse('<a><b><c></c></b></a>')
         root = tree.getroot()

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 08:54:07 2006
@@ -677,32 +677,23 @@
         return result
 
 
-cdef class XPathDocumentEvaluator(XPathEvaluatorBase):
-    """Create an XPath evaluator for a document.
+cdef class XPathElementEvaluator(XPathEvaluatorBase):
+    """Create an XPath evaluator for an element.
+
+    XPath evaluators must not be shared between threads.
     """
     cdef xpath.xmlXPathContext* _c_ctxt
-    cdef _Document _doc
-    
-    def __init__(self, etree, namespaces=None, extensions=None):
+    cdef _Element _element
+    def __init__(self, _NodeBase element not None, namespaces=None, extensions=None):
         cdef xpath.xmlXPathContext* xpathCtxt
         cdef int ns_register_status
         cdef _Document doc
-
-        if isinstance(etree, _Document):
-            doc = <_Document>etree # for internal use only!
-        elif isinstance(etree, _ElementTree):
-            doc = (<_ElementTree>etree)._doc
-        else:
-            raise TypeError, "XPathDocumentEvaluator can only work on ElementTree objects"
-        
+        doc = element._doc
         xpathCtxt = xpath.xmlXPathNewContext(doc._c_doc)
         if xpathCtxt is NULL:
-            # XXX what triggers this exception?
             raise XPathContextError, "Unable to create new XPath context"
-
-        self._doc = doc
+        self._element = element
         self._c_ctxt = xpathCtxt
-        
         XPathEvaluatorBase.__init__(self, namespaces, extensions)
 
     def __dealloc__(self):
@@ -717,55 +708,51 @@
     def registerNamespaces(self, namespaces):
         """Register a prefix -> uri dict.
         """
+        add = self._context.addNamespace
         for prefix, uri in namespaces.items():
-            self.registerNamespace(prefix, uri)
+            add(prefix, uri)
     
     def evaluate(self, _path, **_variables):
-        """Evaluate an XPath expression on the document.  Variables
-        may be given as keyword arguments. Note that namespaces are
-        currently not supported for variables."""
-        return self._evaluate(_path, NULL, _variables)
-
-    cdef object _evaluate(self, path, xmlNode* c_ctxt_node, variable_dict):
+        """Evaluate an XPath expression on the document.  Variables may be
+        provided as keyword arguments. Note that namespaces are currently not
+        supported for variables."""
         cdef xpath.xmlXPathContext* xpathCtxt
         cdef xpath.xmlXPathObject*  xpathObj
         cdef xmlNode* c_node
-        
+        cdef _Document doc
         xpathCtxt = self._c_ctxt
-        # if element context is requested; unfortunately need to modify ctxt
-        xpathCtxt.node = c_ctxt_node
+        xpathCtxt.node = self._element._c_node
+        doc = self._element._doc
 
         self._context._release_temp_refs()
-        self._context.register_context(xpathCtxt, self._doc)
-        self._context.registerVariables(variable_dict)
+        self._context.register_context(xpathCtxt, doc)
+        self._context.registerVariables(_variables)
 
-        path = _utf8(path)
+        path = _utf8(_path)
         xpathObj = xpath.xmlXPathEvalExpression(_cstr(path), xpathCtxt)
         self._context.unregister_context()
 
-        return self._handle_result(xpathObj, self._doc)
+        return self._handle_result(xpathObj, doc)
 
     #def clone(self):
     #    # XXX pretty expensive so calling this from callback is probably
     #    # not desirable
     #    return XPathEvaluator(self._doc, self._namespaces, self._extensions)
 
-cdef class XPathElementEvaluator(XPathDocumentEvaluator):
-    """Create an XPath evaluator for an element.
-    """
-    cdef _Element _element
-    def __init__(self, _Element element not None, namespaces=None, extensions=None):
-        XPathDocumentEvaluator.__init__(
-            self, element._doc, namespaces, extensions)
-        self._element = element
+cdef class XPathDocumentEvaluator(XPathElementEvaluator):
+    """Create an XPath evaluator for an ElementTree.
 
-    def evaluate(self, _path, **_variables):
-        """Evaluate an XPath expression on the element.  Variables may
-        be given as keyword arguments. Note that namespaces are
-        currently not supported for variables."""
-        return self._evaluate(_path, self._element._c_node, _variables)
+    XPath evaluators must not be shared between threads.
+    """
+    def __init__(self, _ElementTree etree not None, namespaces=None, extensions=None):
+        XPathElementEvaluator.__init__(
+            self, etree._context_node, namespaces, extensions)
 
 def XPathEvaluator(etree_or_element, namespaces=None, extensions=None):
+    """Creates and XPath evaluator for an ElementTree or an Element.
+
+    XPath evaluators must not be shared between threads.
+    """
     if isinstance(etree_or_element, _ElementTree):
         return XPathDocumentEvaluator(etree_or_element, namespaces, extensions)
     else:
From scoder at codespeak.net  Fri Apr 28 08:55:49 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 08:55:50 2006
Subject: [Lxml-checkins] r26495 - lxml/trunk/src/lxml/tests
Message-ID: <20060428065549.6BC90100B0@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 08:55:48 2006
New Revision: 26495

Modified:
   lxml/trunk/src/lxml/tests/test_xpathevaluator.py
Log:
better test case for last bug

Modified: lxml/trunk/src/lxml/tests/test_xpathevaluator.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xpathevaluator.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xpathevaluator.py	Fri Apr 28 08:55:48 2006
@@ -111,11 +111,11 @@
         e = etree.XPathEvaluator(child_tree)
         self.assertEquals(
             [],
-            e.evaluate('.//a'))
+            e.evaluate('a'))
         root = child_tree.getroot()
         self.assertEquals(
             [root[0]],
-            e.evaluate('.//c'))
+            e.evaluate('c'))
 
     def test_xpath_evaluator_element(self):
         tree = self.parse('<a><b><c></c></b></a>')
From scoder at codespeak.net  Fri Apr 28 10:33:18 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 10:33:20 2006
Subject: [Lxml-checkins] r26498 - lxml/trunk/doc
Message-ID: <20060428083318.ADD77100A4@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 10:33:17 2006
New Revision: 26498

Modified:
   lxml/trunk/doc/valgrind.txt
Log:
changed cmd line in doc/valgrind.txt to let valgrind do leak-check

Modified: lxml/trunk/doc/valgrind.txt
==============================================================================
--- lxml/trunk/doc/valgrind.txt	(original)
+++ lxml/trunk/doc/valgrind.txt	Fri Apr 28 10:33:17 2006
@@ -1,3 +1,3 @@
 The command used to run the tests with valgrind:
 
-valgrind --tool=memcheck --suppressions=valgrind-python.supp python2.3 test.py
+valgrind --tool=memcheck --leak-check=full --suppressions=valgrind-python.supp python2.3 test.py
From scoder at codespeak.net  Fri Apr 28 10:35:54 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 10:35:56 2006
Subject: [Lxml-checkins] r26500 - lxml/trunk/src/lxml/tests
Message-ID: <20060428083554.D8B3F100A4@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 10:35:53 2006
New Revision: 26500

Modified:
   lxml/trunk/src/lxml/tests/test_xpathevaluator.py
Log:
cosmetics in test case

Modified: lxml/trunk/src/lxml/tests/test_xpathevaluator.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xpathevaluator.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xpathevaluator.py	Fri Apr 28 10:35:53 2006
@@ -139,7 +139,7 @@
             return "hello %s and %s" % (a, b)
         extension = {(None, 'foo'): foo}
         tree = self.parse('<a><b></b></a>')
-        e = etree.XPathEvaluator(tree, None, [extension])
+        e = etree.XPathEvaluator(tree, extensions=[extension])
         self.assertRaises(TypeError, e.evaluate, "foo('you')")
 
     def test_xpath_extensions_error(self):
From scoder at codespeak.net  Fri Apr 28 11:36:47 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 11:36:48 2006
Subject: [Lxml-checkins] r26508 - lxml/trunk/src/lxml
Message-ID: <20060428093647.35A75100AD@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 11:36:46 2006
New Revision: 26508

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
rewrote buggy handling of nodes generated in Python extensions
* now cleans up node-sets the way they were created (instead of a straight 'xmlXPathFreeObject')
* temporary Python refs are cleared after XSLT resp. XPath execution, /before/ returning from user API call
* result node-sets are freed when raising exceptions

Elements from result node sets are now deep-copied if they came from other
documents than the input document in XSLT or the context document in XPath.
This prevents possible multiple-free crashes that could occur before. It
fixes the case where elements were created by extension functions, passed
through the XPath machinery and then garbage collected twice.



Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 11:36:46 2006
@@ -620,7 +620,7 @@
         if xpathCtxt is NULL:
             return
         self._unregisterVariables()
-        self._registered_variables  = []
+        del self._registered_variables[:]
         self._unregister_context()
 
     cdef void _unregisterVariables(self):
@@ -633,7 +633,7 @@
             xpathVarValue = xpath.xmlXPathVariableLookup(xpathCtxt, c_name)
             if xpathVarValue is not NULL:
                 xpath.xmlXPathRegisterVariable(xpathCtxt, c_name, NULL)
-                xpath.xmlXPathFreeObject(xpathVarValue)
+                _freeXPathObject(xpathVarValue)
 
     def registerVariables(self, variable_dict):
         for name, value in variable_dict.items():
@@ -660,20 +660,24 @@
     cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
         if self._context._exc._has_raised():
             if xpathObj is not NULL:
-                xpath.xmlXPathFreeObject(xpathObj)
+                _freeXPathObject(xpathObj)
                 xpathObj = NULL
+            self._context._release_temp_refs()
             self._context._exc._raise_if_stored()
 
         if xpathObj is NULL:
+            self._context._release_temp_refs()
             raise XPathSyntaxError, "Error in xpath expression."
 
         try:
             result = _unwrapXPathObject(xpathObj, doc)
         except XPathResultError:
-            xpath.xmlXPathFreeObject(xpathObj)
+            _freeXPathObject(xpathObj)
+            self._context._release_temp_refs()
             raise
 
-        xpath.xmlXPathFreeObject(xpathObj)
+        _freeXPathObject(xpathObj)
+        self._context._release_temp_refs()
         return result
 
 
@@ -724,7 +728,6 @@
         xpathCtxt.node = self._element._c_node
         doc = self._element._doc
 
-        self._context._release_temp_refs()
         self._context.register_context(xpathCtxt, doc)
         self._context.registerVariables(_variables)
 
@@ -851,7 +854,6 @@
     if python.PyUnicode_Check(obj):
         obj = _utf8(obj)
     if python.PyString_Check(obj):
-        # XXX use the Wrap variant? Or leak...
         return xpath.xmlXPathNewCString(_cstr(obj))
     if python.PyBool_Check(obj):
         return xpath.xmlXPathNewBoolean(obj)
@@ -866,6 +868,7 @@
                 node = <_NodeBase>element
                 xpath.xmlXPathNodeSetAdd(resultSet, node._c_node)
             else:
+                xpath.xmlXPathFreeNodeSet(resultSet)
                 raise XPathResultError, "This is not a node: %s" % element
         return xpath.xmlXPathWrapNodeSet(resultSet)
     else:
@@ -877,7 +880,7 @@
     if xpathObj.type == xpath.XPATH_UNDEFINED:
         raise XPathResultError, "Undefined xpath result"
     elif xpathObj.type == xpath.XPATH_NODESET:
-        return _createNodeSetResult(doc, xpathObj)
+        return _createNodeSetResult(xpathObj, doc)
     elif xpathObj.type == xpath.XPATH_BOOLEAN:
         return bool(xpathObj.boolval)
     elif xpathObj.type == xpath.XPATH_NUMBER:
@@ -897,8 +900,7 @@
     else:
         raise XPathResultError, "Unknown xpath result %s" % str(xpathObj.type)
 
-cdef object _createNodeSetResult(_Document doc,
-                                 xpath.xmlXPathObject* xpathObj):
+cdef object _createNodeSetResult(xpath.xmlXPathObject* xpathObj, _Document doc):
     cdef xmlNode* c_node
     cdef char* s
     cdef _NodeBase element
@@ -909,6 +911,12 @@
     for i from 0 <= i < xpathObj.nodesetval.nodeNr:
         c_node = xpathObj.nodesetval.nodeTab[i]
         if _isElement(c_node):
+            if c_node.doc != doc._c_doc:
+                # XXX: works, but maybe not always the right thing to do?
+                # XPath: only runs when extensions create or copy trees
+                #        -> we store Python refs to these, so that is OK
+                # XSLT: can it leak when merging trees from multiple sources?
+                c_node = tree.xmlDocCopyNode(c_node, doc._c_doc, 1)
             element = _elementFactory(doc, c_node)
             result.append(element)
         elif c_node.type == tree.XML_TEXT_NODE:
@@ -923,6 +931,14 @@
             raise NotImplementedError
     return result
 
+cdef void _freeXPathObject(xpath.xmlXPathObject* xpathObj):
+    """Free the XPath object, but *never* free the *content* of node sets.
+    Python dealloc will do that for us.
+    """
+    if xpathObj.nodesetval is not NULL:
+        xpath.xmlXPathFreeNodeSet(xpathObj.nodesetval)
+        xpathObj.nodesetval = NULL
+    xpath.xmlXPathFreeObject(xpathObj)
 
 cdef void _xpath_function_call(xpath.xmlXPathParserContext* ctxt, int nargs):
     cdef xpath.xmlXPathContext* rctxt
@@ -946,6 +962,7 @@
 
 cdef void _extension_function_call(_BaseContext context,
                                    xpath.xmlXPathParserContext* ctxt, int nargs):
+    cdef _NodeBase node
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
     cdef int i
@@ -956,17 +973,14 @@
     args.reverse()
 
     try:
-        # call the function
         res = context._called_function(None, *args)
-        # hold python objects temporarily so that they won't get deallocated
-        # during processing
-        context._hold(res)
-        # now wrap for XPath consumption
+        # wrap result for XPath consumption
         obj = _wrapXPathObject(res)
+        # prevent Python from deallocating elements handed to libxml2
+        context._hold(res)
         xpath.valuePush(ctxt, obj)
     except:
         xpath.xmlXPathErr(
             ctxt,
             xmlerror.XML_XPATH_EXPR_ERROR - xmlerror.XML_XPATH_EXPRESSION_OK)
         context._exc._store_raised()
-        return
From scoder at codespeak.net  Fri Apr 28 11:42:02 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 11:42:03 2006
Subject: [Lxml-checkins] r26509 - lxml/trunk/src/lxml
Message-ID: <20060428094202.A3790100AD@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 11:42:01 2006
New Revision: 26509

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
allow None to be returned from extension function: convert to empty node set

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 11:42:01 2006
@@ -859,7 +859,9 @@
         return xpath.xmlXPathNewBoolean(obj)
     if python.PyNumber_Check(obj):
         return xpath.xmlXPathNewFloat(obj)
-    if isinstance(obj, _NodeBase):
+    if obj is None:
+        obj = ()
+    elif isinstance(obj, _NodeBase):
         obj = (obj,)
     if python.PySequence_Check(obj):
         resultSet = xpath.xmlXPathNodeSetCreate(NULL)
From scoder at codespeak.net  Fri Apr 28 11:45:05 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 11:45:05 2006
Subject: [Lxml-checkins] r26510 - lxml/trunk/src/lxml/tests
Message-ID: <20060428094505.3D97F100AD@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 11:45:04 2006
New Revision: 26510

Modified:
   lxml/trunk/src/lxml/tests/test_xpathevaluator.py
Log:
lengthy test case for mixing XPath extensions, variables and exceptions

Modified: lxml/trunk/src/lxml/tests/test_xpathevaluator.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xpathevaluator.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xpathevaluator.py	Fri Apr 28 11:45:04 2006
@@ -182,6 +182,68 @@
         self.assertEquals(1, len(r))
         self.assertEquals("true", r[0].get('attr'))
 
+    def test_xpath_extensions_mix(self):
+        x = self.parse('<a attr="true"><test/></a>')
+
+        class LocalException(Exception):
+            pass
+
+        def foo(evaluator, a, varval):
+            etree.Element("DUMMY")
+            if varval == 0:
+                raise LocalException
+            elif varval == 1:
+                return ()
+            elif varval == 2:
+                return None
+            elif varval == 3:
+                return a[0][0]
+            a = a[0]
+            if a.get("attr") == str(varval):
+                return a
+            else:
+                return etree.Element("NODE")
+
+        extension = {(None, 'foo'): foo}
+        e = etree.XPathEvaluator(x, extensions=[extension])
+        del x
+
+        self.assertRaises(LocalException, e.evaluate, "foo(., 0)")
+        self.assertRaises(LocalException, e.evaluate, "foo(., $value)", value=0)
+
+        r = e.evaluate("foo(., $value)", value=1)
+        self.assertEqual(len(r), 0)
+
+        r = e.evaluate("foo(.,  1)")
+        self.assertEqual(len(r), 0)
+
+        r = e.evaluate("foo(., $value)", value=2)
+        self.assertEqual(len(r), 0)
+
+        r = e.evaluate("foo(., $value)", value=3)
+        self.assertEqual(len(r), 1)
+        self.assertEqual(r[0].tag, "test")
+
+        r = e.evaluate("foo(., $value)", value="false")
+        self.assertEqual(len(r), 1)
+        self.assertEqual(r[0].tag, "NODE")
+
+        r = e.evaluate("foo(., 'false')")
+        self.assertEqual(len(r), 1)
+        self.assertEqual(r[0].tag, "NODE")
+
+        r = e.evaluate("foo(., 'true')")
+        self.assertEqual(len(r), 1)
+        self.assertEqual(r[0].tag, "a")
+        self.assertEqual(r[0][0].tag, "test")
+
+        r = e.evaluate("foo(., $value)", value="true")
+        self.assertEqual(len(r), 1)
+        self.assertEqual(r[0].tag, "a")
+
+        self.assertRaises(LocalException, e.evaluate, "foo(., 0)")
+        self.assertRaises(LocalException, e.evaluate, "foo(., $value)", value=0)
+
 
 class ETreeXPathClassTestCase(HelperTestCase):
     "Tests for the XPath class"
From scoder at codespeak.net  Fri Apr 28 11:50:23 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 11:50:24 2006
Subject: [Lxml-checkins] r26513 - lxml/trunk/src/lxml
Message-ID: <20060428095023.9F1F010079@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 11:50:22 2006
New Revision: 26513

Modified:
   lxml/trunk/src/lxml/python.pxd
   lxml/trunk/src/lxml/xslt.pxi
Log:
small C-ification

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Fri Apr 28 11:50:22 2006
@@ -20,6 +20,7 @@
 
     cdef int PyList_GET_SIZE(object l)
     cdef int PyList_Append(object l, object obj)
+    cdef int PyList_Reverse(object l)
     cdef int PyDict_SetItemString(object d, char* key, object value)
     cdef int PyDict_SetItem(object d, object key, object value)
     cdef PyObject* PyDict_GetItemString(object d, char* key)

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 11:50:22 2006
@@ -972,7 +972,7 @@
     doc = context._doc
     for i from 0 <= i < nargs:
         python.PyList_Append(args, _unwrapXPathObject(xpath.valuePop(ctxt), doc))
-    args.reverse()
+    python.PyList_Reverse(args)
 
     try:
         res = context._called_function(None, *args)
From scoder at codespeak.net  Fri Apr 28 11:57:28 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 11:57:30 2006
Subject: [Lxml-checkins] r26514 - lxml/trunk/src/lxml
Message-ID: <20060428095728.CDE5F10079@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 11:57:27 2006
New Revision: 26514

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
call XSLT function lookup only from within XSLT, not from XPath evaluation

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 11:57:27 2006
@@ -191,8 +191,10 @@
     function = context._prepare_function_call(ns_uri, c_name)
     if function is not None:
         return _call_prepared_function
-    else:
+    elif isinstance(context, _XSLTContext):
         return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
+    else:
+        return NULL
 
  
 ################################################################################
From scoder at codespeak.net  Fri Apr 28 12:47:32 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 12:47:33 2006
Subject: [Lxml-checkins] r26518 - lxml/trunk/src/lxml
Message-ID: <20060428104732.951BD1009F@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 12:47:31 2006
New Revision: 26518

Modified:
   lxml/trunk/src/lxml/xpath.pxd
   lxml/trunk/src/lxml/xslt.pxi
Log:
cleanup in extension functions

Modified: lxml/trunk/src/lxml/xpath.pxd
==============================================================================
--- lxml/trunk/src/lxml/xpath.pxd	(original)
+++ lxml/trunk/src/lxml/xpath.pxd	Fri Apr 28 12:47:31 2006
@@ -14,6 +14,31 @@
         XPATH_USERS = 8
         XPATH_XSLT_TREE = 9
 
+    ctypedef enum xmlXPathError:
+        XPATH_EXPRESSION_OK = 0
+        XPATH_NUMBER_ERROR = 1
+        XPATH_UNFINISHED_LITERAL_ERROR = 2
+        XPATH_START_LITERAL_ERROR = 3
+        XPATH_VARIABLE_REF_ERROR = 4
+        XPATH_UNDEF_VARIABLE_ERROR = 5
+        XPATH_INVALID_PREDICATE_ERROR = 6
+        XPATH_EXPR_ERROR = 7
+        XPATH_UNCLOSED_ERROR = 8
+        XPATH_UNKNOWN_FUNC_ERROR = 9
+        XPATH_INVALID_OPERAND = 10
+        XPATH_INVALID_TYPE = 11
+        XPATH_INVALID_ARITY = 12
+        XPATH_INVALID_CTXT_SIZE = 13
+        XPATH_INVALID_CTXT_POSITION = 14
+        XPATH_MEMORY_ERROR = 15
+        XPTR_SYNTAX_ERROR = 16
+        XPTR_RESOURCE_ERROR = 17
+        XPTR_SUB_RESOURCE_ERROR = 18
+        XPATH_UNDEF_PREFIX_ERROR = 19
+        XPATH_ENCODING_ERROR = 20
+        XPATH_INVALID_CHAR_ERROR = 21
+        XPATH_INVALID_CTXT = 22
+
     ctypedef struct xmlNodeSet:
         int nodeNr
         int nodeMax

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 12:47:31 2006
@@ -21,6 +21,9 @@
 class XPathContextError(XPathError):
     pass
 
+class XPathFunctionError(XPathError):
+    pass
+
 class XPathResultError(XPathError):
     pass
 
@@ -32,6 +35,7 @@
 
 cdef class _BaseContext:
     cdef xpath.xmlXPathContext* _xpathCtxt
+    cdef xpath.xmlXPathFuncLookupFunc _ext_lookup_function
     cdef _Document _doc
     cdef object _extensions
     cdef object _namespaces
@@ -92,8 +96,8 @@
         namespaces = self._namespaces
         if namespaces is not None:
             self.registerNamespaces(namespaces)
-        xpath.xmlXPathRegisterFuncLookup(self._xpathCtxt, _function_check,
-                                         <python.PyObject*>self)
+        xpath.xmlXPathRegisterFuncLookup(
+            self._xpathCtxt, self._ext_lookup_function, <python.PyObject*>self)
 
     cdef _unregister_context(self):
         self._unregisterNamespaces()
@@ -132,14 +136,14 @@
     
     # extension functions
 
-    cdef _prepare_function_call(self, ns_uri_utf, name_utf):
+    cdef int _prepare_function_call(self, ns_uri_utf, name_utf):
         cdef python.PyObject* dict_result
         key = (ns_uri_utf, name_utf)
         dict_result = python.PyDict_GetItem(self._function_cache, key)
         if dict_result is not NULL:
             function = <object>dict_result
             self._called_function = function
-            return function
+            return function is not None
 
         dict_result = python.PyDict_GetItem(self._extensions, ns_uri_utf)
         if dict_result is not NULL:
@@ -151,7 +155,7 @@
 
         python.PyDict_SetItem(self._function_cache, key, function)
         self._called_function = function
-        return function
+        return function is not None
 
     # Python reference keeping during XPath function evaluation
 
@@ -179,7 +183,8 @@
                 #print "Holding document:", <int>element._doc._c_doc
                 self._temp_refs.add(element._doc)
 
-cdef xpath.xmlXPathFunction _function_check(void* ctxt, char* c_name, char* c_ns_uri):
+cdef xpath.xmlXPathFunction _function_check(void* ctxt,
+                                            char* c_name, char* c_ns_uri):
     cdef _BaseContext context
     if c_name is NULL:
         return NULL
@@ -188,15 +193,21 @@
     else:
         ns_uri = c_ns_uri
     context = <_BaseContext>ctxt
-    function = context._prepare_function_call(ns_uri, c_name)
-    if function is not None:
+    if context._prepare_function_call(ns_uri, c_name):
         return _call_prepared_function
-    elif isinstance(context, _XSLTContext):
-        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
     else:
         return NULL
 
- 
+cdef xpath.xmlXPathFunction _xslt_function_check(void* ctxt,
+                                                 char* c_name, char* c_ns_uri):
+    cdef xpath.xmlXPathFunction result
+    result = _function_check(ctxt, c_name, c_ns_uri)
+    if result is NULL:
+        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
+    else:
+        return result
+
+
 ################################################################################
 # XSLT document loaders
 
@@ -302,8 +313,9 @@
     cdef xslt.xsltTransformContext* _xsltCtxt
     def __init__(self, namespaces, extensions):
         self._xsltCtxt = NULL
+        self._ext_lookup_function = _xslt_function_check
         if extensions and None in extensions:
-            raise XSLTExtensionError, "extensions must have non-empty namespaces"
+            raise XSLTExtensionError, "extensions must not have empty namespaces"
         _BaseContext.__init__(self, namespaces, extensions)
 
     cdef register_context(self, xslt.xsltTransformContext* xsltCtxt,
@@ -603,9 +615,10 @@
     cdef object _variables
     cdef object _registered_variables
     def __init__(self, namespaces, extensions, variables):
-        _BaseContext.__init__(self, namespaces, extensions)
+        self._ext_lookup_function = _function_check
         self._variables = variables
         self._registered_variables  = []
+        _BaseContext.__init__(self, namespaces, extensions)
         
     cdef register_context(self, xpath.xmlXPathContext* xpathCtxt, _Document doc):
         self._set_xpath_context(xpathCtxt)
@@ -781,7 +794,7 @@
         path = _utf8(path)
         self._xpath = xpath.xmlXPathCompile(_cstr(path))
         if self._xpath is NULL:
-            raise XPathSyntaxError, "Error in xpath expression."
+            raise XPathSyntaxError, "Error in XPath expression"
         self._xpathCtxt = xpath.xmlXPathNewContext(NULL)
 
     def __call__(self, _etree_or_element, **_variables):
@@ -954,8 +967,12 @@
         uri = rctxt.functionURI
     else:
         uri = None
-    context._prepare_function_call(uri, name)
-    _extension_function_call(context, ctxt, nargs)
+    if context._prepare_function_call(uri, name):
+        _extension_function_call(context, ctxt, nargs)
+    else:
+        xpath.xmlXPathErr(ctxt, xpath.XPATH_EXPR_ERROR)
+        exception = XPathFunctionError("XPath function {%s}%s not found" % (uri, name))
+        context._exc._store_exception(exception)
 
 cdef void _call_prepared_function(xpath.xmlXPathParserContext* ctxt, int nargs):
     cdef xpath.xmlXPathContext* rctxt
@@ -970,13 +987,14 @@
     cdef _Document doc
     cdef xpath.xmlXPathObject* obj
     cdef int i
-    args = []
     doc = context._doc
-    for i from 0 <= i < nargs:
-        python.PyList_Append(args, _unwrapXPathObject(xpath.valuePop(ctxt), doc))
-    python.PyList_Reverse(args)
-
     try:
+        args = []
+        for i from 0 <= i < nargs:
+            o = _unwrapXPathObject(xpath.valuePop(ctxt), doc)
+            python.PyList_Append(args, o)
+        python.PyList_Reverse(args)
+
         res = context._called_function(None, *args)
         # wrap result for XPath consumption
         obj = _wrapXPathObject(res)
@@ -984,7 +1002,5 @@
         context._hold(res)
         xpath.valuePush(ctxt, obj)
     except:
-        xpath.xmlXPathErr(
-            ctxt,
-            xmlerror.XML_XPATH_EXPR_ERROR - xmlerror.XML_XPATH_EXPRESSION_OK)
+        xpath.xmlXPathErr(ctxt, xpath.XPATH_EXPR_ERROR)
         context._exc._store_raised()
From scoder at codespeak.net  Fri Apr 28 12:51:47 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 12:51:50 2006
Subject: [Lxml-checkins] r26521 - lxml/trunk/src/lxml
Message-ID: <20060428105147.11D49100B5@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 12:51:43 2006
New Revision: 26521

Modified:
   lxml/trunk/src/lxml/nsclasses.pxi
Log:
remove unused XSLTElement class

Modified: lxml/trunk/src/lxml/nsclasses.pxi
==============================================================================
--- lxml/trunk/src/lxml/nsclasses.pxi	(original)
+++ lxml/trunk/src/lxml/nsclasses.pxi	Fri Apr 28 12:51:43 2006
@@ -10,10 +10,6 @@
     persistent state of elements must be stored in the underlying XML."""
     pass
 
-cdef class XSLTElement:
-    "NOT IMPLEMENTED YET!"
-    pass
-
 cdef object __NAMESPACE_REGISTRIES
 __NAMESPACE_REGISTRIES = {}
 
@@ -56,7 +52,6 @@
     cdef object _ns_uri_utf
     cdef object _classes
     cdef object _extensions
-    cdef object _xslt_elements
     cdef char* _c_ns_uri_utf
     def __init__(self, ns_uri):
         self._ns_uri = ns_uri
@@ -68,7 +63,6 @@
             self._c_ns_uri_utf = _cstr(self._ns_uri_utf)
         self._classes = {}
         self._extensions = {}
-        self._xslt_elements = {}
 
     def update(self, class_dict_iterable):
         """Forgivingly update the registry. If registered values are
@@ -89,8 +83,6 @@
             d = self._classes
         elif name is None:
             raise NamespaceRegistryError, "Registered name can only be None for elements."
-        elif python.PyType_Check(item) and issubclass(item, XSLTElement):
-            d = self._xslt_elements
         elif callable(item):
             d = self._extensions
         else:
@@ -115,7 +107,6 @@
     def clear(self):
         self._classes.clear()
         self._extensions.clear()
-        #self.self._xslt_elements.clear()
 
     def __repr__(self):
         return "Namespace(%r)" % self._ns_uri
From scoder at codespeak.net  Fri Apr 28 19:57:28 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 19:57:29 2006
Subject: [Lxml-checkins] r26532 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20060428175728.44BF4100A3@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 19:57:25 2006
New Revision: 26532

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/cstd.pxd
   lxml/trunk/src/lxml/etree.h
   lxml/trunk/src/lxml/tests/test_xslt.py
   lxml/trunk/src/lxml/tree.pxd
   lxml/trunk/src/lxml/xmlerror.pxi
   lxml/trunk/src/lxml/xslt.pxi
Log:
XSLT error reporting

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Fri Apr 28 19:57:25 2006
@@ -7,6 +7,8 @@
 Features added
 --------------
 
+* Error reporting now also works in XSLT
+
 * Support for custom document loaders (URI resolvers) in parsers and XSLT,
   resolvers are registered at parser level
 

Modified: lxml/trunk/src/lxml/cstd.pxd
==============================================================================
--- lxml/trunk/src/lxml/cstd.pxd	(original)
+++ lxml/trunk/src/lxml/cstd.pxd	Fri Apr 28 19:57:25 2006
@@ -3,3 +3,12 @@
     cdef void* malloc(int size)
     void free(void* ptr)
     
+cdef extern from "stdarg.h":
+    ctypedef void *va_list
+    void va_start(va_list ap, void *last)
+    void va_end(va_list ap)
+
+cdef extern from "etree.h":
+    cdef int va_int(va_list ap)
+    cdef char *va_charptr(va_list ap)
+    

Modified: lxml/trunk/src/lxml/etree.h
==============================================================================
--- lxml/trunk/src/lxml/etree.h	(original)
+++ lxml/trunk/src/lxml/etree.h	Fri Apr 28 19:57:25 2006
@@ -12,4 +12,8 @@
         ((c_node)->type == XML_ELEMENT_NODE || \
 	 (c_node)->type == XML_COMMENT_NODE)
 
+/* v_arg functions */
+#define va_int(ap)     va_arg(ap, int)
+#define va_charptr(ap) va_arg(ap, char *)
+
 #endif /*HAS_ETREE_H*/

Modified: lxml/trunk/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xslt.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xslt.py	Fri Apr 28 19:57:25 2006
@@ -75,7 +75,6 @@
         self.assertRaises(TypeError, etree.XSLT, None)
 
     def test_xslt_input_partial_doc(self):
-        tree = self.parse('<a><b>B</b><c>C</c></a>')
         style = self.parse('''\
 <otherroot>
 <xsl:stylesheet version="1.0"
@@ -272,6 +271,7 @@
         style = etree.XSLT(styledoc)
         result = style.apply(source)
         self.assertEqual('', style.tostring(result))
+        self.assertEqual('', str(result))
 
     def test_xslt_shortcut(self):
         tree = self.parse('<a><b>B</b><c>C</c></a>')
@@ -406,6 +406,18 @@
         self.assertEquals(root[0].tag,
                           '{http://www.w3.org/1999/XSL/Transform}stylesheet')
 
+    def test_xslt_document_error(self):
+        # make sure document('') works from parsed strings
+        xslt = etree.XSLT(etree.XML("""\
+<xsl:stylesheet version="1.0"
+   xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:template match="/">
+    <test>TEXT<xsl:copy-of select="document('uri:__junkfood__is__evil__')//test"/></test>
+  </xsl:template>
+</xsl:stylesheet>
+"""))
+        self.assertRaises(etree.XSLTApplyError, xslt, etree.XML('<a/>'))
+
     def test_exslt_regexp_test(self):
         xslt = etree.XSLT(etree.XML("""\
 <xsl:stylesheet version="1.0"

Modified: lxml/trunk/src/lxml/tree.pxd
==============================================================================
--- lxml/trunk/src/lxml/tree.pxd	(original)
+++ lxml/trunk/src/lxml/tree.pxd	Fri Apr 28 19:57:25 2006
@@ -4,6 +4,7 @@
     ctypedef struct FILE
     cdef int strlen(char* s)
     cdef int strcmp(char* s1, char* s2)
+    cdef int strncmp(char* s1, char* s2, int len)
     
 cdef extern from "libxml/encoding.h":
     ctypedef struct xmlCharEncodingHandler

Modified: lxml/trunk/src/lxml/xmlerror.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlerror.pxi	(original)
+++ lxml/trunk/src/lxml/xmlerror.pxi	Fri Apr 28 19:57:25 2006
@@ -35,6 +35,15 @@
         else:
             self.filename = python.PyString_FromString(error.file)
 
+    cdef _setGeneric(self, int domain, int type, int level, int line,
+                     message, filename):
+        self.domain  = domain
+        self.type    = type
+        self.level   = level
+        self.line    = line
+        self.message = message
+        self.filename = filename
+
     def __repr__(self):
         if self.filename:
             return "%s:%d:%s:%s:%s: %s" % (
@@ -146,10 +155,10 @@
 
     cdef void connect(self):
         del self._entries[:]
-        xmlerror.xmlSetStructuredErrorFunc(<void*>self, _localReceiveError)
+        xmlerror.xmlSetStructuredErrorFunc(<void*>self, _receiveError)
 
     cdef void disconnect(self):
-        xmlerror.xmlSetStructuredErrorFunc(NULL, _globalReceiveError)
+        xmlerror.xmlSetStructuredErrorFunc(NULL, _receiveError)
 
     cdef void _receive(self, xmlerror.xmlError* error):
         cdef _LogEntry entry
@@ -159,6 +168,15 @@
             __GLOBAL_ERROR_LOG.receive(entry)
         self.receive(entry)
 
+    cdef void _receiveGeneric(self, int domain, int type, int level, int line,
+                              message, filename):
+        cdef _LogEntry entry
+        entry = _LogEntry()
+        entry._setGeneric(domain, type, level, line, message, filename)
+        if __GLOBAL_ERROR_LOG is not self:
+            __GLOBAL_ERROR_LOG.receive(entry)
+        self.receive(entry)
+
     def receive(self, entry):
         python.PyList_Append(self._entries, entry)
 
@@ -198,7 +216,7 @@
             logger = logging.getLogger(name)
         else:
             logger = logging.getLogger()
-        self._log    = logger.log
+        self._log = logger.log
 
     def copy(self):
         return self
@@ -220,16 +238,56 @@
     return __GLOBAL_ERROR_LOG.copy()
 
 # local log function: forward error to logger object
-cdef void _localReceiveError(void* c_log_handler, xmlerror.xmlError* error):
+cdef void _receiveError(void* c_log_handler, xmlerror.xmlError* error):
     cdef _ErrorLog log_handler
     if __DEBUG != 0:
-        log_handler = <_ErrorLog>c_log_handler
+        if c_log_handler is not NULL:
+            log_handler = <_ErrorLog>c_log_handler
+        else:
+            log_handler = __GLOBAL_ERROR_LOG
         log_handler._receive(error)
 
-# global log functions: overridden by local functions
-cdef void _globalReceiveError(void* userData, xmlerror.xmlError* error):
-    if __DEBUG != 0:
-        __GLOBAL_ERROR_LOG._receive(error)
+cdef void _receiveGenericError(void* c_log_handler, char* msg, ...):
+    cdef cstd.va_list args
+    cdef _ErrorLog log_handler
+    cdef char* c_text
+    cdef char* c_filename
+    cdef char* c_element
+    cdef int c_line
+    if __DEBUG == 0 or msg == NULL or tree.strlen(msg) < 10:
+        return
+    if c_log_handler is not NULL:
+        log_handler = <_ErrorLog>c_log_handler
+    else:
+        log_handler = __GLOBAL_ERROR_LOG
+
+    cstd.va_start(args, msg)
+    c_text     = cstd.va_charptr(args)
+    c_filename = cstd.va_charptr(args)
+    c_line     = cstd.va_int(args)
+    c_element  = cstd.va_charptr(args)
+    cstd.va_end(args)
+
+    if c_text is NULL:
+        message = None
+    elif c_element is NULL:
+        message = funicode(c_text)
+    else:
+        message = "%s (element '%s')" % (
+            funicode(c_text), funicode(c_element))
+
+    if c_filename is not NULL and tree.strlen(c_filename) > 0:
+        if tree.strncmp(c_filename, 'XSLT:', 5) == 0:
+            filename = '<xslt>'
+        else:
+            filename = funicode(c_filename)
+    else:
+        filename = None
+
+    log_handler._receiveGeneric(xmlerror.XML_FROM_XSLT,
+                                xmlerror.XML_ERR_OK,
+                                xmlerror.XML_ERR_ERROR,
+                                c_line, message, filename)
 
 # dummy function: no debug output at all
 cdef void _nullGenericErrorFunc(void* ctxt, char* msg, ...):
@@ -238,11 +296,10 @@
 # setup for global log:
 cdef void _logLibxmlErrors():
     xmlerror.xmlSetGenericErrorFunc(NULL, _nullGenericErrorFunc)
-    xmlerror.xmlSetStructuredErrorFunc(NULL, _globalReceiveError)
+    xmlerror.xmlSetStructuredErrorFunc(NULL, _receiveError)
 
 cdef void _logLibxsltErrors():
-    xslt.xsltSetGenericErrorFunc(NULL, _nullGenericErrorFunc)
-    # xslt.xsltSetTransformErrorFunc
+    xslt.xsltSetGenericErrorFunc(NULL, _receiveGenericError)
 
 # init global logging
 initThreadLogging()

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 19:57:25 2006
@@ -355,6 +355,7 @@
     cdef xslt.xsltStylesheet* _c_style
     cdef _XSLTResolverContext _xslt_resolver_context
     cdef _ExsltRegExp _regexp
+    cdef _ErrorLog _error_log
 
     def __init__(self, xslt_input, extensions=None, regexp=True):
         cdef xslt.xsltStylesheet* c_style
@@ -394,12 +395,13 @@
         self._c_style = c_style
 
         self._context = _XSLTContext(None, extensions)
+        self._error_log = _ErrorLog()
         if regexp:
             self._regexp  = _ExsltRegExp()
         else:
             self._regexp  = None
         # XXX is it worthwile to use xsltPrecomputeStylesheet here?
-        
+
     def __dealloc__(self):
         if self._xslt_resolver_context is not None and \
                self._xslt_resolver_context._c_style_doc is not NULL:
@@ -407,6 +409,10 @@
         # this cleans up copy of doc as well
         xslt.xsltFreeStylesheet(self._c_style)
 
+    property error_log:
+        def __get__(self):
+            return self._error_log.copy()
+
     def __call__(self, _input, **_kw):
         cdef _Document input_doc
         cdef _NodeBase root_node
@@ -432,6 +438,10 @@
             _destroyFakeDoc(input_doc._c_doc, c_doc)
             raise XSLTApplyError, "Error preparing stylesheet run"
 
+        self._error_log.clear()
+        xslt.xsltSetTransformErrorFunc(transform_ctxt, <void*>self._error_log,
+                                       _receiveGenericError)
+
         ptemp = c_doc._private
         c_doc._private = <python.PyObject*>resolver_context
 
@@ -527,7 +537,6 @@
 # enable EXSLT support for XSLT
 xslt.exsltRegisterAll()
 
-
 ################################################################################
 # EXSLT regexp implementation
 
From scoder at codespeak.net  Fri Apr 28 21:11:26 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:11:28 2006
Subject: [Lxml-checkins] r26534 - lxml/trunk/src/lxml
Message-ID: <20060428191126.4C0DD1008E@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:11:25 2006
New Revision: 26534

Modified:
   lxml/trunk/src/lxml/xslt.pxi
Log:
connect XSLT error log during transformation to receive also libxml2 errors

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 21:11:25 2006
@@ -438,7 +438,7 @@
             _destroyFakeDoc(input_doc._c_doc, c_doc)
             raise XSLTApplyError, "Error preparing stylesheet run"
 
-        self._error_log.clear()
+        self._error_log.connect()
         xslt.xsltSetTransformErrorFunc(transform_ctxt, <void*>self._error_log,
                                        _receiveGenericError)
 
@@ -480,6 +480,7 @@
         c_doc._private = ptemp # restore _private before _destroyFakeDoc!
         _destroyFakeDoc(input_doc._c_doc, c_doc)
 
+        self._error_log.disconnect()
         if self._xslt_resolver_context._has_raised():
             if c_result is not NULL:
                 tree.xmlFreeDoc(c_result)
From scoder at codespeak.net  Fri Apr 28 21:11:40 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:11:42 2006
Subject: [Lxml-checkins] r26535 - in lxml/trunk: doc src/lxml src/lxml/tests
Message-ID: <20060428191140.95D9D1008E@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:11:37 2006
New Revision: 26535

Added:
   lxml/trunk/src/lxml/tests/test.xsd
Modified:
   lxml/trunk/doc/api.txt
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/relaxng.pxi
   lxml/trunk/src/lxml/tests/test_xmlschema.py
   lxml/trunk/src/lxml/xmlschema.pxi
Log:
merged APIs of RelaxNG and XMLSchema into new _Validator superclass, updated doctests

Modified: lxml/trunk/doc/api.txt
==============================================================================
--- lxml/trunk/doc/api.txt	(original)
+++ lxml/trunk/doc/api.txt	Fri Apr 28 21:11:37 2006
@@ -67,8 +67,6 @@
   >>> print entry.domain_name, entry.type_name, entry.filename
   PARSER ERR_TAG_NOT_FINISHED <string>
 
-XSLT error messages are not currently available through the lxml API.
-
 
 xpath method on ElementTree, Element
 ------------------------------------
@@ -221,9 +219,9 @@
   >>> relaxng_doc = lxml.etree.parse(f)
   >>> relaxng = lxml.etree.RelaxNG(relaxng_doc)
 
-You can then validate some ElementTree document with this. You'll get
-back true if the document is valid against the Relax NG schema, and
-false if not::
+You can then validate some ElementTree document against the schema. You'll get
+back True if the document is valid against the Relax NG schema, and False if
+not::
 
   >>> valid = StringIO('<a><b></b></a>')
   >>> doc = lxml.etree.parse(valid)
@@ -235,6 +233,23 @@
   >>> relaxng.validate(doc2)
   0
 
+Calling the schema object has the same effect as calling its validate
+method. This is sometimes used in conditional statements::
+
+  >>> invalid = StringIO('<a><c></c></a>')
+  >>> doc2 = lxml.etree.parse(invalid)
+  >>> if not relaxng(doc2):
+  ...     print "invalid!"
+  invalid!
+
+If you prefer getting an exception when validating, you can use the
+assertValid method::
+
+  >>> relaxng.assertValid(doc2)
+  Traceback (most recent call last):
+    [...]
+  DocumentInvalid: Document does not comply with schema
+
 Starting with version 0.9, lxml now has a simple API to report the errors
 generated by libxml2. If you want to find out why the validation failed in the
 second case, you can look up the error log of the validation process and check
@@ -250,7 +265,7 @@
 appeares during the validation.
 
 Similar to XSLT, there's also a less efficient but easier shortcut method to
-do RelaxNG validation::
+do one-shot RelaxNG validation::
 
   >>> doc.relaxng(relaxng_doc)
   1
@@ -293,6 +308,23 @@
   >>> xmlschema.validate(doc2)
   0
 
+Calling the schema object has the same effect as calling its validate
+method. This is sometimes used in conditional statements::
+
+  >>> invalid = StringIO('<a><c></c></a>')
+  >>> doc2 = lxml.etree.parse(invalid)
+  >>> if not xmlschema(doc2):
+  ...     print "invalid!"
+  invalid!
+
+If you prefer getting an exception when validating, you can use the
+assertValid method::
+
+  >>> xmlschema.assertValid(doc2)
+  Traceback (most recent call last):
+    [...]
+  DocumentInvalid: Document does not comply with schema
+
 Error reporting works like for the RelaxNG class::
 
   >>> log = xmlschema.error_log

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Fri Apr 28 21:11:37 2006
@@ -50,6 +50,9 @@
 class LxmlSyntaxError(LxmlError, SyntaxError):
     pass
 
+class DocumentInvalid(LxmlError):
+    pass
+
 class XIncludeError(LxmlError):
     pass
 
@@ -1349,18 +1352,51 @@
 
 
 # include submodules
+include "proxy.pxi"     # Proxy handling (element backpointers/memory/etc.)
 include "xmlerror.pxi"  # error and log handling
-include "xmlid.pxi"     # XMLID and IDDict
 include "nsclasses.pxi" # Namespace implementation and registry
 include "docloader.pxi" # Support for custom document loaders
 include "parser.pxi"    # XML Parser
+include "xmlid.pxi"     # XMLID and IDDict
 include "xslt.pxi"      # XPath and XSLT
+
+
+################################################################################
+# Validation
+
+cdef class _Validator:
+    "Base class for XML validators."
+    cdef _ErrorLog _error_log
+    def __init__(self):
+        self._error_log = _ErrorLog()
+        
+    def validate(self, etree):
+        """Validate the document using this schema.
+
+        Returns true if document is valid, false if not."""
+        return self(etree)
+
+    def assertValid(self, etree):
+        "Raises DocumentInvalid if the document does not comply with the schema."
+        if not self(etree):
+            raise DocumentInvalid, "Document does not comply with schema"
+
+    def assert_(self, etree):
+        "Raises AssertionError if the document does not comply with the schema."
+        if not self(etree):
+            raise AssertionError, "Document does not comply with schema"
+
+    property error_log:
+        def __get__(self):
+            return self._error_log.copy()
+
 include "relaxng.pxi"   # RelaxNG
 include "xmlschema.pxi" # XMLSchema
-include "proxy.pxi"     # Proxy handling (element backpointers/memory/etc.)
 
 
+################################################################################
 # Private helper functions
+
 cdef _Document _documentOrRaise(object input):
     cdef _Document doc
     doc = _documentOf(input)

Modified: lxml/trunk/src/lxml/relaxng.pxi
==============================================================================
--- lxml/trunk/src/lxml/relaxng.pxi	(original)
+++ lxml/trunk/src/lxml/relaxng.pxi	Fri Apr 28 21:11:37 2006
@@ -13,13 +13,11 @@
 ################################################################################
 # RelaxNG
 
-cdef class RelaxNG:
+cdef class RelaxNG(_Validator):
     """Turn a document into a Relax NG validator.
     Can also load from filesystem directly given file object or filename.
     """
     cdef relaxng.xmlRelaxNG* _c_schema
-    cdef _ErrorLog _error_log
-    
     def __init__(self, etree=None, file=None):
         cdef _Document doc
         cdef _NodeBase root_node
@@ -64,12 +62,12 @@
         relaxng.xmlRelaxNGFreeParserCtxt(parser_ctxt)
         if fake_c_doc is not NULL:
             _destroyFakeDoc(doc._c_doc, fake_c_doc)
-        self._error_log = _ErrorLog()
-        
+        _Validator.__init__(self)
+
     def __dealloc__(self):
         relaxng.xmlRelaxNGFree(self._c_schema)
-        
-    def validate(self, etree):
+
+    def __call__(self, etree):
         """Validate doc using Relax NG.
 
         Returns true if document is valid, false if not."""
@@ -94,7 +92,3 @@
         if ret == -1:
             raise RelaxNGValidateError, "Internal error in Relax NG validation"
         return ret == 0
-
-    property error_log:
-        def __get__(self):
-            return self._error_log.copy()

Added: lxml/trunk/src/lxml/tests/test.xsd
==============================================================================
--- (empty file)
+++ lxml/trunk/src/lxml/tests/test.xsd	Fri Apr 28 21:11:37 2006
@@ -0,0 +1,8 @@
+<xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema">
+  <xsd:element name="a" type="AType"/>
+  <xsd:complexType name="AType">
+    <xsd:sequence>
+      <xsd:element name="b" type="xsd:string" />
+    </xsd:sequence>
+  </xsd:complexType>
+</xsd:schema>

Modified: lxml/trunk/src/lxml/tests/test_xmlschema.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xmlschema.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xmlschema.py	Fri Apr 28 21:11:37 2006
@@ -6,7 +6,7 @@
 
 import unittest
 
-from common_imports import etree, HelperTestCase
+from common_imports import etree, HelperTestCase, fileInTestDir
 
 class ETreeXMLSchemaTestCase(HelperTestCase):
     def test_xmlschema(self):
@@ -45,11 +45,13 @@
         self.assertRaises(etree.XMLSchemaParseError,
                           etree.XMLSchema, schema)
 
-##     def test_xmlschema_include(self):
-##         # this will only work if we access the file through path or
-##         # file object..
-##         f = open(fileInTestDir('test1.rng'), 'r')
-##         schema = etree.RelaxNG(file=f)
+    def test_xmlschema_file(self):
+        # this will only work if we access the file through path or
+        # file object..
+        f = open(fileInTestDir('test.xsd'), 'r')
+        schema = etree.XMLSchema(file=f)
+        tree_valid = self.parse('<a><b></b></a>')
+        self.assert_(schema.validate(tree_valid))
 
     def test_xmlschema_shortcut(self):
         tree_valid = self.parse('<a><b></b></a>')

Modified: lxml/trunk/src/lxml/xmlschema.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlschema.pxi	(original)
+++ lxml/trunk/src/lxml/xmlschema.pxi	Fri Apr 28 21:11:37 2006
@@ -13,46 +13,55 @@
 ################################################################################
 # XMLSchema
 
-cdef class XMLSchema:
+cdef class XMLSchema(_Validator):
     """Turn a document into an XML Schema validator.
     """
     cdef xmlschema.xmlSchema* _c_schema
-    cdef _ErrorLog _error_log
-    
-    def __init__(self, etree):
+    def __init__(self, etree=None, file=None):
         cdef _Document doc
         cdef _NodeBase root_node
         cdef xmlDoc* fake_c_doc
         cdef xmlNode* c_node
         cdef xmlschema.xmlSchemaParserCtxt* parser_ctxt
 
-        doc = _documentOrRaise(etree)
-        root_node = _rootNodeOf(etree)
+        if etree is not None:
+            doc = _documentOrRaise(etree)
+            root_node = _rootNodeOf(etree)
+
+            # work around for libxml2 bug if document is not XML schema at all
+            c_node = root_node._c_node
+            if c_node.ns is NULL or c_node.ns.href is NULL or \
+                   tree.strcmp(c_node.ns.href, 'http://www.w3.org/2001/XMLSchema') != 0:
+                raise XMLSchemaParseError, "Document is not XML Schema"
+
+            fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
+            parser_ctxt = xmlschema.xmlSchemaNewDocParserCtxt(fake_c_doc)
+            if parser_ctxt is NULL:
+                _destroyFakeDoc(doc._c_doc, fake_c_doc)
+                raise XMLSchemaParseError, "Document is not parsable as XML Schema"
+            self._c_schema = xmlschema.xmlSchemaParse(parser_ctxt)
 
-        # work around for libxml2 bug if document is not XML schema at all
-        c_node = root_node._c_node
-        if c_node.ns is NULL or c_node.ns.href is NULL or \
-               tree.strcmp(c_node.ns.href, 'http://www.w3.org/2001/XMLSchema') != 0:
-            raise XMLSchemaParseError, "Document is not XML Schema"
-
-        fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
-        parser_ctxt = xmlschema.xmlSchemaNewDocParserCtxt(fake_c_doc)
-        if parser_ctxt is NULL:
+            xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
             _destroyFakeDoc(doc._c_doc, fake_c_doc)
-            raise XMLSchemaParseError, "Document is not parsable as XML Schema"
-        self._c_schema = xmlschema.xmlSchemaParse(parser_ctxt)
-
-        xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
-        _destroyFakeDoc(doc._c_doc, fake_c_doc)
+        elif file is not None:
+            filename = _getFilenameForFile(file)
+            if filename is None:
+                # XXX assume a string object
+                filename = file
+            parser_ctxt = xmlschema.xmlSchemaNewParserCtxt(filename)
+            self._c_schema = xmlschema.xmlSchemaParse(parser_ctxt)
+            xmlschema.xmlSchemaFreeParserCtxt(parser_ctxt)
+        else:
+            raise XMLSchemaParseError, "No tree or file given"
 
         if self._c_schema is NULL:
             raise XMLSchemaParseError, "Document is not valid XML Schema"
-        self._error_log = _ErrorLog()
+        _Validator.__init__(self)
 
     def __dealloc__(self):
         xmlschema.xmlSchemaFree(self._c_schema)
 
-    def validate(self, etree):
+    def __call__(self, etree):
         """Validate doc using XML Schema.
 
         Returns true if document is valid, false if not.
@@ -78,7 +87,3 @@
         if ret == -1:
             raise XMLSchemaValidateError, "Internal error in XML Schema validation."
         return ret == 0
-
-    property error_log:
-        def __get__(self):
-            return self._error_log.copy()
From scoder at codespeak.net  Fri Apr 28 21:15:49 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:15:50 2006
Subject: [Lxml-checkins] r26536 - lxml/trunk/src/lxml
Message-ID: <20060428191549.980FC10094@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:15:48 2006
New Revision: 26536

Added:
   lxml/trunk/src/lxml/xpath.pxi
      - copied unchanged from r26535, lxml/trunk/src/lxml/xslt.pxi
Log:
new source file xpath.pxi to split xslt.pxi into XSLT and XPath part

From scoder at codespeak.net  Fri Apr 28 21:17:38 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:17:40 2006
Subject: [Lxml-checkins] r26537 - lxml/trunk/src/lxml
Message-ID: <20060428191738.8017210094@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:17:37 2006
New Revision: 26537

Added:
   lxml/trunk/src/lxml/extensions.pxi
      - copied unchanged from r26536, lxml/trunk/src/lxml/xslt.pxi
Log:
new source file extensions.pxi to factor extensions out of xslt.pxi/xpath.pxi

From scoder at codespeak.net  Fri Apr 28 21:36:11 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:36:14 2006
Subject: [Lxml-checkins] r26538 - lxml/trunk/src/lxml
Message-ID: <20060428193611.B091010094@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:36:09 2006
New Revision: 26538

Modified:
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/extensions.pxi
   lxml/trunk/src/lxml/xpath.pxi
   lxml/trunk/src/lxml/xslt.pxi
Log:
major split of xslt.pxi: XSLT -> xslt.pxi, XPath -> xpath.pxi, extension functions -> extensions.pxi

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Fri Apr 28 21:36:09 2006
@@ -1352,13 +1352,15 @@
 
 
 # include submodules
-include "proxy.pxi"     # Proxy handling (element backpointers/memory/etc.)
-include "xmlerror.pxi"  # error and log handling
-include "nsclasses.pxi" # Namespace implementation and registry
-include "docloader.pxi" # Support for custom document loaders
-include "parser.pxi"    # XML Parser
-include "xmlid.pxi"     # XMLID and IDDict
-include "xslt.pxi"      # XPath and XSLT
+include "proxy.pxi"      # Proxy handling (element backpointers/memory/etc.)
+include "xmlerror.pxi"   # error and log handling
+include "nsclasses.pxi"  # Namespace implementation and registry
+include "docloader.pxi"  # Support for custom document loaders
+include "parser.pxi"     # XML Parser
+include "xmlid.pxi"      # XMLID and IDDict
+include "extensions.pxi" # XPath/XSLT extension functions
+include "xpath.pxi"      # XPath evaluation
+include "xslt.pxi"       # XSL transformations
 
 
 ################################################################################

Modified: lxml/trunk/src/lxml/extensions.pxi
==============================================================================
--- lxml/trunk/src/lxml/extensions.pxi	(original)
+++ lxml/trunk/src/lxml/extensions.pxi	Fri Apr 28 21:36:09 2006
@@ -1,37 +1,16 @@
-# XSLT and XPath classes, supports for extension functions
-
-class XSLTError(LxmlError):
-    pass
-
-class XSLTParseError(XSLTError):
-    pass
-
-class XSLTApplyError(XSLTError):
-    pass
-
-class XSLTSaveError(XSLTError):
-    pass
-
-class XSLTExtensionError(XSLTError):
-    pass
+# supports for extension functions in XPath and XSLT
 
 class XPathError(LxmlError):
     pass
 
-class XPathContextError(XPathError):
-    pass
-
 class XPathFunctionError(XPathError):
     pass
 
 class XPathResultError(XPathError):
     pass
 
-class XPathSyntaxError(LxmlSyntaxError):
-    pass
-
 ################################################################################
-# support for extension functions in XPath/XSLT
+# Base class for XSLT and XPath evaluation contexts: functions, namespaces, ...
 
 cdef class _BaseContext:
     cdef xpath.xmlXPathContext* _xpathCtxt
@@ -183,8 +162,20 @@
                 #print "Holding document:", <int>element._doc._c_doc
                 self._temp_refs.add(element._doc)
 
+
+def Extension(module, function_mapping, ns_uri=None):
+    functions = []
+    for function_name, xpath_name in function_mapping.items():
+        functions[xpath_name] = getattr(module, function_name)
+    return {ns_uri : functions}
+
+
+################################################################################
+# helper functions
+
 cdef xpath.xmlXPathFunction _function_check(void* ctxt,
                                             char* c_name, char* c_ns_uri):
+    "Module level lookup function for XPath/XSLT functions"
     cdef _BaseContext context
     if c_name is NULL:
         return NULL
@@ -198,681 +189,6 @@
     else:
         return NULL
 
-cdef xpath.xmlXPathFunction _xslt_function_check(void* ctxt,
-                                                 char* c_name, char* c_ns_uri):
-    cdef xpath.xmlXPathFunction result
-    result = _function_check(ctxt, c_name, c_ns_uri)
-    if result is NULL:
-        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
-    else:
-        return result
-
-
-################################################################################
-# XSLT document loaders
-
-cdef class _XSLTResolverContext(_ResolverContext):
-    cdef xmlDoc* _c_style_doc
-    cdef BaseParser _parser
-    def __init__(self, BaseParser parser not None):
-        _ResolverContext.__init__(self, parser.resolvers)
-        self._parser = parser
-        self._c_style_doc = NULL
-
-cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
-                         void* c_ctxt, xslt.xsltLoadType c_type):
-    cdef xmlDoc* c_doc
-    cdef _ResolverRegistry resolvers
-    cdef _InputDocument doc_ref
-    cdef _XSLTResolverContext xslt_resolver_context
-    cdef _XSLTResolverContext doc_resolver_context
-    cdef _XSLTResolverContext resolver_context
-    cdef XMLParser parser
-    # find resolver contexts of stylesheet and transformed doc
-    c_doc = NULL
-    doc_resolver_context = None
-    if c_type == xslt.XSLT_LOAD_DOCUMENT:
-        c_doc = (<xslt.xsltTransformContext*>c_ctxt).document.doc
-        if c_doc is not NULL and c_doc._private is not NULL:
-            if isinstance(<object>c_doc._private, _XSLTResolverContext):
-                doc_resolver_context = <_XSLTResolverContext>c_doc._private
-        c_doc = (<xslt.xsltTransformContext*>c_ctxt).style.doc
-    elif c_type == xslt.XSLT_LOAD_STYLESHEET:
-        c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
-
-    if c_doc is NULL or c_doc._private is NULL or \
-           not isinstance(<object>c_doc._private, _XSLTResolverContext):
-        # can't call Python without context, fall back to default loader
-        return XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, parse_options, c_ctxt, c_type)
-
-    xslt_resolver_context = <_XSLTResolverContext>c_doc._private
-
-    # quick check if we are looking for the current stylesheet
-    c_doc = xslt_resolver_context._c_style_doc
-    if c_doc is not NULL and c_doc.URL is not NULL:
-        if tree.strcmp(c_uri, c_doc.URL) == 0:
-            return tree.xmlCopyDoc(c_doc, 1)
-
-    # call the Python document loaders
-    c_doc = NULL
-    resolver_context = xslt_resolver_context # currently use only XSLT resolvers
-    resolvers = resolver_context._resolvers
-    try:
-        uri = funicode(c_uri)
-        doc_ref = resolvers.resolve(uri, None, resolver_context)
-
-        if doc_ref is not None:
-            if doc_ref._type == PARSER_DATA_EMPTY:
-                c_doc = _newDoc()
-            if doc_ref._type == PARSER_DATA_STRING:
-                c_doc = _internalParseDoc(
-                    _cstr(doc_ref._data_utf), parse_options,
-                    resolver_context)
-            elif doc_ref._type == PARSER_DATA_FILE:
-                data = doc_ref._file.read()
-                c_doc = _internalParseDoc(
-                    _cstr(data), parse_options,
-                    resolver_context)
-            elif doc_ref._type == PARSER_DATA_FILENAME:
-                c_doc = _internalParseDocFromFile(
-                    _cstr(doc_ref._data_utf), parse_options,
-                    resolver_context)
-            if c_doc is not NULL and c_doc.URL is NULL:
-                c_doc.URL = tree.xmlStrdup(c_uri)
-
-    except Exception, e:
-        xslt_resolver_context._store_raised()
-        return NULL
-
-    if c_doc is NULL:
-        c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, parse_options, c_ctxt, c_type)
-        if c_doc is NULL:
-            message = "Cannot resolve URI %s" % funicode(c_uri)
-            if c_type == xslt.XSLT_LOAD_DOCUMENT:
-                exception = XSLTApplyError(message)
-            else:
-                exception = XSLTParseError(message)
-            xslt_resolver_context._store_exception(exception)
-            return NULL
-    if c_doc is not NULL and c_doc._private is NULL:
-        c_doc._private = <python.PyObject*>xslt_resolver_context
-    return c_doc
-
-cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
-XSLT_DOC_DEFAULT_LOADER = xslt.xsltDocDefaultLoader
-
-xslt.xsltSetLoaderFunc(_doc_loader)
-
-
-################################################################################
-# XSLT
-
-cdef class _XSLTContext(_BaseContext):
-    cdef xslt.xsltTransformContext* _xsltCtxt
-    def __init__(self, namespaces, extensions):
-        self._xsltCtxt = NULL
-        self._ext_lookup_function = _xslt_function_check
-        if extensions and None in extensions:
-            raise XSLTExtensionError, "extensions must not have empty namespaces"
-        _BaseContext.__init__(self, namespaces, extensions)
-
-    cdef register_context(self, xslt.xsltTransformContext* xsltCtxt,
-                               _Document doc):
-        self._xsltCtxt = xsltCtxt
-        self._set_xpath_context(xsltCtxt.xpathCtxt)
-        self._register_context(doc, 0)
-        xsltCtxt.xpathCtxt.userData = <void*>self
-
-    cdef free_context(self):
-        cdef xslt.xsltTransformContext* xsltCtxt
-        xsltCtxt = self._xsltCtxt
-        if xsltCtxt is NULL:
-            return
-        self._free_context()
-        self._xsltCtxt = NULL
-        xslt.xsltFreeTransformContext(xsltCtxt)
-        self._release_temp_refs()
-
-    cdef _registerLocalExtensionFunction(self, ns_utf, name_utf, function):
-        extensions = self._extensions
-        if extensions is None:
-            self._extensions = {ns_utf:{name_utf:function}}
-        else:
-            if ns_utf in extensions:
-                ns_extensions = extensions[ns_utf]
-            else:
-                ns_extensions = extensions[ns_utf] = {}
-            python.PyDict_SetItem(ns_extensions, name_utf, function)
-
-cdef class _ExsltRegExp # forward declaration
-
-cdef class XSLT:
-    """Turn a document into an XSLT object.
-    """
-    cdef _XSLTContext _context
-    cdef xslt.xsltStylesheet* _c_style
-    cdef _XSLTResolverContext _xslt_resolver_context
-    cdef _ExsltRegExp _regexp
-    cdef _ErrorLog _error_log
-
-    def __init__(self, xslt_input, extensions=None, regexp=True):
-        cdef xslt.xsltStylesheet* c_style
-        cdef xmlDoc* c_doc
-        cdef xmlDoc* fake_c_doc
-        cdef _Document doc
-        cdef _NodeBase root_node
-
-        doc = _documentOrRaise(xslt_input)
-        root_node = _rootNodeOf(xslt_input)
-
-        # make a copy of the document as stylesheet parsing modifies it
-        fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
-        c_doc = tree.xmlCopyDoc(fake_c_doc, 1)
-        _destroyFakeDoc(doc._c_doc, fake_c_doc)
-
-        # make sure we always have a stylesheet URL
-        if c_doc.URL is not NULL:
-            # handle a bug in older libxml2 versions
-            tree.xmlFree(c_doc.URL)
-        if doc._c_doc.URL is not NULL:
-            c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
-        else:
-            doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
-            c_doc.URL = tree.xmlStrdup(_cstr(doc_url_utf))
-
-        self._xslt_resolver_context = _XSLTResolverContext(doc._parser)
-        # keep a copy in case we need to access the stylesheet via 'document()'
-        self._xslt_resolver_context._c_style_doc = tree.xmlCopyDoc(c_doc, 1)
-        c_doc._private = <python.PyObject*>self._xslt_resolver_context
-
-        c_style = xslt.xsltParseStylesheetDoc(c_doc)
-        if c_style is NULL:
-            tree.xmlFreeDoc(c_doc)
-            self._xslt_resolver_context._raise_if_stored()
-            raise XSLTParseError, "Cannot parse style sheet"
-        self._c_style = c_style
-
-        self._context = _XSLTContext(None, extensions)
-        self._error_log = _ErrorLog()
-        if regexp:
-            self._regexp  = _ExsltRegExp()
-        else:
-            self._regexp  = None
-        # XXX is it worthwile to use xsltPrecomputeStylesheet here?
-
-    def __dealloc__(self):
-        if self._xslt_resolver_context is not None and \
-               self._xslt_resolver_context._c_style_doc is not NULL:
-            tree.xmlFreeDoc(self._xslt_resolver_context._c_style_doc)
-        # this cleans up copy of doc as well
-        xslt.xsltFreeStylesheet(self._c_style)
-
-    property error_log:
-        def __get__(self):
-            return self._error_log.copy()
-
-    def __call__(self, _input, **_kw):
-        cdef _Document input_doc
-        cdef _NodeBase root_node
-        cdef _Document result_doc
-        cdef _XSLTResolverContext resolver_context
-        cdef xslt.xsltTransformContext* transform_ctxt
-        cdef xmlDoc* c_result
-        cdef xmlDoc* c_doc
-        cdef char** params
-        cdef void* ptemp
-        cdef int i
-
-        input_doc = _documentOrRaise(_input)
-        root_node = _rootNodeOf(_input)
-
-        resolver_context = _XSLTResolverContext(input_doc._parser)
-        resolver_context._c_style_doc = self._xslt_resolver_context._c_style_doc
-
-        c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
-
-        transform_ctxt = xslt.xsltNewTransformContext(self._c_style, c_doc)
-        if transform_ctxt is NULL:
-            _destroyFakeDoc(input_doc._c_doc, c_doc)
-            raise XSLTApplyError, "Error preparing stylesheet run"
-
-        self._error_log.connect()
-        xslt.xsltSetTransformErrorFunc(transform_ctxt, <void*>self._error_log,
-                                       _receiveGenericError)
-
-        ptemp = c_doc._private
-        c_doc._private = <python.PyObject*>resolver_context
-
-        if _kw:
-            # allocate space for parameters
-            # * 2 as we want an entry for both key and value,
-            # and + 1 as array is NULL terminated
-            params = <char**>cstd.malloc(sizeof(char*) * (len(_kw) * 2 + 1))
-            i = 0
-            keep_ref = []
-            for key, value in _kw.items():
-                k = _utf8(key)
-                python.PyList_Append(keep_ref, k)
-                v = _utf8(value)
-                python.PyList_Append(keep_ref, v)
-                params[i] = _cstr(k)
-                i = i + 1
-                params[i] = _cstr(v)
-                i = i + 1
-            params[i] = NULL
-        else:
-            params = NULL
-
-        self._context.register_context(transform_ctxt, input_doc)
-        if self._regexp is not None:
-            self._regexp._register_in_context(self._context)
-
-        c_result = xslt.xsltApplyStylesheetUser(self._c_style, c_doc, params,
-                                                NULL, NULL, transform_ctxt)
-
-        if params is not NULL:
-            # deallocate space for parameters
-            cstd.free(params)
-
-        self._context.free_context()
-        c_doc._private = ptemp # restore _private before _destroyFakeDoc!
-        _destroyFakeDoc(input_doc._c_doc, c_doc)
-
-        self._error_log.disconnect()
-        if self._xslt_resolver_context._has_raised():
-            if c_result is not NULL:
-                tree.xmlFreeDoc(c_result)
-            self._xslt_resolver_context._raise_if_stored()
-
-        if c_result is NULL:
-            raise XSLTApplyError, "Error applying stylesheet"
-
-        result_doc = _documentFactory(c_result, input_doc._parser)
-        return _xsltResultTreeFactory(result_doc, self)
-
-    def apply(self, _input, **_kw):
-        return self.__call__(_input, **_kw)
-
-    def tostring(self, _ElementTree result_tree):
-        """Save result doc to string based on stylesheet output method.
-        """
-        return str(result_tree)
-
-cdef class _XSLTResultTree(_ElementTree):
-    cdef XSLT _xslt
-    def __str__(self):
-        cdef char* s
-        cdef int l
-        cdef int r
-        r = xslt.xsltSaveResultToString(&s, &l, self._doc._c_doc,
-                                        self._xslt._c_style)
-        if r == -1:
-            raise XSLTSaveError, "Error saving XSLT result to string"
-        if s is NULL:
-            return ''
-        result = funicode(s)
-        tree.xmlFree(s)
-        return result
-
-cdef _xsltResultTreeFactory(_Document doc, XSLT xslt):
-    cdef _XSLTResultTree result
-    result = <_XSLTResultTree>_newElementTree(doc, None, _XSLTResultTree)
-    result._xslt = xslt
-    return result
-
-# do not register all libxslt extra function, provide only "node-set"
-# functions like "output" and "write" are a potential security risk
-#xslt.xsltRegisterAllExtras()
-xslt.xsltRegisterExtModuleFunction("node-set",
-                                   xslt.XSLT_LIBXSLT_NAMESPACE,
-                                   xslt.xsltFunctionNodeSet)
-xslt.xsltRegisterExtModuleFunction("node-set",
-                                   xslt.XSLT_SAXON_NAMESPACE,
-                                   xslt.xsltFunctionNodeSet)
-xslt.xsltRegisterExtModuleFunction("node-set",
-                                   xslt.XSLT_XT_NAMESPACE,
-                                   xslt.xsltFunctionNodeSet)
-
-# enable EXSLT support for XSLT
-xslt.exsltRegisterAll()
-
-################################################################################
-# EXSLT regexp implementation
-
-cdef object RE_COMPILE
-RE_COMPILE = re.compile
-
-cdef class _ExsltRegExp:
-    cdef object _compile_map
-    def __init__(self):
-        self._compile_map = {}
-
-    cdef _make_string(self, value):
-        if python.PyString_Check(value) or python.PyUnicode_Check(value):
-            return value
-        else:
-            raise TypeError, "Invalid argument type %s" % type(value)
-
-    cdef _compile(self, rexp, ignore_case):
-        cdef python.PyObject* c_result
-        rexp = self._make_string(rexp)
-        key = (rexp, ignore_case)
-        c_result = python.PyDict_GetItem(self._compile_map, key)
-        if c_result is not NULL:
-            return <object>c_result
-        py_flags = re.UNICODE
-        if ignore_case:
-            py_flags = py_flags | re.IGNORECASE
-        rexp_compiled = RE_COMPILE(rexp, py_flags)
-        python.PyDict_SetItem(self._compile_map, key, rexp_compiled)
-        return rexp_compiled
-
-    def test(self, ctxt, s, rexp, flags=''):
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if rexpc.search(s) is None:
-            return False
-        else:
-            return True
-
-    def match(self, ctxt, s, rexp, flags=''):
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if 'g' in flags:
-            results = rexpc.findall(s)
-            if not results:
-                return ()
-            result_list = []
-            root = Element('matches')
-            for s_match in results:
-                elem = SubElement(root, 'match')
-                elem.text = s_match
-                python.PyList_Append(result_list, elem)
-            return result_list
-        else:
-            result = rexpc.search(s)
-            if result is None:
-                return ()
-            root = Element('match')
-            root.text = result.group()
-            return (root,)
-
-    def replace(self, ctxt, s, rexp, flags, replacement):
-        replacement = self._make_string(replacement)
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if 'g' in flags:
-            count = 0
-        else:
-            count = 1
-        return rexpc.sub(replacement, s, count)
-
-    cdef _register_in_context(self, _XSLTContext context):
-        ns = "http://exslt.org/regular-expressions"
-        context._registerLocalExtensionFunction(ns, "test",    self.test)
-        context._registerLocalExtensionFunction(ns, "match",   self.match)
-        context._registerLocalExtensionFunction(ns, "replace", self.replace)
-
-################################################################################
-# XPath
-
-cdef class _XPathContext(_BaseContext):
-    cdef object _variables
-    cdef object _registered_variables
-    def __init__(self, namespaces, extensions, variables):
-        self._ext_lookup_function = _function_check
-        self._variables = variables
-        self._registered_variables  = []
-        _BaseContext.__init__(self, namespaces, extensions)
-        
-    cdef register_context(self, xpath.xmlXPathContext* xpathCtxt, _Document doc):
-        self._set_xpath_context(xpathCtxt)
-        ns_prefixes = _find_all_extension_prefixes()
-        if ns_prefixes:
-            self.registerNamespaces(ns_prefixes)
-        self._register_context(doc, 1)
-        if self._variables is not None:
-            self.registerVariables(self._variables)
-
-    cdef unregister_context(self):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        xpathCtxt = self._xpathCtxt
-        if xpathCtxt is NULL:
-            return
-        self._unregisterVariables()
-        del self._registered_variables[:]
-        self._unregister_context()
-
-    cdef void _unregisterVariables(self):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject* xpathVarValue
-        cdef char* c_name
-        xpathCtxt = self._xpathCtxt
-        for name_utf in self._registered_variables:
-            c_name = _cstr(name_utf)
-            xpathVarValue = xpath.xmlXPathVariableLookup(xpathCtxt, c_name)
-            if xpathVarValue is not NULL:
-                xpath.xmlXPathRegisterVariable(xpathCtxt, c_name, NULL)
-                _freeXPathObject(xpathVarValue)
-
-    def registerVariables(self, variable_dict):
-        for name, value in variable_dict.items():
-            name_utf = self._to_utf(name)
-            self._registerVariable(name_utf, value)
-            python.PyList_Append(self._registered_variables, name_utf)
-
-    def registerVariable(self, name, value):
-        name_utf = self._to_utf(name)
-        self._registerVariable(name_utf, value)
-        python.PyList_Append(self._registered_variables, name_utf)
-
-    cdef void _registerVariable(self, name_utf, value):
-        xpath.xmlXPathRegisterVariable(
-            self._xpathCtxt, _cstr(name_utf), _wrapXPathObject(value))
-
-
-cdef class XPathEvaluatorBase:
-    cdef _XPathContext _context
-
-    def __init__(self, namespaces, extensions, variables=None):
-        self._context = _XPathContext(namespaces, extensions, variables)
-
-    cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
-        if self._context._exc._has_raised():
-            if xpathObj is not NULL:
-                _freeXPathObject(xpathObj)
-                xpathObj = NULL
-            self._context._release_temp_refs()
-            self._context._exc._raise_if_stored()
-
-        if xpathObj is NULL:
-            self._context._release_temp_refs()
-            raise XPathSyntaxError, "Error in xpath expression."
-
-        try:
-            result = _unwrapXPathObject(xpathObj, doc)
-        except XPathResultError:
-            _freeXPathObject(xpathObj)
-            self._context._release_temp_refs()
-            raise
-
-        _freeXPathObject(xpathObj)
-        self._context._release_temp_refs()
-        return result
-
-
-cdef class XPathElementEvaluator(XPathEvaluatorBase):
-    """Create an XPath evaluator for an element.
-
-    XPath evaluators must not be shared between threads.
-    """
-    cdef xpath.xmlXPathContext* _c_ctxt
-    cdef _Element _element
-    def __init__(self, _NodeBase element not None, namespaces=None, extensions=None):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef int ns_register_status
-        cdef _Document doc
-        doc = element._doc
-        xpathCtxt = xpath.xmlXPathNewContext(doc._c_doc)
-        if xpathCtxt is NULL:
-            raise XPathContextError, "Unable to create new XPath context"
-        self._element = element
-        self._c_ctxt = xpathCtxt
-        XPathEvaluatorBase.__init__(self, namespaces, extensions)
-
-    def __dealloc__(self):
-        if self._c_ctxt is not NULL:
-            xpath.xmlXPathFreeContext(self._c_ctxt)
-    
-    def registerNamespace(self, prefix, uri):
-        """Register a namespace with the XPath context.
-        """
-        self._context.addNamespace(prefix, uri)
-
-    def registerNamespaces(self, namespaces):
-        """Register a prefix -> uri dict.
-        """
-        add = self._context.addNamespace
-        for prefix, uri in namespaces.items():
-            add(prefix, uri)
-    
-    def evaluate(self, _path, **_variables):
-        """Evaluate an XPath expression on the document.  Variables may be
-        provided as keyword arguments. Note that namespaces are currently not
-        supported for variables."""
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject*  xpathObj
-        cdef xmlNode* c_node
-        cdef _Document doc
-        xpathCtxt = self._c_ctxt
-        xpathCtxt.node = self._element._c_node
-        doc = self._element._doc
-
-        self._context.register_context(xpathCtxt, doc)
-        self._context.registerVariables(_variables)
-
-        path = _utf8(_path)
-        xpathObj = xpath.xmlXPathEvalExpression(_cstr(path), xpathCtxt)
-        self._context.unregister_context()
-
-        return self._handle_result(xpathObj, doc)
-
-    #def clone(self):
-    #    # XXX pretty expensive so calling this from callback is probably
-    #    # not desirable
-    #    return XPathEvaluator(self._doc, self._namespaces, self._extensions)
-
-cdef class XPathDocumentEvaluator(XPathElementEvaluator):
-    """Create an XPath evaluator for an ElementTree.
-
-    XPath evaluators must not be shared between threads.
-    """
-    def __init__(self, _ElementTree etree not None, namespaces=None, extensions=None):
-        XPathElementEvaluator.__init__(
-            self, etree._context_node, namespaces, extensions)
-
-def XPathEvaluator(etree_or_element, namespaces=None, extensions=None):
-    """Creates and XPath evaluator for an ElementTree or an Element.
-
-    XPath evaluators must not be shared between threads.
-    """
-    if isinstance(etree_or_element, _ElementTree):
-        return XPathDocumentEvaluator(etree_or_element, namespaces, extensions)
-    else:
-        return XPathElementEvaluator(etree_or_element, namespaces, extensions)
-
-def Extension(module, function_mapping, ns_uri=None):
-    functions = []
-    for function_name, xpath_name in function_mapping.items():
-        functions[xpath_name] = getattr(module, function_name)
-    return {ns_uri : functions}
-
-cdef class XPath(XPathEvaluatorBase):
-    cdef xpath.xmlXPathContext* _xpathCtxt
-    cdef xpath.xmlXPathCompExpr* _xpath
-    cdef object _prefix_map
-    cdef readonly object path
-
-    def __init__(self, path, namespaces=None, extensions=None):
-        XPathEvaluatorBase.__init__(self, namespaces, extensions, None)
-        self.path = path
-        path = _utf8(path)
-        self._xpath = xpath.xmlXPathCompile(_cstr(path))
-        if self._xpath is NULL:
-            raise XPathSyntaxError, "Error in XPath expression"
-        self._xpathCtxt = xpath.xmlXPathNewContext(NULL)
-
-    def __call__(self, _etree_or_element, **_variables):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject*  xpathObj
-        cdef _Document document
-        cdef _NodeBase element
-        cdef _XPathContext context
-
-        document = _documentOrRaise(_etree_or_element)
-        element  = _rootNodeOf(_etree_or_element)
-
-        xpathCtxt = self._xpathCtxt
-        xpathCtxt.doc = document._c_doc
-        xpathCtxt.node = element._c_node
-
-        context = self._context
-        context._release_temp_refs()
-        context.register_context(xpathCtxt, document)
-        context.registerVariables(_variables)
-
-        xpathObj = xpath.xmlXPathCompiledEval(self._xpath, xpathCtxt)
-        context.unregister_context()
-        return self._handle_result(xpathObj, document)
-
-    def evaluate(self, _tree, **_variables):
-        return self(_tree, **_variables)
-
-    def __dealloc__(self):
-        if self._xpathCtxt is not NULL:
-            xpath.xmlXPathFreeContext(self._xpathCtxt)
-        if self._xpath is not NULL:
-            xpath.xmlXPathFreeCompExpr(self._xpath)
-
-cdef object _replace_strings
-cdef object _find_namespaces
-_replace_strings = re.compile('("[^"]*")|(\'[^\']*\')').sub
-_find_namespaces = re.compile('({[^}]+})').findall
-
-cdef class ETXPath(XPath):
-    """Special XPath class that supports the ElementTree {uri} notation for
-    namespaces."""
-    def __init__(self, path, extensions=None):
-        path_utf, namespaces = self._nsextract_path(_utf8(path))
-        XPath.__init__(self, funicode(path_utf), namespaces, extensions)
-
-    cdef _nsextract_path(self, path_utf):
-        # replace {namespaces} by new prefixes
-        cdef int i
-        namespaces = {}
-        stripped_path = _replace_strings('', path_utf) # remove string literals
-        namespace_defs = []
-        i = 1
-        for namespace_def in _find_namespaces(stripped_path):
-            if namespace_def not in namespace_defs:
-                prefix = python.PyString_FromFormat("xpp%02d", i)
-                i = i+1
-                python.PyList_Append(namespace_defs, namespace_def)
-                namespace = namespace_def[1:-1] # remove '{}'
-                python.PyDict_SetItem(namespaces, prefix, namespace)
-                prefix_str = prefix + ':'
-                # FIXME: this also replaces {namespaces} within strings!
-                path_utf   = path_utf.replace(namespace_def, prefix_str)
-        return path_utf, namespaces
-
-################################################################################
-# helper functions
-
 cdef xpath.xmlXPathObject* _wrapXPathObject(object obj) except NULL:
     cdef xpath.xmlNodeSet* resultSet
     cdef _NodeBase node

Modified: lxml/trunk/src/lxml/xpath.pxi
==============================================================================
--- lxml/trunk/src/lxml/xpath.pxi	(original)
+++ lxml/trunk/src/lxml/xpath.pxi	Fri Apr 28 21:36:09 2006
@@ -1,624 +1,12 @@
 # XSLT and XPath classes, supports for extension functions
 
-class XSLTError(LxmlError):
-    pass
-
-class XSLTParseError(XSLTError):
-    pass
-
-class XSLTApplyError(XSLTError):
-    pass
-
-class XSLTSaveError(XSLTError):
-    pass
-
-class XSLTExtensionError(XSLTError):
-    pass
-
-class XPathError(LxmlError):
-    pass
-
 class XPathContextError(XPathError):
     pass
 
-class XPathFunctionError(XPathError):
-    pass
-
-class XPathResultError(XPathError):
-    pass
-
 class XPathSyntaxError(LxmlSyntaxError):
     pass
 
 ################################################################################
-# support for extension functions in XPath/XSLT
-
-cdef class _BaseContext:
-    cdef xpath.xmlXPathContext* _xpathCtxt
-    cdef xpath.xmlXPathFuncLookupFunc _ext_lookup_function
-    cdef _Document _doc
-    cdef object _extensions
-    cdef object _namespaces
-    cdef object _registered_namespaces
-    cdef object _utf_refs
-    cdef object _function_cache
-    cdef object _called_function
-    # for exception handling and temporary reference keeping:
-    cdef _TempStore _temp_refs
-    cdef _ExceptionContext _exc
-
-    def __init__(self, namespaces, extensions):
-        self._xpathCtxt = NULL
-        self._utf_refs = {}
-        self._function_cache = {}
-        self._called_function = None
-
-        # convert old format extensions to UTF-8
-        if isinstance(extensions, (list, tuple)):
-            new_extensions = {}
-            for extension in extensions:
-                for (ns_uri, name), function in extension.items():
-                    ns_utf   = self._to_utf(ns_uri)
-                    name_utf = self._to_utf(name)
-                    try:
-                        new_extensions[ns_utf][name_utf] = function
-                    except KeyError:
-                        new_extensions[ns_utf] = {name_utf : function}
-            extensions = new_extensions or None
-
-        self._doc        = None
-        self._exc        = _ExceptionContext()
-        self._extensions = extensions
-        self._namespaces = namespaces
-        self._registered_namespaces = []
-        self._temp_refs = _TempStore()
-
-    cdef object _to_utf(self, s):
-        "Convert to UTF-8 and keep a reference to the encoded string"
-        cdef python.PyObject* dict_result
-        if s is None:
-            return None
-        dict_result = python.PyDict_GetItem(self._utf_refs, s)
-        if dict_result is not NULL:
-            return <object>dict_result
-        utf = _utf8(s)
-        python.PyDict_SetItem(self._utf_refs, s, utf)
-        return utf
-
-    cdef void _set_xpath_context(self, xpath.xmlXPathContext* xpathCtxt):
-        self._xpathCtxt = xpathCtxt
-        xpathCtxt.userData = <void*>self
-
-    cdef _register_context(self, _Document doc, int allow_none_namespace):
-        self._doc = doc
-        self._exc.clear()
-        python.PyDict_Clear(self._function_cache)
-        namespaces = self._namespaces
-        if namespaces is not None:
-            self.registerNamespaces(namespaces)
-        xpath.xmlXPathRegisterFuncLookup(
-            self._xpathCtxt, self._ext_lookup_function, <python.PyObject*>self)
-
-    cdef _unregister_context(self):
-        self._unregisterNamespaces()
-        self._free_context()
-
-    cdef _free_context(self):
-        del self._registered_namespaces[:]
-        python.PyDict_Clear(self._utf_refs)
-        self._doc = None
-        if self._xpathCtxt is not NULL:
-            self._xpathCtxt.userData = NULL
-            self._xpathCtxt = NULL
-
-    # namespaces (internal UTF-8 methods with leading '_')
-
-    def addNamespace(self, prefix, uri):
-        if self._namespaces is None:
-            self._namespaces = {}
-        python.PyDict_SetItem(self._namespaces, prefix, uri)
-
-    def registerNamespaces(self, namespaces):
-        for prefix, uri in namespaces.items():
-            self.registerNamespace(prefix, uri)
-    
-    def registerNamespace(self, prefix, ns_uri):
-        prefix_utf = self._to_utf(prefix)
-        ns_uri_utf = self._to_utf(ns_uri)
-        xpath.xmlXPathRegisterNs(self._xpathCtxt, prefix_utf, ns_uri_utf)
-        python.PyList_Append(self._registered_namespaces, prefix_utf)
-
-    cdef _unregisterNamespaces(self):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        xpathCtxt = self._xpathCtxt
-        for prefix_utf in self._registered_namespaces:
-            xpath.xmlXPathRegisterNs(xpathCtxt, prefix_utf, NULL)
-    
-    # extension functions
-
-    cdef int _prepare_function_call(self, ns_uri_utf, name_utf):
-        cdef python.PyObject* dict_result
-        key = (ns_uri_utf, name_utf)
-        dict_result = python.PyDict_GetItem(self._function_cache, key)
-        if dict_result is not NULL:
-            function = <object>dict_result
-            self._called_function = function
-            return function is not None
-
-        dict_result = python.PyDict_GetItem(self._extensions, ns_uri_utf)
-        if dict_result is not NULL:
-            dict_result = python.PyDict_GetItem(<object>dict_result, name_utf)
-        if dict_result is not NULL:
-            function = <object>dict_result
-        else:
-            function = _find_extension(ns_uri_utf, name_utf)
-
-        python.PyDict_SetItem(self._function_cache, key, function)
-        self._called_function = function
-        return function is not None
-
-    # Python reference keeping during XPath function evaluation
-
-    cdef _release_temp_refs(self):
-        "Free temporarily referenced objects from this context."
-        self._temp_refs.clear()
-        
-    cdef _hold(self, obj):
-        """A way to temporarily hold references to nodes in the evaluator.
-
-        This is needed because otherwise nodes created in XPath extension
-        functions would be reference counted too soon, during the XPath
-        evaluation.  This is most important in the case of exceptions.
-        """
-        cdef _NodeBase element
-        if isinstance(obj, _NodeBase):
-            obj = (obj,)
-        elif not python.PySequence_Check(obj):
-            return
-        for o in obj:
-            if isinstance(o, _NodeBase):
-                element = <_NodeBase>o
-                #print "Holding element:", <int>element._c_node
-                self._temp_refs.add(element)
-                #print "Holding document:", <int>element._doc._c_doc
-                self._temp_refs.add(element._doc)
-
-cdef xpath.xmlXPathFunction _function_check(void* ctxt,
-                                            char* c_name, char* c_ns_uri):
-    cdef _BaseContext context
-    if c_name is NULL:
-        return NULL
-    if c_ns_uri is NULL:
-        ns_uri = None
-    else:
-        ns_uri = c_ns_uri
-    context = <_BaseContext>ctxt
-    if context._prepare_function_call(ns_uri, c_name):
-        return _call_prepared_function
-    else:
-        return NULL
-
-cdef xpath.xmlXPathFunction _xslt_function_check(void* ctxt,
-                                                 char* c_name, char* c_ns_uri):
-    cdef xpath.xmlXPathFunction result
-    result = _function_check(ctxt, c_name, c_ns_uri)
-    if result is NULL:
-        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
-    else:
-        return result
-
-
-################################################################################
-# XSLT document loaders
-
-cdef class _XSLTResolverContext(_ResolverContext):
-    cdef xmlDoc* _c_style_doc
-    cdef BaseParser _parser
-    def __init__(self, BaseParser parser not None):
-        _ResolverContext.__init__(self, parser.resolvers)
-        self._parser = parser
-        self._c_style_doc = NULL
-
-cdef xmlDoc* _doc_loader(char* c_uri, tree.xmlDict* c_dict, int parse_options,
-                         void* c_ctxt, xslt.xsltLoadType c_type):
-    cdef xmlDoc* c_doc
-    cdef _ResolverRegistry resolvers
-    cdef _InputDocument doc_ref
-    cdef _XSLTResolverContext xslt_resolver_context
-    cdef _XSLTResolverContext doc_resolver_context
-    cdef _XSLTResolverContext resolver_context
-    cdef XMLParser parser
-    # find resolver contexts of stylesheet and transformed doc
-    c_doc = NULL
-    doc_resolver_context = None
-    if c_type == xslt.XSLT_LOAD_DOCUMENT:
-        c_doc = (<xslt.xsltTransformContext*>c_ctxt).document.doc
-        if c_doc is not NULL and c_doc._private is not NULL:
-            if isinstance(<object>c_doc._private, _XSLTResolverContext):
-                doc_resolver_context = <_XSLTResolverContext>c_doc._private
-        c_doc = (<xslt.xsltTransformContext*>c_ctxt).style.doc
-    elif c_type == xslt.XSLT_LOAD_STYLESHEET:
-        c_doc = (<xslt.xsltStylesheet*>c_ctxt).doc
-
-    if c_doc is NULL or c_doc._private is NULL or \
-           not isinstance(<object>c_doc._private, _XSLTResolverContext):
-        # can't call Python without context, fall back to default loader
-        return XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, parse_options, c_ctxt, c_type)
-
-    xslt_resolver_context = <_XSLTResolverContext>c_doc._private
-
-    # quick check if we are looking for the current stylesheet
-    c_doc = xslt_resolver_context._c_style_doc
-    if c_doc is not NULL and c_doc.URL is not NULL:
-        if tree.strcmp(c_uri, c_doc.URL) == 0:
-            return tree.xmlCopyDoc(c_doc, 1)
-
-    # call the Python document loaders
-    c_doc = NULL
-    resolver_context = xslt_resolver_context # currently use only XSLT resolvers
-    resolvers = resolver_context._resolvers
-    try:
-        uri = funicode(c_uri)
-        doc_ref = resolvers.resolve(uri, None, resolver_context)
-
-        if doc_ref is not None:
-            if doc_ref._type == PARSER_DATA_EMPTY:
-                c_doc = _newDoc()
-            if doc_ref._type == PARSER_DATA_STRING:
-                c_doc = _internalParseDoc(
-                    _cstr(doc_ref._data_utf), parse_options,
-                    resolver_context)
-            elif doc_ref._type == PARSER_DATA_FILE:
-                data = doc_ref._file.read()
-                c_doc = _internalParseDoc(
-                    _cstr(data), parse_options,
-                    resolver_context)
-            elif doc_ref._type == PARSER_DATA_FILENAME:
-                c_doc = _internalParseDocFromFile(
-                    _cstr(doc_ref._data_utf), parse_options,
-                    resolver_context)
-            if c_doc is not NULL and c_doc.URL is NULL:
-                c_doc.URL = tree.xmlStrdup(c_uri)
-
-    except Exception, e:
-        xslt_resolver_context._store_raised()
-        return NULL
-
-    if c_doc is NULL:
-        c_doc = XSLT_DOC_DEFAULT_LOADER(
-            c_uri, c_dict, parse_options, c_ctxt, c_type)
-        if c_doc is NULL:
-            message = "Cannot resolve URI %s" % funicode(c_uri)
-            if c_type == xslt.XSLT_LOAD_DOCUMENT:
-                exception = XSLTApplyError(message)
-            else:
-                exception = XSLTParseError(message)
-            xslt_resolver_context._store_exception(exception)
-            return NULL
-    if c_doc is not NULL and c_doc._private is NULL:
-        c_doc._private = <python.PyObject*>xslt_resolver_context
-    return c_doc
-
-cdef xslt.xsltDocLoaderFunc XSLT_DOC_DEFAULT_LOADER
-XSLT_DOC_DEFAULT_LOADER = xslt.xsltDocDefaultLoader
-
-xslt.xsltSetLoaderFunc(_doc_loader)
-
-
-################################################################################
-# XSLT
-
-cdef class _XSLTContext(_BaseContext):
-    cdef xslt.xsltTransformContext* _xsltCtxt
-    def __init__(self, namespaces, extensions):
-        self._xsltCtxt = NULL
-        self._ext_lookup_function = _xslt_function_check
-        if extensions and None in extensions:
-            raise XSLTExtensionError, "extensions must not have empty namespaces"
-        _BaseContext.__init__(self, namespaces, extensions)
-
-    cdef register_context(self, xslt.xsltTransformContext* xsltCtxt,
-                               _Document doc):
-        self._xsltCtxt = xsltCtxt
-        self._set_xpath_context(xsltCtxt.xpathCtxt)
-        self._register_context(doc, 0)
-        xsltCtxt.xpathCtxt.userData = <void*>self
-
-    cdef free_context(self):
-        cdef xslt.xsltTransformContext* xsltCtxt
-        xsltCtxt = self._xsltCtxt
-        if xsltCtxt is NULL:
-            return
-        self._free_context()
-        self._xsltCtxt = NULL
-        xslt.xsltFreeTransformContext(xsltCtxt)
-        self._release_temp_refs()
-
-    cdef _registerLocalExtensionFunction(self, ns_utf, name_utf, function):
-        extensions = self._extensions
-        if extensions is None:
-            self._extensions = {ns_utf:{name_utf:function}}
-        else:
-            if ns_utf in extensions:
-                ns_extensions = extensions[ns_utf]
-            else:
-                ns_extensions = extensions[ns_utf] = {}
-            python.PyDict_SetItem(ns_extensions, name_utf, function)
-
-cdef class _ExsltRegExp # forward declaration
-
-cdef class XSLT:
-    """Turn a document into an XSLT object.
-    """
-    cdef _XSLTContext _context
-    cdef xslt.xsltStylesheet* _c_style
-    cdef _XSLTResolverContext _xslt_resolver_context
-    cdef _ExsltRegExp _regexp
-    cdef _ErrorLog _error_log
-
-    def __init__(self, xslt_input, extensions=None, regexp=True):
-        cdef xslt.xsltStylesheet* c_style
-        cdef xmlDoc* c_doc
-        cdef xmlDoc* fake_c_doc
-        cdef _Document doc
-        cdef _NodeBase root_node
-
-        doc = _documentOrRaise(xslt_input)
-        root_node = _rootNodeOf(xslt_input)
-
-        # make a copy of the document as stylesheet parsing modifies it
-        fake_c_doc = _fakeRootDoc(doc._c_doc, root_node._c_node)
-        c_doc = tree.xmlCopyDoc(fake_c_doc, 1)
-        _destroyFakeDoc(doc._c_doc, fake_c_doc)
-
-        # make sure we always have a stylesheet URL
-        if c_doc.URL is not NULL:
-            # handle a bug in older libxml2 versions
-            tree.xmlFree(c_doc.URL)
-        if doc._c_doc.URL is not NULL:
-            c_doc.URL = tree.xmlStrdup(doc._c_doc.URL)
-        else:
-            doc_url_utf = "XSLT:__STRING__XSLT__%s" % id(self)
-            c_doc.URL = tree.xmlStrdup(_cstr(doc_url_utf))
-
-        self._xslt_resolver_context = _XSLTResolverContext(doc._parser)
-        # keep a copy in case we need to access the stylesheet via 'document()'
-        self._xslt_resolver_context._c_style_doc = tree.xmlCopyDoc(c_doc, 1)
-        c_doc._private = <python.PyObject*>self._xslt_resolver_context
-
-        c_style = xslt.xsltParseStylesheetDoc(c_doc)
-        if c_style is NULL:
-            tree.xmlFreeDoc(c_doc)
-            self._xslt_resolver_context._raise_if_stored()
-            raise XSLTParseError, "Cannot parse style sheet"
-        self._c_style = c_style
-
-        self._context = _XSLTContext(None, extensions)
-        self._error_log = _ErrorLog()
-        if regexp:
-            self._regexp  = _ExsltRegExp()
-        else:
-            self._regexp  = None
-        # XXX is it worthwile to use xsltPrecomputeStylesheet here?
-
-    def __dealloc__(self):
-        if self._xslt_resolver_context is not None and \
-               self._xslt_resolver_context._c_style_doc is not NULL:
-            tree.xmlFreeDoc(self._xslt_resolver_context._c_style_doc)
-        # this cleans up copy of doc as well
-        xslt.xsltFreeStylesheet(self._c_style)
-
-    property error_log:
-        def __get__(self):
-            return self._error_log.copy()
-
-    def __call__(self, _input, **_kw):
-        cdef _Document input_doc
-        cdef _NodeBase root_node
-        cdef _Document result_doc
-        cdef _XSLTResolverContext resolver_context
-        cdef xslt.xsltTransformContext* transform_ctxt
-        cdef xmlDoc* c_result
-        cdef xmlDoc* c_doc
-        cdef char** params
-        cdef void* ptemp
-        cdef int i
-
-        input_doc = _documentOrRaise(_input)
-        root_node = _rootNodeOf(_input)
-
-        resolver_context = _XSLTResolverContext(input_doc._parser)
-        resolver_context._c_style_doc = self._xslt_resolver_context._c_style_doc
-
-        c_doc = _fakeRootDoc(input_doc._c_doc, root_node._c_node)
-
-        transform_ctxt = xslt.xsltNewTransformContext(self._c_style, c_doc)
-        if transform_ctxt is NULL:
-            _destroyFakeDoc(input_doc._c_doc, c_doc)
-            raise XSLTApplyError, "Error preparing stylesheet run"
-
-        self._error_log.connect()
-        xslt.xsltSetTransformErrorFunc(transform_ctxt, <void*>self._error_log,
-                                       _receiveGenericError)
-
-        ptemp = c_doc._private
-        c_doc._private = <python.PyObject*>resolver_context
-
-        if _kw:
-            # allocate space for parameters
-            # * 2 as we want an entry for both key and value,
-            # and + 1 as array is NULL terminated
-            params = <char**>cstd.malloc(sizeof(char*) * (len(_kw) * 2 + 1))
-            i = 0
-            keep_ref = []
-            for key, value in _kw.items():
-                k = _utf8(key)
-                python.PyList_Append(keep_ref, k)
-                v = _utf8(value)
-                python.PyList_Append(keep_ref, v)
-                params[i] = _cstr(k)
-                i = i + 1
-                params[i] = _cstr(v)
-                i = i + 1
-            params[i] = NULL
-        else:
-            params = NULL
-
-        self._context.register_context(transform_ctxt, input_doc)
-        if self._regexp is not None:
-            self._regexp._register_in_context(self._context)
-
-        c_result = xslt.xsltApplyStylesheetUser(self._c_style, c_doc, params,
-                                                NULL, NULL, transform_ctxt)
-
-        if params is not NULL:
-            # deallocate space for parameters
-            cstd.free(params)
-
-        self._context.free_context()
-        c_doc._private = ptemp # restore _private before _destroyFakeDoc!
-        _destroyFakeDoc(input_doc._c_doc, c_doc)
-
-        self._error_log.disconnect()
-        if self._xslt_resolver_context._has_raised():
-            if c_result is not NULL:
-                tree.xmlFreeDoc(c_result)
-            self._xslt_resolver_context._raise_if_stored()
-
-        if c_result is NULL:
-            raise XSLTApplyError, "Error applying stylesheet"
-
-        result_doc = _documentFactory(c_result, input_doc._parser)
-        return _xsltResultTreeFactory(result_doc, self)
-
-    def apply(self, _input, **_kw):
-        return self.__call__(_input, **_kw)
-
-    def tostring(self, _ElementTree result_tree):
-        """Save result doc to string based on stylesheet output method.
-        """
-        return str(result_tree)
-
-cdef class _XSLTResultTree(_ElementTree):
-    cdef XSLT _xslt
-    def __str__(self):
-        cdef char* s
-        cdef int l
-        cdef int r
-        r = xslt.xsltSaveResultToString(&s, &l, self._doc._c_doc,
-                                        self._xslt._c_style)
-        if r == -1:
-            raise XSLTSaveError, "Error saving XSLT result to string"
-        if s is NULL:
-            return ''
-        result = funicode(s)
-        tree.xmlFree(s)
-        return result
-
-cdef _xsltResultTreeFactory(_Document doc, XSLT xslt):
-    cdef _XSLTResultTree result
-    result = <_XSLTResultTree>_newElementTree(doc, None, _XSLTResultTree)
-    result._xslt = xslt
-    return result
-
-# do not register all libxslt extra function, provide only "node-set"
-# functions like "output" and "write" are a potential security risk
-#xslt.xsltRegisterAllExtras()
-xslt.xsltRegisterExtModuleFunction("node-set",
-                                   xslt.XSLT_LIBXSLT_NAMESPACE,
-                                   xslt.xsltFunctionNodeSet)
-xslt.xsltRegisterExtModuleFunction("node-set",
-                                   xslt.XSLT_SAXON_NAMESPACE,
-                                   xslt.xsltFunctionNodeSet)
-xslt.xsltRegisterExtModuleFunction("node-set",
-                                   xslt.XSLT_XT_NAMESPACE,
-                                   xslt.xsltFunctionNodeSet)
-
-# enable EXSLT support for XSLT
-xslt.exsltRegisterAll()
-
-################################################################################
-# EXSLT regexp implementation
-
-cdef object RE_COMPILE
-RE_COMPILE = re.compile
-
-cdef class _ExsltRegExp:
-    cdef object _compile_map
-    def __init__(self):
-        self._compile_map = {}
-
-    cdef _make_string(self, value):
-        if python.PyString_Check(value) or python.PyUnicode_Check(value):
-            return value
-        else:
-            raise TypeError, "Invalid argument type %s" % type(value)
-
-    cdef _compile(self, rexp, ignore_case):
-        cdef python.PyObject* c_result
-        rexp = self._make_string(rexp)
-        key = (rexp, ignore_case)
-        c_result = python.PyDict_GetItem(self._compile_map, key)
-        if c_result is not NULL:
-            return <object>c_result
-        py_flags = re.UNICODE
-        if ignore_case:
-            py_flags = py_flags | re.IGNORECASE
-        rexp_compiled = RE_COMPILE(rexp, py_flags)
-        python.PyDict_SetItem(self._compile_map, key, rexp_compiled)
-        return rexp_compiled
-
-    def test(self, ctxt, s, rexp, flags=''):
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if rexpc.search(s) is None:
-            return False
-        else:
-            return True
-
-    def match(self, ctxt, s, rexp, flags=''):
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if 'g' in flags:
-            results = rexpc.findall(s)
-            if not results:
-                return ()
-            result_list = []
-            root = Element('matches')
-            for s_match in results:
-                elem = SubElement(root, 'match')
-                elem.text = s_match
-                python.PyList_Append(result_list, elem)
-            return result_list
-        else:
-            result = rexpc.search(s)
-            if result is None:
-                return ()
-            root = Element('match')
-            root.text = result.group()
-            return (root,)
-
-    def replace(self, ctxt, s, rexp, flags, replacement):
-        replacement = self._make_string(replacement)
-        flags = self._make_string(flags)
-        s = self._make_string(s)
-        rexpc = self._compile(rexp, 'i' in flags)
-        if 'g' in flags:
-            count = 0
-        else:
-            count = 1
-        return rexpc.sub(replacement, s, count)
-
-    cdef _register_in_context(self, _XSLTContext context):
-        ns = "http://exslt.org/regular-expressions"
-        context._registerLocalExtensionFunction(ns, "test",    self.test)
-        context._registerLocalExtensionFunction(ns, "match",   self.match)
-        context._registerLocalExtensionFunction(ns, "replace", self.replace)
-
-################################################################################
 # XPath
 
 cdef class _XPathContext(_BaseContext):
@@ -762,10 +150,6 @@
 
         return self._handle_result(xpathObj, doc)
 
-    #def clone(self):
-    #    # XXX pretty expensive so calling this from callback is probably
-    #    # not desirable
-    #    return XPathEvaluator(self._doc, self._namespaces, self._extensions)
 
 cdef class XPathDocumentEvaluator(XPathElementEvaluator):
     """Create an XPath evaluator for an ElementTree.
@@ -776,6 +160,7 @@
         XPathElementEvaluator.__init__(
             self, etree._context_node, namespaces, extensions)
 
+
 def XPathEvaluator(etree_or_element, namespaces=None, extensions=None):
     """Creates and XPath evaluator for an ElementTree or an Element.
 
@@ -786,11 +171,6 @@
     else:
         return XPathElementEvaluator(etree_or_element, namespaces, extensions)
 
-def Extension(module, function_mapping, ns_uri=None):
-    functions = []
-    for function_name, xpath_name in function_mapping.items():
-        functions[xpath_name] = getattr(module, function_name)
-    return {ns_uri : functions}
 
 cdef class XPath(XPathEvaluatorBase):
     cdef xpath.xmlXPathContext* _xpathCtxt
@@ -839,6 +219,7 @@
         if self._xpath is not NULL:
             xpath.xmlXPathFreeCompExpr(self._xpath)
 
+
 cdef object _replace_strings
 cdef object _find_namespaces
 _replace_strings = re.compile('("[^"]*")|(\'[^\']*\')').sub
@@ -869,148 +250,3 @@
                 # FIXME: this also replaces {namespaces} within strings!
                 path_utf   = path_utf.replace(namespace_def, prefix_str)
         return path_utf, namespaces
-
-################################################################################
-# helper functions
-
-cdef xpath.xmlXPathObject* _wrapXPathObject(object obj) except NULL:
-    cdef xpath.xmlNodeSet* resultSet
-    cdef _NodeBase node
-    if python.PyUnicode_Check(obj):
-        obj = _utf8(obj)
-    if python.PyString_Check(obj):
-        return xpath.xmlXPathNewCString(_cstr(obj))
-    if python.PyBool_Check(obj):
-        return xpath.xmlXPathNewBoolean(obj)
-    if python.PyNumber_Check(obj):
-        return xpath.xmlXPathNewFloat(obj)
-    if obj is None:
-        obj = ()
-    elif isinstance(obj, _NodeBase):
-        obj = (obj,)
-    if python.PySequence_Check(obj):
-        resultSet = xpath.xmlXPathNodeSetCreate(NULL)
-        for element in obj:
-            if isinstance(element, _NodeBase):
-                node = <_NodeBase>element
-                xpath.xmlXPathNodeSetAdd(resultSet, node._c_node)
-            else:
-                xpath.xmlXPathFreeNodeSet(resultSet)
-                raise XPathResultError, "This is not a node: %s" % element
-        return xpath.xmlXPathWrapNodeSet(resultSet)
-    else:
-        raise XPathResultError, "Unknown return type: %s" % obj
-    return NULL
-
-cdef object _unwrapXPathObject(xpath.xmlXPathObject* xpathObj,
-                               _Document doc):
-    if xpathObj.type == xpath.XPATH_UNDEFINED:
-        raise XPathResultError, "Undefined xpath result"
-    elif xpathObj.type == xpath.XPATH_NODESET:
-        return _createNodeSetResult(xpathObj, doc)
-    elif xpathObj.type == xpath.XPATH_BOOLEAN:
-        return bool(xpathObj.boolval)
-    elif xpathObj.type == xpath.XPATH_NUMBER:
-        return xpathObj.floatval
-    elif xpathObj.type == xpath.XPATH_STRING:
-        return funicode(xpathObj.stringval)
-    elif xpathObj.type == xpath.XPATH_POINT:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_RANGE:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_LOCATIONSET:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_USERS:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_XSLT_TREE:
-        raise NotImplementedError
-    else:
-        raise XPathResultError, "Unknown xpath result %s" % str(xpathObj.type)
-
-cdef object _createNodeSetResult(xpath.xmlXPathObject* xpathObj, _Document doc):
-    cdef xmlNode* c_node
-    cdef char* s
-    cdef _NodeBase element
-    cdef int i
-    result = []
-    if xpathObj.nodesetval is NULL:
-        return result
-    for i from 0 <= i < xpathObj.nodesetval.nodeNr:
-        c_node = xpathObj.nodesetval.nodeTab[i]
-        if _isElement(c_node):
-            if c_node.doc != doc._c_doc:
-                # XXX: works, but maybe not always the right thing to do?
-                # XPath: only runs when extensions create or copy trees
-                #        -> we store Python refs to these, so that is OK
-                # XSLT: can it leak when merging trees from multiple sources?
-                c_node = tree.xmlDocCopyNode(c_node, doc._c_doc, 1)
-            element = _elementFactory(doc, c_node)
-            result.append(element)
-        elif c_node.type == tree.XML_TEXT_NODE:
-            result.append(funicode(c_node.content))
-        elif c_node.type == tree.XML_ATTRIBUTE_NODE:
-            s = tree.xmlNodeGetContent(c_node)
-            attr_value = funicode(s)
-            tree.xmlFree(s)
-            result.append(attr_value)
-        else:
-            print "Not yet implemented result node type:", c_node.type
-            raise NotImplementedError
-    return result
-
-cdef void _freeXPathObject(xpath.xmlXPathObject* xpathObj):
-    """Free the XPath object, but *never* free the *content* of node sets.
-    Python dealloc will do that for us.
-    """
-    if xpathObj.nodesetval is not NULL:
-        xpath.xmlXPathFreeNodeSet(xpathObj.nodesetval)
-        xpathObj.nodesetval = NULL
-    xpath.xmlXPathFreeObject(xpathObj)
-
-cdef void _xpath_function_call(xpath.xmlXPathParserContext* ctxt, int nargs):
-    cdef xpath.xmlXPathContext* rctxt
-    cdef _BaseContext context
-    rctxt = ctxt.context
-    context = <_BaseContext>(rctxt.userData)
-    name = rctxt.function
-    if rctxt.functionURI is not NULL:
-        uri = rctxt.functionURI
-    else:
-        uri = None
-    if context._prepare_function_call(uri, name):
-        _extension_function_call(context, ctxt, nargs)
-    else:
-        xpath.xmlXPathErr(ctxt, xpath.XPATH_EXPR_ERROR)
-        exception = XPathFunctionError("XPath function {%s}%s not found" % (uri, name))
-        context._exc._store_exception(exception)
-
-cdef void _call_prepared_function(xpath.xmlXPathParserContext* ctxt, int nargs):
-    cdef xpath.xmlXPathContext* rctxt
-    cdef _BaseContext context
-    rctxt = ctxt.context
-    context = <_BaseContext>(rctxt.userData)
-    _extension_function_call(context, ctxt, nargs)
-
-cdef void _extension_function_call(_BaseContext context,
-                                   xpath.xmlXPathParserContext* ctxt, int nargs):
-    cdef _NodeBase node
-    cdef _Document doc
-    cdef xpath.xmlXPathObject* obj
-    cdef int i
-    doc = context._doc
-    try:
-        args = []
-        for i from 0 <= i < nargs:
-            o = _unwrapXPathObject(xpath.valuePop(ctxt), doc)
-            python.PyList_Append(args, o)
-        python.PyList_Reverse(args)
-
-        res = context._called_function(None, *args)
-        # wrap result for XPath consumption
-        obj = _wrapXPathObject(res)
-        # prevent Python from deallocating elements handed to libxml2
-        context._hold(res)
-        xpath.valuePush(ctxt, obj)
-    except:
-        xpath.xmlXPathErr(ctxt, xpath.XPATH_EXPR_ERROR)
-        context._exc._store_raised()

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 21:36:09 2006
@@ -15,198 +15,6 @@
 class XSLTExtensionError(XSLTError):
     pass
 
-class XPathError(LxmlError):
-    pass
-
-class XPathContextError(XPathError):
-    pass
-
-class XPathFunctionError(XPathError):
-    pass
-
-class XPathResultError(XPathError):
-    pass
-
-class XPathSyntaxError(LxmlSyntaxError):
-    pass
-
-################################################################################
-# support for extension functions in XPath/XSLT
-
-cdef class _BaseContext:
-    cdef xpath.xmlXPathContext* _xpathCtxt
-    cdef xpath.xmlXPathFuncLookupFunc _ext_lookup_function
-    cdef _Document _doc
-    cdef object _extensions
-    cdef object _namespaces
-    cdef object _registered_namespaces
-    cdef object _utf_refs
-    cdef object _function_cache
-    cdef object _called_function
-    # for exception handling and temporary reference keeping:
-    cdef _TempStore _temp_refs
-    cdef _ExceptionContext _exc
-
-    def __init__(self, namespaces, extensions):
-        self._xpathCtxt = NULL
-        self._utf_refs = {}
-        self._function_cache = {}
-        self._called_function = None
-
-        # convert old format extensions to UTF-8
-        if isinstance(extensions, (list, tuple)):
-            new_extensions = {}
-            for extension in extensions:
-                for (ns_uri, name), function in extension.items():
-                    ns_utf   = self._to_utf(ns_uri)
-                    name_utf = self._to_utf(name)
-                    try:
-                        new_extensions[ns_utf][name_utf] = function
-                    except KeyError:
-                        new_extensions[ns_utf] = {name_utf : function}
-            extensions = new_extensions or None
-
-        self._doc        = None
-        self._exc        = _ExceptionContext()
-        self._extensions = extensions
-        self._namespaces = namespaces
-        self._registered_namespaces = []
-        self._temp_refs = _TempStore()
-
-    cdef object _to_utf(self, s):
-        "Convert to UTF-8 and keep a reference to the encoded string"
-        cdef python.PyObject* dict_result
-        if s is None:
-            return None
-        dict_result = python.PyDict_GetItem(self._utf_refs, s)
-        if dict_result is not NULL:
-            return <object>dict_result
-        utf = _utf8(s)
-        python.PyDict_SetItem(self._utf_refs, s, utf)
-        return utf
-
-    cdef void _set_xpath_context(self, xpath.xmlXPathContext* xpathCtxt):
-        self._xpathCtxt = xpathCtxt
-        xpathCtxt.userData = <void*>self
-
-    cdef _register_context(self, _Document doc, int allow_none_namespace):
-        self._doc = doc
-        self._exc.clear()
-        python.PyDict_Clear(self._function_cache)
-        namespaces = self._namespaces
-        if namespaces is not None:
-            self.registerNamespaces(namespaces)
-        xpath.xmlXPathRegisterFuncLookup(
-            self._xpathCtxt, self._ext_lookup_function, <python.PyObject*>self)
-
-    cdef _unregister_context(self):
-        self._unregisterNamespaces()
-        self._free_context()
-
-    cdef _free_context(self):
-        del self._registered_namespaces[:]
-        python.PyDict_Clear(self._utf_refs)
-        self._doc = None
-        if self._xpathCtxt is not NULL:
-            self._xpathCtxt.userData = NULL
-            self._xpathCtxt = NULL
-
-    # namespaces (internal UTF-8 methods with leading '_')
-
-    def addNamespace(self, prefix, uri):
-        if self._namespaces is None:
-            self._namespaces = {}
-        python.PyDict_SetItem(self._namespaces, prefix, uri)
-
-    def registerNamespaces(self, namespaces):
-        for prefix, uri in namespaces.items():
-            self.registerNamespace(prefix, uri)
-    
-    def registerNamespace(self, prefix, ns_uri):
-        prefix_utf = self._to_utf(prefix)
-        ns_uri_utf = self._to_utf(ns_uri)
-        xpath.xmlXPathRegisterNs(self._xpathCtxt, prefix_utf, ns_uri_utf)
-        python.PyList_Append(self._registered_namespaces, prefix_utf)
-
-    cdef _unregisterNamespaces(self):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        xpathCtxt = self._xpathCtxt
-        for prefix_utf in self._registered_namespaces:
-            xpath.xmlXPathRegisterNs(xpathCtxt, prefix_utf, NULL)
-    
-    # extension functions
-
-    cdef int _prepare_function_call(self, ns_uri_utf, name_utf):
-        cdef python.PyObject* dict_result
-        key = (ns_uri_utf, name_utf)
-        dict_result = python.PyDict_GetItem(self._function_cache, key)
-        if dict_result is not NULL:
-            function = <object>dict_result
-            self._called_function = function
-            return function is not None
-
-        dict_result = python.PyDict_GetItem(self._extensions, ns_uri_utf)
-        if dict_result is not NULL:
-            dict_result = python.PyDict_GetItem(<object>dict_result, name_utf)
-        if dict_result is not NULL:
-            function = <object>dict_result
-        else:
-            function = _find_extension(ns_uri_utf, name_utf)
-
-        python.PyDict_SetItem(self._function_cache, key, function)
-        self._called_function = function
-        return function is not None
-
-    # Python reference keeping during XPath function evaluation
-
-    cdef _release_temp_refs(self):
-        "Free temporarily referenced objects from this context."
-        self._temp_refs.clear()
-        
-    cdef _hold(self, obj):
-        """A way to temporarily hold references to nodes in the evaluator.
-
-        This is needed because otherwise nodes created in XPath extension
-        functions would be reference counted too soon, during the XPath
-        evaluation.  This is most important in the case of exceptions.
-        """
-        cdef _NodeBase element
-        if isinstance(obj, _NodeBase):
-            obj = (obj,)
-        elif not python.PySequence_Check(obj):
-            return
-        for o in obj:
-            if isinstance(o, _NodeBase):
-                element = <_NodeBase>o
-                #print "Holding element:", <int>element._c_node
-                self._temp_refs.add(element)
-                #print "Holding document:", <int>element._doc._c_doc
-                self._temp_refs.add(element._doc)
-
-cdef xpath.xmlXPathFunction _function_check(void* ctxt,
-                                            char* c_name, char* c_ns_uri):
-    cdef _BaseContext context
-    if c_name is NULL:
-        return NULL
-    if c_ns_uri is NULL:
-        ns_uri = None
-    else:
-        ns_uri = c_ns_uri
-    context = <_BaseContext>ctxt
-    if context._prepare_function_call(ns_uri, c_name):
-        return _call_prepared_function
-    else:
-        return NULL
-
-cdef xpath.xmlXPathFunction _xslt_function_check(void* ctxt,
-                                                 char* c_name, char* c_ns_uri):
-    cdef xpath.xmlXPathFunction result
-    result = _function_check(ctxt, c_name, c_ns_uri)
-    if result is NULL:
-        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
-    else:
-        return result
-
 
 ################################################################################
 # XSLT document loaders
@@ -538,6 +346,16 @@
 # enable EXSLT support for XSLT
 xslt.exsltRegisterAll()
 
+cdef xpath.xmlXPathFunction _xslt_function_check(void* ctxt,
+                                                 char* c_name, char* c_ns_uri):
+    "Find XSLT extension function from set of XPath and XSLT functions"
+    cdef xpath.xmlXPathFunction result
+    result = _function_check(ctxt, c_name, c_ns_uri)
+    if result is NULL:
+        return xslt.xsltExtModuleFunctionLookup(c_name, c_ns_uri)
+    else:
+        return result
+
 ################################################################################
 # EXSLT regexp implementation
 
@@ -617,400 +435,3 @@
         context._registerLocalExtensionFunction(ns, "test",    self.test)
         context._registerLocalExtensionFunction(ns, "match",   self.match)
         context._registerLocalExtensionFunction(ns, "replace", self.replace)
-
-################################################################################
-# XPath
-
-cdef class _XPathContext(_BaseContext):
-    cdef object _variables
-    cdef object _registered_variables
-    def __init__(self, namespaces, extensions, variables):
-        self._ext_lookup_function = _function_check
-        self._variables = variables
-        self._registered_variables  = []
-        _BaseContext.__init__(self, namespaces, extensions)
-        
-    cdef register_context(self, xpath.xmlXPathContext* xpathCtxt, _Document doc):
-        self._set_xpath_context(xpathCtxt)
-        ns_prefixes = _find_all_extension_prefixes()
-        if ns_prefixes:
-            self.registerNamespaces(ns_prefixes)
-        self._register_context(doc, 1)
-        if self._variables is not None:
-            self.registerVariables(self._variables)
-
-    cdef unregister_context(self):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        xpathCtxt = self._xpathCtxt
-        if xpathCtxt is NULL:
-            return
-        self._unregisterVariables()
-        del self._registered_variables[:]
-        self._unregister_context()
-
-    cdef void _unregisterVariables(self):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject* xpathVarValue
-        cdef char* c_name
-        xpathCtxt = self._xpathCtxt
-        for name_utf in self._registered_variables:
-            c_name = _cstr(name_utf)
-            xpathVarValue = xpath.xmlXPathVariableLookup(xpathCtxt, c_name)
-            if xpathVarValue is not NULL:
-                xpath.xmlXPathRegisterVariable(xpathCtxt, c_name, NULL)
-                _freeXPathObject(xpathVarValue)
-
-    def registerVariables(self, variable_dict):
-        for name, value in variable_dict.items():
-            name_utf = self._to_utf(name)
-            self._registerVariable(name_utf, value)
-            python.PyList_Append(self._registered_variables, name_utf)
-
-    def registerVariable(self, name, value):
-        name_utf = self._to_utf(name)
-        self._registerVariable(name_utf, value)
-        python.PyList_Append(self._registered_variables, name_utf)
-
-    cdef void _registerVariable(self, name_utf, value):
-        xpath.xmlXPathRegisterVariable(
-            self._xpathCtxt, _cstr(name_utf), _wrapXPathObject(value))
-
-
-cdef class XPathEvaluatorBase:
-    cdef _XPathContext _context
-
-    def __init__(self, namespaces, extensions, variables=None):
-        self._context = _XPathContext(namespaces, extensions, variables)
-
-    cdef object _handle_result(self, xpath.xmlXPathObject* xpathObj, _Document doc):
-        if self._context._exc._has_raised():
-            if xpathObj is not NULL:
-                _freeXPathObject(xpathObj)
-                xpathObj = NULL
-            self._context._release_temp_refs()
-            self._context._exc._raise_if_stored()
-
-        if xpathObj is NULL:
-            self._context._release_temp_refs()
-            raise XPathSyntaxError, "Error in xpath expression."
-
-        try:
-            result = _unwrapXPathObject(xpathObj, doc)
-        except XPathResultError:
-            _freeXPathObject(xpathObj)
-            self._context._release_temp_refs()
-            raise
-
-        _freeXPathObject(xpathObj)
-        self._context._release_temp_refs()
-        return result
-
-
-cdef class XPathElementEvaluator(XPathEvaluatorBase):
-    """Create an XPath evaluator for an element.
-
-    XPath evaluators must not be shared between threads.
-    """
-    cdef xpath.xmlXPathContext* _c_ctxt
-    cdef _Element _element
-    def __init__(self, _NodeBase element not None, namespaces=None, extensions=None):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef int ns_register_status
-        cdef _Document doc
-        doc = element._doc
-        xpathCtxt = xpath.xmlXPathNewContext(doc._c_doc)
-        if xpathCtxt is NULL:
-            raise XPathContextError, "Unable to create new XPath context"
-        self._element = element
-        self._c_ctxt = xpathCtxt
-        XPathEvaluatorBase.__init__(self, namespaces, extensions)
-
-    def __dealloc__(self):
-        if self._c_ctxt is not NULL:
-            xpath.xmlXPathFreeContext(self._c_ctxt)
-    
-    def registerNamespace(self, prefix, uri):
-        """Register a namespace with the XPath context.
-        """
-        self._context.addNamespace(prefix, uri)
-
-    def registerNamespaces(self, namespaces):
-        """Register a prefix -> uri dict.
-        """
-        add = self._context.addNamespace
-        for prefix, uri in namespaces.items():
-            add(prefix, uri)
-    
-    def evaluate(self, _path, **_variables):
-        """Evaluate an XPath expression on the document.  Variables may be
-        provided as keyword arguments. Note that namespaces are currently not
-        supported for variables."""
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject*  xpathObj
-        cdef xmlNode* c_node
-        cdef _Document doc
-        xpathCtxt = self._c_ctxt
-        xpathCtxt.node = self._element._c_node
-        doc = self._element._doc
-
-        self._context.register_context(xpathCtxt, doc)
-        self._context.registerVariables(_variables)
-
-        path = _utf8(_path)
-        xpathObj = xpath.xmlXPathEvalExpression(_cstr(path), xpathCtxt)
-        self._context.unregister_context()
-
-        return self._handle_result(xpathObj, doc)
-
-    #def clone(self):
-    #    # XXX pretty expensive so calling this from callback is probably
-    #    # not desirable
-    #    return XPathEvaluator(self._doc, self._namespaces, self._extensions)
-
-cdef class XPathDocumentEvaluator(XPathElementEvaluator):
-    """Create an XPath evaluator for an ElementTree.
-
-    XPath evaluators must not be shared between threads.
-    """
-    def __init__(self, _ElementTree etree not None, namespaces=None, extensions=None):
-        XPathElementEvaluator.__init__(
-            self, etree._context_node, namespaces, extensions)
-
-def XPathEvaluator(etree_or_element, namespaces=None, extensions=None):
-    """Creates and XPath evaluator for an ElementTree or an Element.
-
-    XPath evaluators must not be shared between threads.
-    """
-    if isinstance(etree_or_element, _ElementTree):
-        return XPathDocumentEvaluator(etree_or_element, namespaces, extensions)
-    else:
-        return XPathElementEvaluator(etree_or_element, namespaces, extensions)
-
-def Extension(module, function_mapping, ns_uri=None):
-    functions = []
-    for function_name, xpath_name in function_mapping.items():
-        functions[xpath_name] = getattr(module, function_name)
-    return {ns_uri : functions}
-
-cdef class XPath(XPathEvaluatorBase):
-    cdef xpath.xmlXPathContext* _xpathCtxt
-    cdef xpath.xmlXPathCompExpr* _xpath
-    cdef object _prefix_map
-    cdef readonly object path
-
-    def __init__(self, path, namespaces=None, extensions=None):
-        XPathEvaluatorBase.__init__(self, namespaces, extensions, None)
-        self.path = path
-        path = _utf8(path)
-        self._xpath = xpath.xmlXPathCompile(_cstr(path))
-        if self._xpath is NULL:
-            raise XPathSyntaxError, "Error in XPath expression"
-        self._xpathCtxt = xpath.xmlXPathNewContext(NULL)
-
-    def __call__(self, _etree_or_element, **_variables):
-        cdef xpath.xmlXPathContext* xpathCtxt
-        cdef xpath.xmlXPathObject*  xpathObj
-        cdef _Document document
-        cdef _NodeBase element
-        cdef _XPathContext context
-
-        document = _documentOrRaise(_etree_or_element)
-        element  = _rootNodeOf(_etree_or_element)
-
-        xpathCtxt = self._xpathCtxt
-        xpathCtxt.doc = document._c_doc
-        xpathCtxt.node = element._c_node
-
-        context = self._context
-        context._release_temp_refs()
-        context.register_context(xpathCtxt, document)
-        context.registerVariables(_variables)
-
-        xpathObj = xpath.xmlXPathCompiledEval(self._xpath, xpathCtxt)
-        context.unregister_context()
-        return self._handle_result(xpathObj, document)
-
-    def evaluate(self, _tree, **_variables):
-        return self(_tree, **_variables)
-
-    def __dealloc__(self):
-        if self._xpathCtxt is not NULL:
-            xpath.xmlXPathFreeContext(self._xpathCtxt)
-        if self._xpath is not NULL:
-            xpath.xmlXPathFreeCompExpr(self._xpath)
-
-cdef object _replace_strings
-cdef object _find_namespaces
-_replace_strings = re.compile('("[^"]*")|(\'[^\']*\')').sub
-_find_namespaces = re.compile('({[^}]+})').findall
-
-cdef class ETXPath(XPath):
-    """Special XPath class that supports the ElementTree {uri} notation for
-    namespaces."""
-    def __init__(self, path, extensions=None):
-        path_utf, namespaces = self._nsextract_path(_utf8(path))
-        XPath.__init__(self, funicode(path_utf), namespaces, extensions)
-
-    cdef _nsextract_path(self, path_utf):
-        # replace {namespaces} by new prefixes
-        cdef int i
-        namespaces = {}
-        stripped_path = _replace_strings('', path_utf) # remove string literals
-        namespace_defs = []
-        i = 1
-        for namespace_def in _find_namespaces(stripped_path):
-            if namespace_def not in namespace_defs:
-                prefix = python.PyString_FromFormat("xpp%02d", i)
-                i = i+1
-                python.PyList_Append(namespace_defs, namespace_def)
-                namespace = namespace_def[1:-1] # remove '{}'
-                python.PyDict_SetItem(namespaces, prefix, namespace)
-                prefix_str = prefix + ':'
-                # FIXME: this also replaces {namespaces} within strings!
-                path_utf   = path_utf.replace(namespace_def, prefix_str)
-        return path_utf, namespaces
-
-################################################################################
-# helper functions
-
-cdef xpath.xmlXPathObject* _wrapXPathObject(object obj) except NULL:
-    cdef xpath.xmlNodeSet* resultSet
-    cdef _NodeBase node
-    if python.PyUnicode_Check(obj):
-        obj = _utf8(obj)
-    if python.PyString_Check(obj):
-        return xpath.xmlXPathNewCString(_cstr(obj))
-    if python.PyBool_Check(obj):
-        return xpath.xmlXPathNewBoolean(obj)
-    if python.PyNumber_Check(obj):
-        return xpath.xmlXPathNewFloat(obj)
-    if obj is None:
-        obj = ()
-    elif isinstance(obj, _NodeBase):
-        obj = (obj,)
-    if python.PySequence_Check(obj):
-        resultSet = xpath.xmlXPathNodeSetCreate(NULL)
-        for element in obj:
-            if isinstance(element, _NodeBase):
-                node = <_NodeBase>element
-                xpath.xmlXPathNodeSetAdd(resultSet, node._c_node)
-            else:
-                xpath.xmlXPathFreeNodeSet(resultSet)
-                raise XPathResultError, "This is not a node: %s" % element
-        return xpath.xmlXPathWrapNodeSet(resultSet)
-    else:
-        raise XPathResultError, "Unknown return type: %s" % obj
-    return NULL
-
-cdef object _unwrapXPathObject(xpath.xmlXPathObject* xpathObj,
-                               _Document doc):
-    if xpathObj.type == xpath.XPATH_UNDEFINED:
-        raise XPathResultError, "Undefined xpath result"
-    elif xpathObj.type == xpath.XPATH_NODESET:
-        return _createNodeSetResult(xpathObj, doc)
-    elif xpathObj.type == xpath.XPATH_BOOLEAN:
-        return bool(xpathObj.boolval)
-    elif xpathObj.type == xpath.XPATH_NUMBER:
-        return xpathObj.floatval
-    elif xpathObj.type == xpath.XPATH_STRING:
-        return funicode(xpathObj.stringval)
-    elif xpathObj.type == xpath.XPATH_POINT:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_RANGE:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_LOCATIONSET:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_USERS:
-        raise NotImplementedError
-    elif xpathObj.type == xpath.XPATH_XSLT_TREE:
-        raise NotImplementedError
-    else:
-        raise XPathResultError, "Unknown xpath result %s" % str(xpathObj.type)
-
-cdef object _createNodeSetResult(xpath.xmlXPathObject* xpathObj, _Document doc):
-    cdef xmlNode* c_node
-    cdef char* s
-    cdef _NodeBase element
-    cdef int i
-    result = []
-    if xpathObj.nodesetval is NULL:
-        return result
-    for i from 0 <= i < xpathObj.nodesetval.nodeNr:
-        c_node = xpathObj.nodesetval.nodeTab[i]
-        if _isElement(c_node):
-            if c_node.doc != doc._c_doc:
-                # XXX: works, but maybe not always the right thing to do?
-                # XPath: only runs when extensions create or copy trees
-                #        -> we store Python refs to these, so that is OK
-                # XSLT: can it leak when merging trees from multiple sources?
-                c_node = tree.xmlDocCopyNode(c_node, doc._c_doc, 1)
-            element = _elementFactory(doc, c_node)
-            result.append(element)
-        elif c_node.type == tree.XML_TEXT_NODE:
-            result.append(funicode(c_node.content))
-        elif c_node.type == tree.XML_ATTRIBUTE_NODE:
-            s = tree.xmlNodeGetContent(c_node)
-            attr_value = funicode(s)
-            tree.xmlFree(s)
-            result.append(attr_value)
-        else:
-            print "Not yet implemented result node type:", c_node.type
-            raise NotImplementedError
-    return result
-
-cdef void _freeXPathObject(xpath.xmlXPathObject* xpathObj):
-    """Free the XPath object, but *never* free the *content* of node sets.
-    Python dealloc will do that for us.
-    """
-    if xpathObj.nodesetval is not NULL:
-        xpath.xmlXPathFreeNodeSet(xpathObj.nodesetval)
-        xpathObj.nodesetval = NULL
-    xpath.xmlXPathFreeObject(xpathObj)
-
-cdef void _xpath_function_call(xpath.xmlXPathParserContext* ctxt, int nargs):
-    cdef xpath.xmlXPathContext* rctxt
-    cdef _BaseContext context
-    rctxt = ctxt.context
-    context = <_BaseContext>(rctxt.userData)
-    name = rctxt.function
-    if rctxt.functionURI is not NULL:
-        uri = rctxt.functionURI
-    else:
-        uri = None
-    if context._prepare_function_call(uri, name):
-        _extension_function_call(context, ctxt, nargs)
-    else:
-        xpath.xmlXPathErr(ctxt, xpath.XPATH_EXPR_ERROR)
-        exception = XPathFunctionError("XPath function {%s}%s not found" % (uri, name))
-        context._exc._store_exception(exception)
-
-cdef void _call_prepared_function(xpath.xmlXPathParserContext* ctxt, int nargs):
-    cdef xpath.xmlXPathContext* rctxt
-    cdef _BaseContext context
-    rctxt = ctxt.context
-    context = <_BaseContext>(rctxt.userData)
-    _extension_function_call(context, ctxt, nargs)
-
-cdef void _extension_function_call(_BaseContext context,
-                                   xpath.xmlXPathParserContext* ctxt, int nargs):
-    cdef _NodeBase node
-    cdef _Document doc
-    cdef xpath.xmlXPathObject* obj
-    cdef int i
-    doc = context._doc
-    try:
-        args = []
-        for i from 0 <= i < nargs:
-            o = _unwrapXPathObject(xpath.valuePop(ctxt), doc)
-            python.PyList_Append(args, o)
-        python.PyList_Reverse(args)
-
-        res = context._called_function(None, *args)
-        # wrap result for XPath consumption
-        obj = _wrapXPathObject(res)
-        # prevent Python from deallocating elements handed to libxml2
-        context._hold(res)
-        xpath.valuePush(ctxt, obj)
-    except:
-        xpath.xmlXPathErr(ctxt, xpath.XPATH_EXPR_ERROR)
-        context._exc._store_raised()
From scoder at codespeak.net  Fri Apr 28 21:41:18 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:41:19 2006
Subject: [Lxml-checkins] r26539 - lxml/trunk/src/lxml
Message-ID: <20060428194118.D9EEB10094@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:41:17 2006
New Revision: 26539

Added:
   lxml/trunk/src/lxml/apihelpers.pxi
      - copied unchanged from r26538, lxml/trunk/src/lxml/etree.pyx
Log:
new source file apihelpers.pxi to move helper functions out of API source file etree.pyx

From scoder at codespeak.net  Fri Apr 28 21:56:37 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Fri Apr 28 21:56:39 2006
Subject: [Lxml-checkins] r26540 - lxml/trunk/src/lxml
Message-ID: <20060428195637.BC25E10083@code0.codespeak.net>

Author: scoder
Date: Fri Apr 28 21:56:36 2006
New Revision: 26540

Modified:
   lxml/trunk/src/lxml/apihelpers.pxi
   lxml/trunk/src/lxml/etree.pyx
   lxml/trunk/src/lxml/extensions.pxi
   lxml/trunk/src/lxml/parser.pxi
   lxml/trunk/src/lxml/xmlerror.pxi
   lxml/trunk/src/lxml/xpath.pxi
   lxml/trunk/src/lxml/xslt.pxi
Log:
moved helper functions from etree.pyx to apihelpers.pxi, some cleanup

Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Fri Apr 28 21:56:36 2006
@@ -1,1403 +1,13 @@
-cimport tree, python
-from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
-from python cimport isinstance, issubclass, hasattr, callable
-from python cimport iter, str, _cstr
-cimport xpath
-cimport xslt
-cimport xmlerror
-cimport xinclude
-cimport c14n
-cimport cstd
-import re
-
-import _elementpath
-from StringIO import StringIO
-import sys
-
-# the rules
-# any libxml C argument/variable is prefixed with c_
-# any non-public function/class is prefixed with an underscore
-# instance creation is always through factories
-
-ctypedef enum LXML_PROXY_TYPE:
-    PROXY_ELEMENT
-    PROXY_ATTRIB
-
-# what to do with libxml2/libxslt error messages?
-# 0 : drop
-# 1 : use log
-cdef int __DEBUG
-__DEBUG = 1
-
-# maximum number of lines in the libxml2/xslt log if __DEBUG == 1
-cdef int __MAX_LOG_SIZE
-__MAX_LOG_SIZE = 100
-
-# make the compiled-in debug state publicly available
-DEBUG = __DEBUG
-
-# Error superclass for ElementTree compatibility
-class Error(Exception):
-    pass
-
-# module level superclass for all exceptions
-class LxmlError(Error):
-    def __init__(self, *args):
-        Error.__init__(self, *args)
-        self.error_log = __copyGlobalErrorLog()
-
-# superclass for all syntax errors
-class LxmlSyntaxError(LxmlError, SyntaxError):
-    pass
-
-class DocumentInvalid(LxmlError):
-    pass
-
-class XIncludeError(LxmlError):
-    pass
-
-class C14NError(LxmlError):
-    pass
-
-
-# class for temporary storage of Python references
-cdef class _TempStore:
-    cdef object _storage
-    def __init__(self):
-        self._storage = {}
-
-    cdef void add(self, obj):
-        python.PyDict_SetItem(self._storage, id(obj), obj)
-
-    cdef void clear(self):
-        python.PyDict_Clear(self._storage)
-
-    cdef object dictcopy(self):
-        return self._storage.copy()
-
-# class for temporarily storing exceptions raised in extensions
-cdef class _ExceptionContext:
-    cdef object _exc_info
-    def __init__(self):
-        self._exc_info = None
-
-    cdef void clear(self):
-        self._exc_info = None
-
-    cdef void _store_raised(self):
-        self._exc_info = sys.exc_info()
-
-    cdef void _store_exception(self, exception):
-        self._exc_info = (exception, None, None)
-
-    cdef _has_raised(self):
-        return self._exc_info is not None
-
-    cdef _raise_if_stored(self):
-        _exc_info = self._exc_info
-        if _exc_info is not None:
-            self._exc_info = None
-            type, value, traceback = _exc_info
-            if traceback is None and value is None:
-                raise type
-            else:
-                raise type, value, traceback
-
-
-cdef class BaseParser # forward declaration
-
-cdef class _Document:
-    """Internal base class to reference a libxml document.
-
-    When instances of this class are garbage collected, the libxml
-    document is cleaned up.
-    """
-    cdef int _ns_counter
-    cdef xmlDoc* _c_doc
-    cdef BaseParser _parser
-    
-    def __dealloc__(self):
-        # if there are no more references to the document, it is safe
-        # to clean the whole thing up, as all nodes have a reference to
-        # the document
-        #print "freeing document:", <int>self._c_doc
-        #displayNode(<xmlNode*>self._c_doc, 0)
-        #print <int>self._c_doc, self._c_doc.dict is __GLOBAL_PARSER_CONTEXT._c_dict
-        tree.xmlFreeDoc(self._c_doc)
-
-    cdef getroot(self):
-        cdef xmlNode* c_node
-        c_node = tree.xmlDocGetRootElement(self._c_doc)
-        if c_node is NULL:
-            return None
-        return _elementFactory(self, c_node)
-
-    cdef buildNewPrefix(self):
-        ns = python.PyString_FromFormat("ns%d", self._ns_counter)
-        self._ns_counter = self._ns_counter + 1
-        return ns
-
-    cdef xmlNs* _findOrBuildNodeNs(self, xmlNode* c_node, char* href):
-        """Get or create namespace structure for a node.
-        """
-        cdef xmlNs* c_ns
-        # look for existing ns
-        c_ns = tree.xmlSearchNsByHref(self._c_doc, c_node, href)
-        if c_ns is not NULL:
-            return c_ns
-        # create ns if existing ns cannot be found
-        # try to simulate ElementTree's namespace prefix creation
-        prefix = self.buildNewPrefix()
-        c_ns = tree.xmlNewNs(c_node, href, _cstr(prefix))
-        return c_ns
-
-    cdef void _setNodeNs(self, xmlNode* c_node, char* href):
-        "Lookup namespace structure and set it for the node."
-        cdef xmlNs* c_ns
-        c_ns = self._findOrBuildNodeNs(c_node, href)
-        tree.xmlSetNs(c_node, c_ns)
-
-    cdef void _setNodeNamespaces(self, xmlNode* c_node,
-                                 object node_ns_utf, object nsmap):
-        """Lookup current namespace prefixes, then set namespace structure for
-        node and register new ns-prefix mappings.
-        """
-        cdef xmlNs*  c_ns
-        cdef xmlDoc* c_doc
-        cdef char*   c_prefix
-        cdef char*   c_href
-        if not nsmap:
-            if node_ns_utf is not None:
-                self._setNodeNs(c_node, node_ns_utf)
-            return
-
-        c_doc  = self._c_doc
-        for prefix, href in nsmap.items():
-            href_utf = _utf8(href)
-            c_href = _cstr(href_utf)
-            if prefix is not None:
-                prefix_utf = _utf8(prefix)
-                c_prefix = _cstr(prefix_utf)
-            else:
-                c_prefix = NULL
-            # add namespace with prefix if ns is not already known
-            c_ns = tree.xmlSearchNsByHref(c_doc, c_node, c_href)
-            if c_ns is NULL:
-                c_ns = tree.xmlNewNs(c_node, c_href, c_prefix)
-            if href_utf == node_ns_utf:
-                tree.xmlSetNs(c_node, c_ns)
-                node_ns_utf = None
-
-        if node_ns_utf is not None:
-            self._setNodeNs(c_node, node_ns_utf)
-
-cdef _Document _parseDocument(source, parser):
-    cdef xmlDoc* c_doc
-    filename = _getFilenameForFile(source)
-    # Support for unamed file-like object (StringIO, urlgrabber.urlopen, ...)
-    if not filename and hasattr(source, 'read'):
-        return _parseMemoryDocument(source.read(), parser)
-
-    # Otherwise parse the file directly from the filesystem
-    if filename is None:
-        filename = source
-    # open filename
-    c_doc = _parseDocFromFile(_utf8(filename), parser)
-    return _documentFactory(c_doc, parser)
-
-cdef _Document _parseMemoryDocument(text, parser):
-    cdef xmlDoc* c_doc
-    if python.PyUnicode_Check(text):
-        text = _stripDeclaration(_utf8(text))
-    c_doc = _parseDoc(text, parser)
-    return _documentFactory(c_doc, parser)
-
-cdef _Document _documentFactory(xmlDoc* c_doc, parser):
-    cdef _Document result
-    result = _Document()
-    result._c_doc = c_doc
-    result._ns_counter = 0
-    if parser is None:
-        parser = __DEFAULT_PARSER
-    result._parser = parser.copy()
-    return result
+# Private helper functions
 
-# to help with debugging
 cdef void displayNode(xmlNode* c_node, indent):
+    # to help with debugging
     cdef xmlNode* c_child
     print indent * ' ', <int>c_node
     c_child = c_node.children
     while c_child is not NULL:
         displayNode(c_child, indent + 1)
         c_child = c_child.next
-        
-cdef class _NodeBase:
-    """Base class to reference a document object and a libxml node.
-
-    By pointing to a Document instance, a reference is kept to
-    _Document as long as there is some pointer to a node in it.
-    """
-    cdef _Document _doc
-    cdef xmlNode* _c_node
-    cdef int _proxy_type
-    
-    def __dealloc__(self):
-        #print "trying to free node:", <int>self._c_node
-        #displayNode(self._c_node, 0)
-        if self._c_node is not NULL:
-            unregisterProxy(self)
-            attemptDeallocation(self._c_node)
-
-    def _init(self):
-        """Called after object initialisation. Subclasses may override
-        this if they recursively call _init() in the superclasses.
-        """
-
-cdef class _ElementTree:
-    cdef _Document _doc
-    cdef _NodeBase _context_node
-
-    def parse(self, source, parser=None):
-        """Updates self with the content of source and returns its root
-        """
-        self._doc          = _parseDocument(source, parser)
-        self._context_node = self._doc.getroot()
-        return self._context_node
-    
-    def getroot(self):
-        return self._context_node
-    
-    def write(self, file, encoding='us-ascii'):
-        if not hasattr(file, 'write'):
-            # file is a filename, we want a file object
-            file = open(file, 'wb')
-
-        m = tostring(self._context_node, encoding)
-        # XXX this is purely for ElementTree compatibility..
-        if encoding == 'UTF-8' or encoding == 'us-ascii':
-            m = _stripDeclaration(m)
-            if m[-1:] == '\n':
-                m = m[:-1]
-        file.write(m)
-
-    def getiterator(self, tag=None):
-        root = self.getroot()
-        if root is None:
-            return ()
-        return root.getiterator(tag)
-
-    def find(self, path):
-        root = self.getroot()
-        assert root is not None
-        if path[:1] == "/":
-            path = "." + path
-        return root.find(path)
-
-    def findtext(self, path, default=None):
-        root = self.getroot()
-        assert root is not None
-        if path[:1] == "/":
-            path = "." + path
-        return root.findtext(path, default)
-
-    def findall(self, path):
-        root = self.getroot()
-        assert root is not None
-        if path[:1] == "/":
-            path = "." + path
-        return root.findall(path)
-    
-    # extensions to ElementTree API
-    def xpath(self, _path, namespaces=None, **_variables):
-        """XPath evaluate in context of document.
-
-        namespaces is an optional dictionary with prefix to namespace URI
-        mappings, used by XPath.
-        
-        Returns a list (nodeset), or bool, float or string.
-
-        In case of a list result, return Element for element nodes,
-        string for text and attribute values.
-
-        Note: if you are going to apply multiple XPath expressions
-        against the same document, it is more efficient to use
-        XPathEvaluator directly.
-        """
-        evaluator = XPathElementEvaluator(self._context_node, namespaces)
-        return evaluator.evaluate(_path, **_variables)
-
-    def xslt(self, _xslt, extensions=None, **_kw):
-        """Transform this document using other document.
-
-        xslt is a tree that should be XSLT
-        keyword parameters are XSLT transformation parameters.
-
-        Returns the transformed tree.
-
-        Note: if you are going to apply the same XSLT stylesheet against
-        multiple documents, it is more efficient to use the XSLT
-        class directly.
-        """
-        style = XSLT(_xslt, extensions)
-        return style(self, **_kw)
-
-    def relaxng(self, relaxng):
-        """Validate this document using other document.
-
-        relaxng is a tree that should contain Relax NG XML
-
-        Returns True or False, depending on whether validation
-        succeeded.
-
-        Note: if you are going to apply the same Relax NG schema against
-        multiple documents, it is more efficient to use the RelaxNG
-        class directly.
-        """
-        schema = RelaxNG(relaxng)
-        return schema.validate(self)
-
-    def xmlschema(self, xmlschema):
-        """Validate this document using other doucment.
-
-        xmlschema is a tree that should contain XML Schema XML.
-
-        Returns True or False, depending on whether validation
-        succeeded.
-
-        Note: If you are going to applyt he same XML Schema against
-        multiple documents, it is more efficient to use the XMLSchema
-        class directly.
-        """
-        schema = XMLSchema(xmlschema)
-        return schema.validate(self)
-        
-    def xinclude(self):
-        """Process this document, including using XInclude.
-        """
-        cdef int result
-        # XXX what happens memory-wise with the original XInclude nodes?
-        # they seem to be still accessible if a reference to them has
-        # been made previously, but I have no idea whether they get freed
-        # at all. The XInclude nodes appear to be still being in the same
-        # parent and same document, but they must not be connected to the
-        # tree..
-        result = xinclude.xmlXIncludeProcessTree(self._context_node._c_node)
-        if result == -1:
-            raise XIncludeError, "XInclude processing failed"
-        
-    def write_c14n(self, file):
-        """C14N write of document. Always writes UTF-8.
-        """
-        cdef xmlDoc* c_base_doc
-        cdef xmlDoc* c_doc
-        cdef char* data
-        cdef int bytes
-        c_base_doc = self._doc._c_doc
-
-        c_doc = _fakeRootDoc(c_base_doc, self._context_node._c_node)
-        bytes = c14n.xmlC14NDocDumpMemory(c_doc, NULL, 0, NULL, 1, &data)
-        _destroyFakeDoc(c_base_doc, c_doc)
-
-        if bytes < 0:
-            raise C14NError, "C14N failed"
-        if not hasattr(file, 'write'):
-            file = open(file, 'wb')
-        file.write(data)
-        tree.xmlFree(data)
-    
-cdef _ElementTree _elementTreeFactory(_Document doc,
-                                      _NodeBase context_node):
-    return _newElementTree(doc, context_node, _ElementTree)
-
-cdef _ElementTree _newElementTree(_Document doc, _NodeBase context_node,
-                                  object baseclass):
-    cdef _ElementTree result
-    result = baseclass()
-    result._doc = doc
-    if context_node is None and doc is not None:
-        context_node = doc.getroot()
-    result._context_node = context_node
-    return result
-
-cdef class _Element(_NodeBase):
-    cdef object _tag
-
-    # MANIPULATORS
-
-    def __setitem__(self, index, _NodeBase element):
-        cdef xmlNode* c_node
-        cdef xmlNode* c_next
-        cdef int foreign
-        c_node = _findChild(self._c_node, index)
-        if c_node is NULL:
-            raise IndexError
-        foreign = self._doc is not element._doc
-        c_next = element._c_node.next
-        _removeText(c_node.next)
-        tree.xmlReplaceNode(c_node, element._c_node)
-        _moveTail(c_next, element._c_node)
-        changeDocumentBelow(element, self._doc, foreign)
-        
-    def __delitem__(self, index):
-        cdef xmlNode* c_node
-        c_node = _findChild(self._c_node, index)
-        if c_node is NULL:
-            raise IndexError
-        _removeText(c_node.next)
-        _removeNode(c_node)
-
-    def __delslice__(self, start, stop):
-        cdef xmlNode* c_node
-        c_node = _findChild(self._c_node, start)
-        _deleteSlice(c_node, start, stop)
-        
-    def __setslice__(self, start, stop, value):
-        cdef xmlNode* c_node
-        cdef xmlNode* c_next
-        cdef _Element mynode
-        cdef int foreign
-        # first, find start of slice
-        c_node = _findChild(self._c_node, start)
-        # now delete the slice
-        if start != stop:
-            c_node = _deleteSlice(c_node, start, stop)
-        # if the insertion point is at the end, append there
-        if c_node is NULL:
-            append = self.append
-            for node in value:
-                append(node)
-            return
-        # if the next element is in the list, insert before it
-        for mynode in value:
-            if mynode is None:
-                raise TypeError, "Node must not be None."
-            foreign = self._doc is not mynode._doc
-            # store possible text tail
-            c_next = mynode._c_node.next
-            # now move node previous to insertion point
-            tree.xmlUnlinkNode(mynode._c_node)
-            tree.xmlAddPrevSibling(c_node, mynode._c_node)
-            # and move tail just behind his node
-            _moveTail(c_next, mynode._c_node)
-            # move it into a new document
-            changeDocumentBelow(mynode, self._doc, foreign)
-
-    def __deepcopy__(self, memo):
-        return self.__copy__()
-        
-    def __copy__(self):
-        cdef xmlNode* c_node
-        cdef xmlDoc* c_doc
-        cdef xmlDoc* fake_c_doc
-        cdef _Document doc
-        doc = self._doc
-        fake_c_doc = _fakeRootDoc(doc._c_doc, self._c_node)
-        c_doc = tree.xmlCopyDoc(fake_c_doc, 1) # recursive copy
-        _destroyFakeDoc(doc._c_doc, fake_c_doc)
-        doc = _documentFactory(c_doc, doc._parser)
-        return doc.getroot()
-        
-    def set(self, key, value):
-        self.attrib[key] = value
-        
-    def append(self, _Element element not None):
-        cdef xmlNode* c_next
-        cdef xmlNode* c_node
-        cdef int foreign
-        foreign = self._doc is not element._doc
-        c_node = element._c_node
-        # store possible text node
-        c_next = c_node.next
-        # XXX what if element is coming from a different document?
-        tree.xmlUnlinkNode(c_node)
-        # move node itself
-        tree.xmlAddChild(self._c_node, c_node)
-        _moveTail(c_next, c_node)
-        # uh oh, elements may be pointing to different doc when
-        # parent element has moved; change them too..
-        changeDocumentBelow(element, self._doc, foreign)
-
-    def clear(self):
-        cdef xmlAttr* c_attr
-        cdef xmlAttr* c_attr_next
-        cdef xmlNode* c_node
-        cdef xmlNode* c_node_next
-        c_node = self._c_node
-        # remove self.text and self.tail
-        _removeText(c_node.children)
-        _removeText(c_node.next)
-        # remove all attributes
-        c_attr = c_node.properties
-        while c_attr is not NULL:
-            c_attr_next = c_attr.next
-            tree.xmlRemoveProp(c_attr)
-            c_attr = c_attr_next
-        # remove all subelements
-        c_node = c_node.children
-        while c_node is not NULL:
-            c_node_next = c_node.next
-            if _isElement(c_node):
-                _removeText(c_node_next)
-                c_node_next = c_node.next
-                _removeNode(c_node)
-            c_node = c_node_next
-    
-    def insert(self, index, _Element element not None):
-        cdef xmlNode* c_node
-        cdef xmlNode* c_next
-        cdef int foreign
-        c_node = _findChild(self._c_node, index)
-        if c_node is NULL:
-            self.append(element)
-            return
-        foreign = self._doc is not element._doc
-        c_next = element._c_node.next
-        tree.xmlAddPrevSibling(c_node, element._c_node)
-        _moveTail(c_next, element._c_node)
-        changeDocumentBelow(element, self._doc, foreign)
-
-    def remove(self, _Element element not None):
-        cdef xmlNode* c_node
-        c_node = element._c_node
-        if c_node.parent is not self._c_node:
-            raise ValueError, "Element is not a child of this node."
-        _removeText(c_node.next)
-        tree.xmlUnlinkNode(c_node)
-        
-    # PROPERTIES
-    property tag:
-        def __get__(self):
-            if self._tag is not None:
-                return self._tag
-            self._tag = _namespacedName(self._c_node)
-            return self._tag
-    
-        def __set__(self, value):
-            cdef xmlNs* c_ns
-            ns, text = _getNsTag(value)
-            self._tag = value
-            tree.xmlNodeSetName(self._c_node, _cstr(text))
-            if ns is None:
-                return
-            self._doc._setNodeNs(self._c_node, _cstr(ns))
-
-    # not in ElementTree, read-only
-    property prefix:
-        def __get__(self):
-            if self._c_node.ns is not NULL:
-                if self._c_node.ns.prefix is not NULL:
-                    return funicode(self._c_node.ns.prefix)
-            return None
-        
-    property attrib:
-        def __get__(self):
-            return _attribFactory(self._doc, self._c_node)
-        
-    property text:
-        def __get__(self):
-            return _collectText(self._c_node.children)
-        
-        def __set__(self, value):
-            cdef xmlNode* c_text_node
-            # remove all text nodes at the start first
-            _removeText(self._c_node.children)
-            if value is None:
-                return
-            # now add new text node with value at start
-            text = _utf8(value)
-            c_text_node = tree.xmlNewDocText(self._doc._c_doc,
-                                             _cstr(text))
-            if self._c_node.children is NULL:
-                tree.xmlAddChild(self._c_node, c_text_node)
-            else:
-                tree.xmlAddPrevSibling(self._c_node.children,
-                                       c_text_node)
-        
-    property tail:
-        def __get__(self):
-            return _collectText(self._c_node.next)
-           
-        def __set__(self, value):
-            cdef xmlNode* c_text_node
-            # remove all text nodes at the start first
-            _removeText(self._c_node.next)
-            if value is None:
-                return
-            text = _utf8(value)
-            c_text_node = tree.xmlNewDocText(self._doc._c_doc, _cstr(text))
-            # XXX what if we're the top element?
-            tree.xmlAddNextSibling(self._c_node, c_text_node)
-
-    # ACCESSORS
-    def __repr__(self):
-        return "<Element %s at %x>" % (self.tag, id(self))
-    
-    def __getitem__(self, index):
-        cdef xmlNode* c_node
-        c_node = _findChild(self._c_node, index)
-        if c_node is NULL:
-            raise IndexError, "list index out of range"
-        return _elementFactory(self._doc, c_node)
-
-    def __getslice__(self, start, stop):
-        cdef xmlNode* c_node
-        cdef _Document doc
-        cdef int c, c_stop
-        # this does not work for negative start, stop, however,
-        # python seems to convert these to positive start, stop before
-        # calling, so this all works perfectly (at the cost of a len() call)
-        c_node = _findChild(self._c_node, start)
-        if c_node is NULL:
-            return []
-        c = start
-        c_stop = stop
-        result = []
-        doc = self._doc
-        while c_node is not NULL and c < c_stop:
-            if _isElement(c_node):
-                ret = python.PyList_Append(result, _elementFactory(doc, c_node))
-                if ret:
-                    raise
-                c = c + 1
-            c_node = c_node.next
-        return result
-            
-    def __len__(self):
-        cdef int c
-        cdef xmlNode* c_node
-        c = 0
-        c_node = self._c_node.children
-        while c_node is not NULL:
-            if _isElement(c_node):
-                c = c + 1
-            c_node = c_node.next
-        return c
-
-    def __nonzero__(self):
-        cdef xmlNode* c_node
-        c_node = _findChildBackwards(self._c_node, 0)
-        return c_node != NULL
-
-    def __iter__(self):
-        return ElementChildIterator(self)
-
-    def __reversed__(self):
-        return ElementChildIterator(self, reversed=True)
-
-    def index(self, _Element x not None, start=None, stop=None):
-        cdef int k
-        cdef int l
-        cdef int c_stop
-        cdef int c_start
-        cdef xmlNode* c_child
-        cdef xmlNode* c_start_node
-        c_child = x._c_node
-        if c_child.parent is not self._c_node:
-            raise ValueError, "Element is not a child of this node."
-
-        if start is None:
-            c_start = 0
-        else:
-            c_start = start
-        if stop is None:
-            c_stop = 0
-        else:
-            c_stop = stop
-            if c_stop == 0 or \
-                   c_start >= c_stop and (c_stop > 0 or c_start < 0):
-                raise ValueError, "list.index(x): x not in slice"
-
-        # for negative slice indices, check slice before searching index
-        if c_start < 0 or c_stop < 0:
-            # start from right, at most up to leftmost(c_start, c_stop)
-            if c_start < c_stop:
-                k = -c_start
-            else:
-                k = -c_stop
-            c_start_node = self._c_node.last
-            l = 1
-            while c_start_node != c_child and l < k:
-                if _isElement(c_start_node):
-                    l = l + 1
-                c_start_node = c_start_node.prev
-            if c_start_node == c_child:
-                # found! before slice end?
-                if c_stop < 0 and l <= -c_stop:
-                    raise ValueError, "list.index(x): x not in slice"
-            elif c_start < 0:
-                raise ValueError, "list.index(x): x not in slice"
-
-        # now determine the index backwards from child
-        c_child = c_child.prev
-        k = 0
-        if c_stop > 0:
-            # we can optimize: stop after c_stop elements if not found
-            while c_child != NULL and k < c_stop:
-                if _isElement(c_child):
-                    k = k + 1
-                c_child = c_child.prev
-            if k < c_stop:
-                return k
-        else:
-            # traverse all
-            while c_child != NULL:
-                if _isElement(c_child):
-                    k = k + 1
-                c_child = c_child.prev
-            if c_start > 0:
-                if k >= c_start:
-                    return k
-            else:
-                return k
-        if c_start or c_stop:
-            raise ValueError, "list.index(x): x not in slice"
-        else:
-            raise ValueError, "list.index(x): x not in list"
-
-    def get(self, key, default=None):
-        # XXX more redundancy, but might be slightly faster than
-        #     return self.attrib.get(key, default)
-        cdef char* cresult
-        cdef char* c_tag
-        ns, tag = _getNsTag(key)
-        c_tag = _cstr(tag)
-        if ns is None:
-            cresult = tree.xmlGetNoNsProp(self._c_node, c_tag)
-        else:
-            cresult = tree.xmlGetNsProp(self._c_node, c_tag, _cstr(ns))
-        if cresult is NULL:
-            result = default
-        else:
-            result = funicode(cresult)
-            tree.xmlFree(cresult)
-        return result
-
-    def keys(self):
-        return self.attrib.keys()
-
-    def items(self):
-        return self.attrib.items()
-
-    def getchildren(self):
-        cdef xmlNode* c_node
-        cdef _Document doc
-        cdef int ret
-        result = []
-        doc = self._doc
-        c_node = self._c_node.children
-        while c_node is not NULL:
-            if _isElement(c_node):
-                ret = python.PyList_Append(result, _elementFactory(doc, c_node))
-                if ret:
-                    raise
-            c_node = c_node.next
-        return result
-
-    def getparent(self):
-        cdef xmlNode* c_node
-        c_node = self._c_node.parent
-        if c_node is not NULL and _isElement(c_node):
-            return _elementFactory(self._doc, c_node)
-        return None
-
-    def getiterator(self, tag=None):
-        iterator = ElementDepthFirstIterator(self)
-        if tag is None or tag == '*':
-            return iterator
-        else:
-            return ElementTagFilter(iterator, tag)
-
-    def makeelement(self, _tag, attrib=None, nsmap=None, **_extra):
-        "Creates a new element associated with the same document."
-        # a little code duplication, but less overhead through doc reuse
-        cdef xmlNode*  c_node
-        cdef xmlDoc*   c_doc
-        cdef _Document doc
-        ns_utf, name_utf = _getNsTag(_tag)
-        doc = self._doc
-        c_doc = doc._c_doc
-        c_node = _createElement(c_doc, name_utf, attrib, _extra)
-        # add namespaces to node if necessary
-        doc._setNodeNamespaces(c_node, ns_utf, nsmap)
-        return _elementFactory(doc, c_node)
-
-    def find(self, path):
-        return _elementpath.find(self, path)
-
-    def findtext(self, path, default=None):
-        return _elementpath.findtext(self, path, default)
-
-    def findall(self, path):
-        return _elementpath.findall(self, path)
-
-    def xpath(self, _path, namespaces=None, **_variables):
-        evaluator = XPathElementEvaluator(self, namespaces)
-        return evaluator.evaluate(_path, **_variables)
-
-cdef _Element _elementFactory(_Document doc, xmlNode* c_node):
-    cdef _Element result
-    cdef char* c_ns_href
-    result = getProxy(c_node, PROXY_ELEMENT)
-    if result is not None:
-        return result
-    if c_node is NULL:
-        return None
-    if c_node.type == tree.XML_ELEMENT_NODE:
-        if c_node.ns == NULL:
-            c_ns_href = NULL
-        else:
-            c_ns_href = c_node.ns.href
-        element_class = _find_element_class(c_ns_href, c_node.name)
-    elif c_node.type == tree.XML_COMMENT_NODE:
-        element_class = _Comment
-    else:
-        assert 0, "Unknown node type: %s" % c_node.type
-    result = element_class()
-    result._tag = None
-    result._doc = doc
-    result._c_node = c_node
-    result._proxy_type = PROXY_ELEMENT
-    registerProxy(result, PROXY_ELEMENT)
-    result._init()
-    return result
-
-cdef class _Comment(_Element):
-    def set(self, key, value):
-        pass
-    
-    def append(self, _Element element):
-        pass
-
-    property tag:
-        def __get__(self):
-            return None
-        
-    property attrib:
-        def __get__(self):
-            return {}
-        
-    property text:
-        def __get__(self):
-            return funicode(self._c_node.content)
-
-        def __set__(self, value):
-            pass
-                        
-    # ACCESSORS
-    def __repr__(self):
-        return "<Comment[%s]>" % self.text
-    
-    def __getitem__(self, n):
-        raise IndexError
-
-    def __len__(self):
-        return 0
-
-    def get(self, key, default=None):
-        return None
-
-    def keys(self):
-        return []
-    
-    def items(self):
-        return []
-    
-cdef _Comment _commentFactory(_Document doc, xmlNode* c_node):
-    cdef _Comment result
-    result = getProxy(c_node, PROXY_ELEMENT)
-    if result is not None:
-        return result
-    if c_node is NULL:
-        return None
-    result = _Comment()
-    result._doc = doc
-    result._c_node = c_node
-    result._proxy_type = PROXY_ELEMENT
-    registerProxy(result, PROXY_ELEMENT)
-    return result
-
-cdef class _Attrib(_NodeBase):
-    # MANIPULATORS
-    def __setitem__(self, key, value):
-        cdef xmlNs* c_ns
-        cdef char* c_value
-        cdef char* c_tag
-        ns, tag = _getNsTag(key)
-        c_tag = _cstr(tag)
-        value = _utf8(value)
-        c_value = _cstr(value)
-        if ns is None:
-            tree.xmlSetProp(self._c_node, c_tag, c_value)
-        else:
-            c_ns = self._doc._findOrBuildNodeNs(self._c_node, _cstr(ns))
-            tree.xmlSetNsProp(self._c_node, c_ns, c_tag, c_value)
-
-    def __delitem__(self, key):
-        cdef xmlNs* c_ns
-        cdef xmlAttr* c_attr
-        cdef char* c_tag
-        ns, tag = _getNsTag(key)
-        c_tag = _cstr(tag)
-        if ns is None:
-            c_attr = tree.xmlHasProp(self._c_node, c_tag)
-        else:
-            c_attr = tree.xmlHasNsProp(self._c_node, c_tag, _cstr(ns))
-        if c_attr is NULL:
-            # XXX free namespace that is not in use..?
-            raise KeyError, key
-        tree.xmlRemoveProp(c_attr)
-        
-    # ACCESSORS
-    def __repr__(self):
-        result = {}
-        for key, value in self.items():
-            result[key] = value
-        return repr(result)
-    
-    def __getitem__(self, key):
-        cdef xmlNs* c_ns
-        cdef char* cresult
-        cdef char* c_tag
-        ns, tag = _getNsTag(key)
-        c_tag = _cstr(tag)
-        if ns is None:
-            cresult = tree.xmlGetNoNsProp(self._c_node, c_tag)
-        else:
-            cresult = tree.xmlGetNsProp(self._c_node, c_tag, _cstr(ns))
-        if cresult is NULL:
-            # XXX free namespace that is not in use..?
-            raise KeyError, key
-        result = funicode(cresult)
-        tree.xmlFree(cresult)
-        return result
-
-    def __len__(self):
-        cdef int c
-        cdef xmlNode* c_node
-        c = 0
-        c_node = <xmlNode*>(self._c_node.properties)
-        while c_node is not NULL:
-            if c_node.type == tree.XML_ATTRIBUTE_NODE:
-                c = c + 1
-            c_node = c_node.next
-        return c
-    
-    def get(self, key, default=None):
-        try:
-            return self.__getitem__(key)
-        except KeyError:
-            return default
-
-    def keys(self):
-        result = []
-        cdef xmlNode* c_node
-        c_node = <xmlNode*>(self._c_node.properties)
-        while c_node is not NULL:
-            if c_node.type == tree.XML_ATTRIBUTE_NODE:
-                python.PyList_Append(result, _namespacedName(c_node))
-            c_node = c_node.next
-        return result
-
-    def __iter__(self):
-        return iter(self.keys())
-    
-    def iterkeys(self):
-        return iter(self.keys())
-
-    def values(self):
-        cdef xmlNode* c_node
-        result = []
-        c_node = <xmlNode*>(self._c_node.properties)
-        while c_node is not NULL:
-            if c_node.type == tree.XML_ATTRIBUTE_NODE:
-                python.PyList_Append(
-                    result, _attributeValue(self._c_node, c_node))
-            c_node = c_node.next
-        return result
-
-    def itervalues(self):
-        return iter(self.values())
-
-    def items(self):
-        result = []
-        cdef xmlNode* c_node
-        c_node = <xmlNode*>(self._c_node.properties)
-        while c_node is not NULL:
-            if c_node.type == tree.XML_ATTRIBUTE_NODE:
-                python.PyList_Append(result, (
-                    _namespacedName(c_node),
-                    _attributeValue(self._c_node, c_node)
-                    ))
-            c_node = c_node.next
-        return result
-
-    def iteritems(self):
-        return iter(self.items())
-
-    def has_key(self, key):
-        cdef xmlNs* c_ns
-        cdef char* result
-        cdef char* c_tag
-        ns, tag = _getNsTag(key)
-        c_tag = _cstr(tag)
-        if ns is None:
-            result = tree.xmlGetNoNsProp(self._c_node, c_tag)
-        else:
-            result = tree.xmlGetNsProp(self._c_node, c_tag, _cstr(ns))
-        if result is not NULL:
-            tree.xmlFree(result)
-            return True
-        else:
-            return False
-
-    def __contains__(self, key):
-        cdef xmlNs* c_ns
-        cdef char* result
-        cdef char* c_tag
-        ns, tag = _getNsTag(key)
-        c_tag = _cstr(tag)
-        if ns is None:
-            result = tree.xmlGetNoNsProp(self._c_node, c_tag)
-        else:
-            result = tree.xmlGetNsProp(self._c_node, c_tag, _cstr(ns))
-        if result is not NULL:
-            tree.xmlFree(result)
-            return True
-        else:
-            return False
-  
-cdef _Attrib _attribFactory(_Document doc, xmlNode* c_node):
-    cdef _Attrib result
-    result = getProxy(c_node, PROXY_ATTRIB)
-    if result is not None:
-        return result
-    result = _Attrib()
-    result._doc = doc
-    result._c_node = c_node
-    result._proxy_type = PROXY_ATTRIB
-    registerProxy(result, PROXY_ATTRIB)
-    return result
-
-ctypedef xmlNode* (*_node_to_node_function)(xmlNode*)
-
-cdef class ElementChildIterator:
-    # we keep Python references here to control GC
-    cdef _NodeBase _node
-    cdef _node_to_node_function _next_element
-    def __init__(self, _NodeBase node, reversed=False): # Python ref!
-        cdef xmlNode* c_node
-        if reversed:
-            c_node = _findChildBackwards(node._c_node, 0)
-            self._next_element = _previousElement
-        else:
-            c_node = _findChildForwards(node._c_node, 0)
-            self._next_element = _nextElement
-        if c_node is NULL:
-            self._node = None
-        else:
-            self._node = _elementFactory(node._doc, c_node)
-    def __iter__(self):
-        return self
-    def __next__(self):
-        cdef xmlNode* c_node
-        cdef _NodeBase current_node
-        # Python ref:
-        current_node = self._node
-        if current_node is None:
-            raise StopIteration
-        c_node = self._next_element(current_node._c_node)
-        if c_node is NULL:
-            self._node = None
-        else:
-            # Python ref:
-            self._node = _elementFactory(current_node._doc, c_node)
-        return current_node
-
-cdef class ElementDepthFirstIterator:
-    """Iterates over an element and its sub-elements in document order (depth
-    first pre-order)."""
-    # we keep Python references here to control GC
-    # keep next node to return and a stack of position state in the tree
-    cdef object _stack
-    cdef _NodeBase _next_node
-    def __init__(self, _NodeBase node not None):
-        cdef xmlNode* c_node
-        self._next_node = node
-        self._stack = []
-        self._findAndPushNextNode(node)
-    def __iter__(self):
-        return self
-    def __next__(self):
-        cdef xmlNode* c_node
-        cdef _NodeBase next_node
-        current_node = self._next_node
-        if current_node is None:
-            raise StopIteration
-        stack = self._stack
-        if python.PyList_GET_SIZE(stack) == 0:
-            self._next_node = None
-            return current_node
-        next_node = stack[-1]
-        self._next_node = next_node
-        self._findAndPushNextNode(next_node)
-        return current_node
-
-    cdef void _findAndPushNextNode(self, _NodeBase node):
-        cdef xmlNode* c_node
-        stack = self._stack
-        # try next child level until we hit a leaf
-        c_node = _findChildForwards(node._c_node, 0)
-        if c_node is NULL:
-            pop = stack.pop
-            while c_node is NULL and python.PyList_GET_SIZE(stack):
-                # walk up the stack until we find a sibling
-                node = pop()
-                c_node = _nextElement(node._c_node)
-        if c_node is not NULL:
-            python.PyList_Append(
-                stack, _elementFactory(node._doc, c_node))
-
-cdef class ElementTagFilter:
-    cdef object _iterator
-    cdef object _pystrings
-    cdef char* _href
-    cdef char* _name
-    def __init__(self, element_iterator, tag):
-        self._iterator = iter(element_iterator)
-        ns_href, name = _getNsTag(tag)
-        self._pystrings = (ns_href, name) # keep Python references
-        self._name = _cstr(name)
-        if ns_href is None:
-            self._href = NULL
-        else:
-            self._href = _cstr(ns_href)
-    def __iter__(self):
-        return self
-    def __next__(self):
-        cdef _NodeBase node
-        while 1:
-            node = self._iterator.next()
-            if self._tagMatches(node._c_node):
-                return node
-
-    cdef int _tagMatches(self, xmlNode* c_node):
-        if tree.strcmp(c_node.name, self._name) == 0:
-            if c_node.ns == NULL or c_node.ns.href == NULL:
-                return self._href == NULL
-            else:
-                return tree.strcmp(c_node.ns.href, self._href) == 0
-        return 0
-
-cdef xmlNode* _createElement(xmlDoc* c_doc, object name_utf,
-                             object attrib, object extra) except NULL:
-    cdef xmlNode* c_node
-    if extra:
-        if attrib is None:
-            attrib = extra
-        else:
-            attrib.update(extra)
-    c_node = tree.xmlNewDocNode(c_doc, NULL, _cstr(name_utf), NULL)
-    if attrib:
-        for name, value in attrib.items():
-            attr_name_utf = _utf8(name)
-            value_utf = _utf8(value)
-            tree.xmlNewProp(c_node, _cstr(attr_name_utf), _cstr(value_utf))
-    return c_node
-
-cdef xmlNode* _createComment(xmlDoc* c_doc, char* text):
-    cdef xmlNode* c_node
-    c_node = tree.xmlNewDocComment(c_doc, text)
-    return c_node
-
-
-# module-level API for ElementTree
-
-def Element(_tag, attrib=None, nsmap=None, **_extra):
-    cdef xmlNode*  c_node
-    cdef xmlDoc*   c_doc
-    cdef _Document doc
-    ns_utf, name_utf = _getNsTag(_tag)
-    c_doc = _newDoc()
-    c_node = _createElement(c_doc, name_utf, attrib, _extra)
-    tree.xmlDocSetRootElement(c_doc, c_node)
-    doc = _documentFactory(c_doc, None)
-    # add namespaces to node if necessary
-    doc._setNodeNamespaces(c_node, ns_utf, nsmap)
-    return _elementFactory(doc, c_node)
-
-def Comment(text=None):
-    cdef _Document doc
-    cdef xmlNode*  c_node
-    cdef xmlDoc*   c_doc
-    if text is None:
-        text = '  '
-    else:
-        text = ' %s ' % _utf8(text)
-    c_doc = _newDoc()
-    doc = _documentFactory(c_doc, None)
-    c_node = _createComment(c_doc, text)
-    tree.xmlAddChild(<xmlNode*>c_doc, c_node)
-    return _commentFactory(doc, c_node)
-
-def SubElement(_Element _parent not None, _tag,
-               attrib=None, nsmap=None, **_extra):
-    cdef xmlNode*  c_node
-    cdef _Document doc
-    ns_utf, name_utf = _getNsTag(_tag)
-    doc = _parent._doc
-    c_node = _createElement(doc._c_doc, name_utf, attrib, _extra)
-    tree.xmlAddChild(_parent._c_node, c_node)
-    # add namespaces to node if necessary
-    doc._setNodeNamespaces(c_node, ns_utf, nsmap)
-    return _elementFactory(doc, c_node)
-
-def ElementTree(_Element element=None, file=None, parser=None):
-    cdef xmlNode* c_next
-    cdef xmlNode* c_node
-    cdef xmlNode* c_node_copy
-    cdef xmlDoc*  c_doc
-    cdef _ElementTree etree
-    cdef _Document doc
-
-    if element is not None:
-        doc  = element._doc
-    elif file is not None:
-        doc = _parseDocument(file, parser)
-    else:
-        c_doc = _newDoc()
-        doc = _documentFactory(c_doc, parser)
-
-    etree = _elementTreeFactory(doc, element)
-
-##     # XXX what if element and file are both not None?
-##     if element is not None:
-##         c_next = element._c_node.next
-##         tree.xmlDocSetRootElement(etree._c_doc, element._c_node)
-##         _moveTail(c_next, element._c_node)
-##         changeDocumentBelow(element, etree)
-    
-    return etree
-
-def HTML(text):
-    cdef _Document doc
-    doc = _parseMemoryDocument(text, __DEFAULT_HTML_PARSER)
-    return doc.getroot()
-
-def XML(text):
-    cdef _Document doc
-    doc = _parseMemoryDocument(text, __DEFAULT_XML_PARSER)
-    return doc.getroot()
-
-fromstring = XML
-
-cdef class QName:
-    cdef readonly object text
-    def __init__(self, text_or_uri, tag=None):
-        if tag is not None:
-            text_or_uri = "{%s}%s" % (text_or_uri, tag)
-        elif not python.PyString_Check(text_or_uri) and \
-             not python.PyUnicode_Check(text_or_uri):
-            text_or_uri = str(text_or_uri)
-        self.text = text_or_uri
-    def __str__(self):
-        return self.text
-    def __hash__(self):
-        return self.text.__hash__()
-
-def iselement(element):
-    return isinstance(element, _Element)
-
-def dump(_NodeBase elem):
-    assert elem is not None, "Must supply element."
-    # better, but not ET compatible : "_NodeBase elem not None"
-    _dumpToFile(sys.stdout, elem._doc._c_doc, elem._c_node)
-
-def tostring(_NodeBase element, encoding='us-ascii'):
-    cdef _Document doc
-    cdef tree.xmlOutputBuffer* c_buffer
-    cdef tree.xmlCharEncodingHandler* enchandler
-    cdef char* enc
-
-    assert element is not None
-    # better, but not ET compatible : "_NodeBase element not None"
-    
-    #if encoding is None:
-    #    encoding = 'UTF-8'
-    if encoding in ('utf8', 'UTF8', 'utf-8'):
-        encoding = 'UTF-8'
-    doc = element._doc
-    enc = _cstr(encoding)
-    # it is necessary to *and* find the encoding handler *and* use
-    # encoding during output
-    enchandler = tree.xmlFindCharEncodingHandler(enc)
-    c_buffer = tree.xmlAllocOutputBuffer(enchandler)
-    tree.xmlNodeDumpOutput(c_buffer, doc._c_doc, element._c_node, 0, 0,
-                           enc)
-    _dumpNextNode(c_buffer, doc._c_doc, element._c_node, enc)
-    tree.xmlOutputBufferFlush(c_buffer)
-    if c_buffer.conv is not NULL: 
-        result = tree.xmlBufferContent(c_buffer.conv)
-    else:
-        result = tree.xmlBufferContent(c_buffer.buffer)
-    tree.xmlOutputBufferClose(c_buffer)
-    return result
-
-def parse(source, parser=None):
-    """Return an ElementTree object loaded with source elements.  If no parser
-    is provided as second argument, the default parser is used.
-    """
-    cdef _Document doc
-    doc = _parseDocument(source, parser)
-    return ElementTree(doc.getroot())
-
-
-# include submodules
-include "proxy.pxi"      # Proxy handling (element backpointers/memory/etc.)
-include "xmlerror.pxi"   # error and log handling
-include "nsclasses.pxi"  # Namespace implementation and registry
-include "docloader.pxi"  # Support for custom document loaders
-include "parser.pxi"     # XML Parser
-include "xmlid.pxi"      # XMLID and IDDict
-include "extensions.pxi" # XPath/XSLT extension functions
-include "xpath.pxi"      # XPath evaluation
-include "xslt.pxi"       # XSL transformations
-
-
-################################################################################
-# Validation
-
-cdef class _Validator:
-    "Base class for XML validators."
-    cdef _ErrorLog _error_log
-    def __init__(self):
-        self._error_log = _ErrorLog()
-        
-    def validate(self, etree):
-        """Validate the document using this schema.
-
-        Returns true if document is valid, false if not."""
-        return self(etree)
-
-    def assertValid(self, etree):
-        "Raises DocumentInvalid if the document does not comply with the schema."
-        if not self(etree):
-            raise DocumentInvalid, "Document does not comply with schema"
-
-    def assert_(self, etree):
-        "Raises AssertionError if the document does not comply with the schema."
-        if not self(etree):
-            raise AssertionError, "Document does not comply with schema"
-
-    property error_log:
-        def __get__(self):
-            return self._error_log.copy()
-
-include "relaxng.pxi"   # RelaxNG
-include "xmlschema.pxi" # XMLSchema
-
-
-################################################################################
-# Private helper functions
 
 cdef _Document _documentOrRaise(object input):
     cdef _Document doc

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Fri Apr 28 21:56:36 2006
@@ -2,9 +2,6 @@
 from tree cimport xmlDoc, xmlNode, xmlAttr, xmlNs, _isElement
 from python cimport isinstance, issubclass, hasattr, callable
 from python cimport iter, str, _cstr
-cimport xpath
-cimport xslt
-cimport xmlerror
 cimport xinclude
 cimport c14n
 cimport cstd
@@ -191,27 +188,6 @@
         if node_ns_utf is not None:
             self._setNodeNs(c_node, node_ns_utf)
 
-cdef _Document _parseDocument(source, parser):
-    cdef xmlDoc* c_doc
-    filename = _getFilenameForFile(source)
-    # Support for unamed file-like object (StringIO, urlgrabber.urlopen, ...)
-    if not filename and hasattr(source, 'read'):
-        return _parseMemoryDocument(source.read(), parser)
-
-    # Otherwise parse the file directly from the filesystem
-    if filename is None:
-        filename = source
-    # open filename
-    c_doc = _parseDocFromFile(_utf8(filename), parser)
-    return _documentFactory(c_doc, parser)
-
-cdef _Document _parseMemoryDocument(text, parser):
-    cdef xmlDoc* c_doc
-    if python.PyUnicode_Check(text):
-        text = _stripDeclaration(_utf8(text))
-    c_doc = _parseDoc(text, parser)
-    return _documentFactory(c_doc, parser)
-
 cdef _Document _documentFactory(xmlDoc* c_doc, parser):
     cdef _Document result
     result = _Document()
@@ -221,15 +197,6 @@
         parser = __DEFAULT_PARSER
     result._parser = parser.copy()
     return result
-
-# to help with debugging
-cdef void displayNode(xmlNode* c_node, indent):
-    cdef xmlNode* c_child
-    print indent * ' ', <int>c_node
-    c_child = c_node.children
-    while c_child is not NULL:
-        displayNode(c_child, indent + 1)
-        c_child = c_child.next
         
 cdef class _NodeBase:
     """Base class to reference a document object and a libxml node.
@@ -1353,7 +1320,8 @@
 
 # include submodules
 include "proxy.pxi"      # Proxy handling (element backpointers/memory/etc.)
-include "xmlerror.pxi"   # error and log handling
+include "apihelpers.pxi" # Private helper functions
+include "xmlerror.pxi"   # Error and log handling
 include "nsclasses.pxi"  # Namespace implementation and registry
 include "docloader.pxi"  # Support for custom document loaders
 include "parser.pxi"     # XML Parser
@@ -1394,391 +1362,3 @@
 
 include "relaxng.pxi"   # RelaxNG
 include "xmlschema.pxi" # XMLSchema
-
-
-################################################################################
-# Private helper functions
-
-cdef _Document _documentOrRaise(object input):
-    cdef _Document doc
-    doc = _documentOf(input)
-    if doc is None:
-        raise TypeError, "Invalid input object: %s" % type(input)
-    else:
-        return doc
-
-cdef _Document _documentOf(object input):
-    # call this to get the document of a
-    # _Document, _ElementTree or _NodeBase object
-    if isinstance(input, _ElementTree):
-        return (<_ElementTree>input)._doc
-    elif isinstance(input, _NodeBase):
-        return (<_NodeBase>input)._doc
-    elif isinstance(input, _Document):
-        return <_Document>input
-    else:
-        return None
-
-cdef _NodeBase _rootNodeOf(object input):
-    # call this to get the root node of a
-    # _Document, _ElementTree or _NodeBase object
-    if isinstance(input, _ElementTree):
-        return (<_ElementTree>input)._context_node
-    elif isinstance(input, _NodeBase):
-        return <_NodeBase>input
-    elif isinstance(input, _Document):
-        return (<_Document>input).getroot()
-    else:
-        return None
-
-cdef xmlDoc* _fakeRootDoc(xmlDoc* c_base_doc, xmlNode* c_node):
-    # build a temporary document that has the given node as root node
-    # note that copy and original must not be modified during its lifetime!!
-    # always call _destroyFakeDoc() after use!
-    cdef xmlNode* c_child
-    cdef xmlNode* c_root
-    cdef xmlDoc*  c_doc
-    c_root = tree.xmlDocGetRootElement(c_base_doc)
-    if c_root == c_node:
-        # already the root node
-        return c_base_doc
-
-    c_doc  = tree.xmlCopyDoc(c_base_doc, 0)        # non recursive!
-    c_root = tree.xmlDocCopyNode(c_node, c_doc, 2) # non recursive!
-
-    c_root.children = c_node.children
-    c_root.last = c_node.last
-    c_root.next = c_root.prev = c_root.parent = NULL
-
-    # store original node
-    c_root._private = c_node
-
-    # divert parent pointers of children
-    c_child = c_root.children
-    while c_child is not NULL:
-        c_child.parent = c_root
-        c_child = c_child.next
-
-    c_doc.children = c_root
-    return c_doc
-
-cdef void _destroyFakeDoc(xmlDoc* c_base_doc, xmlDoc* c_doc):
-    # delete a temporary document
-    cdef xmlNode* c_child
-    cdef xmlNode* c_parent
-    cdef xmlNode* c_root
-    if c_doc != c_base_doc:
-        c_root = tree.xmlDocGetRootElement(c_doc)
-
-        # restore parent pointers of children
-        c_parent = <xmlNode*>c_root._private
-        c_child = c_root.children
-        while c_child is not NULL:
-            c_child.parent = c_parent
-            c_child = c_child.next
-
-        # prevent recursive removal of children
-        c_root.children = c_root.last = c_root._private = NULL
-        tree.xmlFreeDoc(c_doc)
-
-cdef object _attributeValue(xmlNode* c_element, xmlNode* c_attrib_node):
-    cdef char* value
-    if c_attrib_node.ns is NULL or c_attrib_node.ns.href is NULL:
-        value = tree.xmlGetNoNsProp(c_element, c_attrib_node.name)
-    else:
-        value = tree.xmlGetNsProp(c_element, c_attrib_node.name,
-                                  c_attrib_node.ns.href)
-    return funicode(value)
-
-cdef _dumpToFile(f, xmlDoc* c_doc, xmlNode* c_node):
-    cdef python.PyObject* o
-    cdef tree.xmlOutputBuffer* c_buffer
-    
-    if not python.PyFile_Check(f):
-        raise ValueError, "Not a file"
-    o = <python.PyObject*>f
-    c_buffer = tree.xmlOutputBufferCreateFile(python.PyFile_AsFile(o), NULL)
-    tree.xmlNodeDumpOutput(c_buffer, c_doc, c_node, 0, 0, NULL)
-    # dump next node if it's a text node
-    _dumpNextNode(c_buffer, c_doc, c_node, NULL)
-    tree.xmlOutputBufferWriteString(c_buffer, '\n')
-    tree.xmlOutputBufferFlush(c_buffer)
-
-cdef _dumpNextNode(tree.xmlOutputBuffer* c_buffer, xmlDoc* c_doc,
-                   xmlNode* c_node, char* encoding):
-    cdef xmlNode* c_next
-    c_next = c_node.next
-    if c_next is not NULL and c_next.type == tree.XML_TEXT_NODE:
-        tree.xmlNodeDumpOutput(c_buffer, c_doc, c_next, 0, 0, encoding)
-
-cdef object _stripDeclaration(object xml_string):
-    xml_string = xml_string.strip()
-    if xml_string[:5] == '<?xml':
-        i = xml_string.find('?>')
-        if i != -1:
-            if xml_string[i+2:i+3] == '\n':
-                i = i+1
-            xml_string = xml_string[i + 2:]
-    return xml_string
-
-cdef _collectText(xmlNode* c_node):
-    """Collect all text nodes and return them as a unicode string.
-
-    Start collecting at c_node.
-    
-    If there was no text to collect, return None
-    """
-    cdef int scount
-    cdef char* text
-    cdef xmlNode* c_node_cur
-    # check for multiple text nodes
-    scount = 0
-    text = NULL
-    c_node_cur = c_node
-    while c_node_cur is not NULL and c_node_cur.type == tree.XML_TEXT_NODE:
-        if c_node_cur.content[0] != c'\0':
-            text = c_node_cur.content
-            scount = scount + 1
-        c_node_cur = c_node_cur.next
-
-    # handle two most common cases first
-    if text is NULL:
-        return None
-    if scount == 1:
-        return funicode(text)
-
-    # the rest is not performance critical anymore
-    result = ''
-    while c_node is not NULL and c_node.type == tree.XML_TEXT_NODE:
-        result = result + c_node.content
-        c_node = c_node.next
-    return funicode(result)
-
-cdef _removeText(xmlNode* c_node):
-    """Remove all text nodes.
-
-    Start removing at c_node.
-    """
-    cdef xmlNode* c_next
-    while c_node is not NULL and c_node.type == tree.XML_TEXT_NODE:
-        c_next = c_node.next
-        tree.xmlUnlinkNode(c_node)
-        # XXX cannot safely free in case of direct text node proxies..
-        tree.xmlFreeNode(c_node)
-        c_node = c_next
-
-cdef xmlNode* _findChild(xmlNode* c_node, int index):
-    if index < 0:
-        return _findChildBackwards(c_node, -index - 1)
-    else:
-        return _findChildForwards(c_node, index)
-    
-cdef xmlNode* _findChildForwards(xmlNode* c_node, int index):
-    """Return child element of c_node with index, or return NULL if not found.
-    """
-    cdef xmlNode* c_child
-    cdef int c
-    c_child = c_node.children
-    c = 0
-    while c_child is not NULL:
-        if _isElement(c_child):
-            if c == index:
-                return c_child
-            c = c + 1
-        c_child = c_child.next
-    else:
-        return NULL
-
-cdef xmlNode* _findChildBackwards(xmlNode* c_node, int index):
-    """Return child element of c_node with index, or return NULL if not found.
-    Search from the end.
-    """
-    cdef xmlNode* c_child
-    cdef int c
-    c_child = c_node.last
-    c = 0
-    while c_child is not NULL:
-        if _isElement(c_child):
-            if c == index:
-                return c_child
-            c = c + 1
-        c_child = c_child.prev
-    else:
-        return NULL
-    
-cdef xmlNode* _nextElement(xmlNode* c_node):
-    """Given a node, find the next sibling that is an element.
-    """
-    c_node = c_node.next
-    while c_node is not NULL:
-        if _isElement(c_node):
-            return c_node
-        c_node = c_node.next
-    return NULL
-
-cdef xmlNode* _previousElement(xmlNode* c_node):
-    """Given a node, find the next sibling that is an element.
-    """
-    c_node = c_node.prev
-    while c_node is not NULL:
-        if _isElement(c_node):
-            return c_node
-        c_node = c_node.prev
-    return NULL
-
-cdef void _removeNode(xmlNode* c_node):
-    """Unlink and free a node and subnodes if possible.
-    """
-    tree.xmlUnlinkNode(c_node)
-    attemptDeallocation(c_node)
-
-cdef void _moveTail(xmlNode* c_tail, xmlNode* c_target):
-    cdef xmlNode* c_next
-    # tail support: look for any text nodes trailing this node and 
-    # move them too
-    while c_tail is not NULL and c_tail.type == tree.XML_TEXT_NODE:
-        c_next = c_tail.next
-        tree.xmlUnlinkNode(c_tail)
-        tree.xmlAddNextSibling(c_target, c_tail)
-        c_target = c_tail
-        c_tail = c_next
-
-### see etree.h:
-## cdef int _isElement(xmlNode* c_node):
-##     return (c_node.type == tree.XML_ELEMENT_NODE or
-##             c_node.type == tree.XML_COMMENT_NODE)
-
-cdef xmlNode* _deleteSlice(xmlNode* c_node, int start, int stop):
-    """Delete slice, starting with c_node, start counting at start, end at stop.
-    """
-    cdef xmlNode* c_next
-    cdef int c
-    if c_node is NULL:
-        return NULL
-    # now start deleting nodes
-    c = start
-    while c_node is not NULL and c < stop:
-        c_next = c_node.next
-        if _isElement(c_node):
-            _removeText(c_node.next)
-            c_next = c_node.next
-            _removeNode(c_node)
-            c = c + 1
-        c_node = c_next
-    return c_node
-
-cdef int isutf8(char* s):
-    cdef char c
-    c = s[0]
-    while c != c'\0':
-        if c & 0x80:
-            return 1
-        s = s + 1
-        c = s[0]
-    return 0
-
-cdef object funicode(char* s):
-    if isutf8(s):
-        return python.PyUnicode_DecodeUTF8(s, tree.strlen(s), NULL)
-    return python.PyString_FromString(s)
-
-cdef object _utf8(object s):
-    if python.PyString_Check(s):
-        assert not isutf8(_cstr(s)), "All strings must be Unicode or ASCII"
-        return s
-    elif python.PyUnicode_Check(s):
-        return python.PyUnicode_AsUTF8String(s)
-    else:
-        raise TypeError, "Argument must be string or unicode."
-
-cdef _getNsTag(tag):
-    """Given a tag, find namespace URI and tag name.
-    Return None for NS uri if no namespace URI available.
-    """
-    cdef char* c_tag
-    cdef char* c_pos
-    cdef int nslen
-    if isinstance(tag, QName):
-        tag = (<QName>tag).text
-    tag = _utf8(tag)
-    c_tag = _cstr(tag)
-    if c_tag[0] == c'{':
-        c_pos = tree.xmlStrchr(c_tag+1, c'}')
-        if c_pos is NULL:
-            raise ValueError, "Invalid tag name"
-        nslen = c_pos - c_tag - 1
-        ns  = python.PyString_FromStringAndSize(c_tag+1, nslen)
-        tag = python.PyString_FromString(c_pos+1)
-    else:
-        ns = None
-    return ns, tag
-    
-cdef object _namespacedName(xmlNode* c_node):
-    cdef char* href
-    cdef char* name
-    name = c_node.name
-    if c_node.ns is NULL or c_node.ns.href is NULL:
-        return funicode(name)
-    else:
-        href = c_node.ns.href
-        s = python.PyString_FromFormat("{%s}%s", href, name)
-        if isutf8(href) or isutf8(name):
-            return python.PyUnicode_FromEncodedObject(s, 'UTF-8', NULL)
-        else:
-            return s
-
-cdef _getFilenameForFile(source):
-    """Given a Python File or Gzip object, give filename back.
-
-    Returns None if not a file object.
-    """
-    # file instances have a name attribute
-    if hasattr(source, 'name'):
-        return source.name
-    # gzip file instances have a filename attribute
-    if hasattr(source, 'filename'):
-        return source.filename
-    return None
-
-cdef void changeDocumentBelow(_NodeBase node, _Document doc, int recursive):
-    """For a node and all nodes below, change document.
-
-    A node can change document in certain operations as an XML
-    subtree can move. This updates all possible proxies in the
-    tree below (including the current node). It also reconciliates
-    namespaces so they're correct inside the new environment.
-    """
-    if recursive:
-        changeDocumentBelowHelper(node._c_node, doc)
-    tree.xmlReconciliateNs(doc._c_doc, node._c_node)
-    
-cdef void changeDocumentBelowHelper(xmlNode* c_node, _Document doc):
-    cdef ProxyRef* ref
-    cdef xmlNode* c_current
-    cdef xmlAttr* c_attr_current
-    cdef _NodeBase proxy
-
-    if c_node is NULL:
-        return
-    # different _c_doc
-    c_node.doc = doc._c_doc
-    
-    if c_node._private is not NULL:
-        ref = <ProxyRef*>c_node._private
-        while ref is not NULL:
-            proxy = <_NodeBase>ref.proxy
-            proxy._doc = doc
-            ref = ref.next
-
-    # adjust all children
-    c_current = c_node.children
-    while c_current is not NULL:
-        changeDocumentBelowHelper(c_current, doc)
-        c_current = c_current.next
-        
-    # adjust all attributes
-    c_attr_current = c_node.properties
-    while c_attr_current is not NULL:
-        changeDocumentBelowHelper(c_current, doc)
-        c_attr_current = c_attr_current.next
-

Modified: lxml/trunk/src/lxml/extensions.pxi
==============================================================================
--- lxml/trunk/src/lxml/extensions.pxi	(original)
+++ lxml/trunk/src/lxml/extensions.pxi	Fri Apr 28 21:56:36 2006
@@ -1,5 +1,7 @@
 # supports for extension functions in XPath and XSLT
 
+cimport xpath
+
 class XPathError(LxmlError):
     pass
 

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Fri Apr 28 21:56:36 2006
@@ -465,3 +465,29 @@
     result = tree.xmlNewDoc("1.0")
     __GLOBAL_PARSER_CONTEXT._initDocDict(result)
     return result
+
+############################################################
+## API level helper functions for _Document creation
+############################################################
+
+cdef _Document _parseDocument(source, parser):
+    cdef xmlDoc* c_doc
+    filename = _getFilenameForFile(source)
+    # Support for unamed file-like object (StringIO, urlgrabber.urlopen, ...)
+    if not filename and hasattr(source, 'read'):
+        return _parseMemoryDocument(source.read(), parser)
+
+    # Otherwise parse the file directly from the filesystem
+    if filename is None:
+        filename = source
+    # open filename
+    c_doc = _parseDocFromFile(_utf8(filename), parser)
+    return _documentFactory(c_doc, parser)
+
+cdef _Document _parseMemoryDocument(text, parser):
+    cdef xmlDoc* c_doc
+    if python.PyUnicode_Check(text):
+        text = _stripDeclaration(_utf8(text))
+    c_doc = _parseDoc(text, parser)
+    return _documentFactory(c_doc, parser)
+

Modified: lxml/trunk/src/lxml/xmlerror.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlerror.pxi	(original)
+++ lxml/trunk/src/lxml/xmlerror.pxi	Fri Apr 28 21:56:36 2006
@@ -1,5 +1,6 @@
-################################################################################
-# DEBUG setup
+# DEBUG and error logging
+
+cimport xmlerror
 
 # module level API functions
 

Modified: lxml/trunk/src/lxml/xpath.pxi
==============================================================================
--- lxml/trunk/src/lxml/xpath.pxi	(original)
+++ lxml/trunk/src/lxml/xpath.pxi	Fri Apr 28 21:56:36 2006
@@ -1,4 +1,4 @@
-# XSLT and XPath classes, supports for extension functions
+# XPath evaluation
 
 class XPathContextError(XPathError):
     pass

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Fri Apr 28 21:56:36 2006
@@ -1,5 +1,7 @@
 # XSLT and XPath classes, supports for extension functions
 
+cimport xslt
+
 class XSLTError(LxmlError):
     pass
 
From scoder at codespeak.net  Sun Apr 30 09:17:20 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 30 09:17:24 2006
Subject: [Lxml-checkins] r26590 - lxml/trunk/src/lxml
Message-ID: <20060430071720.1CC3E10081@code0.codespeak.net>

Author: scoder
Date: Sun Apr 30 09:17:17 2006
New Revision: 26590

Modified:
   lxml/trunk/src/lxml/xmlerror.pxi
   lxml/trunk/src/lxml/xslt.pxi
Log:
make etree.pyx compilable without libxslt by uncommenting 'include xslt.pxi' line, will still have ElementTree.xslt() function that raises NameError for 'XSLT', but otherwise imports cleanly and works

Modified: lxml/trunk/src/lxml/xmlerror.pxi
==============================================================================
--- lxml/trunk/src/lxml/xmlerror.pxi	(original)
+++ lxml/trunk/src/lxml/xmlerror.pxi	Sun Apr 30 09:17:17 2006
@@ -12,7 +12,11 @@
 def initThreadLogging():
     "Setup logging for the current thread."
     _logLibxmlErrors()
-    _logLibxsltErrors()
+    try:
+        _logLibxsltErrors()
+    except NameError:
+        # compiled without libxslt
+        pass
 
 
 # Logging classes
@@ -299,9 +303,6 @@
     xmlerror.xmlSetGenericErrorFunc(NULL, _nullGenericErrorFunc)
     xmlerror.xmlSetStructuredErrorFunc(NULL, _receiveError)
 
-cdef void _logLibxsltErrors():
-    xslt.xsltSetGenericErrorFunc(NULL, _receiveGenericError)
-
 # init global logging
 initThreadLogging()
 

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Sun Apr 30 09:17:17 2006
@@ -18,6 +18,10 @@
     pass
 
 
+cdef void _logLibxsltErrors():
+    xslt.xsltSetGenericErrorFunc(NULL, _receiveGenericError)
+
+
 ################################################################################
 # XSLT document loaders
 
From scoder at codespeak.net  Sun Apr 30 09:24:06 2006
From: scoder at codespeak.net (scoder@codespeak.net)
Date: Sun Apr 30 09:24:07 2006
Subject: [Lxml-checkins] r26591 - in lxml/trunk: . src/lxml
Message-ID: <20060430072406.CFB8E10081@code0.codespeak.net>

Author: scoder
Date: Sun Apr 30 09:24:04 2006
New Revision: 26591

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/etree.pyx
Log:
CHANGES.txt: mention possibility of compiling without libxslt

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Sun Apr 30 09:24:04 2006
@@ -7,6 +7,9 @@
 Features added
 --------------
 
+* etree module can be compiled without libxslt by commenting out the line
+  'include "xslt.pxi"' at the end of the etree.pyx source file
+
 * Error reporting now also works in XSLT
 
 * Support for custom document loaders (URI resolvers) in parsers and XSLT,

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Sun Apr 30 09:24:04 2006
@@ -1328,7 +1328,10 @@
 include "xmlid.pxi"      # XMLID and IDDict
 include "extensions.pxi" # XPath/XSLT extension functions
 include "xpath.pxi"      # XPath evaluation
-include "xslt.pxi"       # XSL transformations
+
+# XSL transformations
+# comment out to compile without libxslt
+include "xslt.pxi"
 
 
 ################################################################################
From scoder at codespeak.net  Tue Apr 25 15:43:33 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 25 Apr 2006 13:43:33 -0000
Subject: [Lxml-checkins] r26334 - in lxml/pyrex: . Demos Demos/callback Doc
	Pyrex Pyrex/Compiler Pyrex/Distutils Pyrex/Mac
Message-ID: <20060425134333.B42A610091@code0.codespeak.net>

Author: scoder
Date: Tue Apr 25 15:43:21 2006
New Revision: 26334

Modified:
   lxml/pyrex/CHANGES.txt
   lxml/pyrex/Demos/callback/cheese.pyx
   lxml/pyrex/Demos/primes.pyx
   lxml/pyrex/Doc/FAQ.html
   lxml/pyrex/Doc/extension_types.html
   lxml/pyrex/Doc/overview.html
   lxml/pyrex/Doc/sharing.html
   lxml/pyrex/Doc/special_methods.html
   lxml/pyrex/Pyrex/Compiler/CmdLine.py
   lxml/pyrex/Pyrex/Compiler/Code.py
   lxml/pyrex/Pyrex/Compiler/ExprNodes.py
   lxml/pyrex/Pyrex/Compiler/Lexicon.py
   lxml/pyrex/Pyrex/Compiler/Main.py
   lxml/pyrex/Pyrex/Compiler/Naming.py
   lxml/pyrex/Pyrex/Compiler/Nodes.py
   lxml/pyrex/Pyrex/Compiler/Parsing.py
   lxml/pyrex/Pyrex/Compiler/PyrexTypes.py
   lxml/pyrex/Pyrex/Compiler/Scanning.py
   lxml/pyrex/Pyrex/Compiler/Symtab.py
   lxml/pyrex/Pyrex/Compiler/TypeSlots.py
   lxml/pyrex/Pyrex/Compiler/Version.py
   lxml/pyrex/Pyrex/Distutils/build_ext.py
   lxml/pyrex/Pyrex/Mac/DarwinSystem.py
   lxml/pyrex/Pyrex/Utils.py
   lxml/pyrex/ToDo.txt
Log:
updated to Pyrex 0.9.4.1, added support for Python 2.5

Modified: lxml/pyrex/CHANGES.txt
==============================================================================
--- lxml/pyrex/CHANGES.txt	(original)
+++ lxml/pyrex/CHANGES.txt	Tue Apr 25 15:43:21 2006
@@ -1,293 +1,403 @@
-0.9.3.1
+0.9.4.1
 -------
 
 Bug fixes:
 
-	- Fix generated code to compile with GCC 4.0 (remove usage of the
-		cast-as-lvalue extension).
-		["Jeremy Katz" <katzj at redhat.com>]
+  - Fixed indentation problem in Pyrex.Distutils.build_ext.
+    [Oliver Grisel <ogrisel at nuxeo.com>]
 
-	- Fix interoperability with Python 2.4.
-		["Bob Ippolito" <bob at redivi.com>]
 
-0.9.3
+0.9.4
 -----
 
-Enhancements:
+Improvements:
 
-	- Types defined with a ctypedef in a 'cdef extern from' block
-		are now referred to by the typedef name in generated C code,
-		so it is no longer necessary to match the type in the C
-		header file exactly.
+  - All use of lvalue casts has been eliminated, for
+    compatibility with gcc4.
+  
+  - PyMODINIT_FUNC now used to declare the module init function.
+  
+  - Generated code should be compilable as either C or C++.
+    When compiling as C++, "extern C" is used where appropriate
+    to preserve linkage semantics. C++ functions still cannot
+    be called yet.
+  
+  - An extension type can be made weak-referenceable by
+    giving it a C attribute of type object called __weakref__.
+  
+  - Source files opened in universal newlines mode.
+  
+  - Support for public extension type C attributes of type
+    long long and unsigned long long added (but not tested).
+    [Sam Rushing <sam at rushing.nightmare.com>]
+  
+  - Distutils include directories now passed to Pyrex compiler.
+    [Konrad Hinsen <khinsen at cea.fr>]
+  
+  - Integer constants with an "L" suffix are now allowed
+    and are converted to Python long integers.
+    [Rainer Deyke <rainerd at eldwood.com>]
+  
+  - A broken .c file is no longer left behind if there are
+    compilation errors.
+  
+  - Using the result of a Python indexing or attribute access
+    operation as a char * is no longer considered an error in
+    most cases, as the former behaviour proved to be more
+    annoying than helpful.
+  
+Bug fixes:
+
+  - Fixed problems with conversion from Python integers to
+    C unsigned longs. Now use PyInt_AsUnsignedLongMask and
+    PyInt_AsUnsignedLongLongMask instead of the PyLong_*
+    functions (which only work on Python longs).
+    [Wim Vree <wim at vree.be>]
+  
+  - C unsigned ints now converted to/from Python longs intead
+    of Python ints to avoid overflow problems.
+    [Heiko Wundram <me+pyrex at modelnine.org>]
+  
+  - Correct PyArg_ParseTuple format characters now used for
+    unsigned types. [Jeff Bowden <jlb at houseofdistraction.com>]
+  
+  - Nonzero return value from a base class tp_traverse call
+    is handled.
+  
+  - Taking sizeof an incomplete type caused a crash while
+    producing an error message. [Drew Perttula <dperttula at pdi.com>]
+  
+  - If a module cimported itself, definitions of global variables
+    were generated twice. [Parzival Herzog <parzp at shaw.ca>]
+  
+  - Distutils extension updated to handle changed signature of
+    swig_sources(). [David M. Cooke <cookedm at physics.mcmaster.ca>]
+  
+  - Incorrect C code generated for a raw string containing a double
+    quote preceded by a backslash. [Thomas Drake <bistroy at mac.com>]
+  
+  - Declaration of public C function with an exception value written
+    to generated .pxi file without the except clause.
+    [Robby Dermody <robbyd at u20.org>]
+  
+  - __delitem__ method of an extension type with no __setitem__
+    did not get called. [Richard Boulton <richard at tartarus.org>]
+  
+  - A spurious Py_INCREF was generated when a return statement
+    required a type test. [Jonathan Doda <jdoda at sympatico.ca>]
+  
+  - Casting a value to a function pointer and then immediately
+    calling it generated a cast to a function instead of a cast
+    to a function pointer. [Simon Burton <simon at arrowtheory.com>]
+  
+  - Py_TPFLAGS_HAVE_GC was not being set on an extension type that
+    inherited from an external extension type that used GC but did
+    not itself have any PyObject* attributes.
+    [Michael Hordijk <hoffbrinkle at hotmail.com>]
+  
+  - A return statement inside a for statement leaked a reference
+    to the loop's iterator.
+    [J?rgen Kartnaller <j.kartnaller at robotech.at>]
+  
+  - Full module name now appears in __module__ attribute of classes
+    and extension types, provided a correct dotted name is used
+    for the .pyx file. [Giovanni Bajo <rasky at develer.com>]
+
+  - Public extension type with no C attributes produced an
+    invalid .pxi file. [Simon Burton <simon at arrowtheory.com>]
+  
+  - Using a dict constructor as the second operand of a boolean
+    expression crashed the Pyrex compiler.
+    [Stefan Behnel <behnel_ml at gkec.informatik.tu-darmstadt.de>]
+  
+  - A C declaration list ending with a comma resulted in invalid
+    C code being generated. [Alex Coventry <alex_c at mit.edu>]
+  
+  - A raw string containing two consecutive backslashes produced
+    incorrect C code. [Helmut Jarausch <jarausch at igpm.rwth-aachen.de>]
+  
+  - An error is reported if you attempt to declare a special
+    method of an extension type using 'cdef' instead of 'def'.
+    [Sam Rushing <srushing at ironport.com>]
 
-	- Conversion to/from unsigned long now done with
-		PyLong_AsUnsignedLong and PyLong_FromUnsignedLong.
-		[Dug Song <dugsong at monkey.org>]
+0.9.3
+-----
 
-	- A struct, union or enum definition in a 'cdef extern from'
-		block may now be left empty (using 'pass'). This can be useful
-		if you need to declare a variable of that type, but don't need
-		to refer to any of its members.
+Enhancements:
 
-	- More flexible about ordering of qualifiers such as 'long' and
-		'unsigned'.
-		["John (J5) Palmieri" <johnp at redhat.com>]
+  - Types defined with a ctypedef in a 'cdef extern from' block
+    are now referred to by the typedef name in generated C code,
+    so it is no longer necessary to match the type in the C
+    header file exactly.
+
+  - Conversion to/from unsigned long now done with
+    PyLong_AsUnsignedLong and PyLong_FromUnsignedLong.
+    [Dug Song <dugsong at monkey.org>]
+  
+  - A struct, union or enum definition in a 'cdef extern from'
+    block may now be left empty (using 'pass'). This can be useful
+    if you need to declare a variable of that type, but don't need
+    to refer to any of its members.
+  
+  - More flexible about ordering of qualifiers such as 'long' and
+    'unsigned'.
+    ["John (J5) Palmieri" <johnp at redhat.com>]
 
 
 Bug fixes:
 
-	- Non-interned string literals used in a Python class
-		definition did not work.
-		[Atsuo Ishimoto <ishimoto at gembook.org>]
-		[Andreas Kostyrka <andreas at kostyrka.org>]
-
-	- Return types of the buffer interface functions for extension
-		types have been corrected.
-		[Dug Song <dugsong at monkey.org>]
-
-	- Added 'static' to declarations of string literals.
-		[Phil Frost <indigo at bitglue.com>]
-
-	- Float literals are now copied directly to the C code as written,
-		to avoid problems with loss of precision.
-		[Mario Pernici <Mario.Pernici at mi.infn.it>]
-
-	- Inheriting from an extension type with C methods defined in
-		another Pyrex module did not work.
-		[Itamar Shtull-Trauring <itamar at itamarst.org>]
+  - Non-interned string literals used in a Python class
+    definition did not work.
+    [Atsuo Ishimoto <ishimoto at gembook.org>]
+    [Andreas Kostyrka <andreas at kostyrka.org>]
+  
+  - Return types of the buffer interface functions for extension
+    types have been corrected.
+    [Dug Song <dugsong at monkey.org>]
+  
+  - Added 'static' to declarations of string literals.
+    [Phil Frost <indigo at bitglue.com>]
+  
+  - Float literals are now copied directly to the C code as written,
+    to avoid problems with loss of precision.
+    [Mario Pernici <Mario.Pernici at mi.infn.it>]
+  
+  - Inheriting from an extension type with C methods defined in
+    another Pyrex module did not work.
+    [Itamar Shtull-Trauring <itamar at itamarst.org>]
 
 0.9.2.1
 -------
 
 Bug fixes:
 
-	- Corrected an import statement setup.py, and made it
-		check for a unix platform in a more reliable way.
+  - Corrected an import statement setup.py, and made it
+    check for a unix platform in a more reliable way.
 
 0.9.2
 -----
 
 Enhancements:
 
-	- Names of Python global variables and attributes are now
-		interned, and PyObject_GetAttr/SetAttr are used instead
-		of PyObject_GetAttrString/SetAttrString. String literals
-		which resemble Python identifiers are also interned.
-
-	- String literals are now converted to Python objects only
-		once instead of every time they are used.
-
-	- NUL characters are now allowed in Python string literals.
-
-	- Added some missing error checking code to the beginning
-		of module init functions. It's unlikely the operations
-		involved would ever fail, but you never know.
+  - Names of Python global variables and attributes are now
+    interned, and PyObject_GetAttr/SetAttr are used instead
+    of PyObject_GetAttrString/SetAttrString. String literals
+    which resemble Python identifiers are also interned.
+  
+  - String literals are now converted to Python objects only
+    once instead of every time they are used.
+  
+  - NUL characters are now allowed in Python string literals.
+
+  - Added some missing error checking code to the beginning
+    of module init functions. It's unlikely the operations
+    involved would ever fail, but you never know.
 
 Bug fixes:
 
-	- Corrected some problems introduced by moving the Plex
-		package.
+  - Corrected some problems introduced by moving the Plex
+    package.
 
 0.9.1.1
 -------
 
 Bug fixes:
 
-	- Corrected a problem in the setup.py (pyrexc script incorrectly
-		named).
-
-	- Updated the distutils extension to match changes in the
-		Pyrex compiler calling interface.
-
-	- Doing 'make clean' in Demos/callback was removing a little too
-		much (that's why cheesefinder.c kept disappearing).
+  - Corrected a problem in the setup.py (pyrexc script incorrectly
+    named).
+  
+  - Updated the distutils extension to match changes in the
+    Pyrex compiler calling interface.
+  
+  - Doing 'make clean' in Demos/callback was removing a little too
+    much (that's why cheesefinder.c kept disappearing).
 
 0.9.1
 -----
 
 Enhancements:
 
-	- A C method can now call an inherited C method by the usual
-		Python technique.
-		[Jiba <jiba at tuxfamily.org>]
-
-	- The __modname__ of a Python class is now set correctly.
-		[Paul Prescod <paul at prescod.net>]
-
-	- A MANIFEST.in file has been added to the distribution to
-		facilitate building rpms.
-		[contributed by Konrad Hinsen <hinsen at cnrs-orleans.fr>]
+  - A C method can now call an inherited C method by the usual
+    Python technique.
+    [Jiba <jiba at tuxfamily.org>]
+  
+  - The __modname__ of a Python class is now set correctly.
+    [Paul Prescod <paul at prescod.net>]
+  
+  - A MANIFEST.in file has been added to the distribution to
+    facilitate building rpms.
+    [contributed by Konrad Hinsen <hinsen at cnrs-orleans.fr>]
 
 Bug fixes:
 
   - Conditional code now generated to allow for the renaming of LONG_LONG
     to PY_LONG_LONG that occurred between Python 2.2 and 2.3.
 
-	- Header files referenced in cimported modules were not being included.
-		[Tom Popovich <tompopovich at aol.com>]
-
-	- References to C functions and variables in a cimported module were
-		not being recognised if made from within a local scope.
-		[Tom Popovich <tompopovich at aol.com>]
-
-	- Spurious declarations in code generated for a "finally" block.
-		[Brandon Long <blong at fiction.net>]
-
-	- Attempting to return a value from a __contains__ method didn't work.
-		[Andreas Kostyrka <andreas at mtg.co.at>]
-
-	- Incorrect code generated for an extension type with C methods
-		inheriting from a base type with no C methods.
-		[Robin Becker <robin at reportlab.com>]
-
-	- Failure to report an error if a C method was defined in the
-		implementation part of an extension type that was not declared
-		in the corresponding definition part. Documentation also updated
-		to explain that this is necessary.
-		[Jiba <jiba at tuxfamily.org>]
-
-	- Made it an error to forward-declare an extension type with
-		a different base class specification from its subsequent
-		definition.
-		[Jiba <jiba at tuxfamily.org>]
-
-	- C attributes of an extension type were not being propagated
-		through more than one level of inheritance.
-		[Jiba <jiba at tuxfamily.org>]
-
-	- If a garbage collection occurred early enough in the __new__
-		method of an extension type with Python-valued C attributes,
-		a crash could occur in its tp_traverse function.
-		[reported by Jiba <jiba at tuxfamily.org>]
-		[fix suggested by Paul Prescod <paul at prescod.net>]
-
-	- An empty vtable struct is no longer generated for extension
-		types with no C methods.
-		[Robin Becker <robin at reportlab.com>]
-
-	- Memory was leaked in the sq_item function of an extension
-		type with a __getitem__ method.
-		[Atsuo Ishimoto <ishimoto at gembook.org>]
-
-	- Code generated to work around a bug in some versions of Python
-		2.2 which fails to initialise the tp_free slot correctly in
-		some circumstances.
-		[Matthias Baas <baas at ira.uka.de>]
-
-	- Compiler crash when defining an extension type with a base
-		class specified by a dotted name.
-		[Alain Pointdexter <alainpoint at yahoo.fr>]
-
-	- Referencing an extension type defined in a cimported module
-		at run time did not work correctly.
-		[Alain Pointdexter <alainpoint at yahoo.fr>]
-
-	- Incorrect object struct code generated for an extension type
-		whose base class was defined in a .pxd file.
-		[Alain Pointdexter <alainpoint at yahoo.fr>]
-
-	- Redeclaring a type that wasn't previously an extension type
-		as an extension type caused a compiler crash.
-		[Scott Robinson <scott at tranzoa.com>]
-
-	- Incorrect code was generated for return statements in a
-		special method with no return value.
-		[Gary Bishop <gb at cs.unc.edu>]
-
-	- Single-line def statement did not work.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
-
+  - Header files referenced in cimported modules were not being included.
+    [Tom Popovich <tompopovich at aol.com>]
+  
+  - References to C functions and variables in a cimported module were
+    not being recognised if made from within a local scope.
+    [Tom Popovich <tompopovich at aol.com>]
+  
+  - Spurious declarations in code generated for a "finally" block.
+    [Brandon Long <blong at fiction.net>]
+  
+  - Attempting to return a value from a __contains__ method didn't work.
+    [Andreas Kostyrka <andreas at mtg.co.at>]
+  
+  - Incorrect code generated for an extension type with C methods
+    inheriting from a base type with no C methods.
+    [Robin Becker <robin at reportlab.com>]
+  
+  - Failure to report an error if a C method was defined in the
+    implementation part of an extension type that was not declared
+    in the corresponding definition part. Documentation also updated
+    to explain that this is necessary.
+    [Jiba <jiba at tuxfamily.org>]
+  
+  - Made it an error to forward-declare an extension type with
+    a different base class specification from its subsequent
+    definition.
+    [Jiba <jiba at tuxfamily.org>]
+  
+  - C attributes of an extension type were not being propagated
+    through more than one level of inheritance.
+    [Jiba <jiba at tuxfamily.org>]
+  
+  - If a garbage collection occurred early enough in the __new__
+    method of an extension type with Python-valued C attributes,
+    a crash could occur in its tp_traverse function.
+    [reported by Jiba <jiba at tuxfamily.org>]
+    [fix suggested by Paul Prescod <paul at prescod.net>]
+  
+  - An empty vtable struct is no longer generated for extension
+    types with no C methods.
+    [Robin Becker <robin at reportlab.com>]
+  
+  - Memory was leaked in the sq_item function of an extension
+    type with a __getitem__ method.
+    [Atsuo Ishimoto <ishimoto at gembook.org>]
+  
+  - Code generated to work around a bug in some versions of Python
+    2.2 which fails to initialise the tp_free slot correctly in
+    some circumstances.
+    [Matthias Baas <baas at ira.uka.de>]
+  
+  - Compiler crash when defining an extension type with a base
+    class specified by a dotted name.
+    [Alain Pointdexter <alainpoint at yahoo.fr>]
+  
+  - Referencing an extension type defined in a cimported module
+    at run time did not work correctly.
+    [Alain Pointdexter <alainpoint at yahoo.fr>]
+  
+  - Incorrect object struct code generated for an extension type
+    whose base class was defined in a .pxd file.
+    [Alain Pointdexter <alainpoint at yahoo.fr>]
+  
+  - Redeclaring a type that wasn't previously an extension type
+    as an extension type caused a compiler crash.
+    [Scott Robinson <scott at tranzoa.com>]
+  
+  - Incorrect code was generated for return statements in a
+    special method with no return value.
+    [Gary Bishop <gb at cs.unc.edu>]
+  
+  - Single-line def statement did not work.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
+  
 Modifications:
-
-	- Only the last pathname component of the .pyx file is reported in
-		backtraces now.
-		[Bryan Weingarten <bryan.weingarten at pobox.com>]
-
-	- Documentation corrected to remove the erroneous statement that
-		extension classes can have a __del__ method.
-		[Bryan Weingarten <bryan.weingarten at pobox.com>]
-
-	- Note added to documentation explaining that it is not possible
-		for an extension type's __new__ method to explicitly call the
-		inherited __new__ method.
-
-	- The version of Plex included with Pyrex is now installed
-		as a subpackage of the Pyrex package, rather than as a
-		top-level package, so as not to interfere with any other
-		version of Plex the user may have installed.
+  
+  - Only the last pathname component of the .pyx file is reported in
+    backtraces now.
+    [Bryan Weingarten <bryan.weingarten at pobox.com>]
+  
+  - Documentation corrected to remove the erroneous statement that
+    extension classes can have a __del__ method.
+    [Bryan Weingarten <bryan.weingarten at pobox.com>]
+  
+  - Note added to documentation explaining that it is not possible
+    for an extension type's __new__ method to explicitly call the
+    inherited __new__ method.
+  
+  - The version of Plex included with Pyrex is now installed
+    as a subpackage of the Pyrex package, rather than as a
+    top-level package, so as not to interfere with any other
+    version of Plex the user may have installed.
 
 0.9
 ---
 
 New features:
 
-	- Extension types can have properties. See the new "Properties"
-		section in the "Extension Types" page.
-
-	- An extension type can inherit from a builtin type or another
-		extension type. See "Subclassing" in the "Extension Types" page.
-
-	- Extension types can have C methods, which can be overridden
-		in derived extension types. See "C Methods" in the "Extension Types"
-		page.
+  - Extension types can have properties. See the new "Properties"
+    section in the "Extension Types" page.
+  
+  - An extension type can inherit from a builtin type or another
+    extension type. See "Subclassing" in the "Extension Types" page.
+  
+  - Extension types can have C methods, which can be overridden
+    in derived extension types. See "C Methods" in the "Extension Types"
+    page.
 
 Enhancements:
-
-	- Conversion is now performed between C long longs and Python
-		long integers without chopping to the size of a C long.
-		Also the Python PY_LONG_LONG type is now used for long longs
-		for greater portability.
+  
+  - Conversion is now performed between C long longs and Python
+    long integers without chopping to the size of a C long.
+    Also the Python PY_LONG_LONG type is now used for long longs
+    for greater portability.
 
 Bug fixes:
 
-	- Names were sometimes being generated that were insufficiently
-		unique in the presence of cimported declarations.
-
-	- Changed the way the included filename table is declared from
-		char *[] to char **, to stop MSVC from complaining about it
-		having an unknown size.
-		[Alexander A Naanou <alex_nanou at pochta.ru>]
-
-	- Second argument of assert statement was not being coerced
-		to a Python value.
-		[Francois Pinard   http://www.iro.umontreal.ca/~pinard]
-
-	- Return statement without value wasn't accepted in some
-		extension type special methods when it should have been.
-		[Francois Pinard   http://www.iro.umontreal.ca/~pinard]
-
-	- Attempting to call a non-function C value crashed the
-		compiler.
-		[John J Lee <jjl at pobox.com>]
-
-	- Functions declared as "except *" were not returning exceptions.
-		[John J Lee <jjl at pobox.com>]
-
-	- A syntax warning from Plex about assignment to None has
-		been eliminated.
-		[Gordon Williams <g_will at cyberus.ca>]
-
-	- Public function declaration with empty argument list was
-		producing (void) in .pxi file.
-		[Michael P. Dubner <dubnerm at mindless.com>]
-
-	- Incorrect error signalling code was being generated in the
-		__hash__ special method of an extension type.
-
+  - Names were sometimes being generated that were insufficiently 
+    unique in the presence of cimported declarations.
+  
+  - Changed the way the included filename table is declared from
+    char *[] to char **, to stop MSVC from complaining about it
+    having an unknown size.
+    [Alexander A Naanou <alex_nanou at pochta.ru>]
+  
+  - Second argument of assert statement was not being coerced
+    to a Python value.
+    [Francois Pinard   http://www.iro.umontreal.ca/~pinard]
+  
+  - Return statement without value wasn't accepted in some
+    extension type special methods when it should have been.
+    [Francois Pinard   http://www.iro.umontreal.ca/~pinard]
+
+  - Attempting to call a non-function C value crashed the
+    compiler.
+    [John J Lee <jjl at pobox.com>]
+  
+  - Functions declared as "except *" were not returning exceptions.
+    [John J Lee <jjl at pobox.com>]
+    
+  - A syntax warning from Plex about assignment to None has
+    been eliminated.
+    [Gordon Williams <g_will at cyberus.ca>]
+  
+  - Public function declaration with empty argument list was
+    producing (void) in .pxi file.
+    [Michael P. Dubner <dubnerm at mindless.com>]
+  
+  - Incorrect error signalling code was being generated in the
+    __hash__ special method of an extension type.
+    
 
 0.8.1
 -----
 
 Bug fixes:
 
-	- Names of structs, unions and enums in external header
-		files were getting mangled when they shouldn't have been.
-		[Norman Shelley <Norman.Shelley at motorola.com>]
-
-	- Modified distutils extension so that it will stop before
-		compiling the C file if the Pyrex compiler reports errors.
-		[John J Lee <jjl at pobox.com>]
-
+  - Names of structs, unions and enums in external header
+    files were getting mangled when they shouldn't have been.
+    [Norman Shelley <Norman.Shelley at motorola.com>]
+  
+  - Modified distutils extension so that it will stop before
+    compiling the C file if the Pyrex compiler reports errors.
+    [John J Lee <jjl at pobox.com>]
+    
 
 0.8
 ---
@@ -300,54 +410,54 @@
   	declaration must DECLARE THE MODULE from which the extension
   	type originates. See the new version of the "Extension Types"
   	documentation for details.
-
-  	This change was made to eliminate the need for Pyrex to be
-  	told the C name of the type object, or for the Pyrex module
+  	
+  	This change was made to eliminate the need for Pyrex to be 
+  	told the C name of the type object, or for the Pyrex module 
   	to be linked against the object code providing the type object.
-
+  
   	You will have to update any existing external extension type
   	declarations that you are using. I'm sorry about that, but it
   	was too hard to support both the old and new ways.
-
-  - Compile-time importing: A Pyrex module can now import declarations
-  	from another Pyrex module using the new 'cimport' statement. See
-  	the new section on "Sharing Declarations Between Pyrex Modules" in
+  
+  - Compile-time importing: A Pyrex module can now import declarations 
+  	from another Pyrex module using the new 'cimport' statement. See 
+  	the new section on "Sharing Declarations Between Pyrex Modules" in 
   	the documentation.
 
 Minor improvements:
 
-	- An error is reported if you declare a struct, union or
-		extension type using 'cdef' in one place and 'ctypedef'
-		in another.
-
-	- Struct, union and extension types can only be forward-
-		declared using 'cdef', not 'ctypedef' (otherwise invalid
-		C code would be generated).
-
-	- The 'global' statement can be used at the module level to
-		declare that a name is a module-level name rather than a
-		builtin. This can be used to access module attributes such
-		as __name__ that would otherwise be assumed to be builtins.
-		[Pat Maupin <pmaupin at speakeasy.net>]
-
-	- The 'assert' statement now accepts a second argument.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
+  - An error is reported if you declare a struct, union or
+    extension type using 'cdef' in one place and 'ctypedef'
+    in another.
+  
+  - Struct, union and extension types can only be forward-
+    declared using 'cdef', not 'ctypedef' (otherwise invalid 
+    C code would be generated).
+  
+  - The 'global' statement can be used at the module level to
+    declare that a name is a module-level name rather than a
+    builtin. This can be used to access module attributes such
+    as __name__ that would otherwise be assumed to be builtins.
+    [Pat Maupin <pmaupin at speakeasy.net>]
+  
+  - The 'assert' statement now accepts a second argument.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
 
 Bug fixes:
 
-	- When using Python 2.3, "True" or "False" could sometimes
-		turn up in generated code instead of "1" or "0".
-		[Adam Hixson <chixson at ou.edu>]
-
-	- Function return value not always converted to or from a
-		Python object when it should have been.
-
-	- Certain kinds of error in a function call expression
-		could crash the compiler.
-		["Edward C. Jones" <edcjones at erols.com>]
-
-	- Fixed memory leak in functions with * or ** args.
-		[Alexander A Naanou <alex_nanou at pochtamt.ru>]
+  - When using Python 2.3, "True" or "False" could sometimes
+    turn up in generated code instead of "1" or "0".
+    [Adam Hixson <chixson at ou.edu>]
+  
+  - Function return value not always converted to or from a
+    Python object when it should have been.
+  
+  - Certain kinds of error in a function call expression
+    could crash the compiler.
+    ["Edward C. Jones" <edcjones at erols.com>]
+  
+  - Fixed memory leak in functions with * or ** args.
+    [Alexander A Naanou <alex_nanou at pochtamt.ru>]
 
 
 0.7.1
@@ -355,86 +465,86 @@
 
 Bug fixes:
 
-	- Calling a function declared as returning an extension
-		type could crash the compiler.
-
-	- A function call with type errors in the argument list
-		could crash the compiler.
-
-	- An 'else' clause on a for-from statement could crash
-		the compiler.
-
-	- Incorrect casting code was generated when a generic
-		object argument of a special method was declared as
-		being of an extension type.
-		[Phillip J. Eby <pje at telecommunity.com>]
-
-	- A blank line that couldn't be interpreted wholly as
-		a valid indentation sequence caused a syntax error.
-		In particular, a formfeed character on an otherwise
-		blank line wasn't accepted.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
-
-	- Parallel assignments were incorrectly optimised.
+  - Calling a function declared as returning an extension
+    type could crash the compiler.
 
-	- A bare tuple constructor with an extra comma at the
-		end of a line caused a syntax error.
+  - A function call with type errors in the argument list 
+    could crash the compiler.
+  
+  - An 'else' clause on a for-from statement could crash
+    the compiler.
+  
+  - Incorrect casting code was generated when a generic
+    object argument of a special method was declared as
+    being of an extension type.
+    [Phillip J. Eby <pje at telecommunity.com>]
+  
+  - A blank line that couldn't be interpreted wholly as
+    a valid indentation sequence caused a syntax error.
+    In particular, a formfeed character on an otherwise
+    blank line wasn't accepted. 
+    [Francois Pinard <pinard at iro.umontreal.ca>]
+
+  - Parallel assignments were incorrectly optimised.
+  
+  - A bare tuple constructor with an extra comma at the
+    end of a line caused a syntax error.
 
 0.7
 ---
 
 New features:
 
-	- Attributes of extension types can be exposed to Python
-		code, either read/write or read-only.
+  - Attributes of extension types can be exposed to Python
+    code, either read/write or read-only.
 
-	- Different internal and external names can be specified
-		for C entities.
-
-	- None is a compile-time constant, and more efficient code
-		is generated to reference it.
-
-	- Command line options for specifying directories to
-		search for include files.
+  - Different internal and external names can be specified
+    for C entities.
+  
+  - None is a compile-time constant, and more efficient code
+    is generated to reference it.
+  
+  - Command line options for specifying directories to
+    search for include files.
 
 Enhancements:
 
-	- More efficient code is generated for access to Python
-		valued C attributes of extension types.
-
-	- Cosmetic code improvement: Less casting back and forth
-		between extension types and PyObject * when referencing
-		C members of the object struct.
-
-	- C arguments and variables declared as an extension type
-		can take the value None.
-
-	- Form feed characters are accepted as whitespace.
-
-	- Function names in tracebacks are qualified with
-		module name and class name.
+  - More efficient code is generated for access to Python
+    valued C attributes of extension types.
+  
+  - Cosmetic code improvement: Less casting back and forth
+    between extension types and PyObject * when referencing
+    C members of the object struct.
+  
+  - C arguments and variables declared as an extension type
+    can take the value None.
+  
+  - Form feed characters are accepted as whitespace.
+  
+  - Function names in tracebacks are qualified with
+    module name and class name.
 
 Bug fixes:
 
-	- A sufficiently complex expression in a boolean context
-		could cause code to be generated twice for the same
-		subexpression.
-
-	- Incorrect casting code was generated when passing an
-		extension type to a function expecting a generic Python
-		object.
-
-	- Executable statements are now disallowed inside a
-		cdef class block (previously they silently caused
-		crazy C code to be generated).
-
-	- Tracebacks should now report the correct filename for
-		functions defined in files included with the 'include'
-		statement.
-
-	- The documentation incorrectly claimed that an extension
-		type can't have a __del__ method. In fact, it can, and
-		it behaves as expected.
+  - A sufficiently complex expression in a boolean context
+    could cause code to be generated twice for the same 
+    subexpression.
+  
+  - Incorrect casting code was generated when passing an
+    extension type to a function expecting a generic Python
+    object.
+  
+  - Executable statements are now disallowed inside a
+    cdef class block (previously they silently caused
+    crazy C code to be generated).
+  
+  - Tracebacks should now report the correct filename for
+    functions defined in files included with the 'include'
+    statement.
+  
+  - The documentation incorrectly claimed that an extension
+    type can't have a __del__ method. In fact, it can, and
+    it behaves as expected.
 
 
 0.6.1
@@ -442,7 +552,7 @@
 
 Bug fixes:
 
-	- Fixed broken distutils extension.
+  - Fixed broken distutils extension.
 
 
 
@@ -451,102 +561,102 @@
 
 New features:
 
-	- Command line options for reporting version number,
-		requesting a listing file and specifying the name of
-		the generated C file.
-
-	- An 'include' statement allows inclusion of declarations
-		from other Pyrex source files.
-
-	- If there are any public declarations, a Pyrex include
-		file is generated (as well as a .h file) containing
-		declarations for them.
-
-	- Extension types can be declared public, so their C
-		attributes are visible to other Pyrex and C code.
-
-	- Try-except statements can now have an 'else' clause.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
-
-	- Multiple simple statements can be placed on one line
-		separated by semicolons.
-
-	- A suite consisting of a simple statement list can now
-		be placed on the same line after the colon in most
-		cases.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
-
-	- The automatic coercion of a C string to a C char has
-		been removed (it proved to be too error-prone).
-		Instead, there is a new form of literal for C
-		character constants: c'X'
-
-	- The __get__ special method (used by descriptor objects)
-		now allows for the possibility of the 2nd or 3rd
-		arguments being NULL. Also the __set__ method has been
-		split into two methods, __set__ and __delete__.
-		[Phillip J. Eby <pje at telecommunity.com>]
-
-Bug fixes:
-
-	- Values unpacked into a non-Python destination variable
-		were not being converted before assignment.
-		[Gareth Watts <gareth at omnipotent.net>]
-
-	- Hex constants greater than 0x7fffffff caused compiler
-		to crash. [Gareth Watts <gareth at omnipotent.net>]
-
-	- Type slots are no longer statically initialised with
-		extern function pointers, to avoid problems with
-		some compilers. The hack in the distutils extension
-		to work around this by compiling as C++ has been
-		disabled. [Phillip J. Eby <pje at telecommunity.com>]
-
-	- Fixed several more instances of the error-reporting
-		routine being called with arguments in the wrong
-		order. Hoping I've *finally* got all of them now...
-
-	- Nested for-from loops used the same control variable.
-		[Sebastien de Menten <sdementen at hotmail.com>]
-
-	- Fixed some other error message related bugs.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
-
-	- Assigning to slice didn't work.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
-
-	- Temp variables were being declared as extension
-		types and then being assigned PyObject *'s. All
-		Python temp vars are now declared as PyObject *.
-		[Francois Pinard <pinard at iro.umontreal.ca>]
+  - Command line options for reporting version number,
+    requesting a listing file and specifying the name of
+    the generated C file.
+
+  - An 'include' statement allows inclusion of declarations
+    from other Pyrex source files.
+  
+  - If there are any public declarations, a Pyrex include
+    file is generated (as well as a .h file) containing 
+    declarations for them.
+  
+  - Extension types can be declared public, so their C
+    attributes are visible to other Pyrex and C code.
+  
+  - Try-except statements can now have an 'else' clause.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
+  
+  - Multiple simple statements can be placed on one line
+    separated by semicolons.
+  
+  - A suite consisting of a simple statement list can now
+    be placed on the same line after the colon in most
+    cases.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
+
+  - The automatic coercion of a C string to a C char has
+    been removed (it proved to be too error-prone).
+    Instead, there is a new form of literal for C
+    character constants: c'X'
+
+  - The __get__ special method (used by descriptor objects)
+    now allows for the possibility of the 2nd or 3rd
+    arguments being NULL. Also the __set__ method has been
+    split into two methods, __set__ and __delete__.
+    [Phillip J. Eby <pje at telecommunity.com>]
+  
+Bug fixes:
+
+  - Values unpacked into a non-Python destination variable
+    were not being converted before assignment.
+    [Gareth Watts <gareth at omnipotent.net>]
+
+  - Hex constants greater than 0x7fffffff caused compiler
+    to crash. [Gareth Watts <gareth at omnipotent.net>]
+  
+  - Type slots are no longer statically initialised with
+    extern function pointers, to avoid problems with
+    some compilers. The hack in the distutils extension
+    to work around this by compiling as C++ has been
+    disabled. [Phillip J. Eby <pje at telecommunity.com>]
+  
+  - Fixed several more instances of the error-reporting
+    routine being called with arguments in the wrong
+    order. Hoping I've *finally* got all of them now...
+  
+  - Nested for-from loops used the same control variable.
+    [Sebastien de Menten <sdementen at hotmail.com>]
+  
+  - Fixed some other error message related bugs.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
+  
+  - Assigning to slice didn't work.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
+  
+  - Temp variables were being declared as extension
+    types and then being assigned PyObject *'s. All
+    Python temp vars are now declared as PyObject *.
+    [Francois Pinard <pinard at iro.umontreal.ca>]
 
 0.5
 ---
 
 Bug fixes:
 
-	- Algorithm for allocating temp variables redesigned
-		to fix various errors concerning temp
-		variable re-use.
-		[Mark Rowe <bdash at gmx.net>]
-
-	- Memory leak occured sometimes when an implicit
-		type test was applied to the result of an
-		expression.
-		[christoph.wiedemann at daimlerchrysler.com]
-
-	- __set__ method of extension types had wrong
-		signature.
-		[Josh Littlefield <joshl at cisco.com>]
-
+  - Algorithm for allocating temp variables redesigned
+    to fix various errors concerning temp
+    variable re-use.
+    [Mark Rowe <bdash at gmx.net>]
+
+  - Memory leak occured sometimes when an implicit
+    type test was applied to the result of an 
+    expression.
+    [christoph.wiedemann at daimlerchrysler.com]
+  
+  - __set__ method of extension types had wrong
+    signature.
+    [Josh Littlefield <joshl at cisco.com>]
+    
 0.4.6
 -----
 
 Bug fixes:
 
-	- Indexing multi-dimensional C arrays didn't
-		work.
-		[Gary Dietachmayer <gary.dietachmayer at dsto.defence.gov.au>]
+  - Indexing multi-dimensional C arrays didn't
+    work.
+    [Gary Dietachmayer <gary.dietachmayer at dsto.defence.gov.au>]
 
 
 0.4.5
@@ -554,48 +664,48 @@
 
 New features:
 
-	- There is now a 'public' declaration for
-		making Pyrex-defined variables and functions
-		available to external C code. A .h file is
-		also generated if there are any public
-		declarations.
+  - There is now a 'public' declaration for
+    making Pyrex-defined variables and functions
+    available to external C code. A .h file is
+    also generated if there are any public
+    declarations.
 
 Enhancements:
 
-	- Defining __len__/__getitem__ methods in an
-		extension class fills sq_length/sq_item slots
-		as well as mp_length/mp_subscript.
-		[Matthias Baas <baas at ira.uka.de>]
-
-	- The Distutils extension now allows .c files
-		to be incorporated along with .pyx files.
-		[Modification to Distutils extension contributed
-		by Darrell Gallion <dgallion1 at yahoo.com>.]
+  - Defining __len__/__getitem__ methods in an
+    extension class fills sq_length/sq_item slots
+    as well as mp_length/mp_subscript.
+    [Matthias Baas <baas at ira.uka.de>]
+  
+  - The Distutils extension now allows .c files
+    to be incorporated along with .pyx files.
+    [Modification to Distutils extension contributed
+    by Darrell Gallion <dgallion1 at yahoo.com>.]
 
 Bug fixes:
 
-	- Float literals without a decimal point
-		work again now.
-		[Mike Rovner <mike at bindkey.com>]
-		[Peter Lepage <gpl at mail.lns.cornell.edu>]
-
-	- Compiler crashed if exception value didn't
-		match function return type.
-		[Michael JasonSmith <mpj17 at smtphost.canterbury.ac.nz>]
-
-	- The setup.py file should now install the
-		Lexicon.pickle file in the right place.
-		[Patch supplied by David M. Cooke
-		<cookedm at arbutus.physics.mcmaster.ca>]
-
-	- Compiler crashed when compiling a C function that
-		returned an extension type.
-		[David M. Cooke
-		<cookedm at arbutus.physics.mcmaster.ca>]
-
-	- Anonymous enum types did not have C code
-		suppressed inside an extern-from block.
-		[Matthew Mueller <donut at azstarnet.com>]
+  - Float literals without a decimal point
+    work again now.
+    [Mike Rovner <mike at bindkey.com>]
+    [Peter Lepage <gpl at mail.lns.cornell.edu>]
+  
+  - Compiler crashed if exception value didn't
+    match function return type.
+    [Michael JasonSmith <mpj17 at smtphost.canterbury.ac.nz>]
+  
+  - The setup.py file should now install the
+    Lexicon.pickle file in the right place.
+    [Patch supplied by David M. Cooke 
+    <cookedm at arbutus.physics.mcmaster.ca>]
+    
+  - Compiler crashed when compiling a C function that
+    returned an extension type.
+    [David M. Cooke 
+    <cookedm at arbutus.physics.mcmaster.ca>]
+  
+  - Anonymous enum types did not have C code
+    suppressed inside an extern-from block.
+    [Matthew Mueller <donut at azstarnet.com>]
 
 
 0.4.4
@@ -603,46 +713,46 @@
 
 Enhancements:
 
-	- Tracebacks now extend into Pyrex function
-		calls and show line numbers in the Pyrex
-		source file.
-
-	- Syntax for float literals made more lenient
-		(no longer requires digits both before and
-		after the point).
-		[Peter Lepage <gpl at mail.lns.cornell.edu>]
-
-	- Method calls can be made on string literals
-		(e.g. ",".join(x)).
-		[pedro_rodriguez at club-internet.fr]
+  - Tracebacks now extend into Pyrex function
+    calls and show line numbers in the Pyrex
+    source file.
+
+  - Syntax for float literals made more lenient
+    (no longer requires digits both before and
+    after the point).
+    [Peter Lepage <gpl at mail.lns.cornell.edu>]
+  
+  - Method calls can be made on string literals
+    (e.g. ",".join(x)).
+    [pedro_rodriguez at club-internet.fr]
 
 Bug fixes:
 
-	- Incorrect refcount code generated when a
-		Python function needing argument type tests
-		had local Python variables.
-		[Matthias Baas <baas at ira.uka.de>]
-
-	- 'self' parameter of __getitem__ method of
-		extension type had wrong implicit type.
-		[Peter Lepage <gpl at mail.lns.cornell.edu>]
-
-	- Repaired breakage introduced by trying to
-		allow an empty parameter list to be written
-		as (void). No longer attempting to allow
-		this (too hard to parse correctly).
-		[Peter Lepage <gpl at mail.lns.cornell.edu>]
-
-	- Found bug in Plex 1.1.2 which was the *real*
-		cause of the two-newlines-in-a-row problem.
-		Removed the Opt(Eol)+Str("\n") hacks in
-		the scanner which were working around this
-		before.
-		[pedro_rodriguez at club-internet.fr]
-
-	- __call__ special method of extension types
-		had wrong signature.
-		[Peter Lepage <gpl at mail.lns.cornell.edu>]
+  - Incorrect refcount code generated when a
+    Python function needing argument type tests
+    had local Python variables.
+    [Matthias Baas <baas at ira.uka.de>]
+  
+  - 'self' parameter of __getitem__ method of
+    extension type had wrong implicit type.
+    [Peter Lepage <gpl at mail.lns.cornell.edu>]
+  
+  - Repaired breakage introduced by trying to
+    allow an empty parameter list to be written
+    as (void). No longer attempting to allow
+    this (too hard to parse correctly).
+    [Peter Lepage <gpl at mail.lns.cornell.edu>]
+  
+  - Found bug in Plex 1.1.2 which was the *real*
+    cause of the two-newlines-in-a-row problem.
+    Removed the Opt(Eol)+Str("\n") hacks in
+    the scanner which were working around this
+    before.
+    [pedro_rodriguez at club-internet.fr]
+  
+  - __call__ special method of extension types
+    had wrong signature.
+    [Peter Lepage <gpl at mail.lns.cornell.edu>]
 
 
 0.4.3
@@ -650,26 +760,26 @@
 
 New language features:
 
-	- For-from loop for iterating over integer
-		ranges, using pure C loop where possible.
+  - For-from loop for iterating over integer
+    ranges, using pure C loop where possible.
 
 Enhancements:
 
-	- sizeof() can now be applied to types as
-		well as variables.
-
-	- Improved handling of forward-declared
-		extension types.
+  - sizeof() can now be applied to types as
+    well as variables.
+  
+  - Improved handling of forward-declared
+    extension types.
 
 Bug fixes:
 
-	- Two newlines in a row in a triple quoted
-		string caused a parse error on some
-		platforms.
-		[Matthias Baas <baas at ira.uka.de>]
-
-	- Fixed problem with break and continue in
-		the else-clause of a loop.
+  - Two newlines in a row in a triple quoted
+    string caused a parse error on some
+    platforms.
+    [Matthias Baas <baas at ira.uka.de>]
+  
+  - Fixed problem with break and continue in
+    the else-clause of a loop.
 
 
 0.4.2
@@ -677,118 +787,118 @@
 
 New language features:
 
-	- C functions can be declared as having an
-		exception return value, which is checked
-		whenever the function is called. If an
-		exception is detected inside a C function
-		for which no exception value is declared,
-		a warning message is printed and the
-		exception is cleared.
-
-	- Cascaded assignments (i.e. a = b = c
-		are now supported.
-
-	- Anonymous enum declarations are allowed,
-		for when you just want to declare constants.
-
-	- The C types "long long" and "long double"
-		are now understood. Also, "int" is optional
-		after "short" or "long".
-
+  - C functions can be declared as having an
+    exception return value, which is checked
+    whenever the function is called. If an
+    exception is detected inside a C function
+    for which no exception value is declared,
+    a warning message is printed and the
+    exception is cleared.
+  
+  - Cascaded assignments (i.e. a = b = c 
+    are now supported.
+  
+  - Anonymous enum declarations are allowed,
+    for when you just want to declare constants.
+  
+  - The C types "long long" and "long double"
+    are now understood. Also, "int" is optional
+    after "short" or "long".
+  
 Enhancements:
 
-	- A * argument in a function call can now be
-		any sequence, not just a tuple.
-
-	-	A C char* or char[] will be turned into
-		a char by taking its first character if
-		used in a context where a char is required,
-		thus allowing a string literal to be used as
-		a char literal.
-
-	- C string * C int or vice versa is now
-		interpreted as Python string replication.
+  - A * argument in a function call can now be
+    any sequence, not just a tuple.
 
-	- Function arguments are checked for void or
-		incomplete type.
+  -	A C char* or char[] will be turned into
+    a char by taking its first character if
+    used in a context where a char is required,
+    thus allowing a string literal to be used as 
+    a char literal.
+  
+  - C string * C int or vice versa is now
+    interpreted as Python string replication.
+  
+  - Function arguments are checked for void or
+    incomplete type.
 
 Bug fixes:
 
-	- Non-external extension types show up in the
-		module dict once more (this got broken in
-		0.4.1).
-
-	- A spurious decref has been removed from the
-		runtime support code for the "import" statement.
-		Hopefully this will prevent the crashes some
-		people have been experiencing when importing
-		builtin modules.
-		[Mathew Yeates <mathew at comma.jpl.nasa.gov>]
+  - Non-external extension types show up in the
+    module dict once more (this got broken in
+    0.4.1).
+  
+  - A spurious decref has been removed from the
+    runtime support code for the "import" statement.
+    Hopefully this will prevent the crashes some
+    people have been experiencing when importing
+    builtin modules.
+    [Mathew Yeates <mathew at comma.jpl.nasa.gov>]
 
 0.4.1
 -----
 
 New language features:
 
-	- "ctypedef struct/union/enum/class" statements
-		added, for use in extern-from blocks when a
-		header file uses a ctypedef to declare a
-		tagless struct, union or enum type.
-
-	- "pass" allowed in an extern-from block.
-
-	- "cdef extern from *" for when you don't want
-	  to specify an include file name.
-
-	- Argument names may be omitted in function
-	  signatures when they're not needed.
-
-	- New reserved word NULL for the null C pointer.
+  - "ctypedef struct/union/enum/class" statements
+    added, for use in extern-from blocks when a
+    header file uses a ctypedef to declare a
+    tagless struct, union or enum type.
+  
+  - "pass" allowed in an extern-from block.
+  
+  - "cdef extern from *" for when you don't want
+    to specify an include file name.
+  
+  - Argument names may be omitted in function
+    signatures when they're not needed.
+  
+  - New reserved word NULL for the null C pointer.
 
 Compiler enhancements:
-
-	- Lexicon is now picked in binary format, so
-		startup should be much faster on slower
-		machines.
-
-	- If Pyrex decides to rebuild the lexicon and
-		then finds that it can't write a pickle file,
-		it now prints a warning and carries on
-		instead of crashing.
-
-	- Chat about hash codes and lexicon pickling
-		now turned off by default except when creating
-		a new lexicon (which ought never happen now
-		unless you change the scanner).
+  
+  - Lexicon is now picked in binary format, so
+    startup should be much faster on slower
+    machines.
+  
+  - If Pyrex decides to rebuild the lexicon and
+    then finds that it can't write a pickle file,
+    it now prints a warning and carries on
+    instead of crashing.
+  
+  - Chat about hash codes and lexicon pickling
+    now turned off by default except when creating
+    a new lexicon (which ought never happen now
+    unless you change the scanner).
 
 Bug fixes:
 
-	- Modified the runtime support code for "import"
-		statements, hopefully fixing problem with using
-		a Pyrex module in conjunction with py2exe.
-
-	- DL_EXPORT now used in both the prototype and
-	  definition of the module init function.
-
-	- Exception state is now saved and restored around
-	  calls to an extension type __dealloc__ method,
-	  to avoid screwing up if the object is deallocated
-	  while an exception is being propagated.
-
-	- Making an attribute reference to a method of
-	  an extension type caused a compiler crash.
-
-	- Doc string in new-style class definition
-	  caused a run-time error.
-
-	- Insufficient parentheses were put around C type
-	  casts.
-
-	- Constructors for extension types are now read-only
-	  C global variables instead of entries in the
-	  module dict. This change was needed to prevent
-	  Numeric from blowing up due to touching its
-	  typeobject before import_numeric() could be called.
+  - Modified the runtime support code for "import"
+    statements, hopefully fixing problem with using
+    a Pyrex module in conjunction with py2exe.
+  
+  - DL_EXPORT now used in both the prototype and
+    definition of the module init function.
+  
+  - Exception state is now saved and restored around
+    calls to an extension type __dealloc__ method,
+    to avoid screwing up if the object is deallocated
+    while an exception is being propagated.
+  
+  - Making an attribute reference to a method of
+    an extension type caused a compiler crash.
+  
+  - Doc string in new-style class definition
+    caused a run-time error.
+  
+  - Insufficient parentheses were put around C type 
+    casts.
+
+  - Constructors for extension types are now read-only
+    C global variables instead of entries in the
+    module dict. This change was needed to prevent
+    Numeric from blowing up due to touching its
+    typeobject before import_numeric() could be called.
 
 0.4
 ---
@@ -805,7 +915,7 @@
       - Helping deal with types such as "size_t"
       - Helping deal with functions defined as
         macros
-
+  
   - Access to internals of pre-existing extension
     types is now possible by placing an extension
     type declaration inside a "cdef extern from"
@@ -817,14 +927,14 @@
     wrong number of args to certain special
     methods of extension types.
     [Mitch Chapman <mitchchapman at earthlink.net>]
-
+  
   - Compile-time crash when defining an extension
     type with a __hash__ method.
 
 Minor enhancements:
 
   - Hashing of the scanner source file made more
-    platform-independent, making spurious regeneration
+    platform-independent, making spurious regeneration 
     of the pickle less likely.
 
 
@@ -835,11 +945,11 @@
 
   - Runtime crash when using * or ** args in
     a method of an extension type fixed.
-		[Matthew Mueller <donut at azstarnet.com>]
-
-	- Compiler crash when using default argument
-	  values in a method of a Python class.
-	  [Mike Rovner <mike at bindkey.com>]
+    [Matthew Mueller <donut at azstarnet.com>]
+  
+  - Compiler crash when using default argument
+    values in a method of a Python class.
+    [Mike Rovner <mike at bindkey.com>]
 
 Enhancements:
 
@@ -849,12 +959,12 @@
     "initialiser is not constant" problems experienced
     on Windows.
     [Marek Baczek <imbaczek at poczta.fm>]
-
+  
   - On Windows, __declspec(dllexport) is now used for
     the module init func declaration (or should be --
     I haven't tested this).
     [Marek Baczek <imbaczek at poczta.fm>]
-
+  
   - The compiler shouldn't attempt to rewrite the
     Lexicon.pickle file unless the source has been
     changed (hashing is used now instead of comparing
@@ -981,7 +1091,7 @@
 
 * Extension types can now be forward-declared.
 
-* All permutations of (non-Unicode) string literals
+* All permutations of (non-Unicode) string literals 
 and escape codes should work now.
 
 * Hex and octal integer literals.
@@ -1032,13 +1142,13 @@
     - Class definitions are only allowed at the top
       level of a module, not inside a control structure
       or function or another class definition.
-
+    
     - Assigning a Pyrex-defined Python function to a
       class attribute outside of the class definition
-      will not create a method (because it's not an
+      will not create a method (because it's not an 
       interpreted Python function and therefore
       won't trigger the bound-method creation magic).
-
+    
     - The __metaclass__ mechanism and the creation of
       new-style classes is not (yet) supported.
 

Modified: lxml/pyrex/Demos/callback/cheese.pyx
==============================================================================
--- lxml/pyrex/Demos/callback/cheese.pyx	(original)
+++ lxml/pyrex/Demos/callback/cheese.pyx	Tue Apr 25 15:43:21 2006
@@ -3,11 +3,11 @@
 #
 
 cdef extern from "cheesefinder.h":
-	ctypedef void (*cheesefunc)(char *name, void *user_data)
-	void find_cheeses(cheesefunc user_func, void *user_data)
+    ctypedef void (*cheesefunc)(char *name, void *user_data)
+    void find_cheeses(cheesefunc user_func, void *user_data)
 
 def find(f):
-	find_cheeses(callback, <void*>f)
-	
+    find_cheeses(callback, <void*>f)
+    
 cdef void callback(char *name, void *f):
-	(<object>f)(name)
+    (<object>f)(name)

Modified: lxml/pyrex/Demos/primes.pyx
==============================================================================
--- lxml/pyrex/Demos/primes.pyx	(original)
+++ lxml/pyrex/Demos/primes.pyx	Tue Apr 25 15:43:21 2006
@@ -1,18 +1,18 @@
 def primes(int kmax):
-	cdef int n, k, i
-	cdef int p[1000]
-	result = []
-	if kmax > 1000:
-		kmax = 1000
-	k = 0
-	n = 2
-	while k < kmax:
-		i = 0
-		while i < k and n % p[i] <> 0:
-			i = i + 1
-		if i == k:
-			p[k] = n
-			k = k + 1
-			result.append(n)
-		n = n + 1
-	return result
+    cdef int n, k, i
+    cdef int p[1000]
+    result = []
+    if kmax > 1000:
+        kmax = 1000
+    k = 0
+    n = 2
+    while k < kmax:
+        i = 0
+        while i < k and n % p[i] <> 0:
+            i = i + 1
+        if i == k:
+            p[k] = n
+            k = k + 1
+            result.append(n)
+        n = n + 1
+    return result

Modified: lxml/pyrex/Doc/FAQ.html
==============================================================================
--- lxml/pyrex/Doc/FAQ.html	(original)
+++ lxml/pyrex/Doc/FAQ.html	Tue Apr 25 15:43:21 2006
@@ -1 +1,77 @@
-<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
            
  <meta http-equiv="Content-Type"
 content="text/html; charset=iso-8859-1">
            
  <meta name="GENERATOR"
 content="Mozilla/4.51 (Macintosh; I; PPC) [Netscape]">
  <title>FAQ.html</title>
    
</head>
  <body>
      
<center>   
<h1>    
<hr width="100%">Pyrex FAQ   
<hr width="100%"></h1>
  </center>
      
<h2> Contents</h2>
      
<ul>
   <li> <b><a href="#CallCAPI">How do I call Python/C API routines?</a></b></li>
    <li> <b><a href="#NullBytes">How do I convert a C string containing null
 bytes to a Python string?</a></b></li>
    <li> <b><a href="#NumericAccess">How do I access the data inside a Numeric
 array object?</a></b></li>
  <li><b><a href="#Rhubarb">Pyrex says my extension type object has no attribute
'rhubarb', but I know it does. What gives?</a></b><br>
    <br>
  </li>
     
</ul>
      
<hr width="100%">   
<h2> <a name="CallCAPI"></a>How do I call Python/C API routines?</h2>
   Declare them as C functions inside a <tt>cdef extern from</tt> block.
Use  the type name <tt>object</tt> for any parameters and return types which 
are Python object references. Don't use the word <tt>const</tt> anywhere. 
Here is an example which defines and uses the <tt>PyString_FromStringAndSize</tt> 
 routine:   
<blockquote><tt>cdef extern from "Python.h":</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; object PyString_FromStringAndSize(char *, int)</tt> 
      
  <p><tt>cdef char buf[42]</tt> <br>
    <tt>my_string = PyString_FromStringAndSize(buf, 42)</tt></p>
  </blockquote>
      
<h2> <a name="NullBytes"></a>How do I convert a C string containing null
bytes to a Python string?</h2>
   Put in a declaration for the <tt>PyString_FromStringAndSize</tt> API routine 
 and use that<tt>.</tt> See <a href="#CallCAPI">How do I call Python/C API 
 routines?</a>   
<h2> <a name="NumericAccess"></a>How do I access the data inside a Numeric
 array object?</h2>
   Use a <tt>cdef extern from</tt> block to include the Numeric header file 
 and declare the array object as an external extension type. The following 
 code illustrates how to do this:   
<blockquote><tt>cdef extern from "Numeric/arrayobject.h":</tt>       
  <p><tt>&nbsp;&nbsp;&nbsp; struct PyArray_Descr:</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int type_num, elsize</tt> 
   <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; char type</tt> </p>
       
  <p><tt>&nbsp;&nbsp;&nbsp; ctypedef class Numeric.ArrayType [object PyArrayObject]</tt><tt>:</tt> 
   <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef char *data</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int nd</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int *dimensions,
*strides</tt>    <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef object base</tt>
  <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef PyArray_Descr *descr</tt> 
   <br>
    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int flags<br>
    </tt></p>
</blockquote>
<p>For more information about external extension types, see the <a
 href="extension_types.html#ExternalExtTypes">"External Extension Types"</a>
section of the <a href="extension_types.html">"Extension Types"</a> documentation
page.<br>
<tt>    </tt> </p>
    
<h2><a name="Rhubarb"></a>Pyrex says my extension type object has no attribute
'rhubarb', but I know it does. What gives?</h2>
You're probably trying to access it through a reference which Pyrex thinks
is a generic Python object. You need to tell Pyrex that it's a reference
to your extension type by means of a declaration,<br>
for example,<br>
<blockquote><tt>cdef class Vegetables:</tt><br>
  <tt>&nbsp; &nbsp; cdef int rhubarb</tt><br>
  <br>
  <tt>...</tt><br>
  <tt>cdef Vegetables veg</tt><br>
  <tt>veg.rhubarb = 42</tt><br>
</blockquote>
Also see the <a href="extension_types.html#ExtTypeAttrs">"Attributes"</a>
section of the <a
 href="file:///Macintosh%20HD/Local/Projects/C/Pyrex/Doc/extension_types.html">"Extension
Types"</a> documentation page.<br>
--- 
</body>
</html>
\ No newline at end of file
+<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
+<html><head>
+              <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
+              <meta name="GENERATOR" content="Mozilla/4.51 (Macintosh; I; PPC) [Netscape]"><title>FAQ.html</title></head>
+
+<body>
+      <center>   <h1>    <hr width="100%">Pyrex FAQ   
+<hr width="100%"></h1>
+  </center>
+      <h2> Contents</h2>
+      <ul>
+   <li> <b><a href="#CallCAPI">How do I call Python/C API routines?</a></b></li>
+    <li> <b><a href="#NullBytes">How do I convert a C string containing null
+ bytes to a Python string?</a></b></li>
+    <li> <b><a href="#NumericAccess">How do I access the data inside a Numeric
+ array object?</a></b></li>
+  <li><b><a href="#Rhubarb">Pyrex says my extension type object has no attribute
+'rhubarb', but I know it does. What gives?</a></b></li><li><a style="font-weight: bold;" href="#Quack">Python says my extension type has no method called 'quack', but I know it does. What gives?</a><br>
+  </li>
+
+     </ul>
+      <hr width="100%">   <h2> <a name="CallCAPI"></a>How do I call Python/C API routines?</h2>
+   Declare them as C functions inside a <tt>cdef extern from</tt> block.
+Use  the type name <tt>object</tt> for any parameters and return types which 
+are Python object references. Don't use the word <tt>const</tt> anywhere. 
+Here is an example which defines and uses the <tt>PyString_FromStringAndSize</tt>  routine:   
+<blockquote><tt>cdef extern from "Python.h":</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; object PyString_FromStringAndSize(char *, int)</tt>         <p><tt>cdef char buf[42]</tt> <br>
+    <tt>my_string = PyString_FromStringAndSize(buf, 42)</tt></p>
+  </blockquote>
+      <h2> <a name="NullBytes"></a>How do I convert a C string containing null
+bytes to a Python string?</h2>
+   Put in a declaration for the <tt>PyString_FromStringAndSize</tt> API routine 
+ and use that<tt>.</tt> See <a href="#CallCAPI">How do I call Python/C API 
+ routines?</a>   <h2> <a name="NumericAccess"></a>How do I access the data inside a Numeric
+ array object?</h2>
+   Use a <tt>cdef extern from</tt> block to include the Numeric header file 
+ and declare the array object as an external extension type. The following 
+ code illustrates how to do this:   
+<blockquote><tt>cdef extern from "Numeric/arrayobject.h":</tt>         <p><tt>&nbsp;&nbsp;&nbsp; struct PyArray_Descr:</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int type_num, elsize</tt>    <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; char type</tt> </p>
+         <p><tt>&nbsp;&nbsp;&nbsp; ctypedef class Numeric.ArrayType [object PyArrayObject]</tt><tt>:</tt>    <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef char *data</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int nd</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int *dimensions,
+*strides</tt>    <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef object base</tt>
+  <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef PyArray_Descr *descr</tt>    <br>
+    <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int flags<br>
+    </tt></p>
+</blockquote>
+<p>For more information about external extension types, see the <a href="extension_types.html#ExternalExtTypes">"External Extension Types"</a>
+section of the <a href="extension_types.html">"Extension Types"</a> documentation
+page.<br>
+<tt>    </tt> </p>
+    <h2><a name="Rhubarb"></a>Pyrex says my extension type object has no attribute
+'rhubarb', but I know it does. What gives?</h2>
+You're probably trying to access it through a reference which Pyrex thinks
+is a generic Python object. You need to tell Pyrex that it's a reference
+to your extension type by means of a declaration,<br>
+for example,<br>
+<blockquote><tt>cdef class Vegetables:</tt><br>
+  <tt>&nbsp; &nbsp; cdef int rhubarb</tt><br>
+  <br>
+  <tt>...</tt><br>
+  <tt>cdef Vegetables veg</tt><br>
+  <tt>veg.rhubarb = 42</tt><br>
+</blockquote>
+Also see the <a href="extension_types.html#ExtTypeAttrs">"Attributes"</a>
+section of the <a href="extension_types.html">"Extension
+Types"</a> documentation page.<br>
+<h2><a name="Quack"></a>Python says my extension type has no method called 'quack', but I know it does. What gives?</h2>
+You may have declared the method using <span style="font-family: monospace;">cdef</span> instead of <span style="font-family: monospace;">def</span>. Only functions and methods declared with <span style="font-family: monospace;">def</span> are callable from Python code.<br>
+--- 
+</body></html>
\ No newline at end of file

Modified: lxml/pyrex/Doc/extension_types.html
==============================================================================
--- lxml/pyrex/Doc/extension_types.html	(original)
+++ lxml/pyrex/Doc/extension_types.html	Tue Apr 25 15:43:21 2006
@@ -1 +1,444 @@
-<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
       
  <meta http-equiv="Content-Type"
 content="text/html; charset=iso-8859-1">
       
  <meta name="GENERATOR"
 content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]">
  <title>Extension Types</title>
</head>
  <body>
   
<h1>  
<hr width="100%">Extension Types  
<hr width="100%"></h1>
   
<h2> Contents</h2>
   
<ul>
  <li> <a href="#Introduction">Introduction</a></li>
   <li> <a href="#ExtTypeAttrs">Attributes</a></li>
   <li> <a href="#NotNone">Extension types and None</a></li>
   <li> <a href="special_methods.html">Special methods</a></li>
   <li> <a href="#Properties">Properties</a> <font color="#ed181e">(NEW in 
0.9)</font></li>
   <li><a href="#SubclassingExtTypes">Subclassing</a></li>
   <li> <a href="#CMethods">C Methods</a> <font color="#ff0000">(NEW in 0.9)</font><br>
   <a href="#ForwardDeclaringExtTypes">Forward-declaring extension types</a></li>
   <li> <a href="#PublicAndExtern">Public and external extension types</a><font
 color="#2f8b20"> (Section revised in 0.8)</font></li>
     
  <ul>
  <li> <a href="#ExternalExtTypes">External extension types</a></li>
   <li> <a href="#ImplicitImport">Implicit importing</a><font
 color="#2f8b20"> (NEW in 0.8)</font></li>
   <li> <a href="#TypeVsConstructor">Type names vs. constructor names</a></li>
   <li> <a href="#PublicExtensionTypes">Public extension types</a></li>
   <li> <a href="#NameSpecClause">Name specification clause</a></li>
    
  </ul>
  
</ul>
   
<h2> <a name="Introduction"></a>Introduction</h2>
  As well as creating normal user-defined classes with the Python <b>class</b>
statement, Pyrex also lets you create new built-in Python types, known as 
<i>extension types</i>. You define an extension type using the <b>cdef class</b> 
statement. Here's an example:  
<blockquote><tt>cdef class Shrubbery:</tt>   
  <p><tt>&nbsp;&nbsp;&nbsp; cdef int width, height</tt> </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp; def __init__(self, w, h):</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.width = w</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.height = h</tt> </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp; def describe(self):</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "This shrubbery is", 
self.width, \</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
"by", self.height, "cubits."</tt></p>
 </blockquote>
  As you can see, a Pyrex extension type definition looks a lot like a Python
 class definition. Within it, you use the <b>def</b> statement to define
methods that can be called from Python code. You can even define many of
the special methods such as <tt>__init__</tt> as you would in Python.  
<p>The main difference is that you can use the <b>cdef</b> statement to define 
attributes. The attributes may be Python objects (either generic or of a particular
extension type), or they may be of any C data type. So you can use extension
types to wrap arbitrary C data structures and provide a Python-like interface
to them. </p>
 
<h2> <a name="ExtTypeAttrs"></a>Attributes</h2>
  Attributes of an extension type are stored directly in the object's C struct.
 The set of attributes is fixed at compile time; you can't add attributes
to an extension type instance at run time simply by assigning to them, as
you could with a Python class instance. (You can subclass the extension type 
in Python and add attributes to instances of the subclass, however.)  
<p>There are two ways that attributes of an extension type can be accessed:
 by Python attribute lookup, or by direct access to the C struct from Pyrex
 code. Python code is only able to access attributes of an extension type
by the first method, but Pyrex code can use either method. </p>
 
<p>By default, extension type attributes are only accessible by direct access, 
not Python access, which means that they are not accessible from Python code. 
To make them accessible from Python code, you need to declare them as <tt>public</tt> 
or <tt>readonly</tt>. For example, </p>
 
<blockquote><tt>cdef class Shrubbery:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; cdef public int width, height</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; cdef readonly float depth</tt></blockquote>
  makes the <tt>width</tt> and <tt>height</tt> attributes readable and writable
 from Python code, and the <tt>depth</tt> attribute readable but not writable.
 
<p>Note that you can only expose simple C types, such as ints, floats and
 strings, for Python access. You can also expose Python-valued attributes,
 although read-write exposure is only possible for generic Python attributes
 (of type <tt>object</tt>). If the attribute is declared to be of an extension
 type, it must be exposed <tt>readonly</tt>. </p>
 
<p>Note also that the <tt>public</tt> and <tt>readonly</tt> options apply
 only to <i>Python</i> access, not direct access. All the attributes of an 
extension type are always readable and writable by direct access. </p>
 
<p>Howerver, for direct access to be possible, the Pyrex compiler must know 
that you have an instance of that type, and not just a generic Python object. 
It knows this already in the case of the "self" parameter of the methods of
that type, but in other cases you will have to tell it by means of a declaration. 
For example, </p>
 
<blockquote><tt>cdef widen_shrubbery(Shrubbery sh, extra_width):</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
  If you attempt to access an extension type attribute through a generic
object reference, Pyrex will use a Python attribute lookup. If the attribute
is exposed for Python access (using <tt>public</tt> or <tt>readonly</tt>)
then this will work, but it will be much slower than direct access.  
<h2> <a name="NotNone"></a>Extension types and None</h2>
  When you declare a parameter or C variable as being of an extension type,
 Pyrex will allow it to take on the value None as well as values of its declared 
type. This is analogous to the way a C pointer can take on the value NULL, 
and you need to exercise the same caution because of it. There is no problem 
as long as you are performing Python operations on it, because full dynamic 
type checking will be applied. However, when you access C attributes of an 
extension type (as in the <tt>widen_shrubbery</tt> function above), it's up
to you to make sure the reference you're using is not None -- in the interests 
of efficiency, Pyrex does <i>not</i> check this.  
<p>You need to be particularly careful when exposing Python functions which
 take extension types as arguments. If we wanted to make <tt>widen_shrubbery</tt>
a Python function, for example, if we simply wrote </p>
 
<blockquote><tt>def widen_shrubbery(Shrubbery sh, extra_width): # <font
 color="#ed181e">This is</font></tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# <font color="#ed181e">dangerous!</font></tt></blockquote>
  then users of our module could crash it by passing None for the <tt>sh</tt>
parameter.  
<p>One way to fix this would be </p>
 
<blockquote><tt>def widen_shrubbery(Shrubbery sh, extra_width):</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; if sh is None:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; raise TypeError</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
  but since this is anticipated to be such a frequent requirement, Pyrex
provides a more convenient way. Parameters of a Python function declared
as an extension type can have a <b><tt>not None</tt></b> clause:  
<blockquote><tt>def widen_shrubbery(Shrubbery sh not None, extra_width):</tt>
  <br>
   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
  Now the function will automatically check that <tt>sh</tt> is not None
along with checking that it has the right type.  
<p>Note, however that the <tt>not None</tt> clause can <i>only</i> be used
 in Python functions (defined with <tt>def</tt>) and not C functions (defined
 with <tt>cdef</tt>). If you need to check whether a parameter to a C function
 is None, you will need to do it yourself. </p>
 
<p>Some more things to note: </p>
 
<ul>
  <li> The <b>self</b> parameter of a method of an extension type is guaranteed
 never to be None.</li>
  
</ul>
   
<ul>
  <li> When comparing a value with None, keep in mind that, if <tt>x</tt> 
is a Python object, <tt>x is None</tt> and <tt>x is not None</tt> are very 
efficient because they translate directly to C pointer comparisons, whereas 
    <tt>x == None</tt> and <tt>x != None</tt>, or simply using <tt>x</tt> 
as a boolean value (as in <tt>if x: ...</tt>) will invoke Python operations 
and therefore be much slower.</li>
  
</ul>
   
<h2> <a name="ExtTypeSpecialMethods"></a>Special methods</h2>
  Although the principles are similar, there are substantial differences
between many of the __xxx___ special methods of extension types and their
Python counterparts. There is a <a href="special_methods.html">separate page</a> 
devoted to this subject, and you should read it carefully before attempting 
to use any special methods in your extension types.  
<h2> <a name="Properties"></a>Properties</h2>
  There is a special syntax for defining <b>properties</b> in an extension
 class:  
<blockquote><tt>cdef class Spam:</tt>   
  <p><tt>&nbsp;&nbsp;&nbsp; property cheese:</tt> </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; "A doc string can go
here."</tt>   </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __get__(self):</tt>
  <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# This is called when the property is read.</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
...</tt>   </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __set__(self, value):</tt>
  <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# This is called when the property is written.</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
...</tt>   </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __del__(self):</tt>
  <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# This is called when the property is deleted.</tt> <br>
 &nbsp;</p>
 </blockquote>
  The <tt>__get__</tt>, <tt>__set__</tt> and <tt>__del__</tt> methods are
all optional; if they are omitted, an exception will be raised when the corresponding 
operation is attempted.  
<p>Here's a complete example. It defines a property which adds to a list
each time it is written to, returns the list when it is read, and empties
the list when it is deleted. <br>
 &nbsp; </p>
 
<center> 
<table cellpadding="5" align="center">
  <tbody>
     <tr>
  <td bgcolor="#ffaf18"><b><tt>cheesy.pyx</tt></b></td>
   <td bgcolor="#5dbaca"><b>Test input</b></td>
  </tr>
   <tr>
  <td valign="top" rowspan="3" bgcolor="#ffaf18"><tt>cdef class CheeseShop:</tt>
      
      <p><tt>&nbsp; cdef object cheeses</tt> </p>
       
      <p><tt>&nbsp; def __new__(self):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; self.cheeses = []</tt> </p>
       
      <p><tt>&nbsp; property cheese:</tt> </p>
       
      <p><tt>&nbsp;&nbsp;&nbsp; def __get__(self):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; return "We don't have: %s" % self.cheeses</tt>
      </p>
       
      <p><tt>&nbsp;&nbsp;&nbsp; def __set__(self, value):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.cheeses.append(value)</tt>
      </p>
       
      <p><tt>&nbsp;&nbsp;&nbsp; def __del__(self):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; del self.cheeses[:]</tt></p>
       </td>
   <td valign="top" bgcolor="#5dbaca"><tt>from cheesy import CheeseShop</tt>
      
      <p><tt>shop = CheeseShop()</tt> <br>
       <tt>print shop.cheese</tt> </p>
       
      <p><tt>shop.cheese = "camembert"</tt> <br>
       <tt>print shop.cheese</tt> </p>
       
      <p><tt>shop.cheese = "cheddar"</tt> <br>
       <tt>print shop.cheese</tt> </p>
       
      <p><tt>del shop.cheese</tt> <br>
       <tt>print shop.cheese</tt></p>
       </td>
  </tr>
   <tr>
  <td bgcolor="#8cbc1c"><b>Test output</b></td>
  </tr>
   <tr>
  <td bgcolor="#8cbc1c"><tt>We don't have: []</tt> <br>
       <tt>We don't have: ['camembert']</tt> <br>
       <tt>We don't have: ['camembert', 'cheddar']</tt> <br>
       <tt>We don't have: []</tt></td>
  </tr>
    
  </tbody> 
</table>
 </center>
   
<h2> <a name="SubclassingExtTypes"></a>Subclassing</h2>
  An extension type may inherit from a built-in type or another extension
type:  
<blockquote><tt>cdef class Parrot:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; ...</tt><tt></tt>   
  <p><tt>cdef class Norwegian(Parrot):</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; ...</tt></p>
 </blockquote>
   
<p><br>
 A complete definition of the base type must be available to Pyrex, so if 
the base type is a built-in type, it must have been previously declared as 
an <b>extern</b> extension type. If the base type is defined in another Pyrex 
module, it must either be declared as an extern extension type or imported 
using the <b><a href="sharing.html">cimport</a></b> statement. </p>
 
<p>An extension type can only have one base class (no multiple inheritance).
 </p>
 
<p>Pyrex extension types can also be subclassed in Python. A Python class
 can inherit from multiple extension types provided that the usual Python
rules for multiple inheritance are followed (i.e. the C layouts of all the
base classes must be compatible).<br>
 </p>
 
<h2><a name="CMethods"></a>C methods</h2>
 Extension types can have C methods as well as Python methods. Like C functions, 
C methods are declared using <tt>cdef</tt> instead of <tt>def</tt>. C methods 
are "virtual", and may be overridden in derived extension types.<br>
 <br>
 
<table cellpadding="5" align="center">
   <tbody>
     <tr>
       <td valign="top" width="50%" bgcolor="#ffaf18"><b><tt>pets.pyx</tt></b><br>
       </td>
       <td valign="top" bgcolor="#8cbc1c" width="30%"><b>Output</b><br>
       </td>
     </tr>
     <tr>
       <td valign="top" width="50%" bgcolor="#ffaf18"><tt>cdef class Parrot:<br>
       <br>
  &nbsp; cdef void describe(self):<br>
  &nbsp; &nbsp; print "This parrot is resting."<br>
       <br>
  cdef class Norwegian(Parrot):<br>
       <br>
  &nbsp; cdef void describe(self):<br>
&nbsp; &nbsp; Parrot.describe(self)<br>
   &nbsp; &nbsp; print "Lovely plumage!"<br>
       <br>
       <br>
 cdef Parrot p1, p2<br>
    p1 = Parrot()<br>
    p2 = Norwegian()<br>
print "p1:"<br>
    p1.describe()<br>
print "p2:"<br>
    p2.describe()</tt> <br>
       </td>
       <td valign="top" bgcolor="#8cbc1c" width="30%"><tt>p1:<br>
This parrot is resting.<br>
p2:<br>
      </tt><tt>This parrot is resting.<br>
 </tt><tt>  Lovely plumage!</tt><br>
       </td>
     </tr>
   
  </tbody> 
</table>
 <br>
 The above example also illustrates that a C method can call an inherited
C method using the usual Python technique, i.e.<br>
<blockquote><tt>Parrot.describe(self)</tt><br>
</blockquote>
 
<h2><a name="ForwardDeclaringExtTypes"></a>Forward-declaring extension types</h2>
  Extension types can be forward-declared, like struct and union types. This
 will be necessary if you have two extension types that need to refer to
each other, e.g.  
<blockquote><tt>cdef class Shrubbery # forward declaration</tt>   
  <p><tt>cdef class Shrubber:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; cdef Shrubbery work_in_progress</tt> </p>
   
  <p><tt>cdef class Shrubbery:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; cdef Shrubber creator</tt></p>
 </blockquote>
   If you are forward-declaring an exension type that has a base class, you
must specify the base class in both the forward declaration and its subsequent
definition, for example,<br>
<blockquote><tt>cdef class A(B)<br>
  <br>
...<br>
  <br>
cdef class A(B):<br>
&nbsp; &nbsp; # attributes and methods</tt><br>
</blockquote>
 
<h2><a name="PublicAndExtern"></a>Public and external extension types</h2>
  Extension types can be declared <b>extern</b> or <b>public</b>. An <a
 href="#ExternalExtTypes"><b>extern</b> extension type declaration</a> makes 
an extension type defined in external C code available to a Pyrex module. 
A <a href="#PublicExtensionTypes"><b>public</b> extension type declaration</a> 
makes an extension type defined in a Pyrex module available to external C 
code.  
<h3> <a name="ExternalExtTypes"></a>External extension types</h3>
  An <b>extern</b> extension type allows you to gain access to the internals
 of Python objects defined in the Python core or in a non-Pyrex extension
module.  
<blockquote><b>NOTE:</b> In Pyrex versions before 0.8, <b>extern</b> extension
 types were also used to reference extension types defined in another Pyrex
 module. While you can still do that, Pyrex 0.8 and later provides a better
 mechanism for this. See <a href="sharing.html">Sharing C Declarations Between
 Pyrex Modules</a>.</blockquote>
  Here is an example which will let you get at the C-level members of the
built-in <i>complex</i> object.  
<blockquote><tt>cdef extern from "complexobject.h":</tt>   
  <p><tt>&nbsp;&nbsp;&nbsp; struct Py_complex:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; double real</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; double imag</tt> </p>
   
  <p><tt>&nbsp;&nbsp;&nbsp; ctypedef class __builtin__.complex [object PyComplexObject]:</tt>
  <br>
   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef Py_complex cval</tt>
  </p>
   
  <p><tt># A function which uses the above type</tt> <br>
   <tt>def spam(complex c):</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; print "Real:", c.cval.real</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; print "Imag:", c.cval.imag</tt></p>
 </blockquote>
  Some important things to note are:  
<ol>
  <li> In this example, <b>ctypedef class</b> has been used. This is because,
 in the Python header files, the <tt>PyComplexObject</tt> struct is declared
 with</li>
  
</ol>
   
<blockquote>   
  <blockquote><tt>ctypedef struct {</tt> <br>
     <tt>&nbsp;&nbsp;&nbsp; ...</tt> <br>
     <tt>} PyComplexObject;</tt></blockquote>
  </blockquote>
   
<ol>
  <li> As well as the name of the extension type, the <i>module</i> in which 
its type object can be found is also specified. See the <a
 href="#ImplicitImport">implicit importing</a> section below.</li>
   <br>
 &nbsp; <li> The part in square brackets tells Pyrex the name to use for
the object's C struct, so it can generate code that matches what is declared 
in the header file. See the <a href="#NameSpecClause">name specification clause</a>
section below.</li>
   <br>
 &nbsp; <li> When declaring an external extension type, you don't declare 
any methods. Declaration of methods is not required in order to call them, 
because the calls are Python method calls. Also, as with structs and unions, 
if your extension class declaration is inside a <i>cdef extern from</i> block,
 you only need to declare those C members which you wish to access.</li>
  
</ol>
   
<h3> <a name="ImplicitImport"></a>Implicit importing</h3>
   
<blockquote><font color="#ef1f1d">Backwards Incompatibility Note</font>:
You will have to update any pre-0.8 Pyrex modules you have which use <b>extern</b>
extension types. I apologise for this, but for complicated reasons it proved
 to be too difficult to continue supporting the old way of doing these while
 introducing the new features that I wanted.</blockquote>
  Pyrex 0.8 and later requires you to include a module name in an extern
extension class declaration, for example,  
<blockquote><tt>cdef extern class MyModule.Spam:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
  The type object will be implicitly imported from the specified module and
 bound to the corresponding name in this module. In other words, in this
example an implicit  
<ol>
    
  <pre>from <tt>MyModule</tt> import Spam</pre>
  
</ol>
  statement will be executed at module load time.  
<p>The module name can be a dotted name to refer to a module inside a package
 hierarchy, for example, </p>
 
<blockquote><tt>cdef extern class My.Nested.Package.Spam:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
  You can also specify an alternative name under which to import the type
using an <b>as</b> clause, for example,  
<ol>
   <tt>cdef extern class My.Nested.Package.Spam as Yummy:</tt> <br>
   <tt>&nbsp;&nbsp; ...</tt> 
</ol>
  which corresponds to the implicit import statement  
<ol>
    
  <pre>from <tt>My.Nested.Package</tt> import <tt>Spam</tt> as <tt>Yummy</tt></pre>
  
</ol>
   
<h3> <a name="TypeVsConstructor"></a>Type names vs. constructor names</h3>
  Inside a Pyrex module, the name of an extension type serves two distinct
 purposes. When used in an expression, it refers to a module-level global
variable holding the type's constructor (i.e. its type-object). However,
it can also be used as a C type name to declare variables, arguments and
return values of that type.  
<p>When you declare </p>
 
<blockquote><tt>cdef extern class MyModule.Spam:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
  the name <tt>Spam</tt> serves both these roles. There may be other names
 by which you can refer to the constructor, but only <tt>Spam</tt> can be
used as a type name. For example, if you were to explicity <tt>import MyModule</tt>,
 you could use<tt> MyModule.Spam()</tt> to create a Spam instance, but you
 wouldn't be able to use <tt>MyModule.Spam</tt> as a type name.  
<p>When an <b>as</b> clause is used, the name specified in the <b>as</b>
clause also takes over both roles. So if you declare </p>
 
<blockquote><tt>cdef extern class MyModule.Spam as Yummy:</tt> <br>
   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
  then <tt>Yummy</tt> becomes both the type name and a name for the constructor.
 Again, there are other ways that you could get hold of the constructor,
but only <tt>Yummy</tt> is usable as a type name.  
<h3> <a name="PublicExtensionTypes"></a>Public extension types</h3>
  An extension type can be declared <b>public</b>, in which case a <b>.h</b>
file is generated containing declarations for its object struct and type
object. By including the <b>.h</b> file in external C code that you write,
that code can access the attributes of the extension type.  
<h3> <a name="NameSpecClause"></a>Name specification clause</h3>
  The part of the class declaration in square brackets is a special feature
 only available for <b>extern</b> or <b>public</b> extension types. The full 
form of this clause is  
<blockquote><tt>[object </tt><i>object_struct_name</i><tt>, type </tt><i>type_object_name</i><tt>]</tt></blockquote>
  where <i>object_struct_name</i> is the name to assume for the type's C
struct, and <i>type_object_name</i> is the name to assume for the type's
statically declared type object. (The object and type clauses can be written
in either order.)  
<p>If the extension type declaration is inside a <b>cdef extern from</b>
block, the <b>object</b> clause is required, because Pyrex must be able to
generate code that is compatible with the declarations in the header file.
Otherwise, for <b>extern</b> extension types, the <b>object</b> clause is
optional. </p>
 
<p>For <b>public</b> extension types, the <b>object</b> and <b>type</b> clauses 
are both required, because Pyrex must be able to generate code that is compatible 
with external C code. </p>
 
<p> </p>
 
<hr width="100%"> <br>
 Back to the <a href="overview.html">Language Overview</a> <br>
 &nbsp; <br>
 <br>
</body>
</html>
\ No newline at end of file
+<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
+<html><head>
+         <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
+         <meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Extension Types</title></head>
+<body>
+   <h1>  <hr width="100%">Extension Types  
+<hr width="100%"></h1>
+   <h2> Contents</h2>
+   <ul>
+  <li> <a href="#Introduction">Introduction</a></li>
+   <li> <a href="#ExtTypeAttrs">Attributes</a></li>
+   <li> <a href="#NotNone">Extension types and None</a></li>
+   <li> <a href="special_methods.html">Special methods</a></li>
+   <li> <a href="#Properties">Properties</a> <font style="color: rgb(0, 153, 0);" color="#ed181e">(NEW in 
+0.9)</font></li>
+   <li><a href="#SubclassingExtTypes">Subclassing</a></li>
+   <li> <a href="#CMethods">C Methods</a> <font style="color: rgb(0, 153, 0);" color="#ff0000">(NEW in 0.9)</font><br>
+   <a href="#ForwardDeclaringExtTypes">Forward-declaring extension types</a></li><li><a href="#WeakRefs">Making extension types weak-referenceable</a> <span style="color: rgb(255, 0, 0);">(NEW in 0.9.4)</span><br>
+  </li>
+
+   <li> <a href="#PublicAndExtern">Public and external extension types</a><font color="#2f8b20"><br>
+</font></li>
+       <ul>
+  <li> <a href="#ExternalExtTypes">External extension types</a></li>
+   <li> <a href="#ImplicitImport">Implicit importing</a><font color="#2f8b20"><br>
+</font></li>
+   <li> <a href="#TypeVsConstructor">Type names vs. constructor names</a></li>
+   <li> <a href="#PublicExtensionTypes">Public extension types</a></li>
+   <li> <a href="#NameSpecClause">Name specification clause</a></li>
+      </ul>
+  </ul>
+   <h2> <a name="Introduction"></a>Introduction</h2>
+  As well as creating normal user-defined classes with the Python <b>class</b>
+statement, Pyrex also lets you create new built-in Python types, known as 
+<i>extension types</i>. You define an extension type using the <b>cdef class</b> statement. Here's an example:  
+<blockquote><tt>cdef class Shrubbery:</tt>     <p><tt>&nbsp;&nbsp;&nbsp; cdef int width, height</tt> </p>
+     <p><tt>&nbsp;&nbsp;&nbsp; def __init__(self, w, h):</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.width = w</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.height = h</tt> </p>
+     <p><tt>&nbsp;&nbsp;&nbsp; def describe(self):</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "This shrubbery is", 
+self.width, \</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+"by", self.height, "cubits."</tt></p>
+ </blockquote>
+  As you can see, a Pyrex extension type definition looks a lot like a Python
+ class definition. Within it, you use the <b>def</b> statement to define
+methods that can be called from Python code. You can even define many of
+the special methods such as <tt>__init__</tt> as you would in Python.  
+<p>The main difference is that you can use the <b>cdef</b> statement to define 
+attributes. The attributes may be Python objects (either generic or of a particular
+extension type), or they may be of any C data type. So you can use extension
+types to wrap arbitrary C data structures and provide a Python-like interface
+to them. </p>
+ <h2> <a name="ExtTypeAttrs"></a>Attributes</h2>
+  Attributes of an extension type are stored directly in the object's C struct.
+ The set of attributes is fixed at compile time; you can't add attributes
+to an extension type instance at run time simply by assigning to them, as
+you could with a Python class instance. (You can subclass the extension type 
+in Python and add attributes to instances of the subclass, however.)  
+<p>There are two ways that attributes of an extension type can be accessed:
+ by Python attribute lookup, or by direct access to the C struct from Pyrex
+ code. Python code is only able to access attributes of an extension type
+by the first method, but Pyrex code can use either method. </p>
+ <p>By default, extension type attributes are only accessible by direct access, 
+not Python access, which means that they are not accessible from Python code. 
+To make them accessible from Python code, you need to declare them as <tt>public</tt> or <tt>readonly</tt>. For example, </p>
+ <blockquote><tt>cdef class Shrubbery:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; cdef public int width, height</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; cdef readonly float depth</tt></blockquote>
+  makes the <tt>width</tt> and <tt>height</tt> attributes readable and writable
+ from Python code, and the <tt>depth</tt> attribute readable but not writable.
+ 
+<p>Note that you can only expose simple C types, such as ints, floats and
+ strings, for Python access. You can also expose Python-valued attributes,
+ although read-write exposure is only possible for generic Python attributes
+ (of type <tt>object</tt>). If the attribute is declared to be of an extension
+ type, it must be exposed <tt>readonly</tt>. </p>
+ <p>Note also that the <tt>public</tt> and <tt>readonly</tt> options apply
+ only to <i>Python</i> access, not direct access. All the attributes of an 
+extension type are always readable and writable by direct access. </p>
+ <p>Howerver, for direct access to be possible, the Pyrex compiler must know 
+that you have an instance of that type, and not just a generic Python object. 
+It knows this already in the case of the "self" parameter of the methods of
+that type, but in other cases you will have to tell it by means of a declaration. 
+For example, </p>
+ <blockquote><tt>cdef widen_shrubbery(Shrubbery sh, extra_width):</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
+  If you attempt to access an extension type attribute through a generic
+object reference, Pyrex will use a Python attribute lookup. If the attribute
+is exposed for Python access (using <tt>public</tt> or <tt>readonly</tt>)
+then this will work, but it will be much slower than direct access.  
+<h2> <a name="NotNone"></a>Extension types and None</h2>
+  When you declare a parameter or C variable as being of an extension type,
+ Pyrex will allow it to take on the value None as well as values of its declared 
+type. This is analogous to the way a C pointer can take on the value NULL, 
+and you need to exercise the same caution because of it. There is no problem 
+as long as you are performing Python operations on it, because full dynamic 
+type checking will be applied. However, when you access C attributes of an 
+extension type (as in the <tt>widen_shrubbery</tt> function above), it's up
+to you to make sure the reference you're using is not None -- in the interests 
+of efficiency, Pyrex does <i>not</i> check this.  
+<p>You need to be particularly careful when exposing Python functions which
+ take extension types as arguments. If we wanted to make <tt>widen_shrubbery</tt>
+a Python function, for example, if we simply wrote </p>
+ <blockquote><tt>def widen_shrubbery(Shrubbery sh, extra_width): # <font color="#ed181e">This is</font></tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+# <font color="#ed181e">dangerous!</font></tt></blockquote>
+  then users of our module could crash it by passing None for the <tt>sh</tt>
+parameter.  
+<p>One way to fix this would be </p>
+ <blockquote><tt>def widen_shrubbery(Shrubbery sh, extra_width):</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; if sh is None:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; raise TypeError</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
+  but since this is anticipated to be such a frequent requirement, Pyrex
+provides a more convenient way. Parameters of a Python function declared
+as an extension type can have a <b><tt>not None</tt></b> clause:  
+<blockquote><tt>def widen_shrubbery(Shrubbery sh not None, extra_width):</tt>
+  <br>
+   <tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
+  Now the function will automatically check that <tt>sh</tt> is not None
+along with checking that it has the right type.  
+<p>Note, however that the <tt>not None</tt> clause can <i>only</i> be used
+ in Python functions (defined with <tt>def</tt>) and not C functions (defined
+ with <tt>cdef</tt>). If you need to check whether a parameter to a C function
+ is None, you will need to do it yourself. </p>
+ <p>Some more things to note: </p>
+ <ul>
+  <li> The <b>self</b> parameter of a method of an extension type is guaranteed
+ never to be None.</li>
+  </ul>
+   <ul>
+  <li> When comparing a value with None, keep in mind that, if <tt>x</tt> is a Python object, <tt>x is None</tt> and <tt>x is not None</tt> are very 
+efficient because they translate directly to C pointer comparisons, whereas 
+    <tt>x == None</tt> and <tt>x != None</tt>, or simply using <tt>x</tt> as a boolean value (as in <tt>if x: ...</tt>) will invoke Python operations 
+and therefore be much slower.</li>
+  </ul>
+   <h2> <a name="ExtTypeSpecialMethods"></a>Special methods</h2>
+  Although the principles are similar, there are substantial differences
+between many of the <span style="font-family: monospace;">__xxx__</span> special methods of extension types and their
+Python counterparts. There is a <a href="special_methods.html">separate page</a> devoted to this subject, and you should read it carefully before attempting 
+to use any special methods in your extension types.  
+<h2> <a name="Properties"></a>Properties</h2>
+  There is a special syntax for defining <b>properties</b> in an extension
+ class:  
+<blockquote><tt>cdef class Spam:</tt>     <p><tt>&nbsp;&nbsp;&nbsp; property cheese:</tt> </p>
+     <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; "A doc string can go
+here."</tt>   </p>
+     <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __get__(self):</tt>
+  <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+# This is called when the property is read.</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+...</tt>   </p>
+     <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __set__(self, value):</tt>
+  <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+# This is called when the property is written.</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+...</tt>   </p>
+     <p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __del__(self):</tt>
+  <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+# This is called when the property is deleted.</tt> <br>
+ &nbsp;</p>
+ </blockquote>
+  The <tt>__get__</tt>, <tt>__set__</tt> and <tt>__del__</tt> methods are
+all optional; if they are omitted, an exception will be raised when the corresponding 
+operation is attempted.  
+<p>Here's a complete example. It defines a property which adds to a list
+each time it is written to, returns the list when it is read, and empties
+the list when it is deleted. <br>
+ &nbsp; </p>
+ <center> <table align="center" cellpadding="5">
+  <tbody>
+     <tr>
+  <td bgcolor="#ffaf18"><b><tt>cheesy.pyx</tt></b></td>
+   <td bgcolor="#5dbaca"><b>Test input</b></td>
+  </tr>
+   <tr>
+  <td rowspan="3" bgcolor="#ffaf18" valign="top"><tt>cdef class CheeseShop:</tt>
+            <p><tt>&nbsp; cdef object cheeses</tt> </p>
+             <p><tt>&nbsp; def __new__(self):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; self.cheeses = []</tt> </p>
+             <p><tt>&nbsp; property cheese:</tt> </p>
+             <p><tt>&nbsp;&nbsp;&nbsp; def __get__(self):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; return "We don't have: %s" % self.cheeses</tt>
+      </p>
+             <p><tt>&nbsp;&nbsp;&nbsp; def __set__(self, value):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.cheeses.append(value)</tt>
+      </p>
+             <p><tt>&nbsp;&nbsp;&nbsp; def __del__(self):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; del self.cheeses[:]</tt></p>
+       </td>
+   <td bgcolor="#5dbaca" valign="top"><tt>from cheesy import CheeseShop</tt>
+            <p><tt>shop = CheeseShop()</tt> <br>
+       <tt>print shop.cheese</tt> </p>
+             <p><tt>shop.cheese = "camembert"</tt> <br>
+       <tt>print shop.cheese</tt> </p>
+             <p><tt>shop.cheese = "cheddar"</tt> <br>
+       <tt>print shop.cheese</tt> </p>
+             <p><tt>del shop.cheese</tt> <br>
+       <tt>print shop.cheese</tt></p>
+       </td>
+  </tr>
+   <tr>
+  <td bgcolor="#8cbc1c"><b>Test output</b></td>
+  </tr>
+   <tr>
+  <td bgcolor="#8cbc1c"><tt>We don't have: []</tt> <br>
+       <tt>We don't have: ['camembert']</tt> <br>
+       <tt>We don't have: ['camembert', 'cheddar']</tt> <br>
+       <tt>We don't have: []</tt></td>
+  </tr>
+      </tbody> </table>
+ </center>
+   <h2> <a name="SubclassingExtTypes"></a>Subclassing</h2>
+  An extension type may inherit from a built-in type or another extension
+type:  
+<blockquote><tt>cdef class Parrot:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; ...</tt><tt></tt>     <p><tt>cdef class Norwegian(Parrot):</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; ...</tt></p>
+ </blockquote>
+   <p><br>
+ A complete definition of the base type must be available to Pyrex, so if 
+the base type is a built-in type, it must have been previously declared as 
+an <b>extern</b> extension type. If the base type is defined in another Pyrex 
+module, it must either be declared as an extern extension type or imported 
+using the <b><a href="sharing.html">cimport</a></b> statement. </p>
+ <p>An extension type can only have one base class (no multiple inheritance).
+ </p>
+ <p>Pyrex extension types can also be subclassed in Python. A Python class
+ can inherit from multiple extension types provided that the usual Python
+rules for multiple inheritance are followed (i.e. the C layouts of all the
+base classes must be compatible).<br>
+ </p>
+ <h2><a name="CMethods"></a>C methods</h2>
+ Extension types can have C methods as well as Python methods. Like C functions, 
+C methods are declared using <tt>cdef</tt> instead of <tt>def</tt>. C methods 
+are "virtual", and may be overridden in derived extension types.<br>
+ <br>
+ <table align="center" cellpadding="5">
+   <tbody>
+     <tr>
+       <td bgcolor="#ffaf18" valign="top" width="50%"><b><tt>pets.pyx</tt></b><br>
+       </td>
+       <td bgcolor="#8cbc1c" valign="top" width="30%"><b>Output</b><br>
+       </td>
+     </tr>
+     <tr>
+       <td bgcolor="#ffaf18" valign="top" width="50%"><tt>cdef class Parrot:<br>
+       <br>
+  &nbsp; cdef void describe(self):<br>
+  &nbsp; &nbsp; print "This parrot is resting."<br>
+       <br>
+  cdef class Norwegian(Parrot):<br>
+       <br>
+  &nbsp; cdef void describe(self):<br>
+&nbsp; &nbsp; Parrot.describe(self)<br>
+   &nbsp; &nbsp; print "Lovely plumage!"<br>
+       <br>
+       <br>
+ cdef Parrot p1, p2<br>
+    p1 = Parrot()<br>
+    p2 = Norwegian()<br>
+print "p1:"<br>
+    p1.describe()<br>
+print "p2:"<br>
+    p2.describe()</tt> <br>
+       </td>
+       <td bgcolor="#8cbc1c" valign="top" width="30%"><tt>p1:<br>
+This parrot is resting.<br>
+p2:<br>
+      </tt><tt>This parrot is resting.<br>
+ </tt><tt>  Lovely plumage!</tt><br>
+       </td>
+     </tr>
+     </tbody> </table>
+ <br>
+ The above example also illustrates that a C method can call an inherited
+C method using the usual Python technique, i.e.<br>
+<blockquote><tt>Parrot.describe(self)</tt><br>
+</blockquote>
+ <h2><a name="ForwardDeclaringExtTypes"></a>Forward-declaring extension types</h2>
+  Extension types can be forward-declared, like struct and union types. This
+ will be necessary if you have two extension types that need to refer to
+each other, e.g.  
+<blockquote><tt>cdef class Shrubbery # forward declaration</tt>     <p><tt>cdef class Shrubber:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; cdef Shrubbery work_in_progress</tt> </p>
+     <p><tt>cdef class Shrubbery:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; cdef Shrubber creator</tt></p>
+ </blockquote>
+   If you are forward-declaring an exension type that has a base class, you
+must specify the base class in both the forward declaration and its subsequent
+definition, for example,<br>
+<blockquote><tt>cdef class A(B)<br>
+  <br>
+...<br>
+  <br>
+cdef class A(B):<br>
+&nbsp; &nbsp; # attributes and methods</tt><br>
+</blockquote>
+ <h2><a name="WeakRefs"></a>Making extension types weak-referenceable</h2>By
+default, extension types do not support having weak references made to
+them. You can enable weak referencing by declaring a C attribute of
+type <span style="font-family: monospace;">object</span> called <span style="font-family: monospace; font-weight: bold;">__weakref__</span>. For example,<br>
+<br>
+<div style="margin-left: 40px;"><span style="font-family: monospace;">cdef class ExplodingAnimal:</span><br style="font-family: monospace;">
+<span style="font-family: monospace;">&nbsp;&nbsp;&nbsp; """This animal will self-destruct when it is</span><br>
+<span style="font-family: monospace;">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; no longer strongly referenced."""</span><br>
+<span style="font-family: monospace;">&nbsp;&nbsp;&nbsp; </span><br style="font-family: monospace;">
+<span style="font-family: monospace;"></span><span style="font-family: monospace;">&nbsp;&nbsp;&nbsp; cdef object __weakref__</span><br>
+</div>
+<br>
+<h2><a name="PublicAndExtern"></a>Public and external extension types</h2>
+
+  Extension types can be declared <b>extern</b> or <b>public</b>. An <a href="#ExternalExtTypes"><b>extern</b> extension type declaration</a> makes 
+an extension type defined in external C code available to a Pyrex module. 
+A <a href="#PublicExtensionTypes"><b>public</b> extension type declaration</a> makes an extension type defined in a Pyrex module available to external C 
+code.  
+<h3> <a name="ExternalExtTypes"></a>External extension types</h3>
+  An <b>extern</b> extension type allows you to gain access to the internals
+ of Python objects defined in the Python core or in a non-Pyrex extension
+module.  
+<blockquote><b>NOTE:</b> In Pyrex versions before 0.8, <b>extern</b> extension
+ types were also used to reference extension types defined in another Pyrex
+ module. While you can still do that, Pyrex 0.8 and later provides a better
+ mechanism for this. See <a href="sharing.html">Sharing C Declarations Between
+ Pyrex Modules</a>.</blockquote>
+  Here is an example which will let you get at the C-level members of the
+built-in <i>complex</i> object.  
+<blockquote><tt>cdef extern from "complexobject.h":</tt>     <p><tt>&nbsp;&nbsp;&nbsp; struct Py_complex:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; double real</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; double imag</tt> </p>
+     <p><tt>&nbsp;&nbsp;&nbsp; ctypedef class __builtin__.complex [object PyComplexObject]:</tt>
+  <br>
+   <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef Py_complex cval</tt>
+  </p>
+     <p><tt># A function which uses the above type</tt> <br>
+   <tt>def spam(complex c):</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; print "Real:", c.cval.real</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; print "Imag:", c.cval.imag</tt></p>
+ </blockquote>
+  Some important things to note are:  
+<ol>
+  <li> In this example, <b>ctypedef class</b> has been used. This is because,
+ in the Python header files, the <tt>PyComplexObject</tt> struct is declared
+ with<br>
+    <br>
+    <div style="margin-left: 40px;"><tt>ctypedef struct {</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt> <br>
+    <tt>} PyComplexObject;<br>
+    <br>
+    </tt></div>
+</li><li>As well as the name of the extension type, the <i>module</i> in which 
+its type object can be found is also specified. See the <a href="#ImplicitImport">implicit importing</a> section below.&nbsp; <br>
+    <br>
+  </li>
+<li> When declaring an external extension type, you don't declare 
+any methods. Declaration of methods is not required in order to call them, 
+because the calls are Python method calls. Also, as with structs and unions, 
+if your extension class declaration is inside a <i>cdef extern from</i> block,
+ you only need to declare those C members which you wish to access.</li>
+  </ol>
+   <h3> <a name="ImplicitImport"></a>Implicit importing</h3>
+   <blockquote><font color="#ef1f1d">Backwards Incompatibility Note</font>:
+You will have to update any pre-0.8 Pyrex modules you have which use <b>extern</b>
+extension types. I apologise for this, but for complicated reasons it proved
+ to be too difficult to continue supporting the old way of doing these while
+ introducing the new features that I wanted.</blockquote>
+  Pyrex 0.8 and later requires you to include a module name in an extern
+extension class declaration, for example,  
+<blockquote><tt>cdef extern class MyModule.Spam:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+  The type object will be implicitly imported from the specified module and
+ bound to the corresponding name in this module. In other words, in this
+example an implicit  
+<ol>
+      <pre>from <tt>MyModule</tt> import Spam</pre>
+  </ol>
+  statement will be executed at module load time.  
+<p>The module name can be a dotted name to refer to a module inside a package
+ hierarchy, for example, </p>
+ <blockquote><tt>cdef extern class My.Nested.Package.Spam:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+  You can also specify an alternative name under which to import the type
+using an <b>as</b> clause, for example,  
+<ol>
+   <tt>cdef extern class My.Nested.Package.Spam as Yummy:</tt> <br>
+   <tt>&nbsp;&nbsp; ...</tt> </ol>
+  which corresponds to the implicit import statement  
+<ol>
+      <pre>from <tt>My.Nested.Package</tt> import <tt>Spam</tt> as <tt>Yummy</tt></pre>
+  </ol>
+   <h3> <a name="TypeVsConstructor"></a>Type names vs. constructor names</h3>
+  Inside a Pyrex module, the name of an extension type serves two distinct
+ purposes. When used in an expression, it refers to a module-level global
+variable holding the type's constructor (i.e. its type-object). However,
+it can also be used as a C type name to declare variables, arguments and
+return values of that type.  
+<p>When you declare </p>
+ <blockquote><tt>cdef extern class MyModule.Spam:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+  the name <tt>Spam</tt> serves both these roles. There may be other names
+ by which you can refer to the constructor, but only <tt>Spam</tt> can be
+used as a type name. For example, if you were to explicity <tt>import MyModule</tt>,
+ you could use<tt> MyModule.Spam()</tt> to create a Spam instance, but you
+ wouldn't be able to use <tt>MyModule.Spam</tt> as a type name.  
+<p>When an <b>as</b> clause is used, the name specified in the <b>as</b>
+clause also takes over both roles. So if you declare </p>
+ <blockquote><tt>cdef extern class MyModule.Spam as Yummy:</tt> <br>
+   <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+  then <tt>Yummy</tt> becomes both the type name and a name for the constructor.
+ Again, there are other ways that you could get hold of the constructor,
+but only <tt>Yummy</tt> is usable as a type name.  
+<h3> <a name="PublicExtensionTypes"></a>Public extension types</h3>
+  An extension type can be declared <b>public</b>, in which case a <b>.h</b>
+file is generated containing declarations for its object struct and type
+object. By including the <b>.h</b> file in external C code that you write,
+that code can access the attributes of the extension type.  
+<h3> <a name="NameSpecClause"></a>Name specification clause</h3>
+  The part of the class declaration in square brackets is a special feature
+ only available for <b>extern</b> or <b>public</b> extension types. The full 
+form of this clause is  
+<blockquote><tt>[object </tt><i>object_struct_name</i><tt>, type </tt><i>type_object_name</i><span style="font-family: monospace;"> ]</span></blockquote>
+  where <i>object_struct_name</i> is the name to assume for the type's C
+struct, and <i>type_object_name</i> is the name to assume for the type's
+statically declared type object. (The object and type clauses can be written
+in either order.)  
+<p>If the extension type declaration is inside a <b>cdef extern from</b>
+block, the <b>object</b> clause is required, because Pyrex must be able to
+generate code that is compatible with the declarations in the header file.
+Otherwise, for <b>extern</b> extension types, the <b>object</b> clause is
+optional. </p>
+ <p>For <b>public</b> extension types, the <b>object</b> and <b>type</b> clauses 
+are both required, because Pyrex must be able to generate code that is compatible 
+with external C code. </p>
+ <p> </p>
+ <hr width="100%"> <br>
+ Back to the <a href="overview.html">Language Overview</a> <br>
+ &nbsp; <br>
+ <br>
+</body></html>
\ No newline at end of file

Modified: lxml/pyrex/Doc/overview.html
==============================================================================
--- lxml/pyrex/Doc/overview.html	(original)
+++ lxml/pyrex/Doc/overview.html	Tue Apr 25 15:43:21 2006
@@ -1 +1,960 @@
-<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
           
  <meta http-equiv="Content-Type"
 content="text/html; charset=iso-8859-1">
           
  <meta name="GENERATOR"
 content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]">
  <title>Pyrex Language Overview</title>
   
</head>
  <body>
     
<h1>   
<hr width="100%">Overview of the Pyrex Language&nbsp;  
<hr width="100%"></h1>
   This document informally describes the extensions to the Python language 
 made by Pyrex. Some day there will be a reference manual covering everything 
 in more detail. <br>
  &nbsp;  
<h2> Contents</h2>
     
<ul>
   <li> <a href="#Basics">Basics</a></li>
         
  <ul>
   <li> <a href="#PyFuncsVsCFuncs">Python functions vs. C functions</a></li>
    <li> <a href="#PyObjParams">Python objects as parameters</a></li>
    <li> <a href="#CVarAndTypeDecls">C variable and type definitions</a></li>
    <li> <a href="#ScopeRules">Scope rules</a></li>
    <li> <a href="#StatsAndExprs">Statements and expressions</a></li>
             
    <ul>
   <li> <a href="#ExprSyntaxDifferences">Differences between C and Pyrex
expression  syntax</a></li>
    <li> <a href="#ForFromLoop">Integer for-loops</a></li>
            
    </ul>
    <li> <a href="#ExceptionValues">Error return values</a></li>
             
    <ul>
   <li> <a href="#CheckingReturnValues">Checking return values of non-Pyrex
 functions</a></li>
            
    </ul>
    <li> <a href="#IncludeStatement">The <tt>include</tt> statement</a></li>
        
  </ul>
    <li> <a href="#InterfacingWithExternal">Interfacing with External C Code</a></li>
         
  <ul>
   <li> <a href="#ExternDecls">External declarations</a></li>
             
    <ul>
   <li> <a href="#ReferencingHeaders">Referencing C header files</a></li>
    <li> <a href="#StructDeclStyles">Styles of struct/union/enum declaration</a></li>
    <li> <a href="#AccessingAPI">Accessing Python/C API routines</a></li>
    <li> <a href="#CNameSpecs">Resolving naming conflicts - C name specifications</a></li>
            
    </ul>
    <li> <a href="#PublicDecls">Public declarations</a></li>
        
  </ul>
    <li> <a href="extension_types.html">Extension Types</a> <font
 color="#006600">(Section revised in 0.9)</font></li>
    <li> <a href="sharing.html">Sharing Declarations Between Pyrex Modules</a>
     <font color="#006600">(NEW in 0.8)</font></li>
    <li> <a href="#Limitations">Limitations</a></li>
         
  <ul>
   <li> <a href="#Unsupported">Unsupported Python features</a></li>
    <li> <a href="#SemanticDifferences">Semantic differences between Python
 and Pyrex</a></li>
        
  </ul>
    
</ul>
     
<h2>   
<hr width="100%"><a name="Basics"></a>Basics   
<hr width="100%"></h2>
   This section describes the basic features of the Pyrex language. The facilities 
 covered in this section allow you to create Python-callable functions that 
 manipulate C data structures and convert between Python and C data types. 
 Later sections will cover facilities for <a
 href="#InterfacingWithExternal">wrapping external C code</a>, <a
 href="extension_types.html">creating new Python types</a> and <a
 href="sharing.html">cooperation between Pyrex modules</a>.   
<h3> <a name="PyFuncsVsCFuncs"></a>Python functions vs. C functions</h3>
   There are two kinds of function definition in Pyrex:   
<p><b>Python functions</b> are defined using the <b>def</b> statement, as
 in Python. They take Python objects as parameters and return Python objects. 
 </p>
   
<p><b>C functions</b> are defined using the new <b>cdef</b> statement. They
 take either Python objects or C values as parameters, and can return either
 Python objects or C values. </p>
   
<p>Within a Pyrex module, Python functions and C functions can call each other
freely, but only Python functions can be called from outside the module by
interpreted Python code. So, any functions that you want to "export" from
 your Pyrex module must be declared as Python functions. </p>
   
<p>Parameters of either type of function can be declared to have C data types,
 using normal C declaration syntax. For example, </p>
   
<blockquote>      
  <pre>def spam(int i, char *s):<br>&nbsp;&nbsp;&nbsp; ...</pre>
         
  <pre>cdef int eggs(unsigned long l, float f):<br>&nbsp;&nbsp;&nbsp; ...</pre>
   </blockquote>
   When a parameter of a Python function is declared to have a C data type, 
 it is passed in as a Python object and automatically converted to a C value, 
 if possible. Automatic conversion is currently only possible for numeric 
types and string types; attempting to use any other type for the parameter 
of a Python function will result in a compile-time error.   
<p>C functions, on the other hand, can have parameters of any type, since 
 they're passed in directly using a normal C function call. </p>
   
<h3> <a name="PyObjParams"></a>Python objects as parameters and return values</h3>
   If no type is specified for a parameter or return value, <i>it is assumed 
 to be a Python object.</i> (Note that this is different from the C convention, 
 where it would default to <tt>int</tt>.) For example, the following defines 
 a C function that takes two Python objects as parameters and returns a Python
 object:   
<blockquote>      
  <pre>cdef spamobjs(x, y):<br>&nbsp;&nbsp;&nbsp; ...</pre>
   </blockquote>
   Reference counting for these objects is performed automatically according 
 to the standard Python/C API rules (i.e. borrowed references are taken as
 parameters and a new reference is returned).   
<p>The name <b>object</b> can also be used to explicitly declare something 
 as a Python object. This can be useful if the name being declared would otherwise
be taken as the name of a type, for example, </p>
   
<blockquote>      
  <pre>cdef ftang(object int):<br>&nbsp;&nbsp;&nbsp; ...</pre>
   </blockquote>
   declares a parameter called <tt>int</tt> which is a Python object. You 
can also use <b>object </b>as the explicit return type of a function, e.g. 
  
<blockquote>      
  <pre>cdef object ftang(object int):<br>&nbsp;&nbsp;&nbsp; ...</pre>
   </blockquote>
   In the interests of clarity, it is probably a good idea to always be explicit 
 about <b>object </b>parameters in C functions.   
<h3> <a name="CVarAndTypeDecls"></a>C variable and type definitions</h3>
   The <b>cdef</b> statement is also used to declare C variables, either
local  or module-level:   
<blockquote>      
  <pre>cdef int i, j, k<br>cdef float f, g[42], *h</pre>
   </blockquote>
   and C struct, union or enum types:   
<blockquote>      
  <pre>cdef struct Grail:<br>&nbsp;&nbsp;&nbsp; int age<br>&nbsp;&nbsp;&nbsp; float volume</pre>
         
  <pre>cdef union Food:<br>&nbsp;&nbsp;&nbsp; char *spam<br>&nbsp;&nbsp;&nbsp; float *eggs</pre>
         
  <pre>cdef enum CheeseType:<br>&nbsp;&nbsp;&nbsp; cheddar, edam,&nbsp;<br>&nbsp;&nbsp;&nbsp; camembert</pre>
         
  <pre>cdef enum CheeseState:<br>&nbsp;&nbsp;&nbsp; hard = 1<br>&nbsp;&nbsp;&nbsp; soft = 2<br>&nbsp;&nbsp;&nbsp; runny = 3</pre>
   </blockquote>
   There is currently no special syntax for defining a constant, but you
can  use an anonymous enum declaration for this purpose, for example,   
<blockquote><tt>cdef enum:</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; tons_of_spam = 3</tt></blockquote>
   Note that the words struct, union and enum are used only when <i>defining</i> 
a type, not when referring to it. For example, to declare a variable pointing 
 to a Grail you would write   
<blockquote>      
  <pre>cdef Grail *gp</pre>
   </blockquote>
   and <i>not</i>  
<blockquote>      
  <pre>cdef struct Grail *gp <font color="#ed181e"># WRONG</font></pre>
   </blockquote>
   There is also a <b>ctypedef</b> statement for giving names to types, e.g. 
  
<blockquote>      
  <pre>ctypedef unsigned long ULong</pre>
         
  <pre>ctypedef int *IntPtr</pre>
   </blockquote>
     
<h3> <a name="ScopeRules"></a>Scope rules</h3>
   Pyrex determines whether a variable belongs to a local scope, the module 
 scope, or the built-in scope <i>completely statically.</i> As with Python, 
 assigning to a variable which is not otherwise declared implicitly declares 
 it to be a Python variable residing in the scope where it is assigned. Unlike
 Python, however, a name which is referred to but not declared or assigned
 is assumed to reside in the <i>builtin scope, </i>not the module scope.
Names  added to the module dictionary at run time will not shadow such names.
  
<p>You can use a <b>global</b> statement at the module level to explicitly 
 declare a name to be a module-level name when there would otherwise not be
any indication of this, for example, </p>
   
<blockquote><tt>global __name__</tt> <br>
    <tt>print __name__</tt></blockquote>
   Without the <b>global</b> statement, the above would print the name of 
the builtins module.<br>
  <br>
  Note: A consequence of these rules is that the module-level scope behaves
 the same way as a Python local scope if you refer to a variable before assigning
 to it. In particular, tricks such as the following will <i>not</i> work
in  Pyrex:<br>
   
<blockquote>      
  <pre>try:<br>&nbsp; x = True<br>except NameError:<br>&nbsp; True = 1<br></pre>
  </blockquote>
  because, due to the assignment, the True will always be looked up in the
 module-level scope. You would have to do something like this instead:<br>
   
<blockquote>      
  <pre>import __builtin__<br>try:<br>  True = __builtin__.True<br>except AttributeError:<br>  True = 1<br></pre>
  </blockquote>
   
<hr width="100%">  
<h3> <a name="StatsAndExprs"></a>Statements and expressions</h3>
   Control structures and expressions follow Python syntax for the most part. 
 When applied to Python objects, they have the same semantics as in Python 
 (unless otherwise noted). Most of the Python operators can also be applied 
 to C values, with the obvious semantics.   
<p>If Python objects and C values are mixed in an expression, conversions 
 are performed automatically between Python objects and C numeric or string 
 types. </p>
   
<p>Reference counts are maintained automatically for all Python objects, and
all Python operations are automatically checked for errors, with appropriate 
 action taken. </p>
   
<h4> <a name="ExprSyntaxDifferences"></a>Differences between C and Pyrex
expression syntax</h4>
   Pyrex also includes some C operations which have no direct Python equivalent. 
 Some of them are expressed differently in Pyrex than in C.   
<ul>
   <li> There is no <b><tt>-&gt;</tt></b> operator in Pyrex. Instead of <tt>p-&gt;x</tt>, 
 use <tt>p.x</tt></li>
    <br>
  &nbsp; <li> There is no <b><tt>*</tt></b> operator in Pyrex. Instead of 
    <tt>*p</tt>,  use <tt>p[0]</tt></li>
    <br>
  &nbsp; <li> There is an <b><tt>&amp;</tt></b> operator, with the same semantics
 as in C</li>
    <br>
  &nbsp; <li> The null C pointer is called <b><tt>NULL</tt></b>, not 0 (and
     <tt>NULL</tt> is a reserved word).</li>
    <br>
  &nbsp; <li> Character literals are written with a <b>c</b> prefix, for
example:</li>
         
  <ul>
            
    <pre>c'X'</pre>
        
  </ul>
    <li> Type casts are written <b><tt>&lt;type&gt;value</tt></b> , for example:</li>
         
  <ul>
            
    <pre>cdef char *p, float *q<br>p = &lt;char*&gt;q</pre>
        
  </ul>
   <i><b>Warning</b>: Don't attempt to use a typecast to convert between
Python  and C data types -- it won't do the right thing. Leave Pyrex to perform 
the conversion automatically.</i>  
</ul>
     
<h4> <a name="ForFromLoop"></a>Integer for-loops</h4>
   You should be aware that a for-loop such as   
<blockquote><tt>for i in range(n):</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
   won't be very fast, even if <tt>i</tt> and <tt>n</tt> are declared as
C integers, because <tt>range</tt> is a Python function. For iterating over 
ranges of integers, Pyrex has another form of for-loop:   
<blockquote><tt>for i from 0 &lt;= i &lt; n:</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
   If the loop variable and the lower and upper bounds are all C integers, 
this form of loop will be much faster, because Pyrex will translate it into 
pure C code.   
<p>Some things to note about the <tt>for-from</tt> loop: </p>
   
<ul>
   <li> The target expression must be a variable name.</li>
    <li> The name between the lower and upper bounds must be the same as
the  target name.</li>
    <li> The direction of iteration is determined by the relations. If they
 are both from the set {<tt>&lt;</tt>, <tt>&lt;=</tt>} then it is upwards;
 if they are both from the set {<tt>&gt;</tt>, <tt>&gt;=</tt>} then it is
downwards. (Any other combination is disallowed.)</li>
    
</ul>
   Like other Python looping statements, <tt>break</tt> and <tt>continue</tt> 
may be used in the body, and the loop may have an <tt>else</tt> clause. 
 
<h2>   
<hr width="100%"></h2>
     
<h3> <a name="ExceptionValues"></a>Error return values</h3>
   If you don't do anything special, a function declared with <b>cdef</b> 
that does not return a Python object has no way of reporting Python exceptions 
 to its caller. If an exception is detected in such a function, a warning 
message is printed and the exception is ignored.   
<p>If you want a C function that does not return a Python object to be able
 to propagate exceptions to its caller, you need to declare an <b>exception 
 value</b> for it. Here is an example: </p>
   
<blockquote><tt>cdef int spam() except -1:</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
   With this declaration, whenever an exception occurs inside <tt>spam</tt>, 
 it will immediately return with the value <tt>-1</tt>. Furthermore, whenever 
 a call to <tt>spam</tt> returns <tt>-1</tt>, an exception will be assumed 
 to have occurred and will be propagated.   
<p>When you declare an exception value for a function, you should never explicitly
 return that value. If all possible return values are legal and you can't
reserve one entirely for signalling errors, you can use an alternative form
of exception value declaration: </p>
   
<blockquote><tt>cdef int spam() except? -1:</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
   The "?" indicates that the value <tt>-1</tt> only indicates a <i>possible</i> 
error. In this case, Pyrex generates a call to <tt>PyErr_Occurred</tt>if the
exception value is returned, to make sure it really is an error.   
<p>There is also a third form of exception value declaration: </p>
   
<blockquote><tt>cdef int spam() except *:</tt> <br>
    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
   This form causes Pyrex to generate a call to <tt>PyErr_Occurred</tt> after 
 <i>every</i> call to spam, regardless of what value it returns. If you have
 a function returning <tt>void</tt> that needs to propagate errors, you will
 have to use this form, since there isn't any return value to test.   
<p>Some things to note: </p>
   
<ul>
   <li> Currently, exception values can only declared for functions returning
 an integer, float or pointer type, and the value must be a <i>literal</i>, 
 not an expression (although it can be negative). The only possible pointer 
 exception value is <tt>NULL</tt>. Void functions can only use the <tt>except 
 *</tt> form.</li>
    <br>
  &nbsp; <li> The exception value specification is part of the signature
of  the function. If you're passing a pointer to a function as a parameter
or  assigning it to a variable, the declared type of the parameter or variable
 must have the same exception value specification (or lack thereof). Here
is an example of a pointer-to-function declaration with an exception value:</li>
         
  <ul>
            
    <pre><tt>int (*grail)(int, char *) except -1</tt></pre>
        
  </ul>
    <li> You don't need to (and shouldn't) declare exception values for functions 
 which return Python objects. Remember that a function with no declared return
 type implicitly returns a Python object.</li>
    
</ul>
     
<h4> <a name="CheckingReturnValues"></a>Checking return values of non-Pyrex 
 functions</h4>
   It's important to understand that the <tt>except</tt> clause does <i>not</i> 
cause an error to be <i>raised</i> when the specified value is returned. For
example, you can't write something like   
<blockquote>      
  <pre>cdef extern FILE *fopen(char *filename, char *mode) except NULL <font
 color="#ed181e"># WRONG!</font></pre>
   </blockquote>
   and expect an exception to be automatically raised if a call to fopen
returns  NULL. The except clause doesn't work that way; its only purpose
is for <i>propagating</i> exceptions that have already been raised, either
by a Pyrex function or a  C function that calls Python/C API routines. To
get an exception from a non-Python-aware  function such as fopen, you will
have to check the return value and raise  it yourself, for example,   
<blockquote>      
  <pre>cdef FILE *p<br>p = fopen("spam.txt", "r")<br>if p == NULL:<br>&nbsp;&nbsp;&nbsp; raise SpamError("Couldn't open the spam file")</pre>
   </blockquote>
     
<h4>   
<hr width="100%"></h4>
     
<h4> <a name="IncludeStatement"></a>The <tt>include</tt> statement</h4>
   For convenience, a large Pyrex module can be split up into a number of 
files which are put together using the <b>include</b> statement, for example 
  
<blockquote>      
  <pre>include "spamstuff.pxi"</pre>
   </blockquote>
   The contents of the named file are textually included at that point. The 
 included file can contain any complete top-level Pyrex statements, including 
 other <b>include</b> statements. The <b>include</b> statement itself can 
only appear at the top level of a file.   
<p>The <b>include</b> statement can also be used in conjunction with <a
 href="#PublicDecls"><b>public</b> declarations</a> to make C functions and
 variables defined in one Pyrex module accessible to another. However, note
 that some of these uses have been superseded by the facilities described
in <a href="sharing.html">Sharing Declarations Between Pyrex Modules</a>,
and it is expected that use of the <b>include</b> statement for this purpose
will be phased out altogether in future versions. </p>
   
<h2>   
<hr width="100%"><a name="InterfacingWithExternal"></a>Interfacing with External
 C Code   
<hr width="100%"></h2>
   One of the main uses of Pyrex is wrapping existing libraries of C code. 
This is achieved by using <a href="#ExternDecls">external declarations</a> 
to declare the C functions and variables from the library that you want to
 use.   
<p>You can also use <a href="#PublicDecls">public declarations</a> to make 
 C functions and variables defined in a Pyrex module available to external 
 C code. The need for this is expected to be less frequent, but you might 
want to do it, for example, if you are embedding Python in another application 
 as a scripting language. Just as a Pyrex module can be used as a bridge to
allow Python code to call C code, it can also be used to allow C code to
call Python code. </p>
   
<h3> <a name="ExternDecls"></a>External declarations</h3>
   By default, C functions and variables declared at the module level are 
local to the module (i.e. they have the C <b>static</b> storage class). They 
can also be declared <b>extern</b> to specify that they are defined elsewhere,
 for example:   
<blockquote>      
  <pre>cdef extern int spam_counter</pre>
         
  <pre>cdef extern void order_spam(int tons)</pre>
   </blockquote>
     
<blockquote>         </blockquote>
     
<h4> <a name="ReferencingHeaders"></a>Referencing C header files</h4>
   When you use an extern definition on its own as in the examples above, 
Pyrex includes a declaration for it in the generated C file. This can cause 
problems if the declaration doesn't exactly match the declaration that will 
be seen by other C code. If you're wrapping an existing C library, for example, 
it's important that the generated C code is compiled with exactly the same 
declarations as the rest of the library.   
<p>To achieve this, you can tell Pyrex that the declarations are to be found
 in a C header file, like this: </p>
   
<blockquote>      
  <pre>cdef extern from "spam.h":</pre>
         
  <pre>&nbsp;&nbsp;&nbsp; int spam_counter</pre>
         
  <pre>&nbsp;&nbsp;&nbsp; void order_spam(int tons)</pre>
   </blockquote>
   The <b>cdef extern from</b> clause does three things:   
<ol>
   <li> It directs Pyrex to place a <b>#include</b> statement for the named
 header file in the generated C code.<br>
   </li>
  &nbsp; <li> It prevents Pyrex from generating any C code for the declarations
 found in the associated block.<br>
   </li>
  &nbsp; <li> It treats all declarations within the block as though they
started  with <b>cdef extern</b>.</li>
    
</ol>
   It's important to understand that Pyrex does <i>not</i> itself read the 
C header file, so you still need to provide Pyrex versions of any declarations 
 from it that you use. However, the Pyrex declarations don't always have to
exactly match the C ones, and in some cases they shouldn't or can't. In particular:
  
<ol>
   <li> Don't use <b>const</b>. Pyrex doesn't know anything about const,
so  just leave it out. Most of the time this shouldn't cause any problem,
although  on rare occasions you might have to use a cast.<sup><a
 href="#Footnote1"> 1</a></sup><br>
   </li>
  &nbsp; <li> Leave out any platform-specific extensions to C declarations
 such as <b>__declspec()</b>.<br>
   </li>
  &nbsp; <li> If the header file declares a big struct and you only want
to  use a few members, you only need to declare the members you're interested 
in. Leaving the rest out doesn't do any harm, because the C compiler will 
use the full definition from the header file.<br>
     <br>
 In some cases, you might not need <i>any</i> of the struct's members, in
which case you can just put <tt>pass</tt> in the body of the struct declaration,
e.g.<br>
     <br>
     <tt>&nbsp; &nbsp; cdef extern from "foo.h":<br>
 &nbsp; &nbsp; &nbsp; &nbsp; struct spam:<br>
 &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; pass</tt><br>
    <br>
Note that you can only do this inside a <b>cdef extern from</b> block; struct
declarations anywhere else must be non-empty.<br>
    <br>
   </li>
    <li> If the header file uses typedef names such as <b>size_t </b>to refer 
to platform-dependent flavours of numeric types, you will need a corresponding 
     <b>ctypedef</b> statement, but you don't need to match the type exactly,
 just use something of the right general kind (int, float, etc). For example,</li>
         
  <ol>
            
    <pre>ctypedef int size_t</pre>
        
  </ol>
   will work okay whatever the actual size of a size_t is (provided the header 
 file defines it correctly). <br>
  &nbsp; <li> If the header file uses macros to define constants, translate
 them into a dummy <b>enum</b> declaration.<br>
   </li>
  &nbsp; <li> If the header file defines a function using a macro, declare
 it as though it were an ordinary function, with appropriate argument and
result types.</li>
    
</ol>
   A few more tricks and tips:   
<ul>
   <li> If you want to include a C header because it's needed by another
header,  but don't want to use any declarations from it, put <tt><font
 size="+1">pass</font></tt> in the extern-from block:</li>
    
</ul>
     
<ul>
        
  <ul>
      <tt>cdef extern from "spam.h":</tt> <br>
      <tt>&nbsp;&nbsp;&nbsp; pass</tt>      
  </ul>
    
</ul>
     
<ul>
   <li> If you want to include some external declarations, but don't want 
to specify a header file (because it's included by some other header that 
you've already included) you can put <tt>*</tt> in place of the header file 
name:</li>
    
</ul>
     
<blockquote>      
  <blockquote><tt>cdef extern from *:</tt> <br>
      <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
   </blockquote>
     
<h4> <a name="StructDeclStyles"></a>Styles of struct, union and enum declaration</h4>
   There are two main ways that structs, unions and enums can be declared 
in C header files: using a tag name, or using a typedef. There are also some
 variations based on various combinations of these.   
<p>It's important to make the Pyrex declarations match the style used in the
header file, so that Pyrex can emit the right sort of references to the type
in the code it generates. To make this possible, Pyrex provides two different
syntaxes for declaring a struct, union or enum type. The style introduced
above corresponds to the use of a tag name. To get the other style, you prefix
the declaration with <b>ctypedef</b>, as illustrated below. </p>
   
<p>The following table shows the various possible styles that can be found 
 in a header file, and the corresponding Pyrex declaration that you should 
 put in the <b>cdef exern from </b>block. Struct declarations are used as 
an example; the same applies equally to union and enum declarations. </p>
   
<p>Note that in all the cases below, you refer to the type in Pyrex code simply
as <tt><font size="+1">Foo</font></tt>, not <tt><font size="+1">struct Foo</font></tt>.
 <br>
  &nbsp;  
<table cellpadding="5">
   <tbody>
      <tr valign="top" bgcolor="#8cbc1c">
   <td bgcolor="#8cbc1c">&nbsp;</td>
    <td nowrap="nowrap" bgcolor="#ff9933"><b>C code</b></td>
    <td valign="top" bgcolor="#66cccc"><b>Possibilities for corresponding 
Pyrex  code</b></td>
    <td valign="top" bgcolor="#99cc33"><b>Comments</b></td>
   </tr>
    <tr valign="top" bgcolor="#8cbc1c">
   <td>1</td>
    <td bgcolor="#ff9900"><tt>struct Foo {</tt> <br>
        <tt>&nbsp; ...</tt> <br>
        <tt>};</tt></td>
    <td bgcolor="#66cccc"><tt>cdef struct Foo:</tt> <br>
        <tt>&nbsp; ...</tt></td>
    <td>Pyrex will refer to the type as <tt>struct Foo </tt>in the generated 
 C code<tt>.</tt></td>
   </tr>
    <tr valign="top" bgcolor="#8cbc1c">
   <td valign="top">2</td>
    <td nowrap="nowrap" bgcolor="#ff9900"><tt>typedef struct {</tt> <br>
        <tt>&nbsp; ...</tt> <br>
        <tt>} Foo;</tt></td>
    <td valign="top" bgcolor="#66cccc"><tt>ctypedef struct Foo:</tt> <br>
        <tt>&nbsp; ...</tt></td>
    <td valign="top">Pyrex will refer to the type simply as <tt>Foo</tt>
in  the generated C code.</td>
   </tr>
    <tr valign="top" bgcolor="#8cbc1c">
   <td rowspan="2">3</td>
    <td rowspan="2" nowrap="nowrap" bgcolor="#ff9900"><tt>typedef struct
foo  {</tt> <br>
        <tt>&nbsp; ...</tt> <br>
        <tt>} Foo;</tt></td>
    <td valign="top" nowrap="nowrap" bgcolor="#66cccc"><tt>cdef struct foo:</tt> 
      <br>
        <tt>&nbsp; ...</tt> <br>
        <tt>ctypedef foo Foo #optional</tt></td>
    <td valign="top" rowspan="2">If the C header uses both a tag and a typedef 
 with <i>different</i> names, you can use either form of declaration in Pyrex
 (although if you need to forward reference the type, you'll have to use
the  first form).</td>
   </tr>
    <tr>
   <td bgcolor="#66cccc"><tt>ctypedef struct Foo:</tt> <br>
        <tt>&nbsp; ...</tt></td>
   </tr>
    <tr valign="top" bgcolor="#8cbc1c">
   <td>4</td>
    <td nowrap="nowrap" bgcolor="#ff9900"><tt>typedef struct Foo {</tt> <br>
        <tt>&nbsp; ...</tt> <br>
        <tt>} Foo;</tt></td>
    <td valign="top" bgcolor="#66cccc"><tt>cdef struct Foo:</tt> <br>
        <tt>&nbsp; ...</tt></td>
    <td>If the header uses the <i>same</i> name for the tag and the typedef, 
 you won't be able to include a <b>ctypedef</b> for it -- but then, it's not
necessary.</td>
   </tr>
        
  </tbody>  
</table>
    </p>
   
<h4> <a name="AccessingAPI"></a>Accessing Python/C API routines</h4>
   One particular use of the <b>cdef extern from</b> statement is for gaining 
 access to routines in the Python/C API. For example,   
<blockquote>      
  <pre>cdef extern from "Python.h":</pre>
         
  <pre>&nbsp;&nbsp;&nbsp; object PyString_FromStringAndSize(char *s, int len)</pre>
   </blockquote>
   will allow you to create Python strings containing null bytes.   
<p> </p>
   
<hr width="100%">  
<h3> <a name="CNameSpecs"></a>Resolving naming conflicts - C name specifications</h3>
   Each Pyrex module has a single module-level namespace for both Python
and  C names. This can be inconvenient if you want to wrap some external
C functions  and provide the Python user with Python functions of the same
names.   
<p>Pyrex 0.8 provides a couple of different ways of solving this problem. 
 The best way, especially if you have many C functions to wrap, is probably 
 to put the extern C function declarations into a different namespace using 
 the facilities described in the section on <a href="sharing.html">sharing 
 declarations between Pyrex modules</a>. </p>
   
<p>The other way is to use a <b>c name specification</b> to give different 
 Pyrex and C names to the C function. Suppose, for example, that you want 
to wrap an external function called <tt>eject_tomato</tt>. If you declare 
it as </p>
   
<blockquote>      
  <pre>cdef extern void c_eject_tomato "eject_tomato" (float speed)</pre>
   </blockquote>
   then its name inside the Pyrex module will be <tt>c_eject_tomato</tt>, 
whereas its name in C will be <tt>eject_tomato</tt>. You can then wrap it 
with   
<blockquote>      
  <pre>def eject_tomato(speed):<br>&nbsp; c_eject_tomato(speed)</pre>
   </blockquote>
   so that users of your module can refer to it as <tt>eject_tomato</tt>. 
  
<p>Another use for this feature is referring to external names that happen 
 to be Pyrex keywords. For example, if you want to call an external function 
 called <tt>print</tt>, you can rename it to something else in your Pyrex 
module.  </p>
   
<p>As well as functions, C names can be specified for variables, structs, 
 unions, enums, struct and union members, and enum values. For example, </p>
   
<blockquote>      
  <pre>cdef extern int one "ein", two "zwei"<br>cdef extern float three "drei"<br><br>cdef struct spam "SPAM":<br>&nbsp; int i "eye"</pre>
   <tt>cdef enum surprise "inquisition":</tt> <br>
    <tt>&nbsp; first "alpha"</tt> <br>
    <tt>&nbsp; second "beta" = 3</tt></blockquote>
     
<hr width="100%">  
<h3> <a name="PublicDecls"></a>Public Declarations</h3>
   You can make C variables and functions defined in a Pyrex module accessible 
 to external C code (or another Pyrex module) using the <b><tt>public</tt></b> 
keyword, as follows:   
<blockquote><tt>cdef public int spam # public variable declaration</tt>  
   
  <p><tt>cdef public void grail(int num_nuns): # public function declaration</tt> 
  <br>
    <tt>&nbsp;&nbsp;&nbsp; ...</tt></p>
  </blockquote>
   If there are any <tt>public</tt> declarations in a Pyrex module, a <b>.h</b> 
file is generated containing equivalent C declarations for inclusion in other
 C code.   
<p>Pyrex also generates a <b>.pxi</b> file containing Pyrex versions of the
 declarations for inclusion in another Pyrex module using the <b><a
 href="#IncludeStatement">include</a></b> statement. If you use this, you
 will need to arrange for the module using the declarations to be linked
against  the module defining them, and for both modules to be available to
the dynamic  linker at run time. I haven't tested this, so I can't say how
well it will  work on the various platforms. </p>
   
<blockquote>NOTE: If all you want to export is an extension type, there is
 now a better way -- see <a href="sharing.html">Sharing Declarations Between
 Pyrex Modules</a>.</blockquote>
     
<h2>   
<hr width="100%">Extension Types   
<hr width="100%"></h2>
   One of the most powerful features of Pyrex is the ability to easily create 
 new built-in Python types, called <b>extension types</b>. This is a major 
 topic in itself, so there is a&nbsp; <a href="extension_types.html">separate 
 page</a> devoted to it.   
<h2>   
<hr width="100%">Sharing Declarations Between Pyrex Modules   
<hr width="100%"></h2>
   Pyrex 0.8 introduces a substantial new set of facilities allowing a Pyrex 
 module to easily import and use C declarations and extension types from another
Pyrex module. You can now create a set of co-operating Pyrex modules just
as easily as you can create a set of co-operating Python modules. There is
a <a href="sharing.html">separate page</a> devoted to this topic.   
<h2>   
<hr width="100%"><a name="Limitations"></a>Limitations   
<hr width="100%"></h2>
     
<h3> <a name="Unsupported"></a>Unsupported Python features</h3>
   Pyrex is not quite a full superset of Python. The following restrictions 
 apply:   
<blockquote> <li> Function definitions (whether using <b>def</b> or <b>cdef</b>)
 cannot be nested within other function definitions.<br>
   </li>
  &nbsp; <li> Class definitions can only appear at the top level of a module,
 not inside a function.<br>
   </li>
  &nbsp; <li> The<tt> import *</tt> form of import is not allowed anywhere
 (other forms of the import statement are fine, though).<br>
   </li>
  &nbsp; <li> Generators cannot be defined in Pyrex.<br>
     <br>
   </li>
    <li> The <tt>globals()</tt> and <tt>locals()</tt> functions cannot be 
used.</li>
   </blockquote>
   The above restrictions will most likely remain, since removing them would 
 be difficult and they're not really needed for Pyrex's intended applications. 
  
<p>There are also some temporary limitations which may eventually be lifted: 
 </p>
   
<blockquote> <li> Class and function definitions cannot be placed inside
control structures.<br>
   </li>
  &nbsp; <li> In-place arithmetic operators (+=, etc) are not yet supported.<br>
   </li>
  &nbsp; <li> List comprehensions are not yet supported.<br>
   </li>
  &nbsp; <li> There is no support for Unicode.<br>
   </li>
  &nbsp; <li> Special methods of extension types cannot have functioning
docstrings.<br>
     <br>
   </li>
    <li> The use of string literals as comments is not recommended at present,
 because Pyrex doesn't optimize them away, and won't even accept them in
places  where executable statements are not allowed.</li>
   </blockquote>
   There are probably also some other gaps which I can't think of at the
moment.   
<h3> <a name="SemanticDifferences"></a>Semantic differences between Python
 and Pyrex</h3>
     
<h4> Behaviour of class scopes</h4>
   In Python, referring to a method of a class inside the class definition, 
 i.e. while the class is being defined, yields a plain function object, but
 in Pyrex it yields an unbound method<sup><font size="-2"><a
 href="#Footnote2">2</a></font></sup>. A consequence of this is that the
usual idiom for using the classmethod and staticmethod functions, e.g.   
<blockquote>      
  <pre>class Spam:</pre>
         
  <pre>&nbsp; def method(cls):<br>&nbsp;&nbsp;&nbsp; ...</pre>
         
  <pre>&nbsp; method = classmethod(method)</pre>
   </blockquote>
   will not work in Pyrex. This can be worked around by defining the function 
 <i>outside</i> the class, and then assigning the result of classmethod or
 staticmethod inside the class, i.e.   
<blockquote>      
  <pre>def Spam_method(cls):<br>&nbsp; ...</pre>
         
  <pre>class Spam:</pre>
         
  <pre>&nbsp; method = classmethod(Spam_method)</pre>
   </blockquote>
     
<h1>   
<hr width="100%"><font size="+0">Footnotes</font>  
<hr width="100%"></h1>
   <a name="Footnote1"></a>1. A problem with const could arise if you have 
something like   
<blockquote>      
  <pre>cdef extern from "grail.h":<br>&nbsp; char *nun</pre>
   </blockquote>
   where grail.h actually contains   
<blockquote>      
  <pre>extern const char *nun;</pre>
   </blockquote>
   and you do   
<blockquote>      
  <pre>cdef void languissement(char *s):<br>&nbsp; #something that doesn't change s</pre>
         
  <pre>...</pre>
         
  <pre>languissement(nun)</pre>
   </blockquote>
   which will cause the C compiler to complain. You can work around it by 
casting  away the constness:   
<blockquote>      
  <pre>oral(&lt;char *&gt;nun)</pre>
   </blockquote>
     
<hr width="100%"><a name="Footnote2"></a>2. The reason for the different behaviour
of class scopes is that Pyrex-defined Python functions are PyCFunction objects,
not PyFunction objects, and are not recognised by the machinery that creates
a bound or unbound method when a function is extracted from a class. To get
around this, Pyrex wraps each method in an unbound method object itself before
storing it in the class's dictionary. <br>
  &nbsp; <br>
  <br>
 
</body>
</html>
\ No newline at end of file
+<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
+<html><head>
+             
+  <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
+
+             
+  <meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]">
+  <title>Pyrex Language Overview</title>
+</head>
+<body>
+     
+<h1>   <hr width="100%">Overview of the Pyrex Language&nbsp;  <hr width="100%"></h1>
+
+   This document informally describes the extensions to the Python language 
+ made by Pyrex. Some day there will be a reference manual covering everything 
+ in more detail. <br>
+
+  &nbsp;  
+<h2> Contents</h2>
+
+     
+<ul>
+
+   <li> <a href="#Basics">Basics</a></li>
+           <ul>
+   <li> <a href="#PyFuncsVsCFuncs">Python functions vs. C functions</a></li>
+    <li> <a href="#PyObjParams">Python objects as parameters</a></li>
+    <li> <a href="#CVarAndTypeDecls">C variable and type definitions</a></li><li><a href="#AutomaticTypeConversions">Automatic type conversions</a></li>
+    <ul>
+      <li><a href="#PyToCStringCaveats">Caveats when using a Python string in a C context</a></li>
+    </ul>
+
+    <li> <a href="#ScopeRules">Scope rules</a></li>
+    <li> <a href="#StatsAndExprs">Statements and expressions</a></li>
+                 <ul>
+   <li> <a href="#ExprSyntaxDifferences">Differences between C and Pyrex
+expressions<br>
+</a></li>
+    <li> <a href="#ForFromLoop">Integer for-loops</a></li>
+                </ul>
+    <li> <a href="#ExceptionValues">Error return values</a></li>
+                 <ul>
+   <li> <a href="#CheckingReturnValues">Checking return values of non-Pyrex
+ functions</a></li>
+                </ul>
+    <li> <a href="#IncludeStatement">The <tt>include</tt> statement</a></li>
+          </ul>
+    <li> <a href="#InterfacingWithExternal">Interfacing with External C Code</a></li>
+           <ul>
+   <li> <a href="#ExternDecls">External declarations</a></li>
+                 <ul>
+   <li> <a href="#ReferencingHeaders">Referencing C header files</a></li>
+    <li> <a href="#StructDeclStyles">Styles of struct/union/enum declaration</a></li>
+    <li> <a href="#AccessingAPI">Accessing Python/C API routines</a></li>
+    <li> <a href="#CNameSpecs">Resolving naming conflicts - C name specifications</a></li>
+                </ul>
+    <li> <a href="#PublicDecls">Public declarations</a></li>
+          </ul>
+    <li> <a href="extension_types.html">Extension Types</a> <font color="#006600">(Section revised in 0.9)</font></li>
+    <li> <a href="sharing.html">Sharing Declarations Between Pyrex Modules</a>
+     <font color="#006600">(NEW in 0.8)</font></li>
+    <li> <a href="#Limitations">Limitations</a></li>
+           <ul>
+   <li> <a href="#Unsupported">Unsupported Python features</a></li>
+    <li> <a href="#SemanticDifferences">Semantic differences between Python
+ and Pyrex</a></li>
+          </ul>
+    
+</ul>
+
+     
+<h2>   <hr width="100%"><a name="Basics"></a>Basics   
+<hr width="100%"></h2>
+
+   This section describes the basic features of the Pyrex language. The facilities 
+ covered in this section allow you to create Python-callable functions that 
+ manipulate C data structures and convert between Python and C data types. 
+ Later sections will cover facilities for <a href="#InterfacingWithExternal">wrapping external C code</a>, <a href="extension_types.html">creating new Python types</a> and <a href="sharing.html">cooperation between Pyrex modules</a>.   
+<h3> <a name="PyFuncsVsCFuncs"></a>Python functions vs. C functions</h3>
+
+   There are two kinds of function definition in Pyrex:   
+<p><b>Python functions</b> are defined using the <b>def</b> statement, as
+ in Python. They take Python objects as parameters and return Python objects. 
+ </p>
+
+   
+<p><b>C functions</b> are defined using the new <b>cdef</b> statement. They
+ take either Python objects or C values as parameters, and can return either
+ Python objects or C values. </p>
+
+   
+<p>Within a Pyrex module, Python functions and C functions can call each other
+freely, but only Python functions can be called from outside the module by
+interpreted Python code. So, any functions that you want to "export" from
+ your Pyrex module must be declared as Python functions using <span style="font-weight: bold;">def</span>. </p>
+
+   
+<p>Parameters of either type of function can be declared to have C data types,
+ using normal C declaration syntax. For example, </p>
+
+   
+<blockquote>        <pre>def spam(int i, char *s):<br>&nbsp;&nbsp;&nbsp; ...</pre>
+           <pre>cdef int eggs(unsigned long l, float f):<br>&nbsp;&nbsp;&nbsp; ...</pre>
+   </blockquote>
+
+   When a parameter of a Python function is declared to have a C data type, 
+ it is passed in as a Python object and automatically converted to a C value, 
+ if possible. Automatic conversion is currently only possible for numeric 
+types and string types; attempting to use any other type for the parameter 
+of a Python function will result in a compile-time error.   
+<p>C functions, on the other hand, can have parameters of any type, since 
+ they're passed in directly using a normal C function call. </p>
+
+   
+<h3> <a name="PyObjParams"></a>Python objects as parameters and return values</h3>
+
+   If no type is specified for a parameter or return value, <i>it is assumed 
+ to be a Python object.</i> (Note that this is different from the C convention, 
+ where it would default to <tt>int</tt>.) For example, the following defines 
+ a C function that takes two Python objects as parameters and returns a Python
+ object:   
+<blockquote>        <pre>cdef spamobjs(x, y):<br>&nbsp;&nbsp;&nbsp; ...</pre>
+   </blockquote>
+
+   Reference counting for these objects is performed automatically according 
+ to the standard Python/C API rules (i.e. borrowed references are taken as
+ parameters and a new reference is returned).   
+<p>The name <b>object</b> can also be used to explicitly declare something 
+ as a Python object. This can be useful if the name being declared would otherwise
+be taken as the name of a type, for example, </p>
+
+   
+<blockquote>        <pre>cdef ftang(object int):<br>&nbsp;&nbsp;&nbsp; ...</pre>
+   </blockquote>
+
+   declares a parameter called <tt>int</tt> which is a Python object. You 
+can also use <b>object </b>as the explicit return type of a function, e.g. 
+  
+<blockquote>        <pre>cdef object ftang(object int):<br>&nbsp;&nbsp;&nbsp; ...</pre>
+   </blockquote>
+
+   In the interests of clarity, it is probably a good idea to always be explicit 
+ about <b>object </b>parameters in C functions.   
+<h3> <a name="CVarAndTypeDecls"></a>C variable and type definitions</h3>
+
+   The <b>cdef</b> statement is also used to declare C variables, either
+local  or module-level:   
+<blockquote>        <pre>cdef int i, j, k<br>cdef float f, g[42], *h</pre>
+   </blockquote>
+
+   and C struct, union or enum types:   
+<blockquote>        <pre>cdef struct Grail:<br>&nbsp;&nbsp;&nbsp; int age<br>&nbsp;&nbsp;&nbsp; float volume</pre>
+           <pre>cdef union Food:<br>&nbsp;&nbsp;&nbsp; char *spam<br>&nbsp;&nbsp;&nbsp; float *eggs</pre>
+           <pre>cdef enum CheeseType:<br>&nbsp;&nbsp;&nbsp; cheddar, edam,&nbsp;<br>&nbsp;&nbsp;&nbsp; camembert</pre>
+           <pre>cdef enum CheeseState:<br>&nbsp;&nbsp;&nbsp; hard = 1<br>&nbsp;&nbsp;&nbsp; soft = 2<br>&nbsp;&nbsp;&nbsp; runny = 3</pre>
+   </blockquote>
+
+   There is currently no special syntax for defining a constant, but you
+can  use an anonymous enum declaration for this purpose, for example,   
+<blockquote><tt>cdef enum:</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; tons_of_spam = 3</tt></blockquote>
+
+   Note that the words <span style="font-family: monospace;">struct</span>, <span style="font-family: monospace;">union</span> and <span style="font-family: monospace;">enum</span> are used only when <i>defining</i> a type, not when referring to it. For example, to declare a variable pointing 
+ to a Grail you would write   
+<blockquote>        <pre>cdef Grail *gp</pre>
+   </blockquote>
+
+   and <i>not</i>  
+<blockquote>        <pre>cdef struct Grail *gp <font color="#ed181e"># WRONG</font></pre>
+   </blockquote>
+
+   There is also a <b>ctypedef</b> statement for giving names to types, e.g. 
+  
+<blockquote>        <pre>ctypedef unsigned long ULong</pre>
+           <pre>ctypedef int *IntPtr<br></pre></blockquote>
+
+<h3><a name="AutomaticTypeConversions"></a>Automatic type conversions</h3>
+
+In most situations, automatic conversions will be performed for the
+basic numeric and string types when a Python object is used in a
+context requiring a C value, or vice versa. The following table
+summarises the conversion possibilities.<br>
+
+<br>
+
+<table style="margin-left: auto; margin-right: auto; width: 10%; text-align: left;" border="1" cellpadding="4" cellspacing="0">
+
+  <tbody>
+    <tr>
+      <th style="vertical-align: top; width: 40%; white-space: nowrap;">C types<br>
+      </th>
+      <th style="vertical-align: top; width: 150px; white-space: nowrap;">From Python types<br>
+      </th>
+      <th style="vertical-align: top; width: 150px; white-space: nowrap;">To Python types<br>
+      </th>
+    </tr>
+    <tr>
+      <td colspan="1" rowspan="1" style="vertical-align: top; width: 40%; white-space: nowrap;">[unsigned] char<br>
+[unsigned] short<br>
+      int, long</td>
+      <td colspan="1" rowspan="1" style="vertical-align: top; width: 150px; white-space: nowrap;">int, long<br>
+      </td>
+      <td colspan="1" rowspan="1" style="vertical-align: top; width: 150px; white-space: nowrap;">int<br>
+      </td>
+    </tr>
+    <tr>
+    </tr>
+
+    <tr>
+      <td colspan="1" rowspan="1" style="vertical-align: top; width: 40%; white-space: nowrap;">unsigned int<br>
+unsigned long<br>
+      [unsigned] long long<br>
+
+      </td>
+      <td colspan="1" rowspan="1" style="vertical-align: top; white-space: nowrap;">int, long<br>
+      <br>
+
+      </td>
+      <td colspan="1" rowspan="1" style="vertical-align: top; white-space: nowrap;">long<br>
+      <br>
+
+      </td>
+    </tr>
+    <tr>
+      
+      
+      
+    </tr>
+    <tr>
+      <td style="vertical-align: top; width: 40%; white-space: nowrap;">float, double, long double<br>
+      </td>
+      <td style="vertical-align: top; width: 150px; white-space: nowrap;">int, long, float<br>
+      </td>
+      <td style="vertical-align: top; width: 150px; white-space: nowrap;">float<br>
+      </td>
+    </tr>
+    <tr>
+      <td style="vertical-align: top; width: 40%; white-space: nowrap;">char *<br>
+      </td>
+      <td style="vertical-align: top; width: 150px; white-space: nowrap;">str<span style="font-style: italic;"></span><br>
+      </td>
+      <td style="vertical-align: top; width: 150px; white-space: nowrap;">str<br>
+      </td>
+    </tr>
+  </tbody>
+</table>
+
+<br>
+
+<h4><a name="PyToCStringCaveats"></a>Caveats when using a Python string in a C context</h4>
+
+You need to be careful when using a Python string in a context expecting a <span style="font-family: monospace;">char *</span>.
+In this situation, a pointer to the contents of the Python string is
+used, which is only valid as long as the Python string exists. So you
+need to make sure that a reference to the original Python string is
+held for as long as the C string is needed. If you can't guarantee that
+the Python string will live long enough, you will need to copy the C
+string.<br>
+
+<br>
+
+Pyrex detects and prevents <span style="font-style: italic;">some</span> mistakes of this kind. For instance, if you attempt something like<br>
+
+<pre style="margin-left: 40px;">cdef char *s<br>s = pystring1 + pystring2</pre>
+
+then Pyrex will produce the error message "<span style="font-weight: bold;">Obtaining char * from temporary Python value</span>".
+The reason is that concatenating the two Python strings produces a new
+Python string object that is referenced only by a temporary internal
+variable that Pyrex generates. As soon as the statement has finished,
+the temporary variable will be decrefed and the Python string
+deallocated, leaving <span style="font-family: monospace;">s</span> dangling. Since this code could not possibly work, Pyrex refuses to compile it.<br>
+
+<br>
+
+The solution is to assign the result of the concatenation to a Python variable, and then obtain the char * from that, i.e.<br>
+
+<pre style="margin-left: 40px;">cdef char *s<br>p = pystring1 + pystring2<br>s = p<br></pre>
+
+It is then your responsibility to hold the reference <span style="font-family: monospace;">p</span> for as long as necessary.<br>
+
+<br>
+
+Keep in mind that the rules used to detect such errors are only
+heuristics. Sometimes Pyrex will complain unnecessarily, and sometimes
+it will fail to detect a problem that exists. Ultimately, you need to
+understand the issue and be careful what you do.<br>
+
+<blockquote>
+   </blockquote>
+
+
+     
+<h3> <a name="ScopeRules"></a>Scope rules</h3>
+
+   Pyrex determines whether a variable belongs to a local scope, the module 
+ scope, or the built-in scope <i>completely statically.</i> As with Python, 
+ assigning to a variable which is not otherwise declared implicitly declares 
+ it to be a Python variable residing in the scope where it is assigned. Unlike
+ Python, however, a name which is referred to but not declared or assigned
+ is assumed to reside in the <i>builtin scope, </i>not the module scope.
+Names  added to the module dictionary at run time will not shadow such names.
+  
+<p>You can use a <b>global</b> statement at the module level to explicitly 
+ declare a name to be a module-level name when there would otherwise not be
+any indication of this, for example, </p>
+
+   
+<blockquote><tt>global __name__</tt> <br>
+    <tt>print __name__</tt></blockquote>
+
+   Without the <b>global</b> statement, the above would print the name of 
+the builtins module.<br>
+
+  <br>
+
+  Note: A consequence of these rules is that the module-level scope behaves
+ the same way as a Python local scope if you refer to a variable before assigning
+ to it. In particular, tricks such as the following will <i>not</i> work
+in  Pyrex:<br>
+
+   
+<blockquote>        <pre>try:<br>&nbsp; x = True<br>except NameError:<br>&nbsp; True = 1<br></pre>
+  </blockquote>
+
+  because, due to the assignment, the True will always be looked up in the
+ module-level scope. You would have to do something like this instead:<br>
+
+   
+<blockquote>        <pre>import __builtin__<br>try:<br>  True = __builtin__.True<br>except AttributeError:<br>  True = 1<br></pre>
+  </blockquote>
+
+   
+<hr width="100%">  
+<h3> <a name="StatsAndExprs"></a>Statements and expressions</h3>
+
+   Control structures and expressions follow Python syntax for the most part. 
+ When applied to Python objects, they have the same semantics as in Python 
+ (unless otherwise noted). Most of the Python operators can also be applied 
+ to C values, with the obvious semantics.   
+<p>If Python objects and C values are mixed in an expression, conversions 
+ are performed automatically between Python objects and C numeric or string 
+ types. </p>
+
+   
+<p>Reference counts are maintained automatically for all Python objects, and
+all Python operations are automatically checked for errors, with appropriate 
+ action taken. </p>
+
+   
+<h4> <a name="ExprSyntaxDifferences"></a>Differences between C and Pyrex
+expressions</h4>
+There
+are some differences in syntax and semantics between C expressions and
+Pyrex expressions, particularly in the area of C constructs which have
+no direct equivalent in Python.<br>
+
+<ul>
+<li>An integer literal without an <span style="font-family: monospace; font-weight: bold;">L</span> suffix is treated as a C constant, and will be truncated to whatever size your C compiler thinks appropriate. With an <span style="font-family: monospace; font-weight: bold;">L</span> suffix, it will be converted to Python long integer (even if it would be small enough to fit into a C int).<br>
+    <br>
+  </li>
+<li> There is no <b><tt>-&gt;</tt></b> operator in Pyrex. Instead of <tt>p-&gt;x</tt>, 
+ use <tt>p.x</tt></li>
+    
+  &nbsp; <li> There is no <b><tt>*</tt></b> operator in Pyrex. Instead of 
+    <tt>*p</tt>,  use <tt>p[0]</tt></li>
+    
+  &nbsp; <li> There is an <b><tt>&amp;</tt></b> operator, with the same semantics
+ as in C.</li>
+    
+  &nbsp; <li> The null C pointer is called <b><tt>NULL</tt></b>, not 0 (and
+     <tt>NULL</tt> is a reserved word).</li>
+    
+  &nbsp; <li> Character literals are written with a <b>c</b> prefix, for
+example:</li>
+           <ul>
+                <pre>c'X'</pre>
+          </ul>
+    <li> Type casts are written <b><tt>&lt;type&gt;value</tt></b> , for example:</li>
+           <ul>
+                <pre>cdef char *p, float *q<br>p = &lt;char*&gt;q</pre>
+          </ul>
+   <i><b>Warning</b>: Don't attempt to use a typecast to convert between
+Python  and C data types -- it won't do the right thing. Leave Pyrex to perform 
+the conversion automatically.</i>  
+</ul>
+
+     
+<h4> <a name="ForFromLoop"></a>Integer for-loops</h4>
+
+   You should be aware that a for-loop such as   
+<blockquote><tt>for i in range(n):</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+
+   won't be very fast, even if <tt>i</tt> and <tt>n</tt> are declared as
+C integers, because <tt>range</tt> is a Python function. For iterating over 
+ranges of integers, Pyrex has another form of for-loop:   
+<blockquote><tt>for i from 0 &lt;= i &lt; n:</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+
+   If the loop variable and the lower and upper bounds are all C integers, 
+this form of loop will be much faster, because Pyrex will translate it into 
+pure C code.   
+<p>Some things to note about the <tt>for-from</tt> loop: </p>
+
+   
+<ul>
+
+   <li> The target expression must be a variable name.</li>
+    <li> The name between the lower and upper bounds must be the same as
+the  target name.</li>
+    <li> The direction of iteration is determined by the relations. If they
+ are both from the set {<tt>&lt;</tt>, <tt>&lt;=</tt>} then it is upwards;
+ if they are both from the set {<tt>&gt;</tt>, <tt>&gt;=</tt>} then it is
+downwards. (Any other combination is disallowed.)</li>
+    
+</ul>
+
+   Like other Python looping statements, <tt>break</tt> and <tt>continue</tt> may be used in the body, and the loop may have an <tt>else</tt> clause. 
+ 
+<h2>   <hr width="100%"></h2>
+
+     
+<h3> <a name="ExceptionValues"></a>Error return values</h3>
+
+   If you don't do anything special, a function declared with <b>cdef</b> that does not return a Python object has no way of reporting Python exceptions 
+ to its caller. If an exception is detected in such a function, a warning 
+message is printed and the exception is ignored.   
+<p>If you want a C function that does not return a Python object to be able
+ to propagate exceptions to its caller, you need to declare an <b>exception 
+ value</b> for it. Here is an example: </p>
+
+   
+<blockquote><tt>cdef int spam() except -1:</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+
+   With this declaration, whenever an exception occurs inside <tt>spam</tt>, 
+ it will immediately return with the value <tt>-1</tt>. Furthermore, whenever 
+ a call to <tt>spam</tt> returns <tt>-1</tt>, an exception will be assumed 
+ to have occurred and will be propagated.   
+<p>When you declare an exception value for a function, you should never explicitly
+ return that value. If all possible return values are legal and you can't
+reserve one entirely for signalling errors, you can use an alternative form
+of exception value declaration: </p>
+
+   
+<blockquote><tt>cdef int spam() except? -1:</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+
+   The "?" indicates that the value <tt>-1</tt> only indicates a <i>possible</i> error. In this case, Pyrex generates a call to <tt>PyErr_Occurred</tt>if the
+exception value is returned, to make sure it really is an error.   
+<p>There is also a third form of exception value declaration: </p>
+
+   
+<blockquote><tt>cdef int spam() except *:</tt> <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+
+   This form causes Pyrex to generate a call to <tt>PyErr_Occurred</tt> after 
+ <i>every</i> call to spam, regardless of what value it returns. If you have
+ a function returning <tt>void</tt> that needs to propagate errors, you will
+ have to use this form, since there isn't any return value to test.   
+<p>Some things to note: </p>
+
+   
+<ul>
+
+   <li> Currently, exception values can only declared for functions returning
+ an integer, float or pointer type, and the value must be a <i>literal</i>, 
+ not an expression (although it can be negative). The only possible pointer 
+ exception value is <tt>NULL</tt>. Void functions can only use the <tt>except 
+ *</tt> form.</li>
+    <br>
+  &nbsp; <li> The exception value specification is part of the signature
+of  the function. If you're passing a pointer to a function as a parameter
+or  assigning it to a variable, the declared type of the parameter or variable
+ must have the same exception value specification (or lack thereof). Here
+is an example of a pointer-to-function declaration with an exception value:</li>
+           <ul>
+                <pre><tt>int (*grail)(int, char *) except -1</tt></pre>
+          </ul>
+    <li> You don't need to (and shouldn't) declare exception values for functions 
+ which return Python objects. Remember that a function with no declared return
+ type implicitly returns a Python object.</li>
+    
+</ul>
+
+     
+<h4> <a name="CheckingReturnValues"></a>Checking return values of non-Pyrex 
+ functions</h4>
+
+   It's important to understand that the <tt>except</tt> clause does <i>not</i> cause an error to be <i>raised</i> when the specified value is returned. For
+example, you can't write something like   
+<blockquote>        <pre>cdef extern FILE *fopen(char *filename, char *mode) except NULL <font color="#ed181e"># WRONG!</font></pre>
+   </blockquote>
+
+   and expect an exception to be automatically raised if a call to fopen
+returns  NULL. The except clause doesn't work that way; its only purpose
+is for <i>propagating</i> exceptions that have already been raised, either
+by a Pyrex function or a  C function that calls Python/C API routines. To
+get an exception from a non-Python-aware  function such as fopen, you will
+have to check the return value and raise  it yourself, for example,   
+<blockquote>        <pre>cdef FILE *p<br>p = fopen("spam.txt", "r")<br>if p == NULL:<br>&nbsp;&nbsp;&nbsp; raise SpamError("Couldn't open the spam file")</pre>
+   </blockquote>
+
+     
+<h4>   <hr width="100%"></h4>
+
+     
+<h4> <a name="IncludeStatement"></a>The <tt>include</tt> statement</h4>
+
+   For convenience, a large Pyrex module can be split up into a number of 
+files which are put together using the <b>include</b> statement, for example 
+  
+<blockquote>        <pre>include "spamstuff.pxi"</pre>
+   </blockquote>
+
+   The contents of the named file are textually included at that point. The 
+ included file can contain any complete top-level Pyrex statements, including 
+ other <b>include</b> statements. The <b>include</b> statement itself can 
+only appear at the top level of a file.   
+<p>The <b>include</b> statement can also be used in conjunction with <a href="#PublicDecls"><b>public</b> declarations</a> to make C functions and
+ variables defined in one Pyrex module accessible to another. However, note
+ that some of these uses have been superseded by the facilities described
+in <a href="sharing.html">Sharing Declarations Between Pyrex Modules</a>,
+and it is expected that use of the <b>include</b> statement for this purpose
+will be phased out altogether in future versions. </p>
+
+   
+<h2>   <hr width="100%"><a name="InterfacingWithExternal"></a>Interfacing with External
+ C Code   
+<hr width="100%"></h2>
+
+   One of the main uses of Pyrex is wrapping existing libraries of C code. 
+This is achieved by using <a href="#ExternDecls">external declarations</a> to declare the C functions and variables from the library that you want to
+ use.   
+<p>You can also use <a href="#PublicDecls">public declarations</a> to make 
+ C functions and variables defined in a Pyrex module available to external 
+ C code. The need for this is expected to be less frequent, but you might 
+want to do it, for example, if you are embedding Python in another application 
+ as a scripting language. Just as a Pyrex module can be used as a bridge to
+allow Python code to call C code, it can also be used to allow C code to
+call Python code. </p>
+
+   
+<h3> <a name="ExternDecls"></a>External declarations</h3>
+
+   By default, C functions and variables declared at the module level are 
+local to the module (i.e. they have the C <b>static</b> storage class). They 
+can also be declared <b>extern</b> to specify that they are defined elsewhere,
+ for example:   
+<blockquote>        <pre>cdef extern int spam_counter</pre>
+           <pre>cdef extern void order_spam(int tons)</pre>
+   </blockquote>
+
+     
+<blockquote>         </blockquote>
+
+     
+<h4> <a name="ReferencingHeaders"></a>Referencing C header files</h4>
+
+   When you use an extern definition on its own as in the examples above, 
+Pyrex includes a declaration for it in the generated C file. This can cause 
+problems if the declaration doesn't exactly match the declaration that will 
+be seen by other C code. If you're wrapping an existing C library, for example, 
+it's important that the generated C code is compiled with exactly the same 
+declarations as the rest of the library.   
+<p>To achieve this, you can tell Pyrex that the declarations are to be found
+ in a C header file, like this: </p>
+
+   
+<blockquote>        <pre>cdef extern from "spam.h":</pre>
+           <pre>&nbsp;&nbsp;&nbsp; int spam_counter</pre>
+           <pre>&nbsp;&nbsp;&nbsp; void order_spam(int tons)</pre>
+   </blockquote>
+
+   The <b>cdef extern from</b> clause does three things:   
+<ol>
+
+   <li> It directs Pyrex to place a <b>#include</b> statement for the named
+ header file in the generated C code.<br>
+   </li>
+  &nbsp; <li> It prevents Pyrex from generating any C code for the declarations
+ found in the associated block.<br>
+   </li>
+  &nbsp; <li> It treats all declarations within the block as though they
+started  with <b>cdef extern</b>.</li>
+    
+</ol>
+
+   It's important to understand that Pyrex does <i>not</i> itself read the 
+C header file, so you still need to provide Pyrex versions of any declarations 
+ from it that you use. However, the Pyrex declarations don't always have to
+exactly match the C ones, and in some cases they shouldn't or can't. In particular:
+  
+<ol>
+
+   <li> Don't use <b>const</b>. Pyrex doesn't know anything about const,
+so  just leave it out. Most of the time this shouldn't cause any problem,
+although  on rare occasions you might have to use a cast.<sup><a href="#Footnote1"> 1</a></sup><br>
+   </li>
+  &nbsp; <li> Leave out any platform-specific extensions to C declarations
+ such as <b>__declspec()</b>.<br>
+   </li>
+  &nbsp; <li> If the header file declares a big struct and you only want
+to  use a few members, you only need to declare the members you're interested 
+in. Leaving the rest out doesn't do any harm, because the C compiler will 
+use the full definition from the header file.<br>
+     <br>
+ In some cases, you might not need <i>any</i> of the struct's members, in
+which case you can just put <tt>pass</tt> in the body of the struct declaration,
+e.g.<br>
+     <br>
+     <tt>&nbsp; &nbsp; cdef extern from "foo.h":<br>
+ &nbsp; &nbsp; &nbsp; &nbsp; struct spam:<br>
+ &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; pass</tt><br>
+    <br>
+Note that you can only do this inside a <b>cdef extern from</b> block; struct
+declarations anywhere else must be non-empty.<br>
+    <br>
+   </li>
+    <li> If the header file uses typedef names such as <b>size_t </b>to refer 
+to platform-dependent flavours of numeric types, you will need a corresponding 
+     <b>ctypedef</b> statement, but you don't need to match the type exactly,
+ just use something of the right general kind (int, float, etc). For example,</li>
+           <ol>
+                <pre>ctypedef int size_t</pre>
+          </ol>
+   will work okay whatever the actual size of a size_t is (provided the header 
+ file defines it correctly). <br>
+  &nbsp; <li> If the header file uses macros to define constants, translate
+ them into a dummy <b>enum</b> declaration.<br>
+   </li>
+  &nbsp; <li> If the header file defines a function using a macro, declare
+ it as though it were an ordinary function, with appropriate argument and
+result types.</li>
+    
+</ol>
+
+   A few more tricks and tips:   
+<ul>
+
+   <li> If you want to include a C header because it's needed by another
+header,  but don't want to use any declarations from it, put <tt><font size="+1">pass</font></tt> in the extern-from block:</li>
+    
+</ul>
+
+     
+<ul>
+
+          <ul>
+      <tt>cdef extern from "spam.h":</tt> <br>
+      <tt>&nbsp;&nbsp;&nbsp; pass</tt>        </ul>
+    
+</ul>
+
+     
+<ul>
+
+   <li> If you want to include some external declarations, but don't want 
+to specify a header file (because it's included by some other header that 
+you've already included) you can put <tt>*</tt> in place of the header file 
+name:</li>
+    
+</ul>
+
+     
+<blockquote>        <blockquote><tt>cdef extern from *:</tt> <br>
+      <tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
+   </blockquote>
+
+     
+<h4> <a name="StructDeclStyles"></a>Styles of struct, union and enum declaration</h4>
+
+   There are two main ways that structs, unions and enums can be declared 
+in C header files: using a tag name, or using a typedef. There are also some
+ variations based on various combinations of these.   
+<p>It's important to make the Pyrex declarations match the style used in the
+header file, so that Pyrex can emit the right sort of references to the type
+in the code it generates. To make this possible, Pyrex provides two different
+syntaxes for declaring a struct, union or enum type. The style introduced
+above corresponds to the use of a tag name. To get the other style, you prefix
+the declaration with <b>ctypedef</b>, as illustrated below. </p>
+
+   
+<p>The following table shows the various possible styles that can be found 
+ in a header file, and the corresponding Pyrex declaration that you should 
+ put in the <b>cdef exern from </b>block. Struct declarations are used as 
+an example; the same applies equally to union and enum declarations. </p>
+
+   
+<p>Note that in all the cases below, you refer to the type in Pyrex code simply
+as <tt><font size="+1">Foo</font></tt>, not <tt><font size="+1">struct Foo</font></tt>.
+ <br>
+  &nbsp;  <table cellpadding="5">
+   <tbody>
+      <tr bgcolor="#8cbc1c" valign="top">
+   <td bgcolor="#8cbc1c">&nbsp;</td>
+    <td bgcolor="#ff9933" nowrap="nowrap"><b>C code</b></td>
+    <td bgcolor="#66cccc" valign="top"><b>Possibilities for corresponding 
+Pyrex  code</b></td>
+    <td bgcolor="#99cc33" valign="top"><b>Comments</b></td>
+   </tr>
+    <tr bgcolor="#8cbc1c" valign="top">
+   <td>1</td>
+    <td bgcolor="#ff9900"><tt>struct Foo {</tt> <br>
+        <tt>&nbsp; ...</tt> <br>
+        <tt>};</tt></td>
+    <td bgcolor="#66cccc"><tt>cdef struct Foo:</tt> <br>
+        <tt>&nbsp; ...</tt></td>
+    <td>Pyrex will refer to the type as <tt>struct Foo </tt>in the generated 
+ C code<tt>.</tt></td>
+   </tr>
+    <tr bgcolor="#8cbc1c" valign="top">
+   <td valign="top">2</td>
+    <td bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct {</tt> <br>
+        <tt>&nbsp; ...</tt> <br>
+        <tt>} Foo;</tt></td>
+    <td bgcolor="#66cccc" valign="top"><tt>ctypedef struct Foo:</tt> <br>
+        <tt>&nbsp; ...</tt></td>
+    <td valign="top">Pyrex will refer to the type simply as <tt>Foo</tt>
+in  the generated C code.</td>
+   </tr>
+    <tr bgcolor="#8cbc1c" valign="top">
+   <td rowspan="2">3</td>
+    <td rowspan="2" bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct
+foo  {</tt> <br>
+        <tt>&nbsp; ...</tt> <br>
+        <tt>} Foo;</tt></td>
+    <td bgcolor="#66cccc" nowrap="nowrap" valign="top"><tt>cdef struct foo:</tt>       <br>
+        <tt>&nbsp; ...</tt> <br>
+        <tt>ctypedef foo Foo #optional</tt></td>
+    <td rowspan="2" valign="top">If the C header uses both a tag and a typedef 
+ with <i>different</i> names, you can use either form of declaration in Pyrex
+ (although if you need to forward reference the type, you'll have to use
+the  first form).</td>
+   </tr>
+    <tr>
+   <td bgcolor="#66cccc"><tt>ctypedef struct Foo:</tt> <br>
+        <tt>&nbsp; ...</tt></td>
+   </tr>
+    <tr bgcolor="#8cbc1c" valign="top">
+   <td>4</td>
+    <td bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct Foo {</tt> <br>
+        <tt>&nbsp; ...</tt> <br>
+        <tt>} Foo;</tt></td>
+    <td bgcolor="#66cccc" valign="top"><tt>cdef struct Foo:</tt> <br>
+        <tt>&nbsp; ...</tt></td>
+    <td>If the header uses the <i>same</i> name for the tag and the typedef, 
+ you won't be able to include a <b>ctypedef</b> for it -- but then, it's not
+necessary.</td>
+   </tr>
+          </tbody>  </table>
+    </p>
+
+   
+<h4> <a name="AccessingAPI"></a>Accessing Python/C API routines</h4>
+
+   One particular use of the <b>cdef extern from</b> statement is for gaining 
+ access to routines in the Python/C API. For example,   
+<blockquote>        <pre>cdef extern from "Python.h":</pre>
+           <pre>&nbsp;&nbsp;&nbsp; object PyString_FromStringAndSize(char *s, int len)</pre>
+   </blockquote>
+
+   will allow you to create Python strings containing null bytes.   
+<p> </p>
+
+   
+<hr width="100%">  
+<h3> <a name="CNameSpecs"></a>Resolving naming conflicts - C name specifications</h3>
+
+   Each Pyrex module has a single module-level namespace for both Python
+and  C names. This can be inconvenient if you want to wrap some external
+C functions  and provide the Python user with Python functions of the same
+names.   
+<p>Pyrex 0.8 provides a couple of different ways of solving this problem. 
+ The best way, especially if you have many C functions to wrap, is probably 
+ to put the extern C function declarations into a different namespace using 
+ the facilities described in the section on <a href="sharing.html">sharing 
+ declarations between Pyrex modules</a>. </p>
+
+   
+<p>The other way is to use a <b>c name specification</b> to give different 
+ Pyrex and C names to the C function. Suppose, for example, that you want 
+to wrap an external function called <tt>eject_tomato</tt>. If you declare 
+it as </p>
+
+   
+<blockquote>        <pre>cdef extern void c_eject_tomato "eject_tomato" (float speed)</pre>
+   </blockquote>
+
+   then its name inside the Pyrex module will be <tt>c_eject_tomato</tt>, 
+whereas its name in C will be <tt>eject_tomato</tt>. You can then wrap it 
+with   
+<blockquote>        <pre>def eject_tomato(speed):<br>&nbsp; c_eject_tomato(speed)</pre>
+   </blockquote>
+
+   so that users of your module can refer to it as <tt>eject_tomato</tt>. 
+  
+<p>Another use for this feature is referring to external names that happen 
+ to be Pyrex keywords. For example, if you want to call an external function 
+ called <tt>print</tt>, you can rename it to something else in your Pyrex 
+module.  </p>
+
+   
+<p>As well as functions, C names can be specified for variables, structs, 
+ unions, enums, struct and union members, and enum values. For example, </p>
+
+   
+<blockquote>        <pre>cdef extern int one "ein", two "zwei"<br>cdef extern float three "drei"<br><br>cdef struct spam "SPAM":<br>&nbsp; int i "eye"</pre>
+   <tt>cdef enum surprise "inquisition":</tt> <br>
+    <tt>&nbsp; first "alpha"</tt> <br>
+    <tt>&nbsp; second "beta" = 3</tt></blockquote>
+
+     
+<hr width="100%">  
+<h3> <a name="PublicDecls"></a>Public Declarations</h3>
+
+   You can make C variables and functions defined in a Pyrex module accessible 
+ to external C code (or another Pyrex module) using the <b><tt>public</tt></b> keyword, as follows:   
+<blockquote><tt>cdef public int spam # public variable declaration</tt>       <p><tt>cdef public void grail(int num_nuns): # public function declaration</tt>   <br>
+    <tt>&nbsp;&nbsp;&nbsp; ...</tt></p>
+  </blockquote>
+
+   If there are any <tt>public</tt> declarations in a Pyrex module, a <b>.h</b> file is generated containing equivalent C declarations for inclusion in other
+ C code.   
+<p>Pyrex also generates a <b>.pxi</b> file containing Pyrex versions of the
+ declarations for inclusion in another Pyrex module using the <b><a href="#IncludeStatement">include</a></b> statement. If you use this, you
+ will need to arrange for the module using the declarations to be linked
+against  the module defining them, and for both modules to be available to
+the dynamic  linker at run time. I haven't tested this, so I can't say how
+well it will  work on the various platforms. </p>
+
+   
+<blockquote>NOTE: If all you want to export is an extension type, there is
+ now a better way -- see <a href="sharing.html">Sharing Declarations Between
+ Pyrex Modules</a>.</blockquote>
+
+     
+<h2>   <hr width="100%">Extension Types   
+<hr width="100%"></h2>
+
+   One of the most powerful features of Pyrex is the ability to easily create 
+ new built-in Python types, called <b>extension types</b>. This is a major 
+ topic in itself, so there is a&nbsp; <a href="extension_types.html">separate 
+ page</a> devoted to it.   
+<h2>   <hr width="100%">Sharing Declarations Between Pyrex Modules   
+<hr width="100%"></h2>
+
+   Pyrex 0.8 introduces a substantial new set of facilities allowing a Pyrex 
+ module to easily import and use C declarations and extension types from another
+Pyrex module. You can now create a set of co-operating Pyrex modules just
+as easily as you can create a set of co-operating Python modules. There is
+a <a href="sharing.html">separate page</a> devoted to this topic.   
+<h2>   <hr width="100%"><a name="Limitations"></a>Limitations   
+<hr width="100%"></h2>
+
+     
+<h3> <a name="Unsupported"></a>Unsupported Python features</h3>
+
+   Pyrex is not quite a full superset of Python. The following restrictions 
+ apply:   
+<blockquote> <li> Function definitions (whether using <b>def</b> or <b>cdef</b>)
+ cannot be nested within other function definitions.<br>
+   </li>
+  &nbsp; <li> Class definitions can only appear at the top level of a module,
+ not inside a function.<br>
+   </li>
+  &nbsp; <li> The<tt> import *</tt> form of import is not allowed anywhere
+ (other forms of the import statement are fine, though).<br>
+   </li>
+  &nbsp; <li> Generators cannot be defined in Pyrex.<br>
+     <br>
+   </li>
+    <li> The <tt>globals()</tt> and <tt>locals()</tt> functions cannot be 
+used.</li>
+   </blockquote>
+
+   The above restrictions will most likely remain, since removing them would 
+ be difficult and they're not really needed for Pyrex's intended applications. 
+  
+<p>There are also some temporary limitations, which may eventually be lifted, including: 
+ </p>
+
+   
+<blockquote> <li> Class and function definitions cannot be placed inside
+control structures.<br>
+   </li>
+  &nbsp; <li> In-place arithmetic operators (+=, etc) are not yet supported.<br>
+   </li>
+  &nbsp; <li> List comprehensions are not yet supported.<br>
+   </li>
+  &nbsp; <li> There is no support for Unicode.<br>
+   </li>
+  &nbsp; <li> Special methods of extension types cannot have functioning
+docstrings.<br>
+     <br>
+   </li>
+    <li> The use of string literals as comments is not recommended at present,
+ because Pyrex doesn't optimize them away, and won't even accept them in
+places  where executable statements are not allowed.</li>
+   </blockquote>
+   
+<h3> <a name="SemanticDifferences"></a>Semantic differences between Python
+ and Pyrex</h3>
+
+     
+<h4> Behaviour of class scopes</h4>
+
+   In Python, referring to a method of a class inside the class definition, 
+ i.e. while the class is being defined, yields a plain function object, but
+ in Pyrex it yields an unbound method<sup><font size="-2"><a href="#Footnote2">2</a></font></sup>. A consequence of this is that the
+usual idiom for using the classmethod and staticmethod functions, e.g.   
+<blockquote>        <pre>class Spam:</pre>
+           <pre>&nbsp; def method(cls):<br>&nbsp;&nbsp;&nbsp; ...</pre>
+           <pre>&nbsp; method = classmethod(method)</pre>
+   </blockquote>
+
+   will not work in Pyrex. This can be worked around by defining the function 
+ <i>outside</i> the class, and then assigning the result of classmethod or
+ staticmethod inside the class, i.e.   
+<blockquote>        <pre>def Spam_method(cls):<br>&nbsp; ...</pre>
+           <pre>class Spam:</pre>
+           <pre>&nbsp; method = classmethod(Spam_method)</pre>
+   </blockquote>
+
+     
+<h1>   <hr width="100%"><font size="+0">Footnotes</font>  <hr width="100%"></h1>
+
+   <a name="Footnote1"></a>1. A problem with const could arise if you have 
+something like   
+<blockquote>        <pre>cdef extern from "grail.h":<br>&nbsp; char *nun</pre>
+   </blockquote>
+
+   where grail.h actually contains   
+<blockquote>        <pre>extern const char *nun;</pre>
+   </blockquote>
+
+   and you do   
+<blockquote>        <pre>cdef void languissement(char *s):<br>&nbsp; #something that doesn't change s</pre>
+           <pre>...</pre>
+           <pre>languissement(nun)</pre>
+   </blockquote>
+
+   which will cause the C compiler to complain. You can work around it by 
+casting  away the constness:   
+<blockquote>        <pre>languissement(&lt;char *&gt;nun)</pre>
+   </blockquote>
+
+     
+<hr width="100%"><a name="Footnote2"></a>2. The reason for the different behaviour
+of class scopes is that Pyrex-defined Python functions are PyCFunction objects,
+not PyFunction objects, and are not recognised by the machinery that creates
+a bound or unbound method when a function is extracted from a class. To get
+around this, Pyrex wraps each method in an unbound method object itself before
+storing it in the class's dictionary. <br>
+
+  &nbsp; <br>
+
+  <br>
+
+ </body></html>
\ No newline at end of file

Modified: lxml/pyrex/Doc/sharing.html
==============================================================================
--- lxml/pyrex/Doc/sharing.html	(original)
+++ lxml/pyrex/Doc/sharing.html	Tue Apr 25 15:43:21 2006
@@ -1 +1,201 @@
-<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
       
  <meta http-equiv="Content-Type"
 content="text/html; charset=iso-8859-1">
       
  <meta name="GENERATOR"
 content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]">
  <title>Sharing Declarations Between Pyrex Modules</title>
</head>
  <body>
   
<h1>  
<hr width="100%">Sharing Declarations Between Pyrex Modules  
<hr width="100%"></h1>
  This section describes a new set of facilities introduced in Pyrex 0.8
for making C declarations and extension types in one Pyrex module available
for use in another Pyrex module. These facilities are closely modelled on
the Python import mechanism, and can be thought of as a compile-time version 
of it.  
<h2> Contents</h2>
   
<ul>
  <li> <a href="#DefAndImpFiles">Definition and Implementation files</a></li>
     
  <ul>
  <li> <a href="#WhatDefFileContains">What a Definition File contains</a></li>
   <li> <a href="#WhatImpFileContains">What an Implementation File contains</a></li>
    
  </ul>
   <li> <a href="#CImportStatement">The <tt>cimport</tt> statement</a></li>
     
  <ul>
  <li> <a href="#SearchPaths">Search paths for definition files</a></li>
   <li> <a href="#ResolvingNamingConflicts">Using <tt>cimport</tt> to resolve 
naming conflicts</a></li>
    
  </ul>
   <li> <a href="#SharingExtensionTypes">Sharing extension types</a></li>
  
</ul>
   
<h2> <a name="DefAndImpFiles"></a>Definition and Implementation files</h2>
  A Pyrex module can be split into two parts: a <i>definition file</i> with
 a <tt>.pxd</tt> suffix, containing C declarations that are to be available
 to other Pyrex modules, and an <i>implementation file</i> with a <tt>.pyx</tt>
suffix, containing everything else. When a module wants to use something
declared in another module's definition file, it imports it using the <a
 href="#CImportStatement"><b>cimport</b> statement</a>.  
<h3> <a name="WhatDefFileContains"></a>What a Definition File contains</h3>
  A definition file can contain:  
<ul>
  <li> Any kind of C type declaration.</li>
   <li> <b>extern</b> C function or variable declarations.</li>
   <li> The definition part of an extension type (<a
 href="#SharingExtensionTypes">see below</a>).</li>
  
</ul>
  It cannot currently contain any non-extern C function or variable declarations
 (although this may be possible in a future version).  
<p>It cannot contain the implementations of any C or Python functions, or 
any Python class definitions, or any executable statements. </p>
 
<blockquote>NOTE: You don't need to (and shouldn't) declare anything in a 
declaration file <b>public</b> in order to make it available to other Pyrex 
modules; its mere presence in a definition file does that. You only need a
public declaration if you want to make something available to external C code.</blockquote>
   
<h3> <a name="WhatImpFileContains"></a>What an Implementation File contains</h3>
  An implementation file can contain any kind of Pyrex statement, although
 there are some restrictions on the implementation part of an extension type 
if the corresponding definition file also defines that type (see below).
 
<h2> <a name="CImportStatement"></a>The <tt>cimport</tt> statement</h2>
  The <b>cimport</b> statement is used in a definition or implementation
file to gain access to names declared in another definition file. Its syntax
exactly parallels that of the normal Python import statement:  
<blockquote><tt>cimport </tt><i>module</i><tt> [, </tt><i>module</i><tt>...]</tt></blockquote>
   
<blockquote><tt>from </tt><i>module</i><tt> cimport </tt><i>name</i><tt>
[as   </tt><i>name</i><tt>] [, </tt><i>name</i><tt> [as </tt><i>name</i><tt>]
 ...]</tt></blockquote>
  Here is an example. The file on the left is a definition file which exports
 a C data type. The file on the right is an implementation file which imports
 and uses it. <br>
 &nbsp; 
<table cellpadding="5" cols="2" width="100%">
  <tbody>
     <tr>
  <td bgcolor="#ffcc00" width="40%"><b><tt>dishes.pxd</tt></b></td>
   <td bgcolor="#5dbaca"><b><tt>restaurant.pyx</tt></b></td>
  </tr>
   <tr align="left" valign="top">
  <td bgcolor="#ffcc18" width="40%"><tt>cdef enum otherstuff:</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; sausage, eggs, lettuce</tt>       
      <p><tt>cdef struct spamdish:</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; int oz_of_spam</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; otherstuff filler</tt></p>
       </td>
   <td bgcolor="#5dbaca"><tt>cimport dishes</tt> <br>
       <tt>from dishes cimport spamdish</tt>       
      <p><tt>cdef void prepare(spamdish *d):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; d.oz_of_spam = 42</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; d.filler = dishes.sausage</tt> </p>
       
      <p><tt>def serve():</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; spamdish d</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; prepare(&amp;d)</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; print "%d oz spam, filler no. %d" % \</tt>
      <br>
       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (d-&gt;oz_of_spam,
 d-&gt;otherstuff)</tt></p>
       </td>
  </tr>
    
  </tbody> 
</table>
   
<p>It is important to understand that the <b>cimport</b> statement can <i>only</i>
be used to import C data types, external C functions and variables, and extension 
types. It cannot be used to import any Python objects, and (with one exception) 
it doesn't imply any Python import at run time. If you want to refer to any 
Python names from a module that you have cimported, you will have to include 
a regular <b>import</b> statement for it as well. </p>
 
<p>The exception is that when you use <b>cimport</b> to import an extension
 type, its type object is imported at run time and made available by the
name under which you imported it. Using <b>cimport</b> to import extension
types is covered in more detail <a href="#SharingExtensionTypes">below</a>.
</p>
 
<h3> <a name="SearchPaths"></a>Search paths for definition files</h3>
  When you <b>cimport</b> a module called <tt>modulename</tt>, the Pyrex
compiler searches for a file called <tt>modulename.pxd</tt> along the search
path for include files, as specified by <b>-I</b> command line options.  
<p>Also, whenever you compile a file <tt>modulename.pyx</tt>, the corresponding
 definition file <tt>modulename.pxd</tt> is first searched for along the
same path, and if found, it is processed before processing the <tt>.pyx</tt>
file.  </p>
 
<h3> <a name="ResolvingNamingConflicts"></a>Using cimport to resolve naming
 conflicts</h3>
  The cimport mechanism provides a clean and simple way to solve the problem
 of wrapping external C functions with Python functions of the same name.
All you need to do is put the extern C declarations into a .pxd file for
an imaginary module, and cimport that module. You can then refer to the C
functions by qualifying them with the name of the module. Here's an example:
<br>
 &nbsp; 
<table cellpadding="5" cols="2" width="100%">
  <tbody>
     <tr>
  <td bgcolor="#ffcc00" width="50%"><b><tt>c_lunch.pxd</tt></b></td>
   <td bgcolor="#5dbaca"><b><tt>lunch.pyx</tt></b></td>
  </tr>
   <tr align="left" valign="top">
  <td bgcolor="#ffcc18" width="50%"><tt>cdef extern from "lunch.h":</tt>
      <br>
       <tt>&nbsp;&nbsp;&nbsp; void eject_tomato(float)</tt></td>
   <td bgcolor="#5dbaca"><tt>cimport c_lunch</tt>       
      <p><tt>def eject_tomato(float speed):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; c_lunch.eject_tomato(speed)</tt></p>
       </td>
  </tr>
    
  </tbody> 
</table>
   
<p>You don't need any <tt>c_lunch.pyx</tt> file, because the only things
defined in <tt>c_lunch.pxd</tt> are extern C entities. There won't be any
actual <tt>c_lunch</tt> module at run time, but that doesn't matter -- <tt>c_lunch</tt>
has done its job of providing an additional namespace at compile time. </p>
 
<h2> <a name="SharingExtensionTypes"></a>Sharing Extension Types</h2>
  An extension type declaration can also be split into two parts, one in
a definition file and the other in the corresponding implementation file.
<br>
 <br>
 The definition part of the extension type can only declare C attributes
and C methods, not Python methods, and it must declare <i>all</i> of that
type's C attributes and C methods.<br>
 <br>
 The implementation part must implement all of the C methods declared in
the definition part, and may not add any further C attributes. It may also
define Python methods.  
<p>Here is an example of a module which defines and exports an extension
type, and another module which uses it. <br>
 &nbsp; 
<table cellpadding="5" cols="2" width="100%">
  <tbody>
     <tr>
  <td bgcolor="#ffcc18" width="30%"><b><tt>Shrubbing.pxd</tt></b></td>
   <td bgcolor="#5dbaca" width="50%"><b><tt>Shrubbing.pyx</tt></b></td>
  </tr>
   <tr align="left" valign="top">
  <td bgcolor="#ffcc18" width="30%"><tt>cdef class Shrubbery:</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; cdef int width</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; cdef int length</tt></td>
   <td bgcolor="#5dbaca" width="50%"><tt>cdef class Shrubbery:</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; def __new__(self, int w, int l):</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.width = w</tt>
      <br>
       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.length = l</tt>
      
      <p><tt>def standard_shrubbery():</tt> <br>
       <tt>&nbsp;&nbsp;&nbsp; return Shrubbery(3, 7)</tt></p>
       </td>
  </tr>
   <tr>
  <td colspan="2" bgcolor="#8cbc1c" width="30%"><b><tt>Landscaping.pyx</tt></b></td>
  </tr>
   <tr>
  <td colspan="2" bgcolor="#99cc00" width="30%"><tt>cimport Shrubbing</tt>
      <br>
       <tt>import Shrubbing</tt>       
      <p><tt>cdef Shrubbing.Shrubbery sh</tt> <br>
       <tt>sh = Shrubbing.standard_shrubbery()</tt> <br>
       <tt>print "Shrubbery size is %d x %d" % (sh.width, sh.height)</tt>
      <br>
 &nbsp;</p>
       </td>
  </tr>
    
  </tbody> 
</table>
   </p>
 
<p>Some things to note about this example: </p>
 
<ul>
  <li> There is a <tt>cdef class shrubbery</tt> declaration in both Shrubbing.pxd
 and Shrubbing.pyx. When the Shrubbing module is compiled, these two declarations
 are combined into one.</li>
   <br>
 &nbsp; <li> In Landscaping.pyx, the <tt>cimport Shrubbery</tt> declaration 
allows us to refer to the Shrubbery type as <tt>Shrubbing.Shrubbery</tt>. 
But it doesn't bind the name <tt>Shrubbery</tt> in Landscaping's module namespace
 at run time, so to access <tt>Shrubbery.standard_shrubbery</tt> we also
need to <tt>import Shrubbing</tt>.</li>
  
</ul>
   
<hr width="100%">Back to the <a href="overview.html">Language Overview</a>
<br>
 <br>
</body>
</html>
\ No newline at end of file
+<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
+<html><head>
+         <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
+         <meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Sharing Declarations Between Pyrex Modules</title></head>
+
+<body>
+   <h1>  <hr width="100%">Sharing Declarations Between Pyrex Modules  
+<hr width="100%"></h1>
+  This section describes a new set of facilities introduced in Pyrex 0.8
+for making C declarations and extension types in one Pyrex module available
+for use in another Pyrex module. These facilities are closely modelled on
+the Python import mechanism, and can be thought of as a compile-time version 
+of it.  
+<h2> Contents</h2>
+   <ul>
+  <li> <a href="#DefAndImpFiles">Definition and Implementation files</a></li>
+       <ul>
+  <li> <a href="#WhatDefFileContains">What a Definition File contains</a></li>
+   <li> <a href="#WhatImpFileContains">What an Implementation File contains</a></li>
+      </ul>
+   <li> <a href="#CImportStatement">The <tt>cimport</tt> statement</a></li>
+       <ul>
+  <li> <a href="#SearchPaths">Search paths for definition files</a></li>
+   <li> <a href="#ResolvingNamingConflicts">Using <tt>cimport</tt> to resolve 
+naming conflicts</a></li>
+      </ul>
+   <li> <a href="#SharingExtensionTypes">Sharing extension types</a></li>
+  </ul>
+   <h2> <a name="DefAndImpFiles"></a>Definition and Implementation files</h2>
+  A Pyrex module can be split into two parts: a <i>definition file</i> with
+ a <tt>.pxd</tt> suffix, containing C declarations that are to be available
+ to other Pyrex modules, and an <i>implementation file</i> with a <tt>.pyx</tt>
+suffix, containing everything else. When a module wants to use something
+declared in another module's definition file, it imports it using the <a href="#CImportStatement"><b>cimport</b> statement</a>.  
+<h3> <a name="WhatDefFileContains"></a>What a Definition File contains</h3>
+  A definition file can contain:  
+<ul>
+  <li> Any kind of C type declaration.</li>
+   <li> <b>extern</b> C function or variable declarations.</li>
+   <li> The definition part of an extension type (<a href="#SharingExtensionTypes">see below</a>).</li>
+  </ul>
+  It cannot currently contain any non-extern C function or variable declarations
+ (although this may be possible in a future version).  
+<p>It cannot contain the implementations of any C or Python functions, or 
+any Python class definitions, or any executable statements. </p>
+ <blockquote>NOTE: You don't need to (and shouldn't) declare anything in a 
+declaration file <b>public</b> in order to make it available to other Pyrex 
+modules; its mere presence in a definition file does that. You only need a
+public declaration if you want to make something available to external C code.</blockquote>
+   <h3> <a name="WhatImpFileContains"></a>What an Implementation File contains</h3>
+  An implementation file can contain any kind of Pyrex statement, although
+ there are some restrictions on the implementation part of an extension type 
+if the corresponding definition file also defines that type (see below).
+ 
+<h2> <a name="CImportStatement"></a>The <tt>cimport</tt> statement</h2>
+  The <b>cimport</b> statement is used in a definition or implementation
+file to gain access to names declared in another definition file. Its syntax
+exactly parallels that of the normal Python import statement:  
+<blockquote><tt>cimport </tt><i>module</i><tt> [, </tt><i>module</i><tt>...]</tt></blockquote>
+   <blockquote><tt>from </tt><i>module</i><tt> cimport </tt><i>name</i><tt>
+[as   </tt><i>name</i><tt>] [, </tt><i>name</i><tt> [as </tt><i>name</i><tt>]
+ ...]</tt></blockquote>
+  Here is an example. The file on the left is a definition file which exports
+ a C data type. The file on the right is an implementation file which imports
+ and uses it. <br>
+ &nbsp; <table cellpadding="5" cols="2" width="100%">
+  <tbody>
+     <tr>
+  <td bgcolor="#ffcc00" width="40%"><b><tt>dishes.pxd</tt></b></td>
+   <td bgcolor="#5dbaca"><b><tt>restaurant.pyx</tt></b></td>
+  </tr>
+   <tr align="left" valign="top">
+  <td bgcolor="#ffcc18" width="40%"><tt>cdef enum otherstuff:</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; sausage, eggs, lettuce</tt>             <p><tt>cdef struct spamdish:</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; int oz_of_spam</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; otherstuff filler</tt></p>
+       </td>
+   <td bgcolor="#5dbaca"><tt>cimport dishes</tt> <br>
+       <tt>from dishes cimport spamdish</tt>             <p><tt>cdef void prepare(spamdish *d):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; d.oz_of_spam = 42</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; d.filler = dishes.sausage</tt> </p>
+             <p><tt>def serve():</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; spamdish d</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; prepare(&amp;d)</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; print "%d oz spam, filler no. %d" % \</tt>
+      <br>
+       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (d-&gt;oz_of_spam,
+ d-&gt;otherstuff)</tt></p>
+       </td>
+  </tr>
+      </tbody> </table>
+   <p>It is important to understand that the <b>cimport</b> statement can <i>only</i>
+be used to import C data types, external C functions and variables, and extension 
+types. It cannot be used to import any Python objects, and (with one exception) 
+it doesn't imply any Python import at run time. If you want to refer to any 
+Python names from a module that you have cimported, you will have to include 
+a regular <b>import</b> statement for it as well. </p>
+ <p>The exception is that when you use <b>cimport</b> to import an extension
+ type, its type object is imported at run time and made available by the
+name under which you imported it. Using <b>cimport</b> to import extension
+types is covered in more detail <a href="#SharingExtensionTypes">below</a>.
+</p>
+ <h3> <a name="SearchPaths"></a>Search paths for definition files</h3>
+  When you <b>cimport</b> a module called <tt>modulename</tt>, the Pyrex
+compiler searches for a file called <tt>modulename.pxd</tt> along the search
+path for include files, as specified by <b>-I</b> command line options.  
+<p>Also, whenever you compile a file <tt>modulename.pyx</tt>, the corresponding
+ definition file <tt>modulename.pxd</tt> is first searched for along the
+same path, and if found, it is processed before processing the <tt>.pyx</tt>
+file.  </p>
+ <h3> <a name="ResolvingNamingConflicts"></a>Using cimport to resolve naming
+ conflicts</h3>
+  The cimport mechanism provides a clean and simple way to solve the problem
+ of wrapping external C functions with Python functions of the same name.
+All you need to do is put the extern C declarations into a .pxd file for
+an imaginary module, and cimport that module. You can then refer to the C
+functions by qualifying them with the name of the module. Here's an example:
+<br>
+ &nbsp; <table cellpadding="5" cols="2" width="100%">
+  <tbody>
+     <tr>
+  <td bgcolor="#ffcc00" width="50%"><b><tt>c_lunch.pxd</tt></b></td>
+   <td bgcolor="#5dbaca"><b><tt>lunch.pyx</tt></b></td>
+  </tr>
+   <tr align="left" valign="top">
+  <td bgcolor="#ffcc18" width="50%"><tt>cdef extern from "lunch.h":</tt>
+      <br>
+       <tt>&nbsp;&nbsp;&nbsp; void eject_tomato(float)</tt></td>
+   <td bgcolor="#5dbaca"><tt>cimport c_lunch</tt>             <p><tt>def eject_tomato(float speed):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; c_lunch.eject_tomato(speed)</tt></p>
+       </td>
+  </tr>
+      </tbody> </table>
+   <p>You don't need any <tt>c_lunch.pyx</tt> file, because the only things
+defined in <tt>c_lunch.pxd</tt> are extern C entities. There won't be any
+actual <tt>c_lunch</tt> module at run time, but that doesn't matter -- <tt>c_lunch</tt>
+has done its job of providing an additional namespace at compile time. </p>
+ <h2> <a name="SharingExtensionTypes"></a>Sharing Extension Types</h2>
+  An extension type declaration can also be split into two parts, one in
+a definition file and the other in the corresponding implementation file.
+<br>
+ <br>
+ The definition part of the extension type can only declare C attributes
+and C methods, not Python methods, and it must declare <i>all</i> of that
+type's C attributes and C methods.<br>
+ <br>
+ The implementation part must implement all of the C methods declared in
+the definition part, and may not add any further C attributes. It may also
+define Python methods.  
+<p>Here is an example of a module which defines and exports an extension
+type, and another module which uses it. <br>
+ &nbsp; <table cellpadding="5" cols="2" width="100%">
+  <tbody>
+     <tr>
+  <td bgcolor="#ffcc18" width="30%"><b><tt>Shrubbing.pxd</tt></b></td>
+   <td bgcolor="#5dbaca" width="50%"><b><tt>Shrubbing.pyx</tt></b></td>
+  </tr>
+   <tr align="left" valign="top">
+  <td bgcolor="#ffcc18" width="30%"><tt>cdef class Shrubbery:</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; cdef int width</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; cdef int length</tt></td>
+   <td bgcolor="#5dbaca" width="50%"><tt>cdef class Shrubbery:</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; def __new__(self, int w, int l):</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.width = w</tt>
+      <br>
+       <tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.length = l</tt>
+            <p><tt>def standard_shrubbery():</tt> <br>
+       <tt>&nbsp;&nbsp;&nbsp; return Shrubbery(3, 7)</tt></p>
+       </td>
+  </tr>
+   <tr>
+  <td colspan="2" bgcolor="#8cbc1c" width="30%"><b><tt>Landscaping.pyx</tt></b></td>
+  </tr>
+   <tr>
+  <td colspan="2" bgcolor="#99cc00" width="30%"><tt>cimport Shrubbing</tt>
+      <br>
+       <tt>import Shrubbing</tt>             <p><tt>cdef Shrubbing.Shrubbery sh</tt> <br>
+       <tt>sh = Shrubbing.standard_shrubbery()</tt> <br>
+       <tt>print "Shrubbery size is %d x %d" % (sh.width, sh.height)</tt>
+      <br>
+ &nbsp;</p>
+       </td>
+  </tr>
+      </tbody> </table>
+   </p>
+ <p>Some things to note about this example: </p>
+ <ul>
+  <li> There is a <tt>cdef class Shrubbery</tt> declaration in both Shrubbing.pxd
+ and Shrubbing.pyx. When the Shrubbing module is compiled, these two declarations
+ are combined into one.</li>
+   
+ &nbsp; <li> In Landscaping.pyx, the <tt>cimport Shrubbing</tt> declaration 
+allows us to refer to the Shrubbery type as <tt>Shrubbing.Shrubbery</tt>. 
+But it doesn't bind the name <tt>Shrubbery</tt> in Landscaping's module namespace
+ at run time, so to access <tt>Shrubbery.standard_shrubbery</tt> we also
+need to <tt>import Shrubbing</tt>.</li>
+  </ul>
+   <hr width="100%">Back to the <a href="overview.html">Language Overview</a>
+<br>
+ <br>
+</body></html>
\ No newline at end of file

Modified: lxml/pyrex/Doc/special_methods.html
==============================================================================
--- lxml/pyrex/Doc/special_methods.html	(original)
+++ lxml/pyrex/Doc/special_methods.html	Tue Apr 25 15:43:21 2006
@@ -1 +1,598 @@
-<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
       
  <meta http-equiv="Content-Type"
 content="text/html; charset=iso-8859-1">
       
  <meta name="GENERATOR"
 content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]">
  <title>Special Methods of Extenstion Types</title>
</head>
  <body>
   
<h1>  
<hr width="100%">Special Methods of Extension Types  
<hr width="100%"></h1>
  This page describes the special methods currently supported by Pyrex extension
 types. A complete list of all the special methods appears in the table at 
the bottom. Some of these methods behave differently from their Python counterparts 
or have no direct Python counterparts, and require special mention.  
<p><i>Note: Everything said on this page applies only to </i>extension<i>
types, defined with the </i>cdef class<i> statement. It doesn't apply&nbsp;
to classes defined with the Python </i>class<i> statement, where the normal
 Python rules apply.</i> </p>
 
<h2> <font size="+1">Docstrings</font></h2>
  Currently, docstrings are not fully supported in special methods of extension
 types. You can place a docstring in the source to serve as a comment, but
 it won't show up in the corresponding __doc__ attribute at run time. (This
 is a Python limitation -- there's nowhere in the PyTypeObject data structure
 to put such docstrings.)  
<h2> <font size="+1">Initialisation methods: <tt>__new__</tt> and <tt>__init__</tt></font></h2>
  There are two methods concerned with initialising the object<tt>.</tt>
<p>The <b><tt>__new__</tt></b> method is where you should perform basic C-level 
initialisation of the object, including allocation of any C data structures 
that your object will own. You need to be careful what you do in the __new__ 
method, because the object may not yet be a valid Python object when it is 
called. Therefore, you must not invoke any Python operations which might touch
the object; in particular, do not try to call any of its methods. </p>
 
<p>Unlike the corresponding method in Python, your <tt>__new__</tt> method
 is <i>not</i> responsible for <i>creating</i> the object. By the time your
 <tt>__new__</tt> method is called, memory has been allocated for the object 
and any C attributes it has have been initialised to 0 or null. (Any Python 
attributes have also been initialised to <tt>None</tt>, but you probably shouldn't
rely on that.) Your <tt>__new__</tt> method is guaranteed to be called exactly
once.<br>
<br>
If your extension type has a base type, the <tt>__new__</tt> method of the
base type is automatically called <i>before</i> your <tt>__new__</tt> method
is called; you cannot explicitly call the inherited <tt>__new__</tt> method.
If you need to pass a modified argument list to the base type, you will have
to do the relevant part of the initialisation in the <tt>__init__</tt> method
instead (where the normal rules for calling inherited methods apply).<br>
 </p>
 
<p>Note that the first parameter of the <tt>__new__</tt> method is the object 
to be initialised, not the class of the object as it is in Python. </p>
 
<p>Any initialisation which cannot safely be done in the <tt>__new__</tt>
method should be done in the <b><tt>__init__</tt></b> method. By the time
 <tt>__init__</tt> is called, the object is a fully valid Python object and 
all operations are safe. Under some circumstances it is possible for <tt>__init__</tt>
to be called more than once or not to be called at all, so your other methods
 should be designed to be robust in such situations. </p>
 
<p>Keep in mind that any arguments passed to the constructor will be passed
 to the <tt>__new__</tt> method as well as the <tt>__init__</tt> method.
If you anticipate subclassing your extension type in Python, you may find
it useful to give the <tt>__new__</tt> method * and ** arguments so that
it can accept and ignore extra arguments. Otherwise, any Python subclass
which has an <tt>__init__</tt> with a different signature will have to override 
<tt>__new__</tt> as well as <tt>__init__</tt>, which the writer of a Python 
class wouldn't expect to have to do. </p>
 
<h2> <font size="+1">Finalization method: <tt>__dealloc__</tt><tt></tt></font></h2>
  The counterpart to the <tt>__new__</tt> method is the <b><tt>__dealloc__</tt></b>
method, which should perform the inverse of the <tt>__new__</tt> method.
Any C data structures that you allocated in your <tt>__new__</tt> method
should be freed in your <tt>__dealloc__</tt> method.  
<p>You need to be careful what you do in a <tt>__dealloc__</tt> method. By 
the time your <tt>__dealloc__</tt> method is called, the object may already 
have been partially destroyed and may not be in a valid state as far as Python 
is concerned, so you should avoid invoking any Python operations which might 
touch the object. In particular, don't call any other methods of the object 
or do anything which might cause the object to be resurrected. It's best if
you stick to just deallocating C data. </p>
 
<p>You don't need to worry about deallocating Python attributes of your object, 
because that will be done for you by Pyrex after your <tt>__dealloc__</tt>
method returns.<br>
 <br>
 <b>Note:</b> There is no <tt>__del__</tt> method for extension types. (Earlier 
versions of the Pyrex documentation stated that there was, but this turned 
out to be incorrect.)<br>
  </p>
 
<h2><font size="+1">Arithmetic methods</font></h2>
  Arithmetic operator methods, such as <tt>__add__</tt>, behave differently
 from their Python counterparts. There are no separate "reversed" versions
 of these methods (<tt>__radd__</tt>, etc.) Instead, if the first operand
cannot perform the operation, the <i>same</i> method of the second operand
is called, with the operands in the <i>same order</i>.  
<p>This means that you can't rely on the first parameter of these methods
 being "self", and you should test the types of both operands before deciding
 what to do. If you can't handle the combination of types you've been given,
 you should return <tt>NotImplemented</tt>. </p>
 
<p>This also applies to the in-place arithmetic method <tt>__ipow__</tt>.
 It doesn't apply to any of the <i>other</i> in-place methods (<tt>__iadd__</tt>,
 etc.) which always take self as the first argument. </p>
 
<h2> <font size="+1">Rich comparisons</font></h2>
  There are no separate methods for the individual rich comparison operations
 (<tt>__eq__</tt>, <tt>__le__</tt>, etc.) Instead there is a single method
 <tt>__richcmp__</tt> which takes an integer indicating which operation is 
to be performed, as follows:  
<ul>
    
  <ul>
 &nbsp;     
    <table border="0" cellspacing="0" cellpadding="5" nosave="">
  <tbody>
         <tr nosave="">
  <td width="30" bgcolor="#ffcc33" nosave="">           
          <div align="right">&lt;</div>
  </td>
   <td width="30" bgcolor="#66ffff" nosave="">0</td>
   <td><br>
           </td>
   <td width="30" bgcolor="#ffcc33" nosave="">           
          <div align="right">==</div>
  </td>
   <td width="30" bgcolor="#66ffff" nosave="">2</td>
   <td><br>
           </td>
   <td width="30" bgcolor="#ffcc33" nosave="">           
          <div align="right">&gt;</div>
  </td>
   <td width="30" bgcolor="#66ffff" nosave="">4</td>
  </tr>
   <tr nosave="">
  <td bgcolor="#ffcc33" nosave="">           
          <div align="right">&lt;=</div>
  </td>
   <td bgcolor="#66ffff" nosave="">1</td>
   <td><br>
           </td>
   <td bgcolor="#ffcc33" nosave="">           
          <div align="right">!=</div>
  </td>
   <td bgcolor="#66ffff" nosave="">3</td>
   <td><br>
           </td>
   <td bgcolor="#ffcc33" nosave="">           
          <div align="right">&gt;=</div>
  </td>
   <td bgcolor="#66ffff" nosave="">5</td>
  </tr>
        
      </tbody>     
    </table>
    
  </ul>
  
</ul>
   
<h2> <font size="+1">The __new__ method</font></h2>
  Extension types wishing to implement the iterator interface should define
 a method called <b><tt>__new__</tt></b>, <i>not</i> <tt>new</tt>. The Python
 system will automatically supply a <tt>new</tt> method which calls your
__new__.  <b>Do NOT explicitly give your type a <tt>new</tt> method</b>,
or bad things could happen (see note 3).  
<h2> <font size="+1">Special Method Table</font></h2>
  This table lists all of the special methods together with their parameter
 and return types. A parameter named <b>self</b> is of the type the method
 belongs to. Other untyped parameters are generic Python objects.  
<p>You don't have to declare your method as taking these parameter types.
 If you declare different types, conversions will be performed as necessary.
 <br>
 &nbsp; 
<table border="1" cellspacing="0" cellpadding="5" bgcolor="#ccffff"
 nosave="">
  <tbody>
     <tr bgcolor="#ffcc33" nosave="">
  <td nosave=""><b>Name</b></td>
   <td><b>Parameters</b></td>
   <td><b>Return type</b></td>
   <td><b>Description</b></td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>General</b></td>
  </tr>
   <tr>
  <td><tt>__new__</tt></td>
   <td>self, ...</td>
   <td>&nbsp;</td>
   <td>Basic initialisation (no direct Python equivalent)</td>
  </tr>
   <tr>
  <td><tt>__init__</tt></td>
   <td>self, ...</td>
   <td>&nbsp;</td>
   <td>Further initialisation</td>
  </tr>
   <tr>
  <td><tt>__dealloc__</tt></td>
   <td>self</td>
   <td>&nbsp;</td>
   <td>Basic deallocation (no direct Python equivalent)</td>
  </tr>
   <tr>
  <td><tt>__cmp__</tt></td>
   <td>x, y</td>
   <td>int</td>
   <td>3-way comparison</td>
  </tr>
   <tr>
  <td><tt>__richcmp__</tt></td>
   <td>x, y, int op</td>
   <td>object</td>
   <td>Rich comparison (no direct Python equivalent)</td>
  </tr>
   <tr>
  <td><tt>__str__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>str(self)</td>
  </tr>
   <tr>
  <td><tt>__repr__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>repr(self)</td>
  </tr>
   <tr nosave="">
  <td nosave=""><tt>__hash__</tt></td>
   <td>self</td>
   <td>int</td>
   <td>Hash function</td>
  </tr>
   <tr>
  <td><tt>__call__</tt></td>
   <td>self, ...</td>
   <td>object</td>
   <td>self(...)</td>
  </tr>
   <tr>
  <td><tt>__iter__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Return iterator for sequence</td>
  </tr>
   <tr>
  <td><tt>__getattr__</tt></td>
   <td>self, name</td>
   <td>object</td>
   <td>Get attribute</td>
  </tr>
   <tr>
  <td><tt>__setattr__</tt></td>
   <td>self, name, val</td>
   <td>&nbsp;</td>
   <td>Set attribute</td>
  </tr>
   <tr>
  <td><tt>__delattr__</tt></td>
   <td>self, name</td>
   <td>&nbsp;</td>
   <td>Delete attribute</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>Arithmetic operators</b></td>
  </tr>
   <tr>
  <td><tt>__add__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>binary + operator</td>
  </tr>
   <tr>
  <td><tt>__sub__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>binary - operator</td>
  </tr>
   <tr>
  <td><tt>__mul__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>* operator</td>
  </tr>
   <tr>
  <td><tt>__div__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>/&nbsp; operator for old-style division</td>
  </tr>
   <tr>
  <td><tt>__floordiv__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>//&nbsp; operator</td>
  </tr>
   <tr>
  <td><tt>__truediv__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>/&nbsp; operator for new-style division</td>
  </tr>
   <tr>
  <td><tt>__mod__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>% operator</td>
  </tr>
   <tr>
  <td><tt>__divmod__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>combined div and mod</td>
  </tr>
   <tr>
  <td><tt>__pow__</tt></td>
   <td>x, y, z</td>
   <td>object</td>
   <td>** operator or pow(x, y, z)</td>
  </tr>
   <tr>
  <td><tt>__neg__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>unary - operator</td>
  </tr>
   <tr>
  <td><tt>__pos__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>unary + operator</td>
  </tr>
   <tr>
  <td><tt>__abs__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>absolute value</td>
  </tr>
   <tr>
  <td><tt>__nonzero__</tt></td>
   <td>self</td>
   <td>int</td>
   <td>convert to boolean</td>
  </tr>
   <tr>
  <td><tt>__invert__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>~ operator</td>
  </tr>
   <tr>
  <td><tt>__lshift__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>&lt;&lt; operator</td>
  </tr>
   <tr>
  <td><tt>__rshift__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>&gt;&gt; operator</td>
  </tr>
   <tr>
  <td><tt>__and__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>&amp; operator</td>
  </tr>
   <tr>
  <td><tt>__or__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>| operator</td>
  </tr>
   <tr>
  <td><tt>__xor__</tt></td>
   <td>x, y</td>
   <td>object</td>
   <td>^ operator</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>Numeric conversions</b></td>
  </tr>
   <tr>
  <td><tt>__int__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to integer</td>
  </tr>
   <tr>
  <td><tt>__long__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to long integer</td>
  </tr>
   <tr>
  <td><tt>__float__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to float</td>
  </tr>
   <tr>
  <td><tt>__oct__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to octal</td>
  </tr>
   <tr>
  <td><tt>__hex__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Convert to hexadecimal</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>In-place arithmetic operators</b></td>
  </tr>
   <tr>
  <td><tt>__iadd__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>+= operator</td>
  </tr>
   <tr>
  <td><tt>__isub__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>-= operator</td>
  </tr>
   <tr>
  <td><tt>__imul__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>*= operator</td>
  </tr>
   <tr>
  <td><tt>__idiv__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>/= operator for old-style division</td>
  </tr>
   <tr>
  <td><tt>__ifloordiv__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>//= operator</td>
  </tr>
   <tr>
  <td><tt>__itruediv__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>/= operator for new-style division</td>
  </tr>
   <tr>
  <td><tt>__imod__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>%= operator</td>
  </tr>
   <tr>
  <td><tt>__ipow__</tt></td>
   <td>x, y, z</td>
   <td>object</td>
   <td>**= operator</td>
  </tr>
   <tr>
  <td><tt>__ilshift__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>&lt;&lt;= operator</td>
  </tr>
   <tr>
  <td><tt>__irshift__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>&gt;&gt;= operator</td>
  </tr>
   <tr>
  <td><tt>__iand__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>&amp;= operator</td>
  </tr>
   <tr>
  <td><tt>__ior__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>|= operator</td>
  </tr>
   <tr>
  <td><tt>__ixor__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>^= operator</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>Sequences and mappings</b></td>
  </tr>
   <tr>
  <td><tt>__len__</tt></td>
   <td>self</td>
   <td>int</td>
   <td>len(self)</td>
  </tr>
   <tr>
  <td><tt>__getitem__</tt></td>
   <td>self, x</td>
   <td>object</td>
   <td>self[x]</td>
  </tr>
   <tr>
  <td><tt>__setitem__</tt></td>
   <td>self, x, y</td>
   <td>&nbsp;</td>
   <td>self[x] = y</td>
  </tr>
   <tr>
  <td><tt>__delitem__</tt></td>
   <td>self, x</td>
   <td>&nbsp;</td>
   <td>del self[x]</td>
  </tr>
   <tr>
  <td><tt>__getslice__</tt></td>
   <td>self, int i, int j</td>
   <td>object</td>
   <td>self[i:j]</td>
  </tr>
   <tr>
  <td><tt>__setslice__</tt></td>
   <td>self, int i, int j, x</td>
   <td>&nbsp;</td>
   <td>self[i:j] = x</td>
  </tr>
   <tr>
  <td><tt>__delslice__</tt></td>
   <td>self, int i, int j</td>
   <td>&nbsp;</td>
   <td>del self[i:j]</td>
  </tr>
   <tr>
  <td><tt>__contains__</tt></td>
   <td>self, x</td>
   <td>int</td>
   <td>x in self</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>Iterators</b></td>
  </tr>
   <tr>
  <td><tt>__next__</tt></td>
   <td>self</td>
   <td>object</td>
   <td>Get next item (called <tt>next</tt> in Python)</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>Buffer interface</b>&nbsp; (no Python equivalents
 - see note 1)</td>
  </tr>
   <tr>
  <td><tt>__getreadbuffer__</tt></td>
   <td>self, int i, void **p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr>
  <td><tt>__getwritebuffer__</tt></td>
   <td>self, int i, void **p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr>
  <td><tt>__getsegcount__</tt></td>
   <td>self, int *p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr>
  <td><tt>__getcharbuffer__</tt></td>
   <td>self, int i, char **p</td>
   <td>&nbsp;</td>
   <td>&nbsp;</td>
  </tr>
   <tr bgcolor="#66ffff" nosave="">
  <td colspan="4" nosave=""><b>Descriptor objects</b>&nbsp; (no Python equivalents
 - see note 2)</td>
  </tr>
   <tr>
  <td><tt>__get__</tt></td>
   <td>self, instance, class</td>
   <td>object</td>
   <td>Get value of attribute</td>
  </tr>
   <tr>
  <td><tt>__set__</tt></td>
   <td>self, instance, value</td>
   <td>&nbsp;</td>
   <td>Set value of attribute</td>
  </tr>
   <tr>
  <td>__delete__</td>
   <td>self, instance</td>
   <td>&nbsp;</td>
   <td>Delete attribute</td>
  </tr>
    
  </tbody> 
</table>
   </p>
 
<p>Note 1: The buffer interface is intended for use by C code and is not
directly accessible from Python. It is described in the <a
 href="http://www.python.org/doc/current/api/api.html">Python/C API Reference 
Manual</a> under sections <a
 href="http://www.python.org/doc/current/api/abstract-buffer.html">6.6</a>
and <a href="http://www.python.org/doc/current/api/buffer-structs.html">10.6</a>.
 </p>
 
<p>Note 2: Descriptor objects are part of the support mechanism for new-style
 Python classes. See the <a
 href="http://www.python.org/doc/2.2.1/whatsnew/sect-rellinks.html#SECTION000320000000000000000">discussion
 of descriptors in the Python documentation</a>. See also <a
 href="http://www.python.org/peps/pep-0252.html">PEP 252, "Making Types Look 
More Like Classes"</a>, and <a
 href="http://www.python.org/peps/pep-0253.html">PEP 253, "Subtyping Built-In 
Types"</a>. </p>
 
<p>Note 3: If your type defines a <tt>__new__</tt> method, any method called
 <tt>new</tt> that you define will be overwritten with the system-supplied
 <tt>new</tt> at module import time. </p>
 <br>
 <br>
</body>
</html>
\ No newline at end of file
+<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
+<html><head>
+         <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
+         <meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Special Methods of Extenstion Types</title></head>
+<body>
+   <h1>  <hr width="100%">Special Methods of Extension Types  
+<hr width="100%"></h1>
+  This page describes the special methods currently supported by Pyrex extension
+ types. A complete list of all the special methods appears in the table at 
+the bottom. Some of these methods behave differently from their Python counterparts 
+or have no direct Python counterparts, and require special mention.  
+<p><span style="font-weight: bold;">Note:</span><i> Everything said on this page applies only to </i><span style="font-weight: bold;">extension</span><i style="font-weight: bold;">
+</i><span style="font-weight: bold;">types</span><i>, defined with the </i><span style="font-weight: bold; font-family: monospace;">cdef class</span><i> statement. It doesn't apply&nbsp;
+to classes defined with the Python </i><span style="font-family: monospace;">class</span><i><span style="font-family: monospace;"> </span>statement, where the normal
+ Python rules apply.</i> </p>
+ <h2><small>Declaration</small></h2>Special methods of extension types must be declared with <span style="font-family: monospace; font-weight: bold;">def</span>, <span style="font-style: italic;">not</span> <span style="font-family: monospace;">cdef</span>.<br>
+<h2><font size="+1">Docstrings</font></h2>
+
+  Currently, docstrings are not fully supported in special methods of extension
+ types. You can place a docstring in the source to serve as a comment, but
+ it won't show up in the corresponding <span style="font-family: monospace;">__doc__</span> attribute at run time. (This
+ is a Python limitation -- there's nowhere in the PyTypeObject data structure
+ to put such docstrings.)  
+<h2> <font size="+1">Initialisation methods: <tt>__new__</tt> and <tt>__init__</tt></font></h2>
+  There are two methods concerned with initialising the object<tt>.</tt>
+<p>The <b><tt>__new__</tt></b> method is where you should perform basic C-level 
+initialisation of the object, including allocation of any C data structures 
+that your object will own. You need to be careful what you do in the __new__ 
+method, because the object may not yet be a valid Python object when it is 
+called. Therefore, you must not invoke any Python operations which might touch
+the object; in particular, do not try to call any of its methods. </p>
+ <p>Unlike the corresponding method in Python, your <tt>__new__</tt> method
+ is <i>not</i> responsible for <i>creating</i> the object. By the time your
+ <tt>__new__</tt> method is called, memory has been allocated for the object 
+and any C attributes it has have been initialised to 0 or null. (Any Python 
+attributes have also been initialised to <tt>None</tt>, but you probably shouldn't
+rely on that.) Your <tt>__new__</tt> method is guaranteed to be called exactly
+once.<br>
+<br>
+If your extension type has a base type, the <tt>__new__</tt> method of the
+base type is automatically called <i>before</i> your <tt>__new__</tt> method
+is called; you cannot explicitly call the inherited <tt>__new__</tt> method.
+If you need to pass a modified argument list to the base type, you will have
+to do the relevant part of the initialisation in the <tt>__init__</tt> method
+instead (where the normal rules for calling inherited methods apply).<br>
+ </p>
+ <p>Note that the first parameter of the <tt>__new__</tt> method is the object 
+to be initialised, not the class of the object as it is in Python. </p>
+ <p>Any initialisation which cannot safely be done in the <tt>__new__</tt>
+method should be done in the <b><tt>__init__</tt></b> method. By the time
+ <tt>__init__</tt> is called, the object is a fully valid Python object and 
+all operations are safe. Under some circumstances it is possible for <tt>__init__</tt>
+to be called more than once or not to be called at all, so your other methods
+ should be designed to be robust in such situations. </p>
+ <p>Keep in mind that any arguments passed to the constructor will be passed
+ to the <tt>__new__</tt> method as well as the <tt>__init__</tt> method.
+If you anticipate subclassing your extension type in Python, you may find
+it useful to give the <tt>__new__</tt> method * and ** arguments so that
+it can accept and ignore extra arguments. Otherwise, any Python subclass
+which has an <tt>__init__</tt> with a different signature will have to override 
+<tt>__new__</tt> as well as <tt>__init__</tt>, which the writer of a Python 
+class wouldn't expect to have to do. </p>
+ <h2> <font size="+1">Finalization method: <tt>__dealloc__</tt><tt></tt></font></h2>
+  The counterpart to the <tt>__new__</tt> method is the <b><tt>__dealloc__</tt></b>
+method, which should perform the inverse of the <tt>__new__</tt> method.
+Any C data structures that you allocated in your <tt>__new__</tt> method
+should be freed in your <tt>__dealloc__</tt> method.  
+<p>You need to be careful what you do in a <tt>__dealloc__</tt> method. By 
+the time your <tt>__dealloc__</tt> method is called, the object may already 
+have been partially destroyed and may not be in a valid state as far as Python 
+is concerned, so you should avoid invoking any Python operations which might 
+touch the object. In particular, don't call any other methods of the object 
+or do anything which might cause the object to be resurrected. It's best if
+you stick to just deallocating C data. </p>
+ <p>You don't need to worry about deallocating Python attributes of your object, 
+because that will be done for you by Pyrex after your <tt>__dealloc__</tt>
+method returns.<br>
+ <br>
+ <b>Note:</b> There is no <tt>__del__</tt> method for extension types. (Earlier 
+versions of the Pyrex documentation stated that there was, but this turned 
+out to be incorrect.)<br>
+  </p>
+ <h2><font size="+1">Arithmetic methods</font></h2>
+  Arithmetic operator methods, such as <tt>__add__</tt>, behave differently
+ from their Python counterparts. There are no separate "reversed" versions
+ of these methods (<tt>__radd__</tt>, etc.) Instead, if the first operand
+cannot perform the operation, the <i>same</i> method of the second operand
+is called, with the operands in the <i>same order</i>.  
+<p>This means that you can't rely on the first parameter of these methods
+ being "self", and you should test the types of both operands before deciding
+ what to do. If you can't handle the combination of types you've been given,
+ you should return <tt>NotImplemented</tt>. </p>
+ <p>This also applies to the in-place arithmetic method <tt>__ipow__</tt>.
+ It doesn't apply to any of the <i>other</i> in-place methods (<tt>__iadd__</tt>,
+ etc.) which always take self as the first argument. </p>
+ <h2> <font size="+1">Rich comparisons</font></h2>
+  There are no separate methods for the individual rich comparison operations
+ (<tt>__eq__</tt>, <tt>__le__</tt>, etc.) Instead there is a single method
+ <tt>__richcmp__</tt> which takes an integer indicating which operation is 
+to be performed, as follows:  
+<ul>
+      <ul>
+ &nbsp;         <table nosave="" border="0" cellpadding="5" cellspacing="0">
+  <tbody>
+         <tr nosave="">
+  <td nosave="" bgcolor="#ffcc33" width="30">                     <div align="right">&lt;</div>
+  </td>
+   <td nosave="" bgcolor="#66ffff" width="30">0</td>
+   <td><br>
+           </td>
+   <td nosave="" bgcolor="#ffcc33" width="30">                     <div align="right">==</div>
+  </td>
+   <td nosave="" bgcolor="#66ffff" width="30">2</td>
+   <td><br>
+           </td>
+   <td nosave="" bgcolor="#ffcc33" width="30">                     <div align="right">&gt;</div>
+  </td>
+   <td nosave="" bgcolor="#66ffff" width="30">4</td>
+  </tr>
+   <tr nosave="">
+  <td nosave="" bgcolor="#ffcc33">                     <div align="right">&lt;=</div>
+  </td>
+   <td nosave="" bgcolor="#66ffff">1</td>
+   <td><br>
+           </td>
+   <td nosave="" bgcolor="#ffcc33">                     <div align="right">!=</div>
+  </td>
+   <td nosave="" bgcolor="#66ffff">3</td>
+   <td><br>
+           </td>
+   <td nosave="" bgcolor="#ffcc33">                     <div align="right">&gt;=</div>
+  </td>
+   <td nosave="" bgcolor="#66ffff">5</td>
+  </tr>
+              </tbody>         </table>
+      </ul>
+  </ul>
+   <h2> <font size="+1">The __next__ method</font></h2>
+  Extension types wishing to implement the iterator interface should define
+ a method called <b><tt>__next__</tt></b>, <i>not</i> <tt>next</tt>. The Python
+ system will automatically supply a <tt>next</tt> method which calls your
+<span style="font-family: monospace;">__next__</span>.  <b>Do NOT explicitly give your type a <tt>next</tt> method</b>,
+or bad things could happen (see note 3).  
+<h2> <font size="+1">Special Method Table</font></h2>
+  This table lists all of the special methods together with their parameter
+ and return types. A parameter named <b>self</b> is of the type the method
+ belongs to. Other untyped parameters are generic Python objects.  
+<p>You don't have to declare your method as taking these parameter types.
+ If you declare different types, conversions will be performed as necessary.
+ <br>
+ &nbsp; <table nosave="" bgcolor="#ccffff" border="1" cellpadding="5" cellspacing="0">
+  <tbody>
+     <tr nosave="" bgcolor="#ffcc33">
+  <td nosave=""><b>Name</b></td>
+   <td><b>Parameters</b></td>
+   <td><b>Return type</b></td>
+   <td><b>Description</b></td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>General</b></td>
+  </tr>
+   <tr>
+  <td><tt>__new__</tt></td>
+   <td>self, ...</td>
+   <td>&nbsp;</td>
+   <td>Basic initialisation (no direct Python equivalent)</td>
+  </tr>
+   <tr>
+  <td><tt>__init__</tt></td>
+   <td>self, ...</td>
+   <td>&nbsp;</td>
+   <td>Further initialisation</td>
+  </tr>
+   <tr>
+  <td><tt>__dealloc__</tt></td>
+   <td>self</td>
+   <td>&nbsp;</td>
+   <td>Basic deallocation (no direct Python equivalent)</td>
+  </tr>
+   <tr>
+  <td><tt>__cmp__</tt></td>
+   <td>x, y</td>
+   <td>int</td>
+   <td>3-way comparison</td>
+  </tr>
+   <tr>
+  <td><tt>__richcmp__</tt></td>
+   <td>x, y, int op</td>
+   <td>object</td>
+   <td>Rich comparison (no direct Python equivalent)</td>
+  </tr>
+   <tr>
+  <td><tt>__str__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>str(self)</td>
+  </tr>
+   <tr>
+  <td><tt>__repr__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>repr(self)</td>
+  </tr>
+   <tr nosave="">
+  <td nosave=""><tt>__hash__</tt></td>
+   <td>self</td>
+   <td>int</td>
+   <td>Hash function</td>
+  </tr>
+   <tr>
+  <td><tt>__call__</tt></td>
+   <td>self, ...</td>
+   <td>object</td>
+   <td>self(...)</td>
+  </tr>
+   <tr>
+  <td><tt>__iter__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Return iterator for sequence</td>
+  </tr>
+   <tr>
+  <td><tt>__getattr__</tt></td>
+   <td>self, name</td>
+   <td>object</td>
+   <td>Get attribute</td>
+  </tr>
+   <tr>
+  <td><tt>__setattr__</tt></td>
+   <td>self, name, val</td>
+   <td>&nbsp;</td>
+   <td>Set attribute</td>
+  </tr>
+   <tr>
+  <td><tt>__delattr__</tt></td>
+   <td>self, name</td>
+   <td>&nbsp;</td>
+   <td>Delete attribute</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>Arithmetic operators</b></td>
+  </tr>
+   <tr>
+  <td><tt>__add__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>binary + operator</td>
+  </tr>
+   <tr>
+  <td><tt>__sub__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>binary - operator</td>
+  </tr>
+   <tr>
+  <td><tt>__mul__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>* operator</td>
+  </tr>
+   <tr>
+  <td><tt>__div__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>/&nbsp; operator for old-style division</td>
+  </tr>
+   <tr>
+  <td><tt>__floordiv__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>//&nbsp; operator</td>
+  </tr>
+   <tr>
+  <td><tt>__truediv__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>/&nbsp; operator for new-style division</td>
+  </tr>
+   <tr>
+  <td><tt>__mod__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>% operator</td>
+  </tr>
+   <tr>
+  <td><tt>__divmod__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>combined div and mod</td>
+  </tr>
+   <tr>
+  <td><tt>__pow__</tt></td>
+   <td>x, y, z</td>
+   <td>object</td>
+   <td>** operator or pow(x, y, z)</td>
+  </tr>
+   <tr>
+  <td><tt>__neg__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>unary - operator</td>
+  </tr>
+   <tr>
+  <td><tt>__pos__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>unary + operator</td>
+  </tr>
+   <tr>
+  <td><tt>__abs__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>absolute value</td>
+  </tr>
+   <tr>
+  <td><tt>__nonzero__</tt></td>
+   <td>self</td>
+   <td>int</td>
+   <td>convert to boolean</td>
+  </tr>
+   <tr>
+  <td><tt>__invert__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>~ operator</td>
+  </tr>
+   <tr>
+  <td><tt>__lshift__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>&lt;&lt; operator</td>
+  </tr>
+   <tr>
+  <td><tt>__rshift__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>&gt;&gt; operator</td>
+  </tr>
+   <tr>
+  <td><tt>__and__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>&amp; operator</td>
+  </tr>
+   <tr>
+  <td><tt>__or__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>| operator</td>
+  </tr>
+   <tr>
+  <td><tt>__xor__</tt></td>
+   <td>x, y</td>
+   <td>object</td>
+   <td>^ operator</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>Numeric conversions</b></td>
+  </tr>
+   <tr>
+  <td><tt>__int__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Convert to integer</td>
+  </tr>
+   <tr>
+  <td><tt>__long__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Convert to long integer</td>
+  </tr>
+   <tr>
+  <td><tt>__float__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Convert to float</td>
+  </tr>
+   <tr>
+  <td><tt>__oct__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Convert to octal</td>
+  </tr>
+   <tr>
+  <td><tt>__hex__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Convert to hexadecimal</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>In-place arithmetic operators</b></td>
+  </tr>
+   <tr>
+  <td><tt>__iadd__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>+= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__isub__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>-= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__imul__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>*= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__idiv__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>/= operator for old-style division</td>
+  </tr>
+   <tr>
+  <td><tt>__ifloordiv__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>//= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__itruediv__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>/= operator for new-style division</td>
+  </tr>
+   <tr>
+  <td><tt>__imod__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>%= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__ipow__</tt></td>
+   <td>x, y, z</td>
+   <td>object</td>
+   <td>**= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__ilshift__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>&lt;&lt;= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__irshift__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>&gt;&gt;= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__iand__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>&amp;= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__ior__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>|= operator</td>
+  </tr>
+   <tr>
+  <td><tt>__ixor__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>^= operator</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>Sequences and mappings</b></td>
+  </tr>
+   <tr>
+  <td><tt>__len__</tt></td>
+   <td>self</td>
+   <td>int</td>
+   <td>len(self)</td>
+  </tr>
+   <tr>
+  <td><tt>__getitem__</tt></td>
+   <td>self, x</td>
+   <td>object</td>
+   <td>self[x]</td>
+  </tr>
+   <tr>
+  <td><tt>__setitem__</tt></td>
+   <td>self, x, y</td>
+   <td>&nbsp;</td>
+   <td>self[x] = y</td>
+  </tr>
+   <tr>
+  <td><tt>__delitem__</tt></td>
+   <td>self, x</td>
+   <td>&nbsp;</td>
+   <td>del self[x]</td>
+  </tr>
+   <tr>
+  <td><tt>__getslice__</tt></td>
+   <td>self, int i, int j</td>
+   <td>object</td>
+   <td>self[i:j]</td>
+  </tr>
+   <tr>
+  <td><tt>__setslice__</tt></td>
+   <td>self, int i, int j, x</td>
+   <td>&nbsp;</td>
+   <td>self[i:j] = x</td>
+  </tr>
+   <tr>
+  <td><tt>__delslice__</tt></td>
+   <td>self, int i, int j</td>
+   <td>&nbsp;</td>
+   <td>del self[i:j]</td>
+  </tr>
+   <tr>
+  <td><tt>__contains__</tt></td>
+   <td>self, x</td>
+   <td>int</td>
+   <td>x in self</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>Iterators</b></td>
+  </tr>
+   <tr>
+  <td><tt>__next__</tt></td>
+   <td>self</td>
+   <td>object</td>
+   <td>Get next item (called <tt>next</tt> in Python)</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>Buffer interface</b>&nbsp; (no Python equivalents
+ - see note 1)</td>
+  </tr>
+   <tr>
+  <td><tt>__getreadbuffer__</tt></td>
+   <td>self, int i, void **p</td>
+   <td>&nbsp;</td>
+   <td>&nbsp;</td>
+  </tr>
+   <tr>
+  <td><tt>__getwritebuffer__</tt></td>
+   <td>self, int i, void **p</td>
+   <td>&nbsp;</td>
+   <td>&nbsp;</td>
+  </tr>
+   <tr>
+  <td><tt>__getsegcount__</tt></td>
+   <td>self, int *p</td>
+   <td>&nbsp;</td>
+   <td>&nbsp;</td>
+  </tr>
+   <tr>
+  <td><tt>__getcharbuffer__</tt></td>
+   <td>self, int i, char **p</td>
+   <td>&nbsp;</td>
+   <td>&nbsp;</td>
+  </tr>
+   <tr nosave="" bgcolor="#66ffff">
+  <td colspan="4" nosave=""><b>Descriptor objects</b>&nbsp; (no Python equivalents
+ - see note 2)</td>
+  </tr>
+   <tr>
+  <td><tt>__get__</tt></td>
+   <td>self, instance, class</td>
+   <td>object</td>
+   <td>Get value of attribute</td>
+  </tr>
+   <tr>
+  <td><tt>__set__</tt></td>
+   <td>self, instance, value</td>
+   <td>&nbsp;</td>
+   <td>Set value of attribute</td>
+  </tr>
+   <tr>
+  <td style="font-family: monospace;">__delete__</td>
+   <td>self, instance</td>
+   <td>&nbsp;</td>
+   <td>Delete attribute</td>
+  </tr>
+      </tbody> </table>
+   </p>
+ <p>Note 1: The buffer interface is intended for use by C code and is not
+directly accessible from Python. It is described in the <a href="http://www.python.org/doc/current/api/api.html">Python/C API Reference 
+Manual</a> under sections <a href="http://www.python.org/doc/current/api/abstract-buffer.html">6.6</a>
+and <a href="http://www.python.org/doc/current/api/buffer-structs.html">10.6</a>.
+ </p>
+ <p>Note 2: Descriptor objects are part of the support mechanism for new-style
+ Python classes. See the <a href="http://www.python.org/doc/2.2.1/whatsnew/sect-rellinks.html#SECTION000320000000000000000">discussion
+ of descriptors in the Python documentation</a>. See also <a href="http://www.python.org/peps/pep-0252.html">PEP 252, "Making Types Look 
+More Like Classes"</a>, and <a href="http://www.python.org/peps/pep-0253.html">PEP 253, "Subtyping Built-In 
+Types"</a>. </p>
+ <p>Note 3: If your type defines a <tt>__new__</tt> method, any method called
+ <tt>new</tt> that you define will be overwritten with the system-supplied
+ <tt>new</tt> at module import time. </p>
+ <br>
+ <br>
+</body></html>
\ No newline at end of file

Modified: lxml/pyrex/Pyrex/Compiler/CmdLine.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/CmdLine.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/CmdLine.py	Tue Apr 25 15:43:21 2006
@@ -10,7 +10,12 @@
   -v, --version                  Display version number of pyrex compiler
   -l, --create-listing           Write error messages to a listing file
   -I, --include-dir <directory>  Search for include files in named directory
-  -o, --output-file <filename>   Specify name of generated C file"""
+  -o, --output-file <filename>   Specify name of generated C file
+The following experimental options are supported only on MacOSX:
+  -C, --compile    Compile generated .c file to .o file
+  -X, --link       Link .o file to produce extension module (implies -C)
+  -+, --cplus      Use C++ compiler for compiling and linking
+  Additional .o files to link may be supplied when using -X."""
 
 def bad_usage():
     print >>sys.stderr, usage
@@ -47,6 +52,8 @@
             elif option in ("-X", "--link"):
                 options.c_only = 0
                 options.obj_only = 0
+            elif option in ("-+", "--cplus"):
+                options.cplus = 1
             elif option.startswith("-I"):
                 options.include_path.append(get_param(option))
             elif option == "--include-dir":
@@ -56,7 +63,17 @@
             else:
                 bad_usage()
         else:
-            sources.append(pop_arg())
+            arg = pop_arg()
+            if arg.endswith(".pyx"):
+                sources.append(arg)
+            elif arg.endswith(".o"):
+                options.objects.append(arg)
+            else:
+                print >>sys.stderr, \
+                    "pyrexc: %s: Unknown filename suffix" % arg
+    if options.objects and len(sources) > 1:
+        print >>sys.stderr, \
+            "pyrexc: Only one source file allowed together with .o files"
     if options.use_listing_file and len(sources) > 1:
         print >>sys.stderr, \
             "pyrexc: Only one source file allowed when using -o"

Modified: lxml/pyrex/Pyrex/Compiler/Code.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Code.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Code.py	Tue Apr 25 15:43:21 2006
@@ -4,6 +4,7 @@
 
 import Naming
 from Pyrex.Utils import open_new_file
+from PyrexTypes import py_object_type, typecast
 
 class CCodeWriter:
     # f                file            output file
@@ -18,9 +19,9 @@
     # in_try_finally   boolean         inside try of try...finally
     # filename_table   {string : int}  for finding filename table indexes
     # filename_list    [string]        filenames in filename table order
-
+    
     in_try_finally = 0
-
+    
     def __init__(self, outfile_name):
         self.f = open_new_file(outfile_name)
         self.level = 0
@@ -30,7 +31,7 @@
         self.error_label = None
         self.filename_table = {}
         self.filename_list = []
-
+    
     def putln(self, code = ""):
         if self.marker and self.bol:
             self.emit_marker()
@@ -38,7 +39,7 @@
             self.put(code)
         self.f.write("\n");
         self.bol = 1
-
+    
     def emit_marker(self):
         self.f.write("\n");
         self.indent()
@@ -55,74 +56,61 @@
         self.bol = 0
         if dl > 0:
             self.level += dl
-
-    def begin_require_python(self, hex_version):
-        self.putln("#if PY_VERSION_HEX >= %#010X" % hex_version)
-
-    def else_require_python(self, hex_version=None):
-        if hex_version is not None:
-            self.putln("#elif PY_VERSION_HEX >= %#010X" % hex_version)
-        else:
-            self.putln("#else")
-
-    def end_require_python(self):
-        self.putln("#endif")
-
+    
     def increase_indent(self):
         self.level = self.level + 1
-
+    
     def decrease_indent(self):
         self.level = self.level - 1
-
+    
     def begin_block(self):
         self.putln("{")
         self.increase_indent()
-
+    
     def end_block(self):
         self.decrease_indent()
         self.putln("}")
-
+    
     def indent(self):
         self.f.write("  " * self.level)
-
+    
     def mark_pos(self, pos):
         file, line, col = pos
         self.marker = '"%s":%s' % (file, line)
 
     def init_labels(self):
         self.label_counter = 0
-        self.used_labels = {}
         self.return_label = self.new_label()
         self.new_error_label()
         self.continue_label = None
         self.break_label = None
-
+    
     def new_label(self):
         n = self.label_counter
         self.label_counter = n + 1
         return "%s%d" % (Naming.label_prefix, n)
-
+    
     def new_error_label(self):
         old_err_lbl = self.error_label
         self.error_label = self.new_label()
         return old_err_lbl
-
+    
     def get_loop_labels(self):
         return (
             self.continue_label,
             self.break_label)
-
+    
     def set_loop_labels(self, labels):
         (self.continue_label,
          self.break_label) = labels
-
+    
     def new_loop_labels(self):
         old_labels = self.get_loop_labels()
         self.set_loop_labels(
-            (self.new_label(),
+            (self.new_label(), 
              self.new_label()))
         return old_labels
-
+    
     def get_all_labels(self):
         return (
             self.continue_label,
@@ -147,35 +135,47 @@
         self.set_all_labels(new_labels)
         return old_labels
 
-    def mark_label_used(self, lbl):
-        self.used_labels[lbl] = 1
-
     def put_label(self, lbl):
-        if lbl in self.used_labels:
-            self.putln("%s:;" % lbl)
-
-    def put_goto(self, lbl):
-        self.putln("goto %s;" % lbl)
-        self.mark_label_used(lbl)
-
-    def put_var_declarations(self, entries, static = 0, dll_linkage = None):
+        self.putln("%s:;" % lbl)
+    
+    def put_var_declarations(self, entries, static = 0, dll_linkage = None,
+            definition = True):
         for entry in entries:
             if not entry.in_cinclude:
-                self.put_var_declaration(entry, static, dll_linkage)
-
-    def put_var_declaration(self, entry, static = 0, dll_linkage = None):
-        if entry.visibility == 'extern':
-            self.put("extern ")
-        elif static and entry.visibility <> 'public':
-            self.put("static ")
-        if entry.visibility <> 'public':
+                self.put_var_declaration(entry, static, dll_linkage, definition)
+    
+    def put_var_declaration(self, entry, static = 0, dll_linkage = None,
+            definition = True):
+        #print "Code.put_var_declaration:", entry.name, "definition =", definition
+        visibility = entry.visibility
+        if visibility == 'private' and not definition:
+            return
+        if visibility == 'extern':
+            storage_class = Naming.extern_c_macro
+        elif visibility == 'public':
+            if definition:
+                storage_class = ""
+            else:
+                storage_class = Naming.extern_c_macro
+        elif visibility == 'private':
+            if static:
+                storage_class = "static"
+            else:
+                storage_class = ""
+        if storage_class:
+            self.put("%s " % storage_class)
+        #if visibility == 'extern' or visibility == 'public' and not definition:
+        #	self.put("%s " % Naming.extern_c_macro)
+        #elif static and visibility <> 'public':
+        #	self.put("static ")
+        if visibility <> 'public':
             dll_linkage = None
         self.put(entry.type.declaration_code(entry.cname,
             dll_linkage = dll_linkage))
         if entry.init is not None:
             self.put(" = %s" % entry.type.literal_code(entry.init))
         self.putln(";")
-
+    
     def entry_as_pyobject(self, entry):
         type = entry.type
         if (not entry.is_self_arg and not entry.type.is_complete()) \
@@ -183,30 +183,32 @@
             return "(PyObject *)" + entry.cname
         else:
             return entry.cname
-
+    
     def as_pyobject(self, cname, type):
-        if type.is_extension_type and type.base_type:
-            return "(PyObject *)" + cname
-        else:
-            return cname
-
+        return typecast(py_object_type, type, cname)
+        #if type.is_extension_type and type.base_type:
+        #	return "(PyObject *)" + cname
+        #else:
+        #	return cname
+    
     def put_incref(self, cname, type):
         self.putln("Py_INCREF(%s);" % self.as_pyobject(cname, type))
-
+    
     def put_decref(self, cname, type):
         self.putln("Py_DECREF(%s);" % self.as_pyobject(cname, type))
-
+    
     def put_var_incref(self, entry):
         if entry.type.is_pyobject:
             self.putln("Py_INCREF(%s);" % self.entry_as_pyobject(entry))
-
+    
     def put_decref_clear(self, cname, type):
         self.putln("Py_DECREF(%s); %s = 0;" % (
-            self.as_pyobject(cname, type), cname))
-
+            typecast(py_object_type, type, cname), cname))
+            #self.as_pyobject(cname, type), cname))
+    
     def put_xdecref(self, cname, type):
         self.putln("Py_XDECREF(%s);" % self.as_pyobject(cname, type))
-
+    
     def put_xdecref_clear(self, cname, type):
         self.putln("Py_XDECREF(%s); %s = 0;" % (
             self.as_pyobject(cname, type), cname))
@@ -214,44 +216,45 @@
     def put_var_decref(self, entry):
         if entry.type.is_pyobject:
             self.putln("Py_DECREF(%s);" % self.entry_as_pyobject(entry))
-
+    
+    def put_var_decref_clear(self, entry):
+        if entry.type.is_pyobject:
+            self.putln("Py_DECREF(%s); %s = 0;" % (
+                self.entry_as_pyobject(entry), entry.cname))
+    
     def put_var_xdecref(self, entry):
         if entry.type.is_pyobject:
             self.putln("Py_XDECREF(%s);" % self.entry_as_pyobject(entry))
-
+    
     def put_var_xdecref_clear(self, entry):
         if entry.type.is_pyobject:
             self.putln("Py_XDECREF(%s); %s = 0;" % (
                 self.entry_as_pyobject(entry), entry.cname))
-
+    
     def put_var_decrefs(self, entries):
         for entry in entries:
             if entry.xdecref_cleanup:
                 self.put_var_xdecref(entry)
             else:
                 self.put_var_decref(entry)
-
+    
     def put_var_xdecrefs(self, entries):
         for entry in entries:
             self.put_var_xdecref(entry)
-
+    
     def put_var_xdecrefs_clear(self, entries):
         for entry in entries:
             self.put_var_xdecref_clear(entry)
-
-    def put_init_to_py_none(self, cast, cname):
-        if cast:
-            self.putln("%s = (void *)Py_None; Py_INCREF(%s %s);" % (cname, cast, cname))
-        else:
-            self.putln("%s = Py_None; Py_INCREF(%s);" % (cname, cname))
-
+    
+    def put_init_to_py_none(self, cname, type):
+        py_none = typecast(type, py_object_type, "Py_None")
+        self.putln("%s = %s; Py_INCREF(Py_None);" % (cname, py_none))
+    
     def put_init_var_to_py_none(self, entry, template = "%s"):
         code = template % entry.cname
-        if entry.type.is_extension_type:
-            cast = "(PyObject *)"
-        else:
-            cast = None
-        self.put_init_to_py_none(cast, code)
+        #if entry.type.is_extension_type:
+        #	code = "((PyObject*)%s)" % code
+        self.put_init_to_py_none(code, entry.type)
 
     def put_pymethoddef(self, entry, term):
         if entry.doc:
@@ -260,13 +263,12 @@
             doc_code = 0
         self.putln(
             '{"%s", (PyCFunction)%s, METH_VARARGS|METH_KEYWORDS, %s}%s' % (
-                entry.name,
-                entry.func_cname,
+                entry.name, 
+                entry.func_cname, 
                 doc_code,
                 term))
-
+    
     def error_goto(self, pos):
-        self.mark_label_used(self.error_label)
         return "{%s = %s[%s]; %s = %s; goto %s;}" % (
             Naming.filename_cname,
             Naming.filetable_cname,
@@ -274,7 +276,7 @@
             Naming.lineno_cname,
             pos[1],
             self.error_label)
-
+    
     def lookup_filename(self, filename):
         try:
             index = self.filename_table[filename]
@@ -292,13 +294,13 @@
     def __init__(self, outfile_name):
         self.f = open_new_file(outfile_name)
         self.level = 0
-
+    
     def putln(self, code):
         self.f.write("%s%s\n" % (" " * self.level, code))
-
+    
     def indent(self):
         self.level += 1
-
+    
     def dedent(self):
         self.level -= 1
 

Modified: lxml/pyrex/Pyrex/Compiler/ExprNodes.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/ExprNodes.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/ExprNodes.py	Tue Apr 25 15:43:21 2006
@@ -8,7 +8,7 @@
 import Naming
 from Nodes import Node
 import PyrexTypes
-from PyrexTypes import py_object_type
+from PyrexTypes import py_object_type, typecast
 import Symtab
 import Options
 
@@ -17,15 +17,18 @@
     debug_coercion
 
 class ExprNode(Node):
-    #  subexprs    [string]    Class var holding names of subexpr node attrs
-    #  type        PyrexType   Type of the result
-    #  result      string      C code fragment
-    #  is_temp     boolean     Result is in a temporary variable
+    #  subexprs     [string]     Class var holding names of subexpr node attrs
+    #  type         PyrexType    Type of the result
+    #  result_code  string       Code fragment
+    #  result_ctype string       C type of result_code if different from type
+    #  is_temp      boolean      Result is in a temporary variable
     #  is_sequence_constructor  
-    #              boolean     Is a list or tuple constructor expression
+    #               boolean      Is a list or tuple constructor expression
     #  saved_subexpr_nodes
-    #              [ExprNode or [ExprNode or None] or None]
-    #                           Cached result of subexpr_nodes()
+    #               [ExprNode or [ExprNode or None] or None]
+    #                            Cached result of subexpr_nodes()
+    
+    result_ctype = None
 
     #  The Analyse Expressions phase for expressions is split
     #  into two sub-phases:
@@ -39,7 +42,7 @@
     #
     #    Allocate Temps
     #      Allocates temporary variables where needed, and fills
-    #      in the result field of each node.
+    #      in the result_code field of each node.
     #
     #  ExprNode provides some convenience routines which
     #  perform both of the above phases. These should only
@@ -89,7 +92,7 @@
     #        A default implementation of allocate_temps is 
     #        provided which uses the following abstract method:
     #
-    #          result_code
+    #          calculate_result_code
     #            - Return a C code fragment evaluating to
     #              the result. This is only called when the
     #              result is not a temporary.
@@ -142,16 +145,16 @@
     #        - Generate code to perform the deletion.
     #        - Call generate_disposal_code on all sub-expressions.
     #
-    #       result_as_extension_type
-    #         Normally, the results of all nodes whose type
-    #         is a Python object, either generic or an extension
-    #         type, are returned as a generic Python object, so
-    #         that they can be passed directly to Python/C API
-    #         routines. This method is called to obtain the
-    #         result as the actual type of the node. It is only
-    #         called when the type is known to actually be an
-    #         extension type, and nodes whose result can never
-    #         be an extension type need not implement it.
+    #       #result_as_extension_type ### OBSOLETE ###
+    #       #  Normally, the results of all nodes whose type
+    #       #  is a Python object, either generic or an extension
+    #       #  type, are returned as a generic Python object, so
+    #       #  that they can be passed directly to Python/C API
+    #       #  routines. This method is called to obtain the
+    #       #  result as the actual type of the node. It is only
+    #       #  called when the type is known to actually be an
+    #       #  extension type, and nodes whose result can never
+    #       #  be an extension type need not implement it.
     #
     
     is_sequence_constructor = 0
@@ -192,6 +195,19 @@
             self.saved_subexpr_nodes = nodes
         return self.saved_subexpr_nodes
     
+    def result_as(self, type = None):
+        #  Return the result code cast to the specified C type.
+        return typecast(type, self.ctype(), self.result_code)
+    
+    def py_result(self):
+        #  Return the result code cast to PyObject *.
+        return self.result_as(py_object_type)
+    
+    def ctype(self):
+        #  Return the native C type of the result (i.e. the
+        #  C type of the result_code expression).
+        return self.result_ctype or self.type
+    
     # ------------- Declaration Analysis ----------------
     
     def analyse_target_declaration(self, env):
@@ -287,7 +303,7 @@
         if debug_temp_alloc:
             print self, "Allocating target temps"
         self.allocate_subexpr_temps(env)
-        self.result = self.target_code()
+        self.result_code = self.target_code()
     
     def allocate_temps(self, env, result = None):
         #  Allocate temporary variables for this node and
@@ -325,27 +341,26 @@
         if result:
             if not self.is_temp:
                 raise InternalError("Result forced on non-temp node")
-            self.result = result
+            self.result_code = result
         elif self.is_temp:
             type = self.type
             if not type.is_void:
                 if type.is_pyobject:
                     type = PyrexTypes.py_object_type
-                self.result = env.allocate_temp(type)
+                self.result_code = env.allocate_temp(type)
             else:
-                self.result = None
+                self.result_code = None
             if debug_temp_alloc:
-                print self, "Allocated result", self.result
-                #print_call_chain(self, "allocated temp", self.result)
+                print self, "Allocated result", self.result_code
         else:
-            self.result = self.result_code()
+            self.result_code = self.calculate_result_code()
     
     def target_code(self):
         #  Return code fragment for use as LHS of a C assignment.
-        return self.result_code()
+        return self.calculate_result_code()
     
-    def result_code(self):
-        self.not_implemented("result_code")
+    def calculate_result_code(self):
+        self.not_implemented("calculate_result_code")
     
     def release_target_temp(self, env):
         #  Release temporaries used by LHS of an assignment.
@@ -356,8 +371,8 @@
         #  otherwise release results of its sub-expressions.
         if self.is_temp:
             if debug_temp_alloc:
-                print self, "Releasing result", self.result
-            env.release_temp(self.result)
+                print self, "Releasing result", self.result_code
+            env.release_temp(self.result_code)
         else:
             self.release_subexpr_temps(env)
     
@@ -373,10 +388,8 @@
     def make_owned_reference(self, code):
         #  If result is a pyobject, make sure we own
         #  a reference to it.
-        #if self.type.is_pyobject and not self.is_temp:
         if self.type.is_pyobject and not self.result_in_temp():
-            #code.put_incref(self.result, self.type)
-            code.put_incref(self.result, py_object_type)
+            code.put_incref(self.result_code, self.ctype())
     
     def generate_evaluation_code(self, code):
         #  Generate code to evaluate this node and
@@ -399,7 +412,7 @@
         # temporary Python reference.
         if self.is_temp:
             if self.type.is_pyobject:
-                code.put_decref_clear(self.result, self.type)
+                code.put_decref_clear(self.result_code, self.ctype())
         else:
             self.generate_subexpr_disposal_code(code)
     
@@ -415,7 +428,7 @@
         # the result if it is a Python object.
         if self.is_temp:
             if self.type.is_pyobject:
-                code.putln("%s = 0;" % self.result)
+                code.putln("%s = 0;" % self.result_code)
         else:
             self.generate_subexpr_disposal_code(code)
     
@@ -517,9 +530,9 @@
         return 1
     
     def analyse_types(self, env):
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
     
-    def result_code(self):
+    def calculate_result_code(self):
         return self.value
 
     def generate_result_code(self, code):
@@ -554,7 +567,7 @@
     def check_const(self):
         pass
     
-    def result_code(self):
+    def calculate_result_code(self):
         return str(self.value)
 
     def generate_result_code(self, code):
@@ -569,7 +582,7 @@
 class CharNode(ConstNode):
     type = PyrexTypes.c_char_type
     
-    def result_code(self):
+    def calculate_result_code(self):
         return "'%s'" % self.value
 
 
@@ -608,28 +621,46 @@
         env.add_py_string(entry)
         return StringNode(self.pos, entry = entry, type = py_object_type)
             
-    def result_code(self):
+    def calculate_result_code(self):
         if self.type.is_pyobject:
             return self.entry.pystring_cname
         else:
             return self.entry.cname
 
 
+class LongNode(AtomicExprNode):
+    #  Python long integer literal
+    #
+    #  value   string
+    
+    def analyse_types(self, env):
+        self.type = py_object_type
+        self.is_temp = 1
+    
+    def generate_evaluation_code(self, code):
+        code.putln(
+            '%s = PyLong_FromString("%s", 0, 0); if (!%s) %s' % (
+                self.result_code,
+                self.value,
+                self.result_code,
+                code.error_goto(self.pos)))	
+
+
 class ImagNode(AtomicExprNode):
     #  Imaginary number literal
     #
     #  value   float    imaginary part
     
     def analyse_types(self, env):
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_evaluation_code(self, code):
         code.putln(
             "%s = PyComplex_FromDoubles(0.0, %s); if (!%s) %s" % (
-                self.result,
+                self.result_code,
                 self.value,
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -660,8 +691,9 @@
     def analyse_target_declaration(self, env):
         self.entry = env.lookup_here(self.name)
         if not self.entry:
-            self.entry = env.declare_var(self.name, 
-                PyrexTypes.py_object_type, self.pos)
+            #print "NameNode.analyse_target_declaration:", self.name ###
+            #print "...declaring as py_object_type" ###
+            self.entry = env.declare_var(self.name, py_object_type, self.pos)
     
     def analyse_types(self, env):
         self.entry = env.lookup(self.name)
@@ -672,8 +704,10 @@
     def analyse_entry(self, env):
         self.check_identifier_kind()
         self.type = self.entry.type
+        if self.entry.is_declared_generic:
+            self.result_ctype = py_object_type
         # Reference to C array turns into pointer to first element.
-        if self.type.is_array:
+        while self.type.is_array:
             self.type = self.type.element_ptr_type()
         if self.entry.is_pyglobal or self.entry.is_builtin:
             assert self.type.is_pyobject, "Python global or builtin not a Python object"
@@ -729,25 +763,10 @@
         #  result is in a temporary.
         return 0
 
-    def result_code(self):
-        if self.entry is None:
-            return "<error>" # There was an error earlier
-        result = self.entry.cname
-        if self.type.is_extension_type and \
-            not self.entry.is_declared_generic:
-                result = "((PyObject *)%s)" % result
-        return result
-    
-    def result_as_extension_type(self):
+    def calculate_result_code(self):
         if self.entry is None:
             return "<error>" # There was an error earlier
-        #if not self.entry.is_self_arg:
-        if not self.entry.is_declared_generic:
-            return self.entry.cname
-        else:
-            return "((%s)%s)" % (
-                self.type.declaration_code(""), 
-                self.entry.cname)
+        return self.entry.cname
     
     def generate_result_code(self, code):
         if not hasattr(self, 'entry'):
@@ -764,18 +783,18 @@
                 #assert entry.interned_cname is not None
                 code.putln(
                     '%s = __Pyx_GetName(%s, %s); if (!%s) %s' % (
-                    self.result,
+                    self.result_code,
                     namespace, 
                     entry.interned_cname,
-                    self.result, 
+                    self.result_code, 
                     code.error_goto(self.pos)))		
             else:
                 code.putln(
                     '%s = __Pyx_GetName(%s, "%s"); if (!%s) %s' % (
-                    self.result,
+                    self.result_code,
                     namespace, 
                     self.entry.name,
-                    self.result, 
+                    self.result_code, 
                     code.error_goto(self.pos)))		
 
     def generate_assignment_code(self, rhs, code):
@@ -789,14 +808,14 @@
                     'if (PyObject_SetAttr(%s, %s, %s) < 0) %s' % (
                         namespace, 
                         entry.interned_cname,
-                        rhs.result, 
+                        rhs.py_result(), 
                         code.error_goto(self.pos)))
             else:
                 code.putln(
                     'if (PyObject_SetAttrString(%s, "%s", %s) < 0) %s' % (
                         namespace, 
                         entry.name,
-                        rhs.result, 
+                        rhs.py_result(), 
                         code.error_goto(self.pos)))
             if debug_disposal_code:
                 print "NameNode.generate_assignment_code:"
@@ -804,9 +823,13 @@
             rhs.generate_disposal_code(code)
         else:
             if self.type.is_pyobject:
+                #print "NameNode.generate_assignment_code: to", self.name ###
+                #print "...from", rhs ###
+                #print "...LHS type", self.type, "ctype", self.ctype() ###
+                #print "...RHS type", rhs.type, "ctype", rhs.ctype() ###
                 rhs.make_owned_reference(code)
-                code.put_decref(self.result, self.type)
-            code.putln('%s = %s;' % (self.entry.cname, rhs.result))
+                code.put_decref(self.result_code, self.ctype())
+            code.putln('%s = %s;' % (self.result_code, rhs.result_as(self.ctype())))
             if debug_disposal_code:
                 print "NameNode.generate_assignment_code:"
                 print "...generating post-assignment code for", rhs
@@ -835,15 +858,15 @@
     def analyse_types(self, env):
         self.arg.analyse_types(env)
         self.arg = self.arg.coerce_to_pyobject(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PyObject_Repr(%s); if (!%s) %s" % (
-                self.result,
-                self.arg.result,
-                self.result,
+                self.result_code,
+                self.arg.py_result(),
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -862,21 +885,21 @@
         self.module_name = self.module_name.coerce_to_pyobject(env)
         if self.name_list:
             self.name_list.analyse_types(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
         env.use_utility_code(import_utility_code)
     
     def generate_result_code(self, code):
         if self.name_list:
-            name_list_code = self.name_list.result
+            name_list_code = self.name_list.py_result()
         else:
             name_list_code = "0"
         code.putln(
             "%s = __Pyx_Import(%s, %s); if (!%s) %s" % (
-                self.result,
-                self.module_name.result,
+                self.result_code,
+                self.module_name.py_result(),
                 name_list_code,
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -891,15 +914,15 @@
     def analyse_types(self, env):
         self.sequence.analyse_types(env)
         self.sequence = self.sequence.coerce_to_pyobject(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PyObject_GetIter(%s); if (!%s) %s" % (
-                self.result,
-                self.sequence.result,
-                self.result,
+                self.result_code,
+                self.sequence.py_result(),
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -914,17 +937,17 @@
     def __init__(self, iterator, env):
         self.pos = iterator.pos
         self.iterator = iterator
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PyIter_Next(%s);" % (
-                self.result,
-                self.iterator.result))
+                self.result_code,
+                self.iterator.py_result()))
         code.putln(
             "if (!%s) {" %
-                self.result)
+                self.result_code)
         code.putln(
                 "if (PyErr_Occurred()) %s" %
                     code.error_goto(self.pos))
@@ -941,15 +964,15 @@
     
     def __init__(self, pos, env):
         ExprNode.__init__(self, pos)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
         env.use_utility_code(get_exception_utility_code)
     
     def generate_result_code(self, code):
         code.putln(
             "%s = __Pyx_GetExcValue(); if (!%s) %s" % (
-                self.result,
-                self.result,
+                self.result_code,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -960,6 +983,8 @@
     def __init__(self, pos, type, env):
         ExprNode.__init__(self, pos)
         self.type = type
+        if type.is_pyobject:
+            self.result_ctype = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
@@ -987,6 +1012,9 @@
     
     subexprs = ['base', 'index']
     
+    def is_ephemeral(self):
+        return self.base.is_ephemeral()
+    
     def analyse_target_declaration(self, env):
         pass
     
@@ -995,7 +1023,7 @@
         self.index.analyse_types(env)
         if self.base.type.is_pyobject: 
             self.index = self.index.coerce_to_pyobject(env)
-            self.type = PyrexTypes.py_object_type
+            self.type = py_object_type
             self.is_temp = 1
         else:
             if self.base.type.is_ptr or self.base.type.is_array:
@@ -1007,7 +1035,7 @@
                 self.type = PyrexTypes.error_type
             if self.index.type.is_pyobject:
                 self.index = self.index.coerce_to(
-                    PyrexTypes.c_int_type, env)
+                    PyrexTypes.c_py_ssize_t_type, env)
             if not self.index.type.is_int:
                 error(self.pos,
                     "Invalid index type '%s'" %
@@ -1020,18 +1048,18 @@
     def is_lvalue(self):
         return 1
     
-    def result_code(self):
+    def calculate_result_code(self):
         return "(%s[%s])" % (
-            self.base.result, self.index.result)
+            self.base.result_code, self.index.result_code)
     
     def generate_result_code(self, code):
         if self.type.is_pyobject:
             code.putln(
                 "%s = PyObject_GetItem(%s, %s); if (!%s) %s" % (
-                    self.result,
-                    self.base.result,
-                    self.index.result,
-                    self.result,
+                    self.result_code,
+                    self.base.py_result(),
+                    self.index.py_result(),
+                    self.result_code,
                     code.error_goto(self.pos)))
     
     def generate_assignment_code(self, rhs, code):
@@ -1039,23 +1067,23 @@
         if self.type.is_pyobject:
             code.putln(
                 "if (PyObject_SetItem(%s, %s, %s) < 0) %s" % (
-                    self.base.result,
-                    self.index.result,
-                    rhs.result,
+                    self.base.py_result(),
+                    self.index.py_result(),
+                    rhs.py_result(),
                     code.error_goto(self.pos)))
             self.generate_subexpr_disposal_code(code)
         else:
             code.putln(
                 "%s = %s;" % (
-                    self.result, rhs.result))
+                    self.result_code, rhs.result_code))
         rhs.generate_disposal_code(code)
     
     def generate_deletion_code(self, code):
         self.generate_subexpr_evaluation_code(code)
         code.putln(
             "if (PyObject_DelItem(%s, %s) < 0) %s" % (
-                self.base.result,
-                self.index.result,
+                self.base.py_result(),
+                self.index.py_result(),
                 code.error_goto(self.pos)))
         self.generate_subexpr_disposal_code(code)
 
@@ -1079,32 +1107,32 @@
         if self.stop:
             self.stop.analyse_types(env)
         self.base = self.base.coerce_to_pyobject(env)
-        c_int = PyrexTypes.c_int_type
+        c_int = PyrexTypes.c_py_ssize_t_type
         if self.start:
             self.start = self.start.coerce_to(c_int, env)
         if self.stop:
             self.stop = self.stop.coerce_to(c_int, env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PySequence_GetSlice(%s, %s, %s); if (!%s) %s" % (
-                self.result,
-                self.base.result,
+                self.result_code,
+                self.base.py_result(),
                 self.start_code(),
                 self.stop_code(),
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
     
     def generate_assignment_code(self, rhs, code):
         self.generate_subexpr_evaluation_code(code)
         code.putln(
             "if (PySequence_SetSlice(%s, %s, %s, %s) < 0) %s" % (
-                self.base.result,
+                self.base.py_result(),
                 self.start_code(),
                 self.stop_code(),
-                rhs.result,
+                rhs.result_code,
                 code.error_goto(self.pos)))
         self.generate_subexpr_disposal_code(code)
         rhs.generate_disposal_code(code)
@@ -1113,7 +1141,7 @@
         self.generate_subexpr_evaluation_code(code)
         code.putln(
             "if (PySequence_DelSlice(%s, %s, %s) < 0) %s" % (
-                self.base.result,
+                self.base.py_result(),
                 self.start_code(),
                 self.stop_code(),
                 code.error_goto(self.pos)))
@@ -1121,18 +1149,18 @@
     
     def start_code(self):
         if self.start:
-            return self.start.result
+            return self.start.result_code
         else:
             return "0"
     
     def stop_code(self):
         if self.stop:
-            return self.stop.result
+            return self.stop.result_code
         else:
             return "0x7fffffff"
     
-    def result_code(self):
-        # self.result is not used, but this method must exist
+    def calculate_result_code(self):
+        # self.result_code is not used, but this method must exist
         return "<unused>"
     
 
@@ -1152,17 +1180,17 @@
         self.start = self.start.coerce_to_pyobject(env)
         self.stop = self.stop.coerce_to_pyobject(env)
         self.step = self.step.coerce_to_pyobject(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PySlice_New(%s, %s, %s); if (!%s) %s" % (
-                self.result,
-                self.start.result, 
-                self.stop.result, 
-                self.step.result,
-                self.result,
+                self.result_code,
+                self.start.py_result(), 
+                self.stop.py_result(), 
+                self.step.py_result(),
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -1190,37 +1218,42 @@
             # was obtained, because we need to pass it as 'self'.
             self.self = function.obj
             function.obj = CloneNode(self.self)
-        if self.function.type.is_pyobject:
+        func_type = self.function_type()
+        if func_type.is_pyobject:
             self.arg_tuple = TupleNode(self.pos, args = self.args)
             self.args = None
             self.arg_tuple.analyse_types(env)
-            self.type = PyrexTypes.py_object_type
+            self.type = py_object_type
             self.is_temp = 1
         else:
             for arg in self.args:
                 arg.analyse_types(env)
-            if self.self and self.function.type.args:
+            if self.self and func_type.args:
                 # Coerce 'self' to the type expected by the method.
-                expected_type = self.function.type.args[0].type
+                expected_type = func_type.args[0].type
                 self.coerced_self = CloneNode(self.self).coerce_to(
                     expected_type, env)
                 # Insert coerced 'self' argument into argument list.
                 self.args.insert(0, self.coerced_self)
             self.analyse_c_function_call(env)
     
-    def analyse_c_function_call(self, env):
+    def function_type(self):
+        # Return the type of the function being called, coercing a function
+        # pointer to a function if necessary.
         func_type = self.function.type
-        # Coerce function pointer to function
         if func_type.is_ptr:
             func_type = func_type.base_type
-            self.function.type = func_type
+        return func_type
+    
+    def analyse_c_function_call(self, env):
+        func_type = self.function_type()
         # Check function type
         if not func_type.is_cfunction:
             if not func_type.is_error:
                 error(self.pos, "Calling non-function type '%s'" %
                     func_type)
             self.type = PyrexTypes.error_type
-            self.result = "<error>"
+            self.result_code = "<error>"
             return
         # Check no. of args
         expected_nargs = len(func_type.args)
@@ -1235,7 +1268,7 @@
                         % (expected_str, actual_nargs))
                 self.args = None
                 self.type = PyrexTypes.error_type
-                self.result = "<error>"
+                self.result_code = "<error>"
                 return
         # Coerce arguments
         for i in range(expected_nargs):
@@ -1251,67 +1284,64 @@
             or func_type.exception_value is not None \
             or func_type.exception_check:
                 self.is_temp = 1
+                if self.type.is_pyobject:
+                    self.result_ctype = py_object_type
     
-    def result_code(self):
-        return self.c_call_code(as_extension_type = 0)
+    def calculate_result_code(self):
+        return self.c_call_code()
     
-    def result_as_extension_type(self):
-        return self.c_call_code(as_extension_type = 1)
-    
-    def c_call_code(self, as_extension_type):
-        if self.args is None or not self.function.type.is_cfunction:
+    def c_call_code(self):
+        func_type = self.function_type()
+        if self.args is None or not func_type.is_cfunction:
             return "<error>"
-        formal_args = self.function.type.args
+        formal_args = func_type.args
         arg_list_code = []
         for (formal_arg, actual_arg) in \
             zip(formal_args, self.args):
-                if formal_arg.type.is_extension_type:
-                    arg_code = actual_arg.result_as_extension_type()
-                    if not formal_arg.type.same_as(actual_arg.type):
-                        arg_code = "((%s)%s)" % (
-                            formal_arg.type.declaration_code(""),
-                            arg_code)
-                else:
-                    arg_code = actual_arg.result
+                arg_code = actual_arg.result_as(formal_arg.type)
                 arg_list_code.append(arg_code)
         for actual_arg in self.args[len(formal_args):]:
-            arg_list_code.append(actual_arg.result)
-        result = "%s(%s)" % (self.function.result,
+            arg_list_code.append(actual_arg.result_code)
+        result = "%s(%s)" % (self.function.result_code,
             join(arg_list_code, ","))
-        if self.type.is_extension_type and not as_extension_type:
-            result = "((PyObject *)%s)" % result
         return result
     
     def generate_result_code(self, code):
-        #print_call_chain("SimpleCallNode.generate_result_code") ###
-        if self.function.type.is_pyobject:
+        func_type = self.function_type()
+        if func_type.is_pyobject:
             code.putln(
                 "%s = PyObject_Call(%s, %s, 0); if (!%s) %s" % (
-                    self.result,
-                    self.function.result,
-                    self.arg_tuple.result,
-                    self.result,
+                    self.result_code,
+                    self.function.py_result(),
+                    self.arg_tuple.py_result(),
+                    self.result_code,
                     code.error_goto(self.pos)))
-        elif self.function.type.is_cfunction:
+        elif func_type.is_cfunction:
             exc_checks = []
             if self.type.is_pyobject:
-                exc_checks.append("!%s" % self.result)
+                exc_checks.append("!%s" % self.result_code)
             else:
-                exc_val = self.function.type.exception_value
-                exc_check = self.function.type.exception_check
+                exc_val = func_type.exception_value
+                exc_check = func_type.exception_check
                 if exc_val is not None:
-                    exc_checks.append("%s == %s" % (self.result, exc_val))
+                    exc_checks.append("%s == %s" % (self.result_code, exc_val))
                 if exc_check:
                     exc_checks.append("PyErr_Occurred()")
             if self.is_temp or exc_checks:
-                if self.result:
-                    lhs = "%s = " % self.result
+                rhs = self.c_call_code()
+                if self.result_code:
+                    lhs = "%s = " % self.result_code
+                    if self.is_temp and self.type.is_pyobject:
+                        #return_type = self.type # func_type.return_type
+                        #print "SimpleCallNode.generate_result_code: casting", rhs, \
+                        #	"from", return_type, "to pyobject" ###
+                        rhs = typecast(py_object_type, self.type, rhs)
                 else:
                     lhs = ""
                 code.putln(
                     "%s%s; if (%s) %s" % (
                         lhs,
-                        self.c_call_code(as_extension_type = 0),
+                        rhs,
                         " && ".join(exc_checks),
                         code.error_goto(self.pos)))
     
@@ -1340,37 +1370,37 @@
         if self.starstar_arg:
             self.starstar_arg = \
                 self.starstar_arg.coerce_to_pyobject(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
         
     def generate_result_code(self, code):
         if self.keyword_args and self.starstar_arg:
             code.putln(
                 "if (PyDict_Update(%s, %s) < 0) %s" % (
-                    self.keyword_args.result, 
-                    self.starstar_arg.result,
+                    self.keyword_args.py_result(), 
+                    self.starstar_arg.py_result(),
                     code.error_goto(self.pos)))
-            keyword_code = self.keyword_args.result
+            keyword_code = self.keyword_args.py_result()
         elif self.keyword_args:
-            keyword_code = self.keyword_args.result
+            keyword_code = self.keyword_args.py_result()
         elif self.starstar_arg:
-            keyword_code = self.starstar_arg.result
+            keyword_code = self.starstar_arg.py_result()
         else:
             keyword_code = None
         if not keyword_code:
             call_code = "PyObject_CallObject(%s, %s)" % (
-                self.function.result,
-                self.positional_args.result)
+                self.function.py_result(),
+                self.positional_args.py_result())
         else:
             call_code = "PyEval_CallObjectWithKeywords(%s, %s, %s)" % (
-                self.function.result,
-                self.positional_args.result,
+                self.function.py_result(),
+                self.positional_args.py_result(),
                 keyword_code)
         code.putln(
             "%s = %s; if (!%s) %s" % (
-                self.result,
+                self.result_code,
                 call_code,
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -1385,15 +1415,15 @@
     def analyse_types(self, env):
         self.arg.analyse_types(env)
         self.arg = self.arg.coerce_to_pyobject(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PySequence_Tuple(%s); if (!%s) %s" % (
-                self.result,
-                self.arg.result,
-                self.result,
+                self.result_code,
+                self.arg.py_result(),
+                self.result_code,
                 code.error_goto(self.pos)))
     
 
@@ -1506,11 +1536,12 @@
         if self.entry and self.entry.is_cmethod and not self.is_called:
             error(self.pos, "C method can only be called")
         # Reference to C array turns into pointer to first element.
-        if self.type.is_array:
+        while self.type.is_array:
             self.type = self.type.element_ptr_type()
         if self.is_py_attr:
             if not target:
                 self.is_temp = 1
+                self.result_ctype = py_object_type
     
     def analyse_attribute(self, env):
         # Look up attribute and set self.type and self.member.
@@ -1551,7 +1582,7 @@
         # declared or is declared as a Python method. Treat it as a Python
         # attribute reference.
         if obj_type.is_pyobject:
-            self.type = PyrexTypes.py_object_type
+            self.type = py_object_type
             self.is_py_attr = 1
             if Options.intern_names:
                 self.interned_attr_cname = env.intern(self.attribute)
@@ -1575,34 +1606,21 @@
     
     def is_ephemeral(self):
         if self.obj:
-            return ExprNode.is_ephemeral(self)
+            return self.obj.is_ephemeral()
         else:
             return NameNode.is_ephemeral(self)
     
-    def result_code(self):
-        return self.select_code()[0]
-    
-    def result_as_extension_type(self):
-        return self.uncast_select_code()
-    
-    def select_code(self):
-        orig_code = self.uncast_select_code()
-        if self.type.is_extension_type:
-            code = "((PyObject *)%s)" % orig_code
-        else:
-            code = orig_code
-        return code, orig_code
-    
-    def uncast_select_code(self):
-        obj_type = self.obj.type
-        if obj_type.is_extension_type:
-            obj_code = self.obj.result_as_extension_type()
-        else:
-            obj_code = self.obj.result
+    def calculate_result_code(self):
+        #print "AttributeNode.calculate_result_code:", self.member ###
+        #print "...obj node =", self.obj, "code", self.obj.result_code ###
+        #print "...obj type", self.obj.type, "ctype", self.obj.ctype() ###
+        obj = self.obj
+        obj_code = obj.result_as(obj.type)
+        #print "...obj_code =", obj_code ###
         if self.entry and self.entry.is_cmethod:
             return "((struct %s *)%s%s%s)->%s" % (
-                obj_type.vtabstruct_cname, obj_code, self.op, 
-                obj_type.vtabslot_cname, self.member)
+                obj.type.vtabstruct_cname, obj_code, self.op, 
+                obj.type.vtabslot_cname, self.member)
         else:
             return "%s%s%s" % (obj_code, self.op, self.member)
     
@@ -1611,18 +1629,18 @@
             if Options.intern_names:
                 code.putln(
                     '%s = PyObject_GetAttr(%s, %s); if (!%s) %s' % (
-                        self.result,
-                        self.obj.result,
+                        self.result_code,
+                        self.obj.py_result(),
                         self.interned_attr_cname,
-                        self.result,
+                        self.result_code,
                         code.error_goto(self.pos)))
             else:
                 code.putln(
                     '%s = PyObject_GetAttrString(%s, "%s"); if (!%s) %s' % (
-                        self.result,
-                        self.obj.result,
+                        self.result_code,
+                        self.objpy_result(),
                         self.attribute,
-                        self.result,
+                        self.result_code,
                         code.error_goto(self.pos)))
     
     def generate_assignment_code(self, rhs, code):
@@ -1631,38 +1649,46 @@
             if Options.intern_names:
                 code.putln(
                     'if (PyObject_SetAttr(%s, %s, %s) < 0) %s' % (
-                        self.obj.result,
+                        self.obj.py_result(),
                         self.interned_attr_cname,
-                        rhs.result,
+                        rhs.py_result(),
                         code.error_goto(self.pos)))
             else:
                 code.putln(
                     'if (PyObject_SetAttrString(%s, "%s", %s) < 0) %s' % (
-                        self.obj.result,
+                        self.obj.py_result(),
                         self.attribute,
-                        rhs.result,
+                        rhs.py_result(),
                         code.error_goto(self.pos)))
             rhs.generate_disposal_code(code)
         else:
-            select_code, orig_code = self.select_code()
+            #select_code = self.select_code()
+            select_code = self.result_code
             if self.type.is_pyobject:
                 rhs.make_owned_reference(code)
-                code.put_decref(select_code, self.type)
+                code.put_decref(select_code, self.ctype())
             code.putln(
                 "%s = %s;" % (
-                    orig_code,
-                    rhs.result))
+                    select_code,
+                    rhs.result_code))
             rhs.generate_post_assignment_code(code)
         self.obj.generate_disposal_code(code)
     
     def generate_deletion_code(self, code):
         self.obj.generate_evaluation_code(code)
         if self.is_py_attr:
-            code.putln(
-                'if (PyObject_DelAttrString(%s, "%s") < 0) %s' % (
-                    self.obj.result,
-                    self.attribute,
-                    code.error_goto(self.pos)))
+            if Options.intern_names:
+                code.putln(
+                    'if (PyObject_DelAttr(%s, %s) < 0) %s' % (
+                        self.obj.py_result(),
+                        self.interned_attr_cname,
+                        code.error_goto(self.pos)))
+            else:
+                code.putln(
+                    'if (PyObject_DelAttrString(%s, "%s") < 0) %s' % (
+                        self.obj.py_result(),
+                        self.attribute,
+                        code.error_goto(self.pos)))
         else:
             error(self.pos, "Cannot delete C attribute of extension type")
         self.obj.generate_disposal_code(code)
@@ -1695,7 +1721,7 @@
             arg = self.args[i]
             arg.analyse_types(env)
             self.args[i] = arg.coerce_to_pyobject(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def analyse_target_types(self, env):
@@ -1708,7 +1734,7 @@
             coerced_unpacked_item = unpacked_item.coerce_to(arg.type, env)
             self.unpacked_items.append(unpacked_item)
             self.coerced_unpacked_items.append(coerced_unpacked_item)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         env.use_utility_code(unpacking_utility_code)
     
     def allocate_target_temps(self, env):
@@ -1728,20 +1754,22 @@
     
     def generate_assignment_code(self, rhs, code):
         for i in range(len(self.args)):
-            unpack_result = self.unpacked_items[i].result
+            item = self.unpacked_items[i]
+            unpack_code = "__Pyx_UnpackItem(%s, %s)" % (
+                rhs.py_result(),
+                i)
             code.putln(
-                "%s = __Pyx_UnpackItem(%s, %s); if (!%s) %s" % (
-                    unpack_result,
-                    rhs.result,
-                    i,
-                    unpack_result,
+                "%s = %s; if (!%s) %s" % (
+                    item.result_code,
+                    typecast(item.ctype(), py_object_type, unpack_code),
+                    item.result_code,
                     code.error_goto(self.pos)))
             value_node = self.coerced_unpacked_items[i]
             value_node.generate_evaluation_code(code)
             self.args[i].generate_assignment_code(value_node, code)
         code.putln(
             "if (__Pyx_EndUnpack(%s, %s) < 0) %s" % (
-                rhs.result,
+                rhs.py_result(),
                 len(self.args),
                 code.error_goto(self.pos)))
         if debug_disposal_code:
@@ -1756,20 +1784,19 @@
     def generate_operation_code(self, code):
         code.putln(
             "%s = PyTuple_New(%s); if (!%s) %s" % (
-                self.result,
+                self.result_code,
                 len(self.args),
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
         for i in range(len(self.args)):
             arg = self.args[i]
-            result = arg.result
             if not arg.result_in_temp():
-                code.put_incref(result, arg.type)
+                code.put_incref(arg.result_code, arg.ctype())
             code.putln(
                 "PyTuple_SET_ITEM(%s, %s, %s);" % (
-                    self.result,
+                    self.result_code,
                     i,
-                    result))
+                    arg.py_result()))
     
     def generate_subexpr_disposal_code(self, code):
         # We call generate_post_assignment_code here instead
@@ -1784,19 +1811,19 @@
     
     def generate_operation_code(self, code):
         code.putln("%s = PyList_New(%s); if (!%s) %s" %
-            (self.result,
+            (self.result_code,
             len(self.args),
-            self.result,
+            self.result_code,
             code.error_goto(self.pos)))
         for i in range(len(self.args)):
             arg = self.args[i]
             #if not arg.is_temp:
             if not arg.result_in_temp():
-                code.put_incref(arg.result, arg.type)
+                code.put_incref(arg.result_code, arg.ctype())
             code.putln("PyList_SET_ITEM(%s, %s, %s);" %
-                (self.result,
+                (self.result_code,
                 i,
-                arg.result))
+                arg.py_result()))
                 
     def generate_subexpr_disposal_code(self, code):
         # We call generate_post_assignment_code here instead
@@ -1820,13 +1847,13 @@
             value = value.coerce_to_pyobject(env)
             new_pairs.append((key, value))
         self.key_value_pairs = new_pairs
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
-    def allocate_temps(self, env):
+    def allocate_temps(self, env, result = None):
         #  Custom method used here because key-value
         #  pairs are evaluated and used one at a time.
-        self.allocate_temp(env)
+        self.allocate_temp(env, result)
         for key, value in self.key_value_pairs:
             key.allocate_temps(env)
             value.allocate_temps(env)
@@ -1838,17 +1865,17 @@
         #  pairs are evaluated and used one at a time.
         code.putln(
             "%s = PyDict_New(); if (!%s) %s" % (
-                self.result,
-                self.result,
+                self.result_code,
+                self.result_code,
                 code.error_goto(self.pos)))
         for key, value in self.key_value_pairs:
             key.generate_evaluation_code(code)
             value.generate_evaluation_code(code)
             code.putln(
                 "if (PyDict_SetItem(%s, %s, %s) < 0) %s" % (
-                    self.result,
-                    key.result,
-                    value.result,
+                    self.result_code,
+                    key.py_result(),
+                    value.py_result(),
                     code.error_goto(self.pos)))
             key.generate_disposal_code(code)
             value.generate_disposal_code(code)
@@ -1874,8 +1901,8 @@
         if self.doc:
             self.doc.analyse_types(env)
             self.doc = self.doc.coerce_to_pyobject(env)
-        self.module_name = env.global_scope().module_name
-        self.type = PyrexTypes.py_object_type
+        self.module_name = env.global_scope().qualified_name
+        self.type = py_object_type
         self.is_temp = 1
         env.use_utility_code(create_class_utility_code);
     
@@ -1883,25 +1910,17 @@
         if self.doc:
             code.putln(
                 'if (PyDict_SetItemString(%s, "__doc__", %s) < 0) %s' % (
-                    self.dict.result,
-                    self.doc.result,
+                    self.dict.py_result(),
+                    self.doc.py_result(),
                     code.error_goto(self.pos)))
-##		code.putln(
-##			'%s = PyClass_New(%s, %s, %s); if (!%s) %s' % (
-##				self.result,
-##				self.bases.result,
-##				self.dict.result,
-##				self.name.result,
-##				self.result,
-##				code.error_goto(self.pos)))
         code.putln(
             '%s = __Pyx_CreateClass(%s, %s, %s, "%s"); if (!%s) %s' % (
-                self.result,
-                self.bases.result,
-                self.dict.result,
-                self.name.result,
+                self.result_code,
+                self.bases.py_result(),
+                self.dict.py_result(),
+                self.name.py_result(),
                 self.module_name,
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -1917,16 +1936,16 @@
     
     def analyse_types(self, env):
         self.function.analyse_types(env)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PyMethod_New(%s, 0, %s); if (!%s) %s" % (
-                self.result,
-                self.function.result,
+                self.result_code,
+                self.function.py_result(),
                 self.class_cname,
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 
@@ -1938,15 +1957,15 @@
     #  pymethdef_cname   string   PyMethodDef structure
     
     def analyse_types(self, env):
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
         self.is_temp = 1
     
     def generate_result_code(self, code):
         code.putln(
             "%s = PyCFunction_New(&%s, 0); if (!%s) %s" % (
-                self.result,
+                self.result_code,
                 self.pymethdef_cname,
-                self.result,
+                self.result_code,
                 code.error_goto(self.pos)))
 
 #-------------------------------------------------------------------
@@ -1973,7 +1992,7 @@
         self.operand.analyse_types(env)
         if self.is_py_operation():
             self.coerce_operand_to_pyobject(env)
-            self.type = PyrexTypes.py_object_type
+            self.type = py_object_type
             self.is_temp = 1
         else:
             self.analyse_c_operation(env)
@@ -1998,10 +2017,10 @@
         function = self.py_operation_function()
         code.putln(
             "%s = %s(%s); if (!%s) %s" % (
-                self.result, 
+                self.result_code, 
                 function, 
-                self.operand.result,
-                self.result,
+                self.operand.py_result(),
+                self.result_code,
                 code.error_goto(self.pos)))
         
     def type_error(self):
@@ -2023,8 +2042,8 @@
         self.operand = self.operand.coerce_to_boolean(env)
         self.type = PyrexTypes.c_int_type
     
-    def result_code(self):
-        return "(!%s)" % self.operand.result
+    def calculate_result_code(self):
+        return "(!%s)" % self.operand.result_code
     
     def generate_result_code(self, code):
         pass
@@ -2041,8 +2060,8 @@
     def py_operation_function(self):
         return "PyNumber_Positive"
     
-    def result_code(self):
-        return self.operand.result
+    def calculate_result_code(self):
+        return self.operand.result_code
 
 
 class UnaryMinusNode(UnopNode):
@@ -2059,8 +2078,8 @@
     def py_operation_function(self):
         return "PyNumber_Negative"
     
-    def result_code(self):
-        return "(-%s)" % self.operand.result
+    def calculate_result_code(self):
+        return "(-%s)" % self.operand.result_code
 
 
 class TildeNode(UnopNode):
@@ -2075,8 +2094,8 @@
     def py_operation_function(self):
         return "PyNumber_Invert"
     
-    def result_code(self):
-        return "(~%s)" % self.operand.result
+    def calculate_result_code(self):
+        return "(~%s)" % self.operand.result_code
 
 
 class AmpersandNode(ExprNode):
@@ -2103,10 +2122,10 @@
     def error(self, mess):
         error(self.pos, mess)
         self.type = PyrexTypes.error_type
-        self.result = "<error>"
+        self.result_code = "<error>"
     
-    def result_code(self):
-        return "(&%s)" % self.operand.result
+    def calculate_result_code(self):
+        return "(&%s)" % self.operand.result_code
 
     def generate_result_code(self, code):
         pass
@@ -2143,33 +2162,32 @@
         from_py = self.operand.type.is_pyobject
         if from_py and not to_py and self.operand.is_ephemeral():
             error(self.pos, "Casting temporary Python object to non-Python type")
-        #if to_py:
         if to_py and not from_py:
+            self.result_ctype = py_object_type
             self.is_temp = 1			
     
     def check_const(self):
         self.operand.check_const()
     
-    def result_code(self):
-        if self.type.is_pyobject:
-            cast = "PyObject *"
+    def calculate_result_code(self):
+        opnd = self.operand
+        result_code = self.type.cast_code(opnd.result_code)
+        return result_code
+    
+    def result_as(self, type):
+        if self.type.is_pyobject and not self.is_temp:
+            #  Optimise away some unnecessary casting
+            return self.operand.result_as(type)
         else:
-            cast = self.type.declaration_code("")
-        return "((%s)%s)" % (cast, self.operand.result)
-    
-    def result_as_extension_type(self):
-        return "((%s)%s)" % (
-            self.type.declaration_code(""),
-            self.operand.result)
+            return ExprNode.result_as(self, type)
 
     def generate_result_code(self, code):
         if self.is_temp:
             code.putln(
                 "%s = (PyObject *)%s;" % (
-                    self.result,
-                    #self.type.declaration_code(""),
-                    self.operand.result))
-            code.put_incref(self.result, self.type)
+                    self.result_code,
+                    self.operand.result_code))
+            code.put_incref(self.result_code, self.ctype())
 
 
 class SizeofNode(ExprNode):
@@ -2199,10 +2217,10 @@
         elif arg_type.is_void:
             error(self.pos, "Cannot take sizeof void")
         elif not arg_type.is_complete():
-            error(self.pos, "Cannot take sizeof incomplete type '%s'" % arg_code)
+            error(self.pos, "Cannot take sizeof incomplete type '%s'" % arg_type)
         self.type = PyrexTypes.c_int_type
         
-    def result_code(self):
+    def calculate_result_code(self):
         arg_code = self.arg_type.declaration_code("")
         return "(sizeof(%s))" % arg_code
     
@@ -2218,8 +2236,8 @@
         self.operand.analyse_types(env)
         self.type = PyrexTypes.c_int_type
     
-    def result_code(self):
-        return "(sizeof(%s))" % self.operand.result
+    def calculate_result_code(self):
+        return "(sizeof(%s))" % self.operand.result_code
     
     def generate_result_code(self, code):
         pass
@@ -2251,7 +2269,7 @@
         self.operand2.analyse_types(env)
         if self.is_py_operation():
             self.coerce_operands_to_pyobjects(env)
-            self.type = PyrexTypes.py_object_type
+            self.type = py_object_type
             self.is_temp = 1
         else:
             self.analyse_c_operation(env)
@@ -2269,6 +2287,7 @@
         self.operand2.check_const()
     
     def generate_result_code(self, code):
+        #print "BinopNode.generate_result_code:", self.operand1, self.operand2 ###
         if self.operand1.type.is_pyobject:
             function = self.py_operation_function()
             if function == "PyNumber_Power":
@@ -2277,12 +2296,12 @@
                 extra_args = ""
             code.putln(
                 "%s = %s(%s, %s%s); if (!%s) %s" % (
-                    self.result, 
+                    self.result_code, 
                     function, 
-                    self.operand1.result,
-                    self.operand2.result,
+                    self.operand1.py_result(),
+                    self.operand2.py_result(),
                     extra_args,
-                    self.result,
+                    self.result_code,
                     code.error_goto(self.pos)))
         else:
             if self.is_temp:
@@ -2316,11 +2335,11 @@
     def c_types_okay(self, type1, type2):
         return type1.is_numeric and type2.is_numeric
 
-    def result_code(self):
+    def calculate_result_code(self):
         return "(%s %s %s)" % (
-            self.operand1.result, 
+            self.operand1.result_code, 
             self.operator, 
-            self.operand2.result)
+            self.operand2.result_code)
     
     def py_operation_function(self):
         return self.py_functions[self.operator]
@@ -2415,9 +2434,9 @@
         else:
             return None
     
-    def result_code(self):
+    def calculate_result_code(self):
         return "pow(%s, %s)" % (
-            self.operand1.result, self.operand2.result)
+            self.operand1.result_code, self.operand2.result_code)
             
 
 class BoolBinopNode(ExprNode):
@@ -2441,7 +2460,7 @@
             self.operand2 = self.operand2.coerce_to_pyobject(env)
             self.temp_bool = TempNode(self.pos,
                 PyrexTypes.c_int_type, env)
-            self.type = PyrexTypes.py_object_type
+            self.type = py_object_type
         else:
             self.operand1 = self.operand1.coerce_to_boolean(env)
             self.operand2 = self.operand2.coerce_to_boolean(env)
@@ -2455,7 +2474,7 @@
         #self.operand2 = self.operand2.coerce_to_simple(env)
         self.is_temp = 1
     
-    def allocate_temps(self, env, result = None):
+    def allocate_temps(self, env, result_code = None):
         #  We don't need both operands at the same time, and
         #  one of the operands will also be our result. So we
         #  use an allocation strategy here which results in
@@ -2463,12 +2482,12 @@
         #  result variable. This allows us to avoid some 
         #  assignments and increfs/decrefs that would otherwise
         #  be necessary.
-        self.allocate_temp(env, result)
-        self.operand1.allocate_temps(env, self.result)
+        self.allocate_temp(env, result_code)
+        self.operand1.allocate_temps(env, self.result_code)
         if self.temp_bool:
             self.temp_bool.allocate_temp(env)
             self.temp_bool.release_temp(env)
-        self.operand2.allocate_temps(env, self.result)
+        self.operand2.allocate_temps(env, self.result_code)
         #  We haven't called release_temp on either operand,
         #  because although they are temp nodes, they don't own 
         #  their result variable. And because they are temp
@@ -2481,11 +2500,11 @@
         self.operand1.check_const()
         self.operand2.check_const()
     
-    def result_code(self):
+    def calculate_result_code(self):
         return "(%s %s %s)" % (
-            self.operand1.result,
+            self.operand1.result_code,
             self.py_to_c_op[self.operator],
-            self.operand2.result)
+            self.operand2.result_code)
     
     py_to_c_op = {'and': "&&", 'or': "||"}
 
@@ -2508,15 +2527,15 @@
     def generate_operand1_test(self, code):
         #  Generate code to test the truth of the first operand.
         if self.type.is_pyobject:
-            test_result = self.temp_bool.result
+            test_result = self.temp_bool.result_code
             code.putln(
                 "%s = PyObject_IsTrue(%s); if (%s < 0) %s" % (
                     test_result,
-                    self.operand1.result,
+                    self.operand1.py_result(),
                     test_result,
                     code.error_goto(self.pos)))
         else:
-            test_result = self.operand1.result
+            test_result = self.operand1.result_code
         return test_result
 
 
@@ -2550,37 +2569,37 @@
         else:
             return 0
 
-    def generate_operation_code(self, code, result, 
+    def generate_operation_code(self, code, result_code, 
             operand1, op , operand2):
         if op == 'in' or op == 'not_in':
             code.putln(
                 "%s = PySequence_Contains(%s, %s); if (%s < 0) %s" % (
-                    result, 
-                    operand2.result, 
-                    operand1.result, 
-                    result,
+                    result_code, 
+                    operand2.py_result(), 
+                    operand1.py_result(), 
+                    result_code,
                     code.error_goto(self.pos)))
             if op == 'not_in':
                 code.putln(
                     "%s = !%s;" % (
-                        result, result))
+                        result_code, result_code))
         elif (operand1.type.is_pyobject
             and op not in ('is', 'is_not')):
                 code.putln(
                     "if (PyObject_Cmp(%s, %s, &%s) < 0) %s" % (
-                        operand1.result, 
-                        operand2.result, 
-                        result,
+                        operand1.py_result(), 
+                        operand2.py_result(), 
+                        result_code,
                         code.error_goto(self.pos)))
                 code.putln(
                     "%s = %s %s 0;" % (
-                        result, result, op))
+                        result_code, result_code, op))
         else:
             code.putln("%s = %s %s %s;" % (
-                result, 
-                operand1.result, 
+                result_code, 
+                operand1.result_code, 
                 self.c_operator(op), 
-                operand2.result))
+                operand2.result_code))
     
     def c_operator(self, op):
         if op == 'is':
@@ -2658,21 +2677,21 @@
         if self.cascade:
             self.not_const()
 
-    def result_code(self):
+    def calculate_result_code(self):
         return "(%s %s %s)" % (
-            self.operand1.result,
+            self.operand1.result_code,
             self.c_operator(self.operator),
-            self.operand2.result)
+            self.operand2.result_code)
     
     def generate_evaluation_code(self, code):
         self.operand1.generate_evaluation_code(code)
         self.operand2.generate_evaluation_code(code)
         if self.is_temp:
-            self.generate_operation_code(code, self.result, 
+            self.generate_operation_code(code, self.result_code, 
                 self.operand1, self.operator, self.operand2)
             if self.cascade:
                 self.cascade.generate_evaluation_code(code,
-                    self.result, self.operand2)
+                    self.result_code, self.operand2)
             self.operand1.generate_disposal_code(code)
             self.operand2.generate_disposal_code(code)
     
@@ -2799,13 +2818,11 @@
         CoercionNode.__init__(self, arg)
         self.type = new_type
     
-    def result_code(self):
-        return "((%s)%s)" % (
-            self.type.declaration_code(""),
-            self.arg.result)
-
-    def result_as_extension_type(self):
-        return self.result
+    def calculate_result_code(self):
+        #return "((%s)%s)" % (
+        #	self.type.declaration_code(""),
+        #	self.arg.result)
+        return self.arg.result_as(self.type)
 
     def generate_result_code(self, code):
         self.arg.generate_result_code(code)
@@ -2822,6 +2839,7 @@
         assert dst_type.is_extension_type, "PyTypeTest on non extension type"
         CoercionNode.__init__(self, arg)
         self.type = dst_type
+        self.result_ctype = arg.ctype()
         env.use_utility_code(type_test_utility_code)
     
     def result_in_temp(self):
@@ -2830,19 +2848,14 @@
     def is_ephemeral(self):
         return self.arg.is_ephemeral()
     
-    def result_code(self):
-        return self.arg.result
-    
-    def result_as_extension_type(self):
-        return "((%s)%s)" % (
-            self.type.declaration_code(""),
-            self.arg.result)
+    def calculate_result_code(self):
+        return self.arg.result_code
     
     def generate_result_code(self, code):
         if self.type.typeobj_is_available():
             code.putln(
                 "if (!__Pyx_TypeTest(%s, %s)) %s" % (
-                    self.result,
+                    self.arg.py_result(),
                     self.type.typeptr_cname,
                     code.error_goto(self.pos)))
         else:
@@ -2859,19 +2872,19 @@
 
     def __init__(self, arg, env):
         CoercionNode.__init__(self, arg)
-        self.type = PyrexTypes.py_object_type
+        self.type = py_object_type
+        self.is_temp = 1
         if not arg.type.to_py_function:
             error(arg.pos,
                 "Cannot convert '%s' to Python object" % arg.type)
-        self.is_temp = 1
     
     def generate_result_code(self, code):
         function = self.arg.type.to_py_function
         code.putln('%s = %s(%s); if (!%s) %s' % (
-            self.result, 
+            self.result_code, 
             function, 
-            self.arg.result, 
-            self.result, 
+            self.arg.result_code, 
+            self.result_code, 
             code.error_goto(self.pos)))
 
 
@@ -2882,21 +2895,21 @@
     def __init__(self, result_type, arg, env):
         CoercionNode.__init__(self, arg)
         self.type = result_type
+        self.is_temp = 1
         if not result_type.from_py_function:
             error(arg.pos,
                 "Cannot convert Python object to '%s'" % result_type)
         if self.type.is_string and self.arg.is_ephemeral():
             error(arg.pos,
                 "Obtaining char * from temporary Python value")
-        self.is_temp = 1
     
     def generate_result_code(self, code):
-        opnd = self.arg.result
+        #opnd = self.arg.py_result()
         function = self.type.from_py_function
         code.putln('%s = %s(%s); if (PyErr_Occurred()) %s' % (
-            self.result, 
+            self.result_code, 
             function, 
-            self.arg.result, 
+            self.arg.py_result(), 
             code.error_goto(self.pos)))
 
 
@@ -2915,16 +2928,16 @@
             self.not_const()
         self.arg.check_const()
     
-    def result_code(self):
-        return "(%s != 0)" % self.arg.result
+    def calculate_result_code(self):
+        return "(%s != 0)" % self.arg.result_code
 
     def generate_result_code(self, code):
         if self.arg.type.is_pyobject:
             code.putln(
                 "%s = PyObject_IsTrue(%s); if (%s < 0) %s" % (
-                    self.result, 
-                    self.arg.result, 
-                    self.result,
+                    self.result_code, 
+                    self.arg.py_result(), 
+                    self.result_code,
                     code.error_goto(self.pos)))
 
 
@@ -2937,14 +2950,16 @@
         CoercionNode.__init__(self, arg)
         self.type = self.arg.type
         self.is_temp = 1
+        if self.type.is_pyobject:
+            self.result_ctype = py_object_type
     
     def generate_result_code(self, code):
         #self.arg.generate_evaluation_code(code) # Already done
         # by generic generate_subexpr_evaluation_code!
         code.putln("%s = %s;" % (
-            self.result, self.arg.result))
+            self.result_code, self.arg.result_as(self.ctype())))
         if self.type.is_pyobject:
-            code.put_incref(self.result, self.type)
+            code.put_incref(self.result_code, self.ctype())
 
 
 class CloneNode(CoercionNode):
@@ -2960,12 +2975,13 @@
     def __init__(self, arg):
         CoercionNode.__init__(self, arg)
         self.type = arg.type
+        self.result_ctype = arg.result_ctype
     
-    def result_code(self):
-        return self.arg.result
+    def calculate_result_code(self):
+        return self.arg.result_code
     
-    def result_as_extension_type(self):
-        return self.arg.result_as_extension_type()
+    #def result_as_extension_type(self):
+    #	return self.arg.result_as_extension_type()
     
     def generate_evaluation_code(self, code):
         pass
@@ -3085,24 +3101,24 @@
     PyErr_SetString(PyExc_ValueError, "unpack sequence of wrong size");
 }
 
-static PyObject *__Pyx_UnpackItem(PyObject *seq, int i) {
-  PyObject *item = PySequence_GetItem(seq, i);
-  if (!item) {
+static PyObject *__Pyx_UnpackItem(PyObject *seq, Py_ssize_t i) {
+  PyObject *item;
+  if (!(item = PySequence_GetItem(seq, i))) {
     if (PyErr_ExceptionMatches(PyExc_IndexError))
-      __Pyx_UnpackError();
+    	__Pyx_UnpackError();
   }
   return item;
 }
 
-static int __Pyx_EndUnpack(PyObject *seq, int i) {
-  PyObject *item = PySequence_GetItem(seq, i);
-  if (item) {
+static int __Pyx_EndUnpack(PyObject *seq, Py_ssize_t i) {
+  PyObject *item;
+  if (item = PySequence_GetItem(seq, i)) {
     Py_DECREF(item);
     __Pyx_UnpackError();
     return -1;
   }
   PyErr_Clear();
-  return 0;
+    return 0;
 }
 """
 

Modified: lxml/pyrex/Pyrex/Compiler/Lexicon.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Lexicon.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Lexicon.py	Tue Apr 25 15:43:21 2006
@@ -19,7 +19,6 @@
     hexdigit = Any("0123456789ABCDEFabcdef")
     indentation = Bol + Rep(Any(" \t"))
     
-    #resword = apply(Str, reserved_words)
     decimal = Rep1(digit)
     dot = Str(".")
     exponent = Any("Ee") + Opt(Any("+-")) + decimal
@@ -27,6 +26,7 @@
     
     name = letter + Rep(letter | digit)
     intconst = decimal | (Str("0x") + Rep1(hexdigit))
+    longconst = intconst + Str("L")
     fltconst = (decimal_fract + Opt(exponent)) | (decimal + exponent)
     imagconst = (intconst | fltconst) + Any("jJ")
     
@@ -73,9 +73,9 @@
     lineterm = Eol + Opt(Str("\n"))
     
     return Lexicon([
-        #(resword, TEXT),
         (name, 'IDENT'),
         (intconst, 'INT'),
+        (longconst, 'LONG'),
         (fltconst, 'FLOAT'),
         (imagconst, 'IMAG'),
         (punct | diphthong, TEXT),

Modified: lxml/pyrex/Pyrex/Compiler/Main.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Main.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Main.py	Tue Apr 25 15:43:21 2006
@@ -2,7 +2,7 @@
 #   Pyrex Top Level
 #
 
-import sys
+import os, sys
 if sys.version_info[:2] < (2, 2):
     print >>sys.stderr, "Sorry, Pyrex requires Python 2.2 or later"
     sys.exit(1)
@@ -132,7 +132,7 @@
 
     def parse(self, source_filename, type_names, pxd):
         # Parse the given source file and return a parse tree.
-        f = open(source_filename, "r")
+        f = open(source_filename, "rU")
         s = PyrexScanner(f, source_filename, 
             type_names = type_names, context = self)
         try:
@@ -166,23 +166,39 @@
         if options.output_file:
             result.c_file = os.path.join(cwd, options.output_file)
         else:
-            result.c_file = replace_suffix(source, ".c")
+            if options.cplus:
+                c_suffix = ".cpp"
+            else:
+                c_suffix = ".c"
+            result.c_file = replace_suffix(source, c_suffix)
         module_name = self.extract_module_name(source)
         initial_pos = (source, 1, 0)
         scope = self.find_module(module_name, pos = initial_pos, need_pxd = 0)
+        errors_occurred = False
         try:
             tree = self.parse(source, scope.type_names, pxd = 0)
             tree.process_implementation(scope, result)
         except CompileError:
-            result.c_file = None
+            errors_occurred = True
         Errors.close_listing_file()
         result.num_errors = Errors.num_errors
         if result.num_errors > 0:
+            errors_occurred = True
+        if errors_occurred:
+            try:
+                os.unlink(result.c_file)
+            except EnvironmentError:
+                pass
             result.c_file = None
         if result.c_file and not options.c_only and c_compile:
-            result.object_file = c_compile(result.c_file)
+            result.object_file = c_compile(result.c_file,
+                verbose_flag = options.show_version,
+                cplus = options.cplus)
             if not options.obj_only and c_link:
-                result.extension_file = c_link(result.object_file)
+                result.extension_file = c_link(result.object_file,
+                    extra_objects = options.objects,
+                    verbose_flag = options.show_version,
+                    cplus = options.cplus)
         return result
 
 #------------------------------------------------------------------------
@@ -200,10 +216,18 @@
     errors_to_stderr  boolean   Echo errors to stderr when using .lis
     include_path      [string]  Directories to search for include files
     output_file       string    Name of generated .c file
+    
+    Following options are experimental and only used on MacOSX:
+    
+    c_only            boolean   Stop after generating C file (default)
+    obj_only          boolean   Stop after compiling to .o file
+    objects           [string]  Extra .o files to link with
+    cplus             boolean   Compile as c++ code
     """
     
     def __init__(self, defaults = None, **kw):
         self.include_path = []
+        self.objects = []
         if defaults:
             self.__dict__.update(defaults.__dict__)
         self.__dict__.update(kw)
@@ -230,6 +254,7 @@
         self.object_file = None
         self.extension_file = None
 
+
 def compile(source, options = None, c_compile = 0, c_link = 0):
     """
     compile(source, options = default_options)
@@ -288,6 +313,7 @@
     errors_to_stderr = 1,
     c_only = 1,
     obj_only = 1,
+    cplus = 0,
     output_file = None)
     
 if sys.platform == "mac":
@@ -295,6 +321,8 @@
     default_options.use_listing_file = 1
 elif sys.platform == "darwin":
     from Pyrex.Mac.DarwinSystem import c_compile, c_link, CCompilerError
+elif 'linux' in sys.platform:
+    from Pyrex.PC.LinuxSystem import c_compile, c_link, CCompilerError
 else:
     c_compile = None
     c_link = None

Modified: lxml/pyrex/Pyrex/Compiler/Naming.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Naming.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Naming.py	Tue Apr 25 15:43:21 2006
@@ -39,6 +39,7 @@
 filename_cname   = pyrex_prefix + "filename"
 filetable_cname  = pyrex_prefix + "f"
 filenames_cname  = pyrex_prefix + "filenames"
+fileinit_cname   = pyrex_prefix + "init_filenames"
 intern_tab_cname = pyrex_prefix + "intern_tab"
 kwds_cname       = pyrex_prefix + "kwds"
 lineno_cname     = pyrex_prefix + "lineno"
@@ -49,3 +50,5 @@
 self_cname       = pyrex_prefix + "self"
 stringtab_cname  = pyrex_prefix + "string_tab"
 vtabslot_cname   = pyrex_prefix + "vtab"
+
+extern_c_macro  = pyrex_prefix.upper() + "EXTERN_C"

Modified: lxml/pyrex/Pyrex/Compiler/Nodes.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Nodes.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Nodes.py	Tue Apr 25 15:43:21 2006
@@ -44,7 +44,7 @@
     #         'type' attribute of each ExprNode. Insert coercion nodes into the
     #         tree where needed to convert to and from Python objects. 
     #         Allocate temporary locals for intermediate results. Fill
-    #         in the 'result' attribute of each ExprNode with a C code
+    #         in the 'result_code' attribute of each ExprNode with a C code
     #         fragment.
     #
     #   (3) generate_code
@@ -71,7 +71,6 @@
     def generate_const_definitions(self, env, code):
         if env.const_entries:
             code.putln("")
-            #code.put_var_declarations(env.const_entries, static = 1)
             for entry in env.const_entries:
                 if not entry.is_interned:
                     code.put_var_declaration(entry, static = 1)
@@ -127,38 +126,41 @@
             if entry.visibility == 'public':
                 public_extension_types.append(entry)
         if public_vars_and_funcs or public_extension_types:
-            #import os
-            #outname_base, _ = os.path.splitext(result.c_file)
-            #result.h_file = outname_base + ".h"
-            #result.i_file = outname_base + ".pxi"
             result.h_file = replace_suffix(result.c_file, ".h")
             result.i_file = replace_suffix(result.c_file, ".pxi")
             h_code = Code.CCodeWriter(result.h_file)
             i_code = Code.PyrexCodeWriter(result.i_file)
+            self.generate_extern_c_macro_definition(h_code)
             for entry in public_vars_and_funcs:
-                h_code.putln("extern %s;" % 
+                h_code.putln("%s %s;" % (
+                    Naming.extern_c_macro,
                     entry.type.declaration_code(
-                        entry.cname, dll_linkage = "DL_IMPORT"))
+                        entry.cname, dll_linkage = "DL_IMPORT")))
                 i_code.putln("cdef extern %s" % 
                     entry.type.declaration_code(entry.cname, pyrex = 1))
             for entry in public_extension_types:
                 self.generate_cclass_header_code(entry.type, h_code)
                 self.generate_cclass_include_code(entry.type, i_code)
-            h_code.putln("extern DL_IMPORT(void) init%s(void);" % env.module_name)
-            #result.h_file_generated = 1
-            #result.i_file_generated = 1
+            h_code.putln("PyMODINIT_FUNC init%s(void);" % env.module_name)
     
     def generate_cclass_header_code(self, type, h_code):
-        h_code.putln("extern DL_IMPORT(PyTypeObject) %s;" % type.typeobj_cname)
+        #h_code.putln("extern DL_IMPORT(PyTypeObject) %s;" % type.typeobj_cname)
+        h_code.putln("%s DL_IMPORT(PyTypeObject) %s;" % (
+            Naming.extern_c_macro,
+            type.typeobj_cname))
         self.generate_obj_struct_definition(type, h_code)
     
     def generate_cclass_include_code(self, type, i_code):
         i_code.putln("cdef extern class %s.%s:" % (
             type.module_name, type.name))
         i_code.indent()
-        for entry in type.scope.var_entries:
-            i_code.putln("cdef %s" % 
-                entry.type.declaration_code(entry.cname, pyrex = 1))
+        var_entries = type.scope.var_entries
+        if var_entries:
+            for entry in var_entries:
+                i_code.putln("cdef %s" % 
+                    entry.type.declaration_code(entry.cname, pyrex = 1))
+        else:
+            i_code.putln("pass")
         i_code.dedent()
     
     def generate_c_code(self, env, result):
@@ -168,7 +170,8 @@
         code.init_labels()
         self.generate_module_preamble(env, modules, code)
         for module in modules:
-            self.generate_declarations_for_module(module, code)
+            self.generate_declarations_for_module(module, code,
+                definition = module is env)
         code.putln("")
         code.putln("/* Implementation of %s */" % env.qualified_name)
         self.generate_const_definitions(env, code)
@@ -179,27 +182,38 @@
         self.generate_py_string_table(env, code)
         self.generate_typeobj_definitions(env, code)
         self.generate_method_table(env, code)
+        self.generate_filename_init_prototype(code)
         self.generate_module_init_func(modules[:-1], env, code)
         self.generate_filename_table(code)
         self.generate_utility_functions(env, code)
         result.c_file_generated = 1
     
     def find_referenced_modules(self, env, module_list, modules_seen):
-        for imported_module in env.cimported_modules:
-            if imported_module not in modules_seen:
-                modules_seen[imported_module] = 1
+        if env not in modules_seen:
+            modules_seen[env] = 1
+            for imported_module in env.cimported_modules:
                 self.find_referenced_modules(imported_module, module_list, modules_seen)
-        module_list.append(env)
+            module_list.append(env)
         
     def generate_module_preamble(self, env, cimported_modules, code):
         code.putln('/* Generated by Pyrex %s on %s */' % (
             Version.version, time.asctime()))
         code.putln('')
+        code.putln('#define PY_SSIZE_T_CLEAN')
         for filename in env.python_include_files:
             code.putln('#include "%s"' % filename)
         code.putln("#ifndef PY_LONG_LONG")
         code.putln("  #define PY_LONG_LONG LONG_LONG")
         code.putln("#endif")
+        code.putln("#if PY_VERSION_HEX < 0x02050000")
+        code.putln("  typedef int Py_ssize_t;")
+        code.putln("  #define PY_SSIZE_T_MAX INT_MAX")
+        code.putln("  #define PY_SSIZE_T_MIN INT_MIN")
+        code.putln("  #define PyInt_FromSsize_t(z) PyInt_FromLong(z)")
+        code.putln("  #define PyInt_AsSsize_t(o)   PyInt_AsLong(o)")
+        code.putln("#endif")
+        self.generate_extern_c_macro_definition(code)
+        code.putln("%s double pow(double, double);" % Naming.extern_c_macro)
         self.generate_includes(env, cimported_modules, code)
         #for filename in env.include_files:
         #	code.putln('#include "%s"' % filename)
@@ -214,11 +228,19 @@
         code.putln('static PyObject *%s;' % Naming.builtins_cname)
         code.putln('static int %s;' % Naming.lineno_cname)
         code.putln('static char *%s;' % Naming.filename_cname)
-        code.putln('staticforward char **%s;' % Naming.filetable_cname)
+        code.putln('static char **%s;' % Naming.filetable_cname)
         if env.doc:
             code.putln('')
             code.putln('static char %s[] = "%s";' % (env.doc_cname, env.doc))
     
+    def generate_extern_c_macro_definition(self, code):
+        name = Naming.extern_c_macro
+        code.putln("#ifdef __cplusplus")
+        code.putln('#define %s extern "C"' % name)
+        code.putln("#else")
+        code.putln("#define %s extern" % name)
+        code.putln("#endif")
+
     def generate_includes(self, env, cimported_modules, code):
         includes = env.include_files[:]
         for module in cimported_modules:
@@ -241,15 +263,13 @@
             # Some C compilers don't like an empty array
             code.putln("0")
         code.putln("};")
-        code.putln("statichere char **%s = %s;" % 
-            (Naming.filetable_cname, Naming.filenames_cname))
     
-    def generate_declarations_for_module(self, env, code):
+    def generate_declarations_for_module(self, env, code, definition):
         code.putln("")
         code.putln("/* Declarations from %s */" % env.qualified_name)
         self.generate_type_predeclarations(env, code)
         self.generate_type_definitions(env, code)
-        self.generate_global_declarations(env, code)
+        self.generate_global_declarations(env, code, definition)
         self.generate_cfunction_predeclarations(env, code)
 
     def generate_type_predeclarations(self, env, code):
@@ -329,11 +349,17 @@
         name = entry.type.typeobj_cname
         if name:
             if entry.visibility == 'extern' and not entry.in_cinclude:
-                code.putln("extern DL_IMPORT(PyTypeObject) %s;" % name)
+                code.putln("%s DL_IMPORT(PyTypeObject) %s;" % (
+                    Naming.extern_c_macro,
+                    name))
             elif entry.visibility == 'public':
-                code.putln("DL_EXPORT(PyTypeObject) %s;" % name)
-            else:
-                code.putln("staticforward PyTypeObject %s;" % name)
+                #code.putln("DL_EXPORT(PyTypeObject) %s;" % name)
+                code.putln("%s DL_EXPORT(PyTypeObject) %s;" % (
+                    Naming.extern_c_macro,
+                    name))
+            # ??? Do we really need the rest of this? ???
+            #else:
+            #	code.putln("staticforward PyTypeObject %s;" % name)
     
     def generate_exttype_vtable_struct(self, entry, code):
         # Generate struct declaration for an extension type's vtable.
@@ -393,13 +419,13 @@
                     attr.type.declaration_code(attr.cname))
         code.putln(footer)
 
-    def generate_global_declarations(self, env, code):
+    def generate_global_declarations(self, env, code, definition):
         code.putln("")
         for entry in env.c_class_entries:
             code.putln("static PyTypeObject *%s = 0;" % 
                 entry.type.typeptr_cname)
         code.put_var_declarations(env.var_entries, static = 1, 
-            dll_linkage = "DL_EXPORT")
+            dll_linkage = "DL_EXPORT", definition = definition)
         code.put_var_declarations(env.default_entries, static = 1)
     
     def generate_cfunction_predeclarations(self, env, code):
@@ -412,7 +438,7 @@
                 header = entry.type.declaration_code(entry.cname, 
                     dll_linkage = dll_linkage)
                 if entry.visibility <> 'private':
-                    storage_class = ""
+                    storage_class = "%s " % Naming.extern_c_macro
                 else:
                     storage_class = "static "
                 code.putln("%s%s; /*proto*/" % (
@@ -420,7 +446,7 @@
                     header))
     
     def generate_typeobj_definitions(self, env, code):
-        modname = env.module_name
+        full_module_name = env.qualified_name
         for entry in env.c_class_entries:
             #print "generate_typeobj_definitions:", entry.name
             #print "...visibility =", entry.visibility
@@ -451,7 +477,7 @@
                     self.generate_method_table(scope, code)
                     self.generate_member_table(scope, code)
                     self.generate_getset_table(scope, code)
-                    self.generate_typeobj_definition(modname, entry, code)
+                    self.generate_typeobj_definition(full_module_name, entry, code)
     
     def generate_exttype_vtable(self, scope, code):
         # Generate the definition of an extension type's vtable.
@@ -484,9 +510,9 @@
         self.generate_self_cast(scope, code)
         type = scope.parent_type
         if type.vtabslot_cname:
-            code.putln("p->%s = (struct %s *)%s;" % (
-                type.vtabslot_cname,
+            code.putln("*(struct %s **)&p->%s = %s;" % (
                 type.vtabstruct_cname,
+                type.vtabslot_cname,
                 type.vtabptr_cname))
         for entry in scope.var_entries:
             if entry.type.is_pyobject:
@@ -510,12 +536,7 @@
         code.putln(
             "static void %s(PyObject *o) {"
                 % scope.mangle_internal("tp_dealloc"))
-        # only need the object cast to the type if we need to decref
-        # some instance attributes
-        for entry in scope.var_entries:
-            if entry.type.is_pyobject:
-                self.generate_self_cast(scope, code)
-                break
+        self.generate_self_cast(scope, code)
         self.generate_usr_dealloc_call(scope, code)
         for entry in scope.var_entries:
             if entry.type.is_pyobject:
@@ -559,16 +580,12 @@
         code.putln(
             "static int %s(PyObject *o, visitproc v, void *a) {"
                 % scope.mangle_internal("tp_traverse"))
-        # only need e, p if we have object attributes
-        for entry in scope.var_entries:
-            if entry.type.is_pyobject:
-                code.putln(
-                    "int e;")
-                self.generate_self_cast(scope, code)
-                break
+        code.putln(
+                "int e;")
+        self.generate_self_cast(scope, code)
         if base_type:
             code.putln(
-                    "%s->tp_traverse(o, v, a);" %
+                    "e = %s->tp_traverse(o, v, a); if (e) return e;" %
                         base_type.typeptr_cname)
         for entry in scope.var_entries:
             if entry.type.is_pyobject:
@@ -594,11 +611,7 @@
         code.putln(
             "static int %s(PyObject *o) {"
                 % scope.mangle_internal("tp_clear"))
-        # only need cast to self type if have object attributes to dereference
-        for entry in scope.var_entries:
-            if entry.type.is_pyobject:
-                self.generate_self_cast(scope, code)
-                break
+        self.generate_self_cast(scope, code)
         if base_type:
             code.putln(
                 "%s->tp_clear(o);" %
@@ -619,12 +632,12 @@
         # a __getitem__ method is present. It converts its
         # argument to a Python integer and calls mp_subscript.
         code.putln(
-            "static PyObject *%s(PyObject *o, int i) {" %
+            "static PyObject *%s(PyObject *o, Py_ssize_t i) {" %
                 scope.mangle_internal("sq_item"))
         code.putln(
                 "PyObject *r;")
         code.putln(
-                "PyObject *x = PyInt_FromLong(i); if(!x) return 0;")
+                "PyObject *x = PyInt_FromSsize_t(i); if(!x) return 0;")
         code.putln(
                 "r = o->ob_type->tp_as_mapping->mp_subscript(o, x);")
         code.putln(
@@ -710,7 +723,7 @@
         del_entry = scope.lookup_here("__delslice__")
         code.putln("")
         code.putln(
-            "static int %s(PyObject *o, int i, int j, PyObject *v) {" %
+            "static int %s(PyObject *o, Py_ssize_t i, Py_ssize_t j, PyObject *v) {" %
                 scope.mangle_internal("sq_ass_slice"))
         code.putln(
                 "if (v) {")
@@ -957,7 +970,8 @@
         if entry.visibility == 'public':
             header = "DL_EXPORT(PyTypeObject) %s = {"
         else:
-            header = "statichere PyTypeObject %s = {"
+            #header = "statichere PyTypeObject %s = {"
+            header = "PyTypeObject %s = {"
         #code.putln(header % scope.parent_type.typeobj_cname)
         code.putln(header % type.typeobj_cname)
         code.putln(
@@ -1076,28 +1090,44 @@
             code.putln(
                 "};")
     
+    def generate_filename_init_prototype(self, code):
+        code.putln("");
+        code.putln("static void %s(void); /*proto*/" % Naming.fileinit_cname)
+
     def generate_module_init_func(self, imported_modules, env, code):
         code.putln("")
-        header = "DL_EXPORT(void) init%s(void)" % env.module_name
+        header = "PyMODINIT_FUNC init%s(void)" % env.module_name
         code.putln("%s; /*proto*/" % header)
         code.putln("%s {" % header)
         code.put_var_declarations(env.temp_entries)
+        #code.putln("/*--- Libary function declarations ---*/")
         env.generate_library_function_declarations(code)
+        self.generate_filename_init_call(code)
+        #code.putln("/*--- Module creation code ---*/")
         self.generate_module_creation_code(env, code)
+        #code.putln("/*--- Intern code ---*/")
         self.generate_intern_code(env, code)
+        #code.putln("/*--- String init code ---*/")
         self.generate_string_init_code(env, code)
+        #code.putln("/*--- Global init code ---*/")
         self.generate_global_init_code(env, code)
+        #code.putln("/*--- Type import code ---*/")
         for module in imported_modules:
             self.generate_type_import_code_for_module(module, env, code)
+        #code.putln("/*--- Type init code ---*/")
         self.generate_type_init_code(env, code)
+        #code.putln("/*--- Execution code ---*/")
         self.body.generate_execution_code(code)
         code.putln("return;")
         code.put_label(code.error_label)
         code.put_var_xdecrefs(env.temp_entries)
-        code.putln('__Pyx_AddTraceback("%s");' % (env.module_name))
+        code.putln('__Pyx_AddTraceback("%s");' % (env.qualified_name))
         env.use_utility_code(traceback_utility_code)
         code.putln('}')
     
+    def generate_filename_init_call(self, code):
+        code.putln("%s();" % Naming.fileinit_cname)
+    
     def generate_module_creation_code(self, env, code):
         # Generate code to create the module object and
         # install the builtins.
@@ -1227,6 +1257,17 @@
                         scope.class_name,
                         typeobj_cname,
                         code.error_goto(entry.pos)))
+                weakref_entry = scope.lookup_here("__weakref__")
+                if weakref_entry:
+                    if weakref_entry.type is py_object_type:
+                        tp_weaklistoffset = "%s.tp_weaklistoffset" % typeobj_cname
+                        code.putln("if (%s == 0) %s = offsetof(struct %s, %s);" % (
+                            tp_weaklistoffset,
+                            tp_weaklistoffset,
+                            type.objstruct_cname,
+                            weakref_entry.cname))
+                    else:
+                        error(weakref_entry.pos, "__weakref__ slot must be of type 'object'")
     
     def generate_exttype_vtable_init_code(self, entry, code):
         # Generate code to initialise the C method table of an
@@ -1244,10 +1285,9 @@
                         Naming.obj_base_cname,
                         type.base_type.vtabptr_cname))
             for meth_entry in type.scope.cfunc_entries:
-                #if not meth_entry.is_inherited:
                 if meth_entry.func_cname:
                     code.putln(
-                        "%s.%s = (void *)%s;" % (
+                        "*(void **)&%s.%s = (void *)%s;" % (
                             type.vtable_cname,
                             meth_entry.cname,
                             meth_entry.func_cname))
@@ -1264,6 +1304,11 @@
     def generate_utility_functions(self, env, code):
         code.putln("")
         code.putln("/* Runtime support code */")
+        code.putln("")
+        code.putln("static void %s(void) {" % Naming.fileinit_cname)
+        code.putln("%s = %s;" % 
+            (Naming.filetable_cname, Naming.filenames_cname))
+        code.putln("}")
         for utility_code in env.utility_code_used:
             code.put(utility_code)
 
@@ -1375,7 +1420,7 @@
             if not self.dimension.type.is_int:
                 error(self.dimension.pos, "Array dimension not integer")
             #size = self.dimension.value
-            size = self.dimension.result
+            size = self.dimension.result_code
         else:
             size = None
         if not base_type.is_complete():
@@ -1422,7 +1467,7 @@
         else:
             if self.exception_value:
                 self.exception_value.analyse_const_expression(env)
-                exc_val = self.exception_value.result
+                exc_val = self.exception_value.result_code
                 if not return_type.assignable_from(self.exception_value.type):
                     error(self.exception_value.pos,
                         "Exception value incompatible with function return type")
@@ -1606,7 +1651,7 @@
     def analyse_declarations(self, env, enum_entry):
         if self.value:
             self.value.analyse_const_expression(env)
-            value = self.value.result
+            value = self.value.result_code
         else:
             value = self.name
         entry = env.declare_const(self.name, enum_entry.type, 
@@ -1697,17 +1742,16 @@
         # ----- Default return value
         code.putln("")
         if self.return_type.is_pyobject:
-            if self.return_type.is_extension_type:
-                cast = "(PyObject *)"
-            else:
-                cast = None
+            #if self.return_type.is_extension_type:
+            #	lhs = "(PyObject *)%s" % Naming.retval_cname
+            #else:
             lhs = Naming.retval_cname
-            code.put_init_to_py_none(cast, lhs)
+            code.put_init_to_py_none(lhs, self.return_type)
         else:
             val = self.return_type.default_value
             if val:
                 code.putln("%s = %s;" % (Naming.retval_cname, val))
-        code.put_goto(code.return_label)
+        code.putln("goto %s;" % code.return_label)
         # ----- Error cleanup
         code.put_label(code.error_label)
         code.put_var_xdecrefs(lenv.temp_entries)
@@ -1734,10 +1778,10 @@
         self.put_stararg_decrefs(code)
         if not self.return_type.is_void:
             retval_code = Naming.retval_cname
-            if self.return_type.is_extension_type:
-                retval_code = "((%s)%s) " % (
-                    self.return_type.declaration_code(""),
-                    retval_code)
+            #if self.return_type.is_extension_type:
+            #	retval_code = "((%s)%s) " % (
+            #		self.return_type.declaration_code(""),
+            #		retval_code)
             code.putln("return %s;" % retval_code)
         code.putln("}")
     
@@ -1819,7 +1863,7 @@
         header = self.return_type.declaration_code(entity,
             dll_linkage = dll_linkage)
         if self.visibility <> 'private':
-            storage_class = ""
+            storage_class = "%s " % Naming.extern_c_macro
         else:
             storage_class = "static "
         code.putln("%s%s {" % (
@@ -2243,11 +2287,11 @@
                 code.putln(
                     "%s = %s;" % (
                         arg.default_entry.cname,
-                        default.result))
+                        default.result_as(arg.default_entry.type)))
                 if default.is_temp and default.type.is_pyobject:
                     code.putln(
                         "%s = 0;" %
-                            default.result)
+                            default.result_code)
         # For Python class methods, create and store function object
         if self.assmt:
             self.assmt.generate_execution_code(code)
@@ -2298,8 +2342,8 @@
         self.classobj.analyse_expressions(env)
         genv = env.global_scope()
         cenv = PyClassScope(name = self.name, outer_scope = genv)
-        cenv.class_dict_cname = self.dict.result
-        cenv.class_obj_cname = self.classobj.result
+        cenv.class_dict_cname = self.dict.result_code
+        cenv.class_obj_cname = self.classobj.result_code
         self.scope = cenv
         self.body.analyse_declarations(cenv)
         self.body.analyse_expressions(cenv)
@@ -2307,7 +2351,7 @@
         self.dict.release_temp(env)
         self.classobj.release_temp(env)
         self.target.release_target_temp(env)
-        env.recycle_pending_temps()
+        #env.recycle_pending_temps()
     
     def generate_function_definitions(self, env, code):
         self.generate_py_string_decls(self.scope, code)
@@ -2452,12 +2496,12 @@
     def analyse_expressions(self, env):
         self.expr.analyse_expressions(env)
         self.expr.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
     
     def generate_execution_code(self, code):
         self.expr.generate_evaluation_code(code)
-        if not self.expr.is_temp and self.expr.result:
-            code.putln("%s;" % self.expr.result)
+        if not self.expr.is_temp and self.expr.result_code:
+            code.putln("%s;" % self.expr.result_code)
         self.expr.generate_disposal_code(code)
 
 
@@ -2502,15 +2546,6 @@
         self.lhs.allocate_target_temps(env)
         self.lhs.release_target_temp(env)
         self.rhs.release_temp(env)		
-    
-#	def analyse_assignment(self, env, lhs, rhs):
-#		# Returns coerced RHS.
-#		rhs.analyse_types(env)
-#		lhs.analyse_target_types(env)
-#		rhs = rhs.coerce_to(lhs.type, env)
-#		rhs.allocate_temps(env)
-#		lhs.allocate_target_temps(env)
-#		return rhs
 
     def generate_rhs_evaluation_code(self, code):
         self.rhs.generate_evaluation_code(code)
@@ -2640,7 +2675,7 @@
             arg.allocate_temps(env)
             arg.release_temp(env)
             self.args[i] = arg
-            env.recycle_pending_temps() # TEMPORARY
+            #env.recycle_pending_temps() # TEMPORARY
         env.use_utility_code(printing_utility_code)
     
     def generate_execution_code(self, code):
@@ -2648,7 +2683,7 @@
             arg.generate_evaluation_code(code)
             code.putln(
                 "if (__Pyx_PrintItem(%s) < 0) %s" % (
-                    arg.result,
+                    arg.py_result(),
                     code.error_goto(self.pos)))
             arg.generate_disposal_code(code)
         if not self.ends_with_comma:
@@ -2671,7 +2706,7 @@
             arg.analyse_target_expression(env)
             if not arg.type.is_pyobject:
                 error(arg.pos, "Deletion of non-Python object")
-            env.recycle_pending_temps() # TEMPORARY
+            #env.recycle_pending_temps() # TEMPORARY
     
     def generate_execution_code(self, code):
         for arg in self.args:
@@ -2699,7 +2734,9 @@
         if not code.break_label:
             error(self.pos, "break statement not inside loop")
         else:
-            code.put_goto(code.break_label)
+            code.putln(
+                "goto %s;" %
+                    code.break_label)
 
 
 class ContinueStatNode(StatNode):
@@ -2713,7 +2750,9 @@
         elif not code.continue_label:
             error(self.pos, "continue statement not inside loop")
         else:
-            code.put_goto(code.continue_label)
+            code.putln(
+                "goto %s;" %
+                    code.continue_label)
 
 
 class ReturnStatNode(StatNode):
@@ -2721,10 +2760,12 @@
     #
     #  value         ExprNode or None
     #  return_type   PyrexType
+    #  temps_in_use  [Entry]            Temps in use at time of return
     
     def analyse_expressions(self, env):
         return_type = env.return_type
         self.return_type = return_type
+        self.temps_in_use = env.temps_in_use()
         if not return_type:
             error(self.pos, "Return not inside a function body")
             return
@@ -2747,32 +2788,27 @@
         if not self.return_type:
             # error reported earlier
             return
+        for entry in self.temps_in_use:
+            code.put_var_decref_clear(entry)
         if self.value:
             self.value.generate_evaluation_code(code)
-            if self.value.type.is_pyobject and not self.value.is_temp:
-                code.put_incref(self.value.result, self.value.type)
-            if self.return_type.is_extension_type:
-                cast = "(%s)" % self.return_type.declaration_code("")
-            else:
-                cast = ""
+            self.value.make_owned_reference(code)
             code.putln(
-                "%s = %s%s;" % (
+                "%s = %s;" % (
                     Naming.retval_cname,
-                    cast,
-                    self.value.result))
+                    self.value.result_as(self.return_type)))
             self.value.generate_post_assignment_code(code)
         else:
             if self.return_type.is_pyobject:
-                code.putln(
-                    "%s = Py_None; Py_INCREF(%s);" % (
-                        Naming.retval_cname,
-                        Naming.retval_cname))
+                code.put_init_to_py_none(Naming.retval_cname, self.return_type)
             elif self.return_type.is_returncode:
                 code.putln(
                     "%s = %s;" % (
                         Naming.retval_cname,
                         self.return_type.default_value))
-        code.put_goto(code.return_label)
+        code.putln(
+            "goto %s;" %
+                code.return_label)
 
 
 class RaiseStatNode(StatNode):
@@ -2801,7 +2837,7 @@
             self.exc_value.release_temp(env)
         if self.exc_tb:
             self.exc_tb.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
         if not (self.exc_type or self.exc_value or self.exc_tb):
             env.use_utility_code(reraise_utility_code)
         else:
@@ -2810,17 +2846,17 @@
     def generate_execution_code(self, code):
         if self.exc_type:
             self.exc_type.generate_evaluation_code(code)
-            type_code = self.exc_type.result
+            type_code = self.exc_type.py_result()
         else:
             type_code = 0
         if self.exc_value:
             self.exc_value.generate_evaluation_code(code)
-            value_code = self.exc_value.result
+            value_code = self.exc_value.py_result()
         else:
             value_code = "0"
         if self.exc_tb:
             self.exc_tb.generate_evaluation_code(code)
-            tb_code = self.exc_tb.result
+            tb_code = self.exc_tb.py_result()
         else:
             tb_code = "0"
         if self.exc_type or self.exc_value or self.exc_tb:
@@ -2857,7 +2893,7 @@
         self.cond.release_temp(env)
         if self.value:
             self.value.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
     
     def generate_execution_code(self, code):
         self.cond.generate_evaluation_code(code)
@@ -2865,11 +2901,11 @@
             self.value.generate_evaluation_code(code)
         code.putln(
             "if (!%s) {" %
-                self.cond.result)
+                self.cond.result_code)
         if self.value:
             code.putln(
                 "PyErr_SetObject(PyExc_AssertionError, %s);" %
-                    self.value.result)
+                    self.value.py_result())
         else:
             code.putln(
                 "PyErr_SetNone(PyExc_AssertionError);")
@@ -2925,16 +2961,18 @@
         self.condition = \
             self.condition.analyse_temp_boolean_expression(env)
         self.condition.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
         self.body.analyse_expressions(env)
     
     def generate_execution_code(self, code, end_label):
         self.condition.generate_evaluation_code(code)
         code.putln(
             "if (%s) {" %
-                self.condition.result)
+                self.condition.result_code)
         self.body.generate_execution_code(code)
-        code.put_goto(end_label)
+        code.putln(
+            "goto %s;" %
+                end_label)
         code.putln("}")
         
     
@@ -2954,7 +2992,7 @@
         self.condition = \
             self.condition.analyse_temp_boolean_expression(env)
         self.condition.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
         self.body.analyse_expressions(env)
         if self.else_clause:
             self.else_clause.analyse_expressions(env)
@@ -2963,12 +3001,12 @@
         old_loop_labels = code.new_loop_labels()
         code.putln(
             "while (1) {")
+        code.put_label(code.continue_label)
         self.condition.generate_evaluation_code(code)
         code.putln(
             "if (!%s) break;" %
-                self.condition.result)
+                self.condition.result_code)
         self.body.generate_execution_code(code)
-        code.put_label(code.continue_label)
         code.putln("}")
         break_label = code.break_label
         code.set_loop_labels(old_loop_labels)
@@ -3004,22 +3042,22 @@
         self.target.allocate_target_temps(env)
         self.item.release_temp(env)
         self.target.release_target_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
         self.body.analyse_expressions(env)
-        self.iterator.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
         if self.else_clause:
             self.else_clause.analyse_expressions(env)
-    
+        self.iterator.release_temp(env)
+
     def generate_execution_code(self, code):
         old_loop_labels = code.new_loop_labels()
         self.iterator.generate_evaluation_code(code)
         code.putln(
             "for (;;) {")
+        code.put_label(code.continue_label)
         self.item.generate_evaluation_code(code)
         self.target.generate_assignment_code(self.item, code)
         self.body.generate_execution_code(code)
-        code.put_label(code.continue_label)
         code.putln(
             "}")
         break_label = code.break_label
@@ -3075,7 +3113,7 @@
             c_loopvar_node = ExprNodes.TempNode(self.pos, 
                 PyrexTypes.c_long_type, env)
             c_loopvar_node.allocate_temps(env)
-            self.loopvar_name = c_loopvar_node.result
+            self.loopvar_name = c_loopvar_node.result_code
             self.py_loopvar_node = \
                 ExprNodes.CloneNode(c_loopvar_node).coerce_to_pyobject(env)
         self.bound1.allocate_temps(env)
@@ -3093,7 +3131,7 @@
             self.else_clause.analyse_expressions(env)
         self.bound1.release_temp(env)
         self.bound2.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
             
     def generate_execution_code(self, code):
         old_loop_labels = code.new_loop_labels()
@@ -3103,8 +3141,8 @@
         code.putln(
             "for (%s = %s%s; %s %s %s; %s%s) {" % (
                 self.loopvar_name,
-                self.bound1.result, offset,
-                self.loopvar_name, self.relation2, self.bound2.result,
+                self.bound1.result_code, offset,
+                self.loopvar_name, self.relation2, self.bound2.result_code,
                 incop, self.loopvar_name))
         if self.py_loopvar_node:
             self.py_loopvar_node.generate_evaluation_code(code)
@@ -3170,7 +3208,9 @@
             self.else_clause.generate_execution_code(code)
             code.putln(
                 "}")
-        code.put_goto(end_label)
+        code.putln(
+            "goto %s;" %
+                end_label)
         code.put_label(our_error_label)
         code.put_var_xdecrefs_clear(self.cleanup_list)
         default_clause_seen = 0
@@ -3182,7 +3222,9 @@
                     error(except_clause.pos, "Default except clause not last")
             except_clause.generate_handling_code(code, end_label)
         if not default_clause_seen:
-            code.put_goto(code.error_label)
+            code.putln(
+                "goto %s;" %
+                    code.error_label)
         code.put_label(end_label)
 
 
@@ -3218,7 +3260,7 @@
         self.exc_value.release_temp(env)
         if self.target:
             self.target.release_target_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
         self.body.analyse_expressions(env)
     
     def generate_handling_code(self, code, end_label):
@@ -3228,7 +3270,7 @@
             code.putln(
                 "%s = PyErr_ExceptionMatches(%s);" % (
                     self.match_flag,
-                    self.pattern.result))
+                    self.pattern.py_result()))
             self.pattern.generate_disposal_code(code)
             code.putln(
                 "if (%s) {" %
@@ -3247,7 +3289,9 @@
         else:
             self.exc_value.generate_disposal_code(code)
         self.body.generate_execution_code(code)
-        code.put_goto(end_label)
+        code.putln(
+            "goto %s;"
+                % end_label)
         code.putln(
             "}")
 
@@ -3320,7 +3364,6 @@
         code.putln(
                 "__pyx_why = 0; goto %s;" %
                     catch_label)
-        code.mark_label_used(catch_label)
         for i in range(len(new_labels)):
             if new_labels[i] and new_labels[i] <> "<try>":
                 if new_labels[i] == new_error_label:
@@ -3332,7 +3375,6 @@
                             new_labels[i],
                             i+1,
                             catch_label))
-                    code.mark_label_used(catch_label)
         code.put_label(catch_label)
         code.set_all_labels(old_labels)
         self.finally_clause.generate_execution_code(code)
@@ -3347,7 +3389,6 @@
                         "case %s: goto %s;" % (
                             i+1,
                             old_labels[i]))
-                    code.mark_label_used(old_labels[i])
         code.putln(
                 "}")		
         code.putln(
@@ -3367,7 +3408,9 @@
         code.putln(
                 "%s = %s;" % (
                     self.lineno_var, Naming.lineno_cname))
-        code.put_goto(catch_label)
+        code.putln(
+                "goto %s;" %
+                    catch_label)
         code.putln(
             "}")
             
@@ -3385,7 +3428,9 @@
             code.putln(
                 "%s = 0;" %
                     var)
-        code.put_goto(error_label)
+        code.putln(
+                "goto %s;" %
+                    error_label)
         code.putln(
             "}")
 
@@ -3469,7 +3514,7 @@
             target.release_temp(env)
         self.module.release_temp(env)
         self.item.release_temp(env)
-        env.recycle_pending_temps() # TEMPORARY
+        #env.recycle_pending_temps() # TEMPORARY
     
     def generate_execution_code(self, code):
         self.module.generate_evaluation_code(code)
@@ -3477,20 +3522,20 @@
             for cname, target in self.interned_items:
                 code.putln(
                     '%s = PyObject_GetAttr(%s, %s); if (!%s) %s' % (
-                        self.item.result, 
-                        self.module.result,
+                        self.item.result_code, 
+                        self.module.py_result(),
                         cname,
-                        self.item.result,
+                        self.item.result_code,
                         code.error_goto(self.pos)))
                 target.generate_assignment_code(self.item, code)
         else:
             for name, target in self.items:
                 code.putln(
                     '%s = PyObject_GetAttrString(%s, "%s"); if (!%s) %s' % (
-                        self.item.result, 
-                        self.module.result,
+                        self.item.result_code, 
+                        self.module.py_result(),
                         name,
-                        self.item.result,
+                        self.item.result_code,
                         code.error_goto(self.pos)))
                 target.generate_assignment_code(self.item, code)
         self.module.generate_disposal_code(code)
@@ -3505,8 +3550,8 @@
 """
 typedef struct {PyObject **p; char *s;} __Pyx_InternTabEntry; /*proto*/
 typedef struct {PyObject **p; char *s; long n;} __Pyx_StringTabEntry; /*proto*/
-static PyObject *__Pyx_UnpackItem(PyObject *, int); /*proto*/
-static int __Pyx_EndUnpack(PyObject *, int); /*proto*/
+static PyObject *__Pyx_UnpackItem(PyObject *, Py_ssize_t); /*proto*/
+static int __Pyx_EndUnpack(PyObject *, Py_ssize_t); /*proto*/
 static int __Pyx_PrintItem(PyObject *); /*proto*/
 static int __Pyx_PrintNewline(void); /*proto*/
 static void __Pyx_Raise(PyObject *type, PyObject *value, PyObject *tb); /*proto*/
@@ -3516,7 +3561,7 @@
 static int __Pyx_ArgTypeTest(PyObject *obj, PyTypeObject *type, int none_allowed, char *name); /*proto*/
 static int __Pyx_TypeTest(PyObject *obj, PyTypeObject *type); /*proto*/
 static int __Pyx_GetStarArgs(PyObject **args, PyObject **kwds,\
- char *kwd_list[], int nargs, PyObject **args2, PyObject **kwds2); /*proto*/
+ char *kwd_list[], Py_ssize_t nargs, PyObject **args2, PyObject **kwds2); /*proto*/
 static void __Pyx_WriteUnraisable(char *name); /*proto*/
 static void __Pyx_AddTraceback(char *funcname); /*proto*/
 static PyTypeObject *__Pyx_ImportType(char *module_name, char *class_name, long size);  /*proto*/
@@ -3558,7 +3603,7 @@
         return -1;
     if (PyString_Check(v)) {
         char *s = PyString_AsString(v);
-        int len = PyString_Size(v);
+        Py_ssize_t len = PyString_Size(v);
         if (len > 0 &&
             isspace(Py_CHARMASK(s[len-1])) &&
             s[len-1] != ' ')
@@ -3613,7 +3658,7 @@
     }
     if (PyString_Check(type))
         ;
-    else if (PyClass_Check(type))
+    else if (PyType_Check(type) || PyClass_Check(type))
         ; /*PyErr_NormalizeException(&type, &value, &tb);*/
     else if (PyInstance_Check(type)) {
         /* Raising an instance.  The value should be a dummy. */
@@ -3702,7 +3747,7 @@
     PyObject **args, 
     PyObject **kwds,
     char *kwd_list[], 
-    int nargs,
+    Py_ssize_t nargs,
     PyObject **args2, 
     PyObject **kwds2)
 {

Modified: lxml/pyrex/Pyrex/Compiler/Parsing.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Parsing.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Parsing.py	Tue Apr 25 15:43:21 2006
@@ -422,13 +422,15 @@
             value = int(s.systring)
         s.next()
         return ExprNodes.IntNode(pos, value = value)
+    elif sy == 'LONG':
+        value = s.systring
+        s.next()
+        return ExprNodes.LongNode(pos, value = value)
     elif sy == 'FLOAT':
-        #value = float(s.systring)
         value = s.systring
         s.next()
         return ExprNodes.FloatNode(pos, value = value)
     elif sy == 'IMAG':
-        #value = float(s.systring[:-1])
         value = s.systring[:-1]
         s.next()
         return ExprNodes.ImagNode(pos, value = value)
@@ -500,6 +502,10 @@
             if kind == 'r':
                 if systr == '\\\n':
                     chars.append(r'\\\n')
+                elif systr == r'\"':
+                    chars.append(r'\\\"')
+                elif systr == r'\\':
+                    chars.append(r'\\\\')
                 else:
                     chars.append('\\' + systr)
             else:
@@ -1041,7 +1047,7 @@
     s.expect_newline("Syntax error in include statement")
     include_file_path = s.context.find_include_file(include_file_name, pos)
     if include_file_path:
-        f = open(include_file_path, "r")
+        f = open(include_file_path, "rU")
         s2 = PyrexScanner(f, include_file_path, s)
         try:
             tree = p_statement_list(s2, level)
@@ -1251,7 +1257,7 @@
 #	"void", "signed", "unsigned"
 #)
 
-basic_c_type_names = ("void", "char", "int", "float", "double")
+basic_c_type_names = ("void", "char", "int", "float", "double", "Py_ssize_t")
 
 sign_and_longness_words = ("short", "long", "signed", "unsigned")
 
@@ -1377,7 +1383,7 @@
     if s.sy == "-":
         sign = "-"
         s.next()
-    if s.sy in ('INT', 'FLOAT', 'NULL'):
+    if s.sy in ('INT', 'LONG', 'FLOAT', 'NULL'):
         s.systring = sign + s.systring
         return p_atom(s)
     else:
@@ -1575,6 +1581,8 @@
         declarators = [declarator]
         while s.sy == ',':
             s.next()
+            if s.sy == 'NEWLINE':
+                break
             declarator = p_c_declarator(s, cmethod_flag = cmethod_flag)
             declarators.append(declarator)
         s.expect_newline("Syntax error in C variable declaration")

Modified: lxml/pyrex/Pyrex/Compiler/PyrexTypes.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/PyrexTypes.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/PyrexTypes.py	Tue Apr 25 15:43:21 2006
@@ -35,8 +35,8 @@
     #    of this type, given a code fragment for the entity.
     #    * If for_display, this is for reading by a human in an error
     #      message; otherwise it must be valid C code.
-    #    * If dll_linkage is not None, it must be 'DL_IMPORT' or
-    #      'DL_EXPORT', and will be added to the base type part of
+    #    * If dll_linkage is not None, it must be 'DL_EXPORT' or
+    #      'DL_IMPORT', and will be added to the base type part of
     #      the declaration.
     #    * If pyrex = 1, this is for use in a 'cdef extern'
     #      statement of a Pyrex include file.
@@ -111,6 +111,9 @@
         # A type is incomplete if it is an unsized array,
         # a struct whose attributes are not defined, etc.
         return 1
+    
+    def cast_code(self, expr_code):
+        return "((%s)%s)" % (self.declaration_code(""), expr_code)
 
 
 class CTypedefType:
@@ -289,11 +292,15 @@
     is_numeric = 1
     default_value = "0"
     
-    parsetuple_formats = "chilLfd?" # rank -> format
+    parsetuple_formats = ( # rank -> format
+        "?HIkK????", # unsigned
+        "chilL?fd?", # signed
+    )
     
-    def __init__(self, rank, pymemberdef_typecode = None):
+    def __init__(self, rank, signed = 1, pymemberdef_typecode = None):
         self.rank = rank
-        ptf = self.parsetuple_formats[rank]
+        self.signed = signed
+        ptf = self.parsetuple_formats[signed][rank]
         if ptf == '?':
             ptf = None
         self.parsetuple_format = ptf
@@ -329,36 +336,49 @@
     from_py_function = "PyInt_AsLong"
 
     def __init__(self, rank, signed, pymemberdef_typecode = None, is_returncode = 0):
-        CNumericType.__init__(self, rank, pymemberdef_typecode)
-        self.signed = signed
+        CNumericType.__init__(self, rank, signed, pymemberdef_typecode)
         self.is_returncode = is_returncode
 
 
+class CPySSizeTType(CIntType):
+
+    to_py_function = "PyInt_FromSsize_t"
+    from_py_function = "PyInt_AsSsize_t"
+
+
+class CUIntType(CIntType):
+
+    to_py_function = "PyLong_FromUnsignedLong"
+    from_py_function = "PyInt_AsUnsignedLongMask"
+
+
 class CULongType(CIntType):
 
     to_py_function = "PyLong_FromUnsignedLong"
-    from_py_function = "PyLong_AsUnsignedLong"
+    from_py_function = "PyInt_AsUnsignedLongMask"
 
 
 class CLongLongType(CIntType):
 
     to_py_function = "PyLong_FromLongLong"
-    from_py_function = "PyLong_AsLongLong"
+    from_py_function = "PyInt_AsUnsignedLongLongMask"
 
 
 class CULongLongType(CIntType):
 
     to_py_function = "PyLong_FromUnsignedLongLong"
-    from_py_function = "PyLong_AsUnsignedLongLong"
+    from_py_function = "PyInt_AsUnsignedLongLongMask"
 
 
 class CFloatType(CNumericType):
 
     is_float = 1
-    signed = 1
     to_py_function = "PyFloat_FromDouble"
     from_py_function = "PyFloat_AsDouble"
-
+    
+    def __init__(self, rank, pymemberdef_typecode = None):
+        CNumericType.__init__(self, rank, 1, pymemberdef_typecode)
+    
 
 class CArrayType(CType):
     #  base_type     CType              Element type
@@ -423,6 +443,7 @@
     
     def declaration_code(self, entity_code, 
             for_display = 0, dll_linkage = None, pyrex = 0):
+        #print "CPtrType.declaration_code: pointer to", self.base_type ###
         return self.base_type.declaration_code(
             "(*%s)" % entity_code,
             for_display, dll_linkage, pyrex)
@@ -521,7 +542,7 @@
         if not arg_decl_code and not pyrex:
             arg_decl_code = "void"
         exc_clause = ""
-        if for_display:
+        if pyrex or for_display:
             if self.exception_value and self.exception_check:
                 exc_clause = " except? %s" % self.exception_value
             elif self.exception_value:
@@ -683,17 +704,18 @@
 c_short_type =    CIntType(1, 1, "T_SHORT")
 c_int_type =      CIntType(2, 1, "T_INT")
 c_long_type =     CIntType(3, 1, "T_LONG")
-c_longlong_type = CLongLongType(4, 1)
+c_longlong_type = CLongLongType(4, 1, "T_LONGLONG")
+c_py_ssize_t_type = CPySSizeTType(5, 1)
 
 c_uchar_type =     CIntType(0, 0, "T_UBYTE")
 c_ushort_type =    CIntType(1, 0, "T_USHORT")
-c_uint_type =      CIntType(2, 0, "T_UINT")
+c_uint_type =      CUIntType(2, 0, "T_UINT")
 c_ulong_type =     CULongType(3, 0, "T_ULONG")
-c_ulonglong_type = CULongLongType(4, 0)
+c_ulonglong_type = CULongLongType(4, 0, "T_ULONGLONG")
 
-c_float_type =      CFloatType(5, "T_FLOAT")
-c_double_type =     CFloatType(6, "T_DOUBLE")
-c_longdouble_type = CFloatType(7)
+c_float_type =      CFloatType(6, "T_FLOAT")
+c_double_type =     CFloatType(7, "T_DOUBLE")
+c_longdouble_type = CFloatType(8)
 
 c_null_ptr_type =     CNullPtrType(c_void_type)
 c_char_array_type =   CCharArrayType(None)
@@ -705,7 +727,7 @@
 
 error_type =    ErrorType()
 
-lowest_float_rank = 5
+lowest_float_rank = 6
 
 rank_to_type_name = (
     "char",         # 0
@@ -713,9 +735,10 @@
     "int",          # 2
     "long",         # 3
     "PY_LONG_LONG", # 4
-    "float",        # 5
-    "double",       # 6
-    "long double",  # 7
+    "Py_ssize_t",   # 5
+    "float",        # 6
+    "double",       # 7
+    "long double",  # 8
 )
 
 sign_and_rank_to_type = {
@@ -730,9 +753,10 @@
     (1, 2): c_int_type, 
     (1, 3): c_long_type,
     (1, 4): c_longlong_type,
-    (1, 5): c_float_type, 
-    (1, 6): c_double_type,
-    (1, 7): c_longdouble_type,
+    (1, 5): c_py_ssize_t_type,
+    (1, 6): c_float_type, 
+    (1, 7): c_double_type,
+    (1, 8): c_longdouble_type,
 }
 
 modifiers_and_name_to_type = {
@@ -748,6 +772,7 @@
     (1, 0, "int"): c_int_type, 
     (1, 1, "int"): c_long_type,
     (1, 2, "int"): c_longlong_type,
+    (1, 0, "Py_ssize_t"): c_py_ssize_t_type,
     (1, 0, "float"): c_float_type, 
     (1, 0, "double"): c_double_type,
     (1, 1, "double"): c_longdouble_type,
@@ -788,3 +813,19 @@
     else:
         return base
 
+def same_type(type1, type2):
+    return type1.same_as(type2)
+    
+def assignable_from(type1, type2):
+    return type1.assignable_from(type2)
+
+def typecast(to_type, from_type, expr_code):
+    #  Return expr_code cast to a C type which can be
+    #  assigned to to_type, assuming its existing C type
+    #  is from_type.
+    if to_type is from_type or \
+        (not to_type.is_pyobject and assignable_from(to_type, from_type)):
+            return expr_code
+    else:
+        #print "typecast: to", to_type, "from", from_type ###
+        return to_type.cast_code(expr_code)

Modified: lxml/pyrex/Pyrex/Compiler/Scanning.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Scanning.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Scanning.py	Tue Apr 25 15:43:21 2006
@@ -21,7 +21,6 @@
 
 debug_scanner = 0
 trace_scanner = 0
-#scanner_dump_file = open("Lexicon_dump.txt", "w")
 scanner_debug_flags = 0
 scanner_dump_file = None
 binary_lexicon_pickle = 1
@@ -39,7 +38,7 @@
     import md5
     try:
         try:
-            f = open(path)
+            f = open(path, "rU")
             text = f.read()
         except IOError, e:
             print "Unable to hash scanner source file (%s)" % e
@@ -150,11 +149,6 @@
     
     def __call__(self, stream, text):
         return getattr(stream, self.name)(text)
-        
-
-#def make_lexicon():
-#	import Lexicon
-#	return Lexicon.lexicon
 
 #------------------------------------------------------------------
 

Modified: lxml/pyrex/Pyrex/Compiler/Symtab.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Symtab.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Symtab.py	Tue Apr 25 15:43:21 2006
@@ -22,7 +22,6 @@
     # cname            string     C name of entity
     # type             PyrexType  Type of entity
     # doc              string     Doc string
-    # #borrowed         bool       Is a borrowed reference
     # init             string     Initial value
     # visibility       'private' or 'public' or 'extern'
     # is_builtin       boolean    Is a Python builtin name
@@ -171,7 +170,6 @@
         return self.global_scope().intern(name)
     
     def qualifying_scope(self):
-        #return self.outer_scope
         return self.parent_scope
     
     def mangle(self, prefix, name = None):
@@ -208,9 +206,6 @@
     def qualify_name(self, name):
         return "%s.%s" % (self.qualified_name, name)
     
-    #def undeclare(self, name):
-    #	del self.entries[name]
-    
     def declare_const(self, name, type, value, pos, cname = None):
         # Add an entry for a named constant.
         if not cname:
@@ -420,17 +415,23 @@
                 % cname)
         self.free_temp_entries.append(entry)
     
-    def recycle_pending_temps(self):
-        # Obsolete
-        pass
+    def temps_in_use(self):
+        # Return a new list of temp entries currently in use.
+        return [entry for entry in self.temp_entries
+            if entry not in self.free_temp_entries]
+    
+    #def recycle_pending_temps(self):
+    #	# Obsolete
+    #	pass
 
     def use_utility_code(self, new_code):
         self.global_scope().use_utility_code(new_code)
     
     def generate_library_function_declarations(self, code):
         # Generate extern decls for C library funcs used.
-        if self.pow_function_used:
-            code.putln("extern double pow(double, double);")
+        #if self.pow_function_used:
+        #	code.putln("%s double pow(double, double);" % Naming.extern_c_macro)
+        pass
         
     def defines_any(self, names):
         # Test whether any of the given names are
@@ -885,8 +886,8 @@
     def release_temp(self, cname):
         self.outer_scope.release_temp(cname)
 
-    def recycle_pending_temps(self):
-        self.outer_scope.recycle_pending_temps()
+    #def recycle_pending_temps(self):
+    #	self.outer_scope.recycle_pending_temps()
 
     def add_default_value(self, type):
         return self.outer_scope.add_default_value(type)
@@ -977,6 +978,8 @@
             
     def declare_cfunction(self, name, type, pos,
             cname = None, visibility = 'private', defining = 0):
+        if get_special_method_signature(name):
+            error(pos, "Special methods must be declared with 'def', not 'cdef'")
         args = type.args
         if not args:
             error(pos, "C method has no self argument")

Modified: lxml/pyrex/Pyrex/Compiler/TypeSlots.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/TypeSlots.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/TypeSlots.py	Tue Apr 25 15:43:21 2006
@@ -26,6 +26,7 @@
     #    'i'  int
     #    'I'  int *
     #    'l'  long
+    #    'Z'  Py_ssize_t
     #    's'  char *
     #    'S'  char **
     #    'r'  int used only to signal exception
@@ -42,6 +43,7 @@
         'i': PyrexTypes.c_int_type,
         'I': PyrexTypes.c_int_ptr_type,
         'l': PyrexTypes.c_long_type,
+        'Z': PyrexTypes.c_py_ssize_t_type,
         's': PyrexTypes.c_char_ptr_type,
         'S': PyrexTypes.c_char_ptr_ptr_type,
         'r': PyrexTypes.c_returncode_type,
@@ -212,9 +214,10 @@
     #  Descriptor for the type flags slot.
     
     def slot_code(self, scope):
-        value = "Py_TPFLAGS_DEFAULT|Py_TPFLAGS_CHECKTYPES|Py_TPFLAGS_BASETYPE"
-        if scope.has_pyobject_attrs:
-            value += "|Py_TPFLAGS_HAVE_GC"
+        # Always add Py_TPFLAGS_HAVE_GC -- PyType_Ready doesn't seem to inherit it
+        value = "Py_TPFLAGS_DEFAULT|Py_TPFLAGS_CHECKTYPES|Py_TPFLAGS_BASETYPE|Py_TPFLAGS_HAVE_GC"
+        #if scope.has_pyobject_attrs:
+        #	value += "|Py_TPFLAGS_HAVE_GC"
         return value
         
 
@@ -353,18 +356,30 @@
 iternaryfunc = Signature("TOO", "O")       # typedef PyObject * (*ternaryfunc)(PyObject *, PyObject *, PyObject *);
 callfunc = Signature("T*", "O")            # typedef PyObject * (*ternaryfunc)(PyObject *, PyObject *, PyObject *);
 inquiry = Signature("T", "i")              # typedef int (*inquiry)(PyObject *);
+lenfunc = Signature("T", "Z")              # typedef Py_ssize_t (*lenfunc)(PyObject *);
+
                                            # typedef int (*coercion)(PyObject **, PyObject **);
 intargfunc = Signature("Ti", "O")          # typedef PyObject *(*intargfunc)(PyObject *, int);
+ssizeargfunc = Signature("TZ", "O")        # typedef PyObject *(*ssizeargfunc)(PyObject *, Py_ssize_t);
 intintargfunc = Signature("Tii", "O")      # typedef PyObject *(*intintargfunc)(PyObject *, int, int);
+ssizessizeargfunc = Signature("TZZ", "O")  # typedef PyObject *(*ssizessizeargfunc)(PyObject *, Py_ssize_t, Py_ssize_t);
 intobjargproc = Signature("TiO", 'r')      # typedef int(*intobjargproc)(PyObject *, int, PyObject *);
+ssizeobjargproc = Signature("TZO", 'r')    # typedef int(*ssizeobjargproc)(PyObject *, Py_ssize_t, PyObject *);
 intintobjargproc = Signature("TiiO", 'r')  # typedef int(*intintobjargproc)(PyObject *, int, int, PyObject *);
+ssizessizeobjargproc = Signature("TZZO", 'r') # typedef int(*ssizessizeobjargproc)(PyObject *, Py_ssize_t, Py_ssize_t, PyObject *);
+
 intintargproc = Signature("Tii", 'r')
+ssizessizeargproc = Signature("TZZ", 'r')
 objargfunc = Signature("TO", "O")
 objobjargproc = Signature("TOO", 'r')      # typedef int (*objobjargproc)(PyObject *, PyObject *, PyObject *);
 getreadbufferproc = Signature("TiP", 'i')  # typedef int (*getreadbufferproc)(PyObject *, int, void **);
 getwritebufferproc = Signature("TiP", 'i') # typedef int (*getwritebufferproc)(PyObject *, int, void **);
 getsegcountproc = Signature("TI", 'i')     # typedef int (*getsegcountproc)(PyObject *, int *);
 getcharbufferproc = Signature("TiS", 'i')  # typedef int (*getcharbufferproc)(PyObject *, int, const char **);
+readbufferproc = Signature("TZP", "Z")     # typedef Py_ssize_t (*readbufferproc)(PyObject *, Py_ssize_t, void **);
+writebufferproc = Signature("TZP", "Z")    # typedef Py_ssize_t (*writebufferproc)(PyObject *, Py_ssize_t, void **);
+segcountproc = Signature("TZ", "Z")        # typedef Py_ssize_t (*segcountproc)(PyObject *, Py_ssize_t *);
+writebufferproc = Signature("TZS", "Z")    # typedef Py_ssize_t (*charbufferproc)(PyObject *, Py_ssize_t, char **);
 objargproc = Signature("TO", 'r')          # typedef int (*objobjproc)(PyObject *, PyObject *);
                                            # typedef int (*visitproc)(PyObject *, void *);
                                            # typedef int (*traverseproc)(PyObject *, visitproc, void *);
@@ -453,14 +468,17 @@
     MethodSlot(binaryfunc, "nb_true_divide", "__truediv__"),
     MethodSlot(ibinaryfunc, "nb_inplace_floor_divide", "__ifloordiv__"),
     MethodSlot(ibinaryfunc, "nb_inplace_true_divide", "__itruediv__"),
+
+    # Added in release 2.5
+#    MethodSlot(lenfunc, "nb_index", "??"),
 )
 
 PySequenceMethods = (
-    MethodSlot(inquiry, "sq_length", "__len__"),    # EmptySlot("sq_length"), # mp_length used instead
+    MethodSlot(lenfunc, "sq_length", "__len__"),    # EmptySlot("sq_length"), # mp_length used instead
     EmptySlot("sq_concat"), # nb_add used instead
     EmptySlot("sq_repeat"), # nb_multiply used instead
     SyntheticSlot("sq_item", ["__getitem__"], "0"),    #EmptySlot("sq_item"),   # mp_subscript used instead
-    MethodSlot(intintargfunc, "sq_slice", "__getslice__"),
+    MethodSlot(ssizessizeargfunc, "sq_slice", "__getslice__"),
     EmptySlot("sq_ass_item"), # mp_ass_subscript used instead
     SyntheticSlot("sq_ass_slice", ["__setslice__", "__delslice__"], "0"),
     MethodSlot(cmpfunc, "sq_contains", "__contains__"),
@@ -469,9 +487,9 @@
 )
 
 PyMappingMethods = (
-    MethodSlot(inquiry, "mp_length", "__len__"),
+    MethodSlot(lenfunc, "mp_length", "__len__"),
     MethodSlot(objargfunc, "mp_subscript", "__getitem__"),
-    SyntheticSlot("mp_ass_subscript", ["__setitem__"], "0"),
+    SyntheticSlot("mp_ass_subscript", ["__setitem__", "__delitem__"], "0"),
 )
 
 PyBufferProcs = (
@@ -564,8 +582,8 @@
 MethodSlot(destructor, "", "__dealloc__")
 MethodSlot(objobjargproc, "", "__setitem__")
 MethodSlot(objargproc, "", "__delitem__")
-MethodSlot(intintobjargproc, "", "__setslice__")
-MethodSlot(intintargproc, "", "__delslice__")
+MethodSlot(ssizessizeobjargproc, "", "__setslice__")
+MethodSlot(ssizessizeargproc, "", "__delslice__")
 MethodSlot(getattrofunc, "", "__getattr__")
 MethodSlot(setattrofunc, "", "__setattr__")
 MethodSlot(delattrofunc, "", "__delattr__")

Modified: lxml/pyrex/Pyrex/Compiler/Version.py
==============================================================================
--- lxml/pyrex/Pyrex/Compiler/Version.py	(original)
+++ lxml/pyrex/Pyrex/Compiler/Version.py	Tue Apr 25 15:43:21 2006
@@ -1 +1 @@
-version = '0.9.3.1_lxml'
+version = '0.9.4.1'

Modified: lxml/pyrex/Pyrex/Distutils/build_ext.py
==============================================================================
--- lxml/pyrex/Pyrex/Distutils/build_ext.py	(original)
+++ lxml/pyrex/Pyrex/Distutils/build_ext.py	Tue Apr 25 15:43:21 2006
@@ -7,7 +7,8 @@
 # Pyrex is (c) Greg Ewing.
 
 import distutils.command.build_ext
-import Pyrex.Compiler.Main
+#import Pyrex.Compiler.Main
+from Pyrex.Compiler.Main import CompilationOptions, default_options, compile
 from Pyrex.Compiler.Errors import PyrexError
 from distutils.dep_util import newer
 import os
@@ -18,62 +19,45 @@
 
 class build_ext (distutils.command.build_ext.build_ext):
 
-  description = "compile Pyrex scripts, then build C/C++ extensions (compile/link to build directory)"
+    description = "compile Pyrex scripts, then build C/C++ extensions (compile/link to build directory)"
 
-  def finalize_options (self):
-    distutils.command.build_ext.build_ext.finalize_options(self)
+    def finalize_options (self):
+        distutils.command.build_ext.build_ext.finalize_options(self)
+
+        # The following hack should no longer be needed.
+        if 0:
+            # compiling with mingw32 gets an "initializer not a constant" error
+            # doesn't appear to happen with MSVC!
+            # so if we are compiling with mingw32,
+            # switch to C++ mode, to avoid the problem
+            if self.compiler == 'mingw32':
+                self.swig_cpp = 1
+
+    def swig_sources (self, sources, extension = None):
+        if not self.extensions:
+            return
+
+        # collect the names of the source (.pyx) files
+        pyx_sources = []
+        pyx_sources = [source for source in sources if source.endswith('.pyx')]
+        other_sources = [source for source in sources if not source.endswith('.pyx')]
+
+        #suffix = self.swig_cpp and '.cpp' or '.c'
+        suffix = '.c'
+        for pyx in pyx_sources:
+            # should I raise an exception if it doesn't exist?
+            if os.path.exists(pyx):
+                source = pyx
+                target = replace_suffix(source, suffix)
+                if newer(source, target) or self.force:
+                    self.pyrex_compile(source)
+
+        return [replace_suffix(src, suffix) for src in pyx_sources] + other_sources
+
+    def pyrex_compile(self, source):
+        options = CompilationOptions(default_options,
+            include_path = self.include_dirs)
+        result = compile(source, options)
+        if result.num_errors <> 0:
+            sys.exit(1)
 
-    # The following hack should no longer be needed.
-    if 0:
-      # compiling with mingw32 gets an "initializer not a constant" error
-      # doesn't appear to happen with MSVC!
-      # so if we are compiling with mingw32,
-      # switch to C++ mode, to avoid the problem
-      if self.compiler == 'mingw32':
-        self.swig_cpp = 1
-
-  def swig_sources (self, sources, *otherargs):
-    if not self.extensions:
-      return
-
-    # collect the names of the source (.pyx) files
-    pyx_sources = []
-    pyx_sources = [source for source in sources if source.endswith('.pyx')]
-    other_sources = [source for source in sources if not source.endswith('.pyx')]
-
-    extension = self.swig_cpp and '.cpp' or '.c'
-    for pyx in pyx_sources:
-      # should I raise an exception if it doesn't exist?
-      if os.path.exists(pyx):
-        source = pyx
-        #target = source.replace('.pyx', extension)
-        target = replace_suffix(source, extension)
-        if newer(source, target) or self.force:
-          self.pyrex_compile(source)
-
-          if self.swig_cpp:
-            # rename .c to .cpp (Pyrex always builds .c ...)
-            if os.path.exists(target):
-              os.unlink(target)
-            #os.rename(source.replace('.pyx', '.c'), target)
-            os.rename(replace_suffix(source, '.c'), target)
-            # massage the cpp file
-            self.c_to_cpp(target)
-
-    return [replace_suffix(src, extension) for src in pyx_sources] + other_sources
-
-  def pyrex_compile(self, source):
-    result = Pyrex.Compiler.Main.compile(source)
-    if result.num_errors <> 0:
-      sys.exit(1)
-
-  def c_to_cpp(self, filename):
-    """touch up the Pyrex generated c/cpp files to meet mingw32/distutils requirements."""
-    f = open(filename, 'r')
-    lines = [line for line in f.readlines() if not line.startswith('staticforward PyTypeObject __pyx_type_')]
-    f.close()
-    f = open(filename, 'w')
-    lines.insert(1, 'extern "C" {\n')
-    lines.append('}\n')
-    f.write(''.join(lines))
-    f.close()

Modified: lxml/pyrex/Pyrex/Mac/DarwinSystem.py
==============================================================================
--- lxml/pyrex/Pyrex/Mac/DarwinSystem.py	(original)
+++ lxml/pyrex/Pyrex/Mac/DarwinSystem.py	Tue Apr 25 15:43:21 2006
@@ -3,6 +3,8 @@
 #
 
 verbose = 0
+gcc_pendantic = True
+gcc_warnings_are_errors = False
 
 import os
 from Pyrex.Utils import replace_suffix
@@ -12,13 +14,17 @@
     "/Library/Frameworks/Python.framework/Headers"
 ]
 
-compiler = "gcc"
+compilers = ["gcc", "g++"]
 compiler_options = \
     "-g -c -fno-strict-aliasing -Wno-long-double -no-cpp-precomp " \
-    "-mno-fused-madd -fno-common -dynamic" \
+    "-mno-fused-madd -fno-common -dynamic " \
     .split()
+if gcc_pendantic:
+    compiler_options.extend(["-pedantic", "-Wno-long-long"])
+if gcc_warnings_are_errors:
+    compiler_options.append("-Werror")
 
-linker = "gcc"
+linkers = ["gcc", "g++"]
 linker_options = \
     "-Wl,-F.,-w -bundle -framework Python" \
     .split()
@@ -26,15 +32,16 @@
 class CCompilerError(PyrexError):
     pass
 
-def c_compile(c_file, verbose_flag = 0):
+def c_compile(c_file, verbose_flag = 0, cplus = 0, obj_suffix = ".o"):
     #  Compile the given C source file to produce
     #  an object file. Returns the pathname of the
     #  resulting file.
     c_file = os.path.join(os.getcwd(), c_file)
-    o_file = replace_suffix(c_file, ".o")
+    o_file = replace_suffix(c_file, obj_suffix)
     include_options = []
     for dir in py_include_dirs:
         include_options.append("-I%s" % dir)
+    compiler = compilers[bool(cplus)]
     args = [compiler] + compiler_options + include_options + [c_file, "-o", o_file]
     if verbose_flag or verbose:
         print " ".join(args)
@@ -43,14 +50,15 @@
         raise CCompilerError("C compiler returned status %s" % status)
     return o_file
 
-def c_link(obj_file, verbose_flag = 0):
-    return c_link_list([obj_file], verbose_flag)
+def c_link(obj_file, verbose_flag = 0, extra_objects = [], cplus = 0):
+    return c_link_list([obj_file] + extra_objects, verbose_flag, cplus)
 
-def c_link_list(obj_files, verbose_flag = 0):
+def c_link_list(obj_files, verbose_flag = 0, cplus = 0):
     #  Link the given object files into a dynamically
     #  loadable extension file. Returns the pathname
     #  of the resulting file.
     out_file = replace_suffix(obj_files[0], ".so")
+    linker = linkers[bool(cplus)]
     args = [linker] + linker_options + obj_files + ["-o", out_file]
     if verbose_flag or verbose:
         print " ".join(args)

Modified: lxml/pyrex/Pyrex/Utils.py
==============================================================================
--- lxml/pyrex/Pyrex/Utils.py	(original)
+++ lxml/pyrex/Pyrex/Utils.py	Tue Apr 25 15:43:21 2006
@@ -8,12 +8,9 @@
 def replace_suffix(path, newsuf):
     base, _ = os.path.splitext(path)
     return base + newsuf
-    
-def default_open_new_file(path):
-    return open(path, "w")
 
-if sys.platform == "mac":
-    from Pyrex.Mac.MacUtils import open_new_file
-else:
-    open_new_file = default_open_new_file
+def open_new_file(path):
+    #  Open and truncate existing file to
+    #  preserve metadata on the Mac.
+    return open(path, "w+")
 

Modified: lxml/pyrex/ToDo.txt
==============================================================================
--- lxml/pyrex/ToDo.txt	(original)
+++ lxml/pyrex/ToDo.txt	Tue Apr 25 15:43:21 2006
@@ -269,3 +269,16 @@
 Initial values when declaring variables?
 
 Do something about __stdcall.
+
+Support class methods in extension types using METH_CLASS flag.
+
+Disallow defaulting types to 'object' in C declarations?
+
+C globals with static initialisers.
+
+Find a way of providing C-only initialisers for extension types.
+
+Metaclasses for extension types?
+
+Make extension types use Py_TPFLAGS_HEAPTYPE so their __module__
+will get set dynamically?