From scoder at codespeak.net  Fri Dec  1 09:48:13 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  1 Dec 2006 09:48:13 +0100 (CET)
Subject: [Lxml-checkins] r35185 - in lxml/trunk: . src/lxml
Message-ID: <20061201084813.0B1AF1006C@code0.codespeak.net>

Author: scoder
Date: Fri Dec  1 09:47:54 2006
New Revision: 35185

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/etree.pyx
Log:
accept QName objects in Element.find*()

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Fri Dec  1 09:47:54 2006
@@ -15,6 +15,12 @@
   possible to pass an --rpath directly to distutils; previously this was being
   shadowed.
 
+Bugs fixed
+----------
+
+* Element.find*() did not accept QName objects as path
+
+
 1.1.2 (2006-10-30)
 ==================
 

Modified: lxml/trunk/src/lxml/etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/etree.pyx	(original)
+++ lxml/trunk/src/lxml/etree.pyx	Fri Dec  1 09:47:54 2006
@@ -202,6 +202,22 @@
             raise type, value, traceback
 
 
+cdef class QName:
+    """QName wrapper.
+    """
+    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 _isString(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__()
+
+
 # forward declaration of _BaseParser, see parser.pxi
 cdef class _BaseParser
 
@@ -1176,16 +1192,22 @@
     def find(self, path):
         """Finds the first matching subelement, by tag name or path.
         """
+        if isinstance(path, QName):
+            path = (<QName>path).text
         return _elementpath.find(self, path)
 
     def findtext(self, path, default=None):
         """Finds text for the first matching subelement, by tag name or path.
         """
+        if isinstance(path, QName):
+            path = (<QName>path).text
         return _elementpath.findtext(self, path, default)
 
     def findall(self, path):
         """Finds all matching subelements, by tag name or path.
         """
+        if isinstance(path, QName):
+            path = (<QName>path).text
         return _elementpath.findall(self, path)
 
     def xpath(self, _path, namespaces=None, extensions=None, **_variables):
@@ -1699,21 +1721,6 @@
 
 fromstring = XML
 
-cdef class QName:
-    """QName wrapper.
-    """
-    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 _isString(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):
     """Checks if an object appears to be a valid element object.
     """

From scoder at codespeak.net  Fri Dec  1 09:53:57 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  1 Dec 2006 09:53:57 +0100 (CET)
Subject: [Lxml-checkins] r35186 - lxml/branch/nscleanup
Message-ID: <20061201085357.30D091006C@code0.codespeak.net>

Author: scoder
Date: Fri Dec  1 09:53:51 2006
New Revision: 35186

Added:
   lxml/branch/nscleanup/
      - copied from r35185, lxml/trunk/
Log:
new branch to play around with a custom implementation of xmlReconciliateNs()


From scoder at codespeak.net  Fri Dec  1 09:55:25 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  1 Dec 2006 09:55:25 +0100 (CET)
Subject: [Lxml-checkins] r35187 - in lxml/branch/nscleanup/src/lxml: . tests
Message-ID: <20061201085525.D33051006C@code0.codespeak.net>

Author: scoder
Date: Fri Dec  1 09:55:17 2006
New Revision: 35187

Modified:
   lxml/branch/nscleanup/src/lxml/apihelpers.pxi
   lxml/branch/nscleanup/src/lxml/etree.pyx
   lxml/branch/nscleanup/src/lxml/etree_defs.h
   lxml/branch/nscleanup/src/lxml/proxy.pxi
   lxml/branch/nscleanup/src/lxml/public-api.pxi
   lxml/branch/nscleanup/src/lxml/tests/test_etree.py
   lxml/branch/nscleanup/src/lxml/tree.pxd
Log:
preliminary implementation with 3 failing test cases

Modified: lxml/branch/nscleanup/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/branch/nscleanup/src/lxml/apihelpers.pxi	(original)
+++ lxml/branch/nscleanup/src/lxml/apihelpers.pxi	Fri Dec  1 09:55:17 2006
@@ -147,7 +147,7 @@
             if attr_ns_utf is None:
                 tree.xmlNewProp(c_node, _cstr(attr_name_utf), _cstr(value_utf))
             else:
-                c_ns = doc._findOrBuildNodeNs(c_node, _cstr(attr_ns_utf))
+                c_ns = doc._findOrBuildNodeNs(c_node, _cstr(attr_ns_utf), NULL)
                 tree.xmlNewNsProp(c_node, c_ns,
                                   _cstr(attr_name_utf), _cstr(value_utf))
 
@@ -203,7 +203,8 @@
     if ns is None:
         tree.xmlSetProp(element._c_node, c_tag, c_value)
     else:
-        c_ns = element._doc._findOrBuildNodeNs(element._c_node, _cstr(ns))
+        c_ns = element._doc._findOrBuildNodeNs(element._c_node,
+                                               _cstr(ns), NULL)
         tree.xmlSetNsProp(element._c_node, c_ns, c_tag, c_value)
     return 0
 

Modified: lxml/branch/nscleanup/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/nscleanup/src/lxml/etree.pyx	(original)
+++ lxml/branch/nscleanup/src/lxml/etree.pyx	Fri Dec  1 09:55:17 2006
@@ -298,24 +298,46 @@
         self._ns_counter = self._ns_counter + 1
         return ns
 
-    cdef xmlNs* _findOrBuildNodeNs(self, xmlNode* c_node, char* href):
+    cdef xmlNs* _findOrBuildNodeNs(self, xmlNode* c_node,
+                                   char* c_href, char* c_prefix):
         """Get or create namespace structure for a node.
         """
+        cdef int i
         cdef xmlNs* c_ns
+        cdef xmlNs* c_doc_ns
         # look for existing ns
-        c_ns = tree.xmlSearchNsByHref(self._c_doc, c_node, href)
+        c_ns = tree.xmlSearchNsByHref(self._c_doc, c_node, c_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))
+
+        if c_prefix is NULL or \
+               tree.xmlSearchNs(self._c_doc, c_node, c_prefix) is not NULL:
+            # try to simulate ElementTree's namespace prefix creation
+            for i from 0 <= i < 10000:
+                prefix = self.buildNewPrefix()
+                c_prefix = _cstr(prefix)
+                # make sure it's not used already
+                if tree.xmlSearchNs(self._c_doc, c_node, c_prefix) is NULL:
+                    break
+            if i >= 10000:
+                # XXX too many prefixes in use - this is pretty bad!
+                return NULL
+
+        c_ns = tree.xmlNewNs(c_node, c_href, c_prefix)
+        return c_ns
+        c_doc_ns = self._c_doc.oldNs
+        if c_doc_ns is NULL:
+            # this will create the XML namespace:
+            c_doc_ns = tree.xmlSearchNs(self._c_doc, c_node, 'xml')
+        while c_doc_ns.next is not NULL:
+            c_doc_ns = c_doc_ns.next
+        c_doc_ns.next = c_ns
         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)
+        c_ns = self._findOrBuildNodeNs(c_node, href, NULL)
         tree.xmlSetNs(c_node, c_ns)
 
     cdef void _setNodeNamespaces(self, xmlNode* c_node,

Modified: lxml/branch/nscleanup/src/lxml/etree_defs.h
==============================================================================
--- lxml/branch/nscleanup/src/lxml/etree_defs.h	(original)
+++ lxml/branch/nscleanup/src/lxml/etree_defs.h	Fri Dec  1 09:55:17 2006
@@ -59,6 +59,11 @@
 	 ((c_node)->type == XML_COMMENT_NODE) || \
          ((c_node)->type == XML_PI_NODE))
 
+#define _isElementOrXInclude(c_node) \
+        (_isElement(c_node)                     || \
+         ((c_node)->type == XML_XINCLUDE_START) || \
+         ((c_node)->type == XML_XINCLUDE_END))
+
 #define _getNs(c_node) \
         (((c_node)->ns == 0) ? 0 : ((c_node)->ns->href))
 

Modified: lxml/branch/nscleanup/src/lxml/proxy.pxi
==============================================================================
--- lxml/branch/nscleanup/src/lxml/proxy.pxi	(original)
+++ lxml/branch/nscleanup/src/lxml/proxy.pxi	Fri Dec  1 09:55:17 2006
@@ -169,10 +169,11 @@
     tree below (including the current node). It also reconciliates
     namespaces so they're correct inside the new environment.
     """
-    tree.xmlReconciliateNs(doc._c_doc, node._c_node)
-    if node._doc is not doc:
-        node._doc = doc
-        changeDocumentBelow(node._c_node, doc)
+    _moveNodeToDocument(node, doc)
+##     tree.xmlReconciliateNs(doc._c_doc, node._c_node)
+##     if node._doc is not doc:
+##         node._doc = doc
+##         changeDocumentBelow(node._c_node, doc)
 
 cdef void changeDocumentBelow(xmlNode* c_parent, _Document doc):
     """Update the Python references in the tree below the node.
@@ -187,3 +188,136 @@
     if c_node._private is not NULL:
         (<_NodeBase>c_node._private)._doc = doc
     tree.END_FOR_EACH_ELEMENT_FROM(c_node)
+
+
+cdef void _moveNodeToDocument(_NodeBase node, _Document doc):
+    """Fix the xmlNs pointers of a node and its subtree that were moved.
+
+    Mainly copied from libxml2's xmlReconciliateNs().  Expects libxml2 doc
+    pointers of node to be correct already, but fixes _Document references.
+    """
+    cdef xmlDoc* c_doc
+    cdef xmlNode* c_element
+    cdef xmlNode* c_start_node
+    cdef xmlNode* c_node
+    cdef xmlNs** c_ns_new_cache
+    cdef xmlNs** c_ns_old_cache
+    cdef xmlNs* c_ns
+    cdef xmlNs* c_new_ns
+    cdef cstd.size_t i, c_cache_size, c_cache_last
+
+    c_element = node._c_node
+    c_doc = c_element.doc
+
+    if not tree._isElementOrXInclude(c_element):
+        return
+
+    c_start_node = c_element
+    c_ns_new_cache = NULL
+    c_ns_old_cache = NULL
+    c_cache_size = 0
+    c_cache_last = 0
+
+    while c_element is not NULL:
+        # remove namespaces defined here if already known in ancestors
+        if c_element.nsDef is not NULL:
+            while c_element.nsDef is not NULL and \
+                      tree.xmlSearchNsByHref(c_element.doc, c_element.parent,
+                                             c_element.nsDef.href) is not NULL:
+                c_element.nsDef = c_element.nsDef.next
+            if c_element.nsDef is not NULL:
+                c_ns = c_element.nsDef
+                while c_ns.next is not NULL:
+                    if tree.xmlSearchNsByHref(c_element.doc, c_element.parent,
+                                              c_ns.next.href) is not NULL:
+                        c_ns.next = c_ns.next.next
+                    else:
+                        c_ns = c_ns.next
+
+        # make sure ns declaration of element and its attributes is stored
+        # in this document
+        c_node = c_element
+        while c_node is not NULL:
+            if c_node.ns is not NULL:
+                c_ns = c_node.ns
+                for i from 0 <= i < c_cache_last:
+                    if c_ns is c_ns_old_cache[i]:
+                        c_node.ns = c_ns_new_cache[i]
+                        c_ns = NULL
+                        break
+
+                if c_ns is not NULL:
+                    # not in cache, must find a replacement from this document
+                    c_new_ns = doc._findOrBuildNodeNs(c_node, c_ns.href, c_ns.prefix)
+                    print "FOUND:", c_new_ns.href
+                    if c_cache_last >= c_cache_size:
+                        # must resize cache
+                        if c_cache_size == 0:
+                            c_cache_size = 20
+                        else:
+                            c_cache_size = c_cache_size * 2
+                        c_ns_new_cache = <xmlNs**> python.PyMem_Realloc(
+                            c_ns_new_cache, c_cache_size * sizeof(xmlNs*))
+                        if c_ns_new_cache is NULL:
+                            python.PyMem_Free(c_ns_old_cache)
+                            python.PyErr_NoMemory()
+                        c_ns_old_cache = <xmlNs**> python.PyMem_Realloc(
+                            c_ns_old_cache, c_cache_size * sizeof(xmlNs*))
+                        if c_ns_old_cache is NULL:
+                            python.PyMem_Free(c_ns_new_cache)
+                            python.PyErr_NoMemory()
+                    c_ns_new_cache[c_cache_last] = c_new_ns
+                    c_ns_old_cache[c_cache_last] = c_node.ns
+                    c_cache_last = c_cache_last + 1
+                    c_node.ns = c_ns
+            if c_node is c_element:
+                c_node = <xmlNode*>c_element.properties
+            else:
+                c_node = c_node.next
+                
+        # traverse to next element, start with children
+        c_node = c_element.children
+        while c_node is not NULL and \
+              not tree._isElementOrXInclude(c_node):
+            c_node = c_node.next
+
+        if c_node is NULL:
+            # no children => back off and continue with siblings and parents
+
+            # fix _Document reference (may dealloc the original document!)
+            if c_element._private is not NULL:
+                (<_NodeBase>c_element._private)._doc = doc
+
+            if c_element is c_start_node:
+                break
+
+            # continue with siblings
+            c_node = c_element.next
+            while (c_node is not NULL and
+                   not tree._isElementOrXInclude(c_node)):
+                c_node = c_node.next
+            # if that didn't help, back off through parents' siblings
+            while c_node is NULL:
+                c_element = c_element.parent
+                if c_element is NULL or not tree._isElementOrXInclude(c_element):
+                    break
+
+                # fix _Document reference (may dealloc the original document!)
+                if c_element._private is not NULL:
+                    (<_NodeBase>c_element._private)._doc = doc
+
+                if c_element is c_start_node:
+                    break
+                # parents already done -> look for their siblings
+                c_node = c_element.next
+                while (c_node is not NULL and
+                       not tree._isElementOrXInclude(c_node)):
+                    c_node = c_node.next
+        if c_node is c_start_node:
+            break
+        c_element = c_node
+
+    if c_ns_new_cache is not NULL:
+        python.PyMem_Free(c_ns_new_cache)
+    if c_ns_old_cache is not NULL:
+        python.PyMem_Free(c_ns_old_cache)

Modified: lxml/branch/nscleanup/src/lxml/public-api.pxi
==============================================================================
--- lxml/branch/nscleanup/src/lxml/public-api.pxi	(original)
+++ lxml/branch/nscleanup/src/lxml/public-api.pxi	Fri Dec  1 09:55:17 2006
@@ -138,4 +138,4 @@
                                           char* href) except NULL:
     if doc is None:
         raise TypeError
-    return doc._findOrBuildNodeNs(c_node, href)
+    return doc._findOrBuildNodeNs(c_node, href, NULL)

Modified: lxml/branch/nscleanup/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/nscleanup/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/nscleanup/src/lxml/tests/test_etree.py	Fri Dec  1 09:55:17 2006
@@ -893,6 +893,56 @@
             '<z xmlns="http://ns.infrae.com/foo" xmlns:hoi="http://ns.infrae.com/hoi"><hoi:x></hoi:x></z>',
             self._writeElement(e))
 
+    def test_namespaces_default_copy_element(self):
+        etree = self.etree
+
+        r = {None: 'http://ns.infrae.com/foo'}
+        e1 = etree.Element('{http://ns.infrae.com/foo}bar', nsmap=r)
+        e2 = etree.Element('{http://ns.infrae.com/foo}bar', nsmap=r)
+
+        e1.append(e2)
+
+        self.assertEquals(
+            None,
+            e1.prefix)
+        self.assertEquals(
+            None,
+            e2.prefix)
+        self.assertEquals(
+            '{http://ns.infrae.com/foo}bar',
+            e1.tag)
+        self.assertEquals(
+            '{http://ns.infrae.com/foo}bar',
+            e2.tag)
+
+        print etree.tostring(e1)
+        print etree.tostring(e2)
+
+    def test_namespaces_copy_element(self):
+        etree = self.etree
+
+        r = {None: 'http://ns.infrae.com/BAR'}
+        e1 = etree.Element('{http://ns.infrae.com/BAR}bar', nsmap=r)
+        e2 = etree.Element('{http://ns.infrae.com/foo}bar', nsmap=r)
+
+        e1.append(e2)
+
+        self.assertEquals(
+            None,
+            e1.prefix)
+        self.assertNotEquals(
+            None,
+            e2.prefix)
+        self.assertEquals(
+            '{http://ns.infrae.com/BAR}bar',
+            e1.tag)
+        self.assertEquals(
+            '{http://ns.infrae.com/foo}bar',
+            e2.tag)
+
+        print etree.tostring(e1)
+        print etree.tostring(e2)
+
     def test_element_nsmap(self):
         etree = self.etree
 

Modified: lxml/branch/nscleanup/src/lxml/tree.pxd
==============================================================================
--- lxml/branch/nscleanup/src/lxml/tree.pxd	(original)
+++ lxml/branch/nscleanup/src/lxml/tree.pxd	Fri Dec  1 09:55:17 2006
@@ -77,7 +77,6 @@
         XML_NAMESPACE_DECL=         18
         XML_XINCLUDE_START=         19
         XML_XINCLUDE_END=           20
-
     
     ctypedef struct xmlNs:
         char* href
@@ -193,7 +192,7 @@
     cdef xmlAttr* xmlHasProp(xmlNode* node, char* name)
     cdef xmlAttr* xmlHasNsProp(xmlNode* node, char* name, char* nameSpace)
     cdef char* xmlNodeGetContent(xmlNode* cur)
-    cdef xmlNs* xmlSearchNs(xmlDoc* doc, xmlNode* node, char* nameSpace)
+    cdef xmlNs* xmlSearchNs(xmlDoc* doc, xmlNode* node, char* prefix)
     cdef xmlNs* xmlSearchNsByHref(xmlDoc* doc, xmlNode* node, char* href)
     cdef int xmlIsBlankNode(xmlNode* node)
     cdef void xmlElemDump(FILE* f, xmlDoc* doc, xmlNode* cur)
@@ -206,6 +205,7 @@
     cdef xmlNode* xmlCopyNode(xmlNode* node, int extended)
     cdef xmlNode* xmlDocCopyNode(xmlNode* node, xmlDoc* doc, int extended)
     cdef int xmlReconciliateNs(xmlDoc* doc, xmlNode* tree)
+    cdef xmlNs* xmlNewReconciliedNs(xmlDoc* doc, xmlNode* tree, xmlNs* ns)
     cdef xmlBuffer* xmlBufferCreate()
     cdef char* xmlBufferContent(xmlBuffer* buf)
     cdef int xmlBufferLength(xmlBuffer* buf)
@@ -260,6 +260,7 @@
 
 cdef extern from "etree_defs.h":
     cdef int _isElement(xmlNode* node)
+    cdef int _isElementOrXInclude(xmlNode* node)
     cdef char* _getNs(xmlNode* node)
     cdef void BEGIN_FOR_EACH_ELEMENT_FROM(xmlNode* tree_top,
                                           xmlNode* start_node, int inclusive)

From faassen at codespeak.net  Fri Dec  1 13:46:59 2006
From: faassen at codespeak.net (faassen at codespeak.net)
Date: Fri,  1 Dec 2006 13:46:59 +0100 (CET)
Subject: [Lxml-checkins] r35199 - lxml/trunk/src/lxml/tests
Message-ID: <20061201124659.936D310068@code0.codespeak.net>

Author: faassen
Date: Fri Dec  1 13:46:56 2006
New Revision: 35199

Modified:
   lxml/trunk/src/lxml/tests/common_imports.py
   lxml/trunk/src/lxml/tests/test_elementtree.py
   lxml/trunk/src/lxml/tests/test_nsclasses.py
   lxml/trunk/src/lxml/tests/test_objectify.py
Log:
Some backwards compatibility work for Python 2.3.

Modified: lxml/trunk/src/lxml/tests/common_imports.py
==============================================================================
--- lxml/trunk/src/lxml/tests/common_imports.py	(original)
+++ lxml/trunk/src/lxml/tests/common_imports.py	Fri Dec  1 13:46:56 2006
@@ -23,6 +23,11 @@
     # we need our own version to make it work (Python 2.3?)
     import local_doctest as doctest
 
+try:
+    from operator import itemgetter
+except ImportError:
+    def itemgetter(item):
+        return lambda obj: obj[item]
 
 class HelperTestCase(unittest.TestCase):
     def parse(self, text):
@@ -32,6 +37,12 @@
     def _rootstring(self, tree):
         return etree.tostring(tree.getroot()).replace(' ', '').replace('\n', '')
 
+    # assertFalse doesn't exist in Python 2.3
+    try:
+        unittest.TestCase.assertFalse
+    except AttributeError:
+        assertFalse = unittest.TestCase.failIf
+        
 class SillyFileLike:
     def __init__(self, xml_data='<foo><bar/></foo>'):
         self.xml_data = xml_data

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	Fri Dec  1 13:46:56 2006
@@ -11,9 +11,10 @@
 import unittest, doctest
 import os, re, shutil, tempfile, copy
 
-from common_imports import StringIO, etree, ElementTree, HelperTestCase, fileInTestDir, canonicalize
+from common_imports import StringIO, etree, ElementTree
+from common_imports import HelperTestCase, fileInTestDir, canonicalize
 
-class ETreeTestCaseBase(unittest.TestCase):
+class ETreeTestCaseBase(HelperTestCase):
     etree = None
     
     def setUp(self):

Modified: lxml/trunk/src/lxml/tests/test_nsclasses.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_nsclasses.py	(original)
+++ lxml/trunk/src/lxml/tests/test_nsclasses.py	Fri Dec  1 13:46:56 2006
@@ -10,7 +10,6 @@
 from common_imports import etree, HelperTestCase, doctest
 
 class ETreeNamespaceClassesTestCase(HelperTestCase):
-    assertFalse = HelperTestCase.failIf
     
     class default_class(etree.ElementBase):
         pass

Modified: lxml/trunk/src/lxml/tests/test_objectify.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_objectify.py	(original)
+++ lxml/trunk/src/lxml/tests/test_objectify.py	Fri Dec  1 13:46:56 2006
@@ -12,6 +12,7 @@
 
 from common_imports import etree, StringIO, HelperTestCase, fileInTestDir
 from common_imports import SillyFileLike, canonicalize, doctest
+from common_imports import itemgetter
 
 from lxml import objectify
 
@@ -119,7 +120,7 @@
         self.assertEquals("0", root.c1.c2[0].text)
         self.assertEquals("1", root.c1.c2[1].text)
         self.assertEquals("2", root.c1.c2[2].text)
-        self.assertRaises(IndexError, operator.itemgetter(3), root.c1.c2)
+        self.assertRaises(IndexError, itemgetter(3), root.c1.c2)
 
     def test_child_index_neg(self):
         root = self.XML(xml_str)
@@ -127,7 +128,7 @@
         self.assertEquals("0", root.c1.c2[-3].text)
         self.assertEquals("1", root.c1.c2[-2].text)
         self.assertEquals("2", root.c1.c2[-1].text)
-        self.assertRaises(IndexError, operator.itemgetter(-4), root.c1.c2)
+        self.assertRaises(IndexError, itemgetter(-4), root.c1.c2)
 
     def test_child_len(self):
         root = self.XML(xml_str)

From scoder at codespeak.net  Fri Dec  1 14:32:34 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  1 Dec 2006 14:32:34 +0100 (CET)
Subject: [Lxml-checkins] r35201 - lxml/trunk/src/lxml/tests
Message-ID: <20061201133234.0784A10063@code0.codespeak.net>

Author: scoder
Date: Fri Dec  1 14:32:32 2006
New Revision: 35201

Modified:
   lxml/trunk/src/lxml/tests/test_objectify.py
Log:
rewrote test case to check Element identity instead of equality

Modified: lxml/trunk/src/lxml/tests/test_objectify.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_objectify.py	(original)
+++ lxml/trunk/src/lxml/tests/test_objectify.py	Fri Dec  1 14:32:32 2006
@@ -214,8 +214,7 @@
         self.assertEquals(1, len(root.findall("c")))
         self.assertEquals(2, len(root.findall(".//c")))
         self.assertEquals(3, len(root.findall(".//b")))
-        self.assertEquals(root.findall(".//b")[:2],
-                          root.getchildren()[:2])
+        self.assert_(root.findall(".//b")[1] is root.getchildren()[1])
 
     def test_findall_ns(self):
         XML = self.XML

From scoder at codespeak.net  Mon Dec  4 10:20:40 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Mon,  4 Dec 2006 10:20:40 +0100 (CET)
Subject: [Lxml-checkins] r35240 - in lxml/trunk: . src/lxml
Message-ID: <20061204092040.5839610068@code0.codespeak.net>

Author: scoder
Date: Mon Dec  4 10:20:37 2006
New Revision: 35240

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/ElementInclude.py
Log:
small fix to actually check if xi:fallback tag is misplaced before complaining

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Mon Dec  4 10:20:37 2006
@@ -8,6 +8,9 @@
 Features added
 --------------
 
+* ElementInclude module for ElementTree compatible XInclude processing that
+  honours custom resolvers registered with the source document
+
 * ElementTree.parser property holds the parser used to parse the document
 
 * setup.py has been refactored for greater readability and flexibility.

Modified: lxml/trunk/src/lxml/ElementInclude.py
==============================================================================
--- lxml/trunk/src/lxml/ElementInclude.py	(original)
+++ lxml/trunk/src/lxml/ElementInclude.py	Mon Dec  4 10:20:37 2006
@@ -186,9 +186,11 @@
                     "unknown parse type in xi:include tag (%r)" % parse
                 )
         elif e.tag == XINCLUDE_FALLBACK:
-            raise FatalIncludeError(
-                "xi:fallback tag must be child of xi:include (%r)" % e.tag
-                )
+            parent = e.getparent()
+            if parent is not None and parent.tag != XINCLUDE_INCLUDE:
+                raise FatalIncludeError(
+                    "xi:fallback tag must be child of xi:include (%r)" % e.tag
+                    )
         else:
             raise FatalIncludeError(
                 "Invalid element found in XInclude namespace (%r)" % e.tag

From scoder at codespeak.net  Mon Dec  4 10:25:32 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Mon,  4 Dec 2006 10:25:32 +0100 (CET)
Subject: [Lxml-checkins] r35241 - in lxml/branch/nscleanup/src/lxml: . tests
Message-ID: <20061204092532.153BF10068@code0.codespeak.net>

Author: scoder
Date: Mon Dec  4 10:25:29 2006
New Revision: 35241

Modified:
   lxml/branch/nscleanup/src/lxml/etree.pyx
   lxml/branch/nscleanup/src/lxml/proxy.pxi
   lxml/branch/nscleanup/src/lxml/tests/test_etree.py
Log:
cleanup, new test case and bugfix for stupid typo

Modified: lxml/branch/nscleanup/src/lxml/etree.pyx
==============================================================================
--- lxml/branch/nscleanup/src/lxml/etree.pyx	(original)
+++ lxml/branch/nscleanup/src/lxml/etree.pyx	Mon Dec  4 10:25:29 2006
@@ -323,16 +323,7 @@
                 # XXX too many prefixes in use - this is pretty bad!
                 return NULL
 
-        c_ns = tree.xmlNewNs(c_node, c_href, c_prefix)
-        return c_ns
-        c_doc_ns = self._c_doc.oldNs
-        if c_doc_ns is NULL:
-            # this will create the XML namespace:
-            c_doc_ns = tree.xmlSearchNs(self._c_doc, c_node, 'xml')
-        while c_doc_ns.next is not NULL:
-            c_doc_ns = c_doc_ns.next
-        c_doc_ns.next = c_ns
-        return c_ns
+        return tree.xmlNewNs(c_node, c_href, c_prefix)
 
     cdef void _setNodeNs(self, xmlNode* c_node, char* href):
         "Lookup namespace structure and set it for the node."

Modified: lxml/branch/nscleanup/src/lxml/proxy.pxi
==============================================================================
--- lxml/branch/nscleanup/src/lxml/proxy.pxi	(original)
+++ lxml/branch/nscleanup/src/lxml/proxy.pxi	Mon Dec  4 10:25:29 2006
@@ -1,4 +1,4 @@
-# Proxy functions
+# Proxy functions and low level node allocation stuff
 
 # Proxies represent elements, their reference is stored in the C
 # structure of the respective node to avoid multiple instantiation of
@@ -234,7 +234,7 @@
                     else:
                         c_ns = c_ns.next
 
-        # make sure ns declaration of element and its attributes is stored
+        # make sure the namespace of an element and its attributes is declared
         # in this document
         c_node = c_element
         while c_node is not NULL:
@@ -249,7 +249,6 @@
                 if c_ns is not NULL:
                     # not in cache, must find a replacement from this document
                     c_new_ns = doc._findOrBuildNodeNs(c_node, c_ns.href, c_ns.prefix)
-                    print "FOUND:", c_new_ns.href
                     if c_cache_last >= c_cache_size:
                         # must resize cache
                         if c_cache_size == 0:
@@ -269,12 +268,13 @@
                     c_ns_new_cache[c_cache_last] = c_new_ns
                     c_ns_old_cache[c_cache_last] = c_node.ns
                     c_cache_last = c_cache_last + 1
-                    c_node.ns = c_ns
+                    c_node.ns = c_new_ns
             if c_node is c_element:
+                # after the element, continue with its attributes
                 c_node = <xmlNode*>c_element.properties
             else:
                 c_node = c_node.next
-                
+
         # traverse to next element, start with children
         c_node = c_element.children
         while c_node is not NULL and \
@@ -317,6 +317,7 @@
             break
         c_element = c_node
 
+    # cleanup
     if c_ns_new_cache is not NULL:
         python.PyMem_Free(c_ns_new_cache)
     if c_ns_old_cache is not NULL:

Modified: lxml/branch/nscleanup/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/nscleanup/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/nscleanup/src/lxml/tests/test_etree.py	Mon Dec  4 10:25:29 2006
@@ -943,6 +943,25 @@
         print etree.tostring(e1)
         print etree.tostring(e2)
 
+    def test_namespaces_reuse_after_move(self):
+        Element = self.etree.Element
+        ElementTree = self.etree.ElementTree
+
+        ns_href = "http://a.b.c"
+        one = self.etree.parse(
+            StringIO('<foo><bar xmlns:ns="%s"><ns:baz/></bar></foo>' % ns_href))
+        baz = one.getroot()[0][0]
+
+        two = self.etree.parse(
+            StringIO('<root xmlns:ns="%s"/>' % ns_href))
+        two.getroot().append(baz)
+        del one # make sure the source document is deallocated
+
+        self.assertEquals('{%s}baz' % ns_href, baz.tag)
+        self.assertEquals(
+            '<root xmlns:ns="%s"><ns:baz/></root>' % ns_href,
+            self.etree.tostring(two))
+
     def test_element_nsmap(self):
         etree = self.etree
 

From scoder at codespeak.net  Mon Dec  4 17:32:10 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Mon,  4 Dec 2006 17:32:10 +0100 (CET)
Subject: [Lxml-checkins] r35260 - in lxml/trunk: . src/lxml
Message-ID: <20061204163210.1E6CB1007F@code0.codespeak.net>

Author: scoder
Date: Mon Dec  4 17:32:08 2006
New Revision: 35260

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/etree_defs.h
Log:
threading crash also appears under 2.4.1, not only 2.3

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Mon Dec  4 17:32:08 2006
@@ -13,7 +13,7 @@
 
 * ElementTree.parser property holds the parser used to parse the document
 
-* setup.py has been refactored for greater readability and flexibility.
+* setup.py has been refactored for greater readability and flexibility
 
 * --rpath flag to setup.py to induce automatic linking-in of dynamic library
   runtime search paths has been renamed to --auto-rpath. This makes it
@@ -23,6 +23,9 @@
 Bugs fixed
 ----------
 
+* Error handling could crash under Python <= 2.4.1 - fixed by disabling thread
+  support in these environments
+
 * Element.find*() did not accept QName objects as path
 
 

Modified: lxml/trunk/src/lxml/etree_defs.h
==============================================================================
--- lxml/trunk/src/lxml/etree_defs.h	(original)
+++ lxml/trunk/src/lxml/etree_defs.h	Mon Dec  4 17:32:08 2006
@@ -16,8 +16,8 @@
 #endif
 #endif
 
-/* Threading can crash under Python 2.3 */
-#if PY_VERSION_HEX < 0x02040000
+/* Threading can crash under Python <= 2.4.1 */
+#if PY_VERSION_HEX < 0x02040200
 #ifndef WITHOUT_THREADING
   #define WITHOUT_THREADING
 #endif

From scoder at codespeak.net  Thu Dec  7 08:59:56 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Thu,  7 Dec 2006 08:59:56 +0100 (CET)
Subject: [Lxml-checkins] r35415 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20061207075956.18FCC1007B@code0.codespeak.net>

Author: scoder
Date: Thu Dec  7 08:59:52 2006
New Revision: 35415

Modified:
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/python.pxd
   lxml/trunk/src/lxml/serializer.pxi
   lxml/trunk/src/lxml/tests/test_elementtree.py
Log:
ElementTree.write() did not raise an exception when the file was not writable

Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Thu Dec  7 08:59:52 2006
@@ -23,6 +23,8 @@
 Bugs fixed
 ----------
 
+* ElementTree.write() did not raise an exception when the file was not writable
+
 * Error handling could crash under Python <= 2.4.1 - fixed by disabling thread
   support in these environments
 

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Thu Dec  7 08:59:52 2006
@@ -73,7 +73,10 @@
     cdef void* PyMem_Malloc(size_t size)
     cdef void* PyMem_Realloc(void* p, size_t size)
     cdef void PyMem_Free(void* p)
-    cdef object PyErr_NoMemory() # always returns NULL to pass on exception
+
+    # these two always return NULL to pass on the exception
+    cdef object PyErr_NoMemory()
+    cdef object PyErr_SetFromErrno(object type)
 
     ctypedef enum PyGILState_STATE:
         PyGILState_LOCKED

Modified: lxml/trunk/src/lxml/serializer.pxi
==============================================================================
--- lxml/trunk/src/lxml/serializer.pxi	(original)
+++ lxml/trunk/src/lxml/serializer.pxi	Thu Dec  7 08:59:52 2006
@@ -166,6 +166,8 @@
         filename8 = _encodeFilename(f)
         c_buffer = tree.xmlOutputBufferCreateFilename(
             _cstr(filename8), enchandler, 0)
+        if c_buffer is NULL:
+            python.PyErr_SetFromErrno(IOError)
         state = python.PyEval_SaveThread()
     elif hasattr(f, 'write'):
         writer   = _FilelikeWriter(f)
@@ -217,10 +219,11 @@
         writer._exc_context._raise_if_stored()
 
     if bytes < 0:
-        if writer is not None and len(writer.error_log):
-            message = writer.error_log[0].message
-        else:
-            message = "C14N failed"
+        message = "C14N failed"
+        if writer is not None:
+            errors = writer.error_log
+            if len(errors):
+                message = errors[0].message
         raise C14NError, message
 
 # dump node to file (mainly for debug)

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 Dec  7 08:59:52 2006
@@ -623,6 +623,14 @@
             self.assertEquals(
                 '<doc%s>This is a test.</doc%s>' % (i, i),
                 canonicalize(data))
+        
+    def test_write_fail(self):
+        ElementTree = self.etree.ElementTree
+        XML = self.etree.XML
+
+        tree = ElementTree( XML('<doc>This is a test.</doc>') )
+        self.assertRaises(IOError, tree.write,
+                          "definitely////\\-\\nonexisting\\-\\////FILE")
 
     # this could trigger a crash, apparently because the document
     # reference was prematurely garbage collected

From scoder at codespeak.net  Tue Dec 12 10:44:16 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 12 Dec 2006 10:44:16 +0100 (CET)
Subject: [Lxml-checkins] r35615 - lxml/branch/nscleanup/src/lxml
Message-ID: <20061212094416.ABB4B10070@code0.codespeak.net>

Author: scoder
Date: Tue Dec 12 10:44:12 2006
New Revision: 35615

Modified:
   lxml/branch/nscleanup/src/lxml/proxy.pxi
Log:
fix for namespace cleanup

Modified: lxml/branch/nscleanup/src/lxml/proxy.pxi
==============================================================================
--- lxml/branch/nscleanup/src/lxml/proxy.pxi	(original)
+++ lxml/branch/nscleanup/src/lxml/proxy.pxi	Tue Dec 12 10:44:12 2006
@@ -45,30 +45,31 @@
     # always call _destroyFakeDoc() after use!
     cdef xmlNode* c_child
     cdef xmlNode* c_root
+    cdef xmlNode* c_new_root
     cdef xmlDoc*  c_doc
     c_root = tree.xmlDocGetRootElement(c_base_doc)
     if c_root is c_node:
         # already the root node
         return c_base_doc
 
-    c_doc  = _copyDoc(c_base_doc, 0)               # non recursive!
-    c_root = tree.xmlDocCopyNode(c_node, c_doc, 2) # non recursive!
-    tree.xmlDocSetRootElement(c_doc, c_root)
-
-    c_root.children = c_node.children
-    c_root.last = c_node.last
-    c_root.next = c_root.prev = c_root.parent = NULL
+    c_doc  = _copyDoc(c_base_doc, 0)                   # non recursive!
+    c_new_root = tree.xmlDocCopyNode(c_node, c_doc, 2) # non recursive!
+    tree.xmlDocSetRootElement(c_doc, c_new_root)
+
+    c_new_root.children = c_node.children
+    c_new_root.last = c_node.last
+    c_new_root.next = c_new_root.prev = c_new_root.parent = NULL
 
     # store original node
     c_doc._private = c_node
 
     # divert parent pointers of children
-    c_child = c_root.children
+    c_child = c_new_root.children
     while c_child is not NULL:
-        c_child.parent = c_root
+        c_child.parent = c_new_root
         c_child = c_child.next
 
-    c_doc.children = c_root
+    c_doc.children = c_new_root
     return c_doc
 
 cdef void _destroyFakeDoc(xmlDoc* c_base_doc, xmlDoc* c_doc):
@@ -219,20 +220,27 @@
     c_cache_last = 0
 
     while c_element is not NULL:
-        # remove namespaces defined here if already known in ancestors
+        # remove namespaces defined here that are known in the new ancestors
         if c_element.nsDef is not NULL:
-            while c_element.nsDef is not NULL and \
-                      tree.xmlSearchNsByHref(c_element.doc, c_element.parent,
-                                             c_element.nsDef.href) is not NULL:
+            while c_element.nsDef is not NULL:
+                c_ns = tree.xmlSearchNsByHref(
+                    c_element.doc, c_element.parent, c_element.nsDef.href)
+                if c_ns is NULL:
+                    break
                 c_element.nsDef = c_element.nsDef.next
             if c_element.nsDef is not NULL:
-                c_ns = c_element.nsDef
-                while c_ns.next is not NULL:
-                    if tree.xmlSearchNsByHref(c_element.doc, c_element.parent,
-                                              c_ns.next.href) is not NULL:
-                        c_ns.next = c_ns.next.next
+                c_new_ns = c_element.nsDef
+                while c_new_ns.next is not NULL:
+                    if c_new_ns.next is not c_element.ns:
+                        c_ns = tree.xmlSearchNsByHref(
+                            c_element.doc, c_element.parent, c_new_ns.next.href)
+                        if c_ns is not NULL:
+                            # not known or at least not different
+                            c_new_ns.next = c_new_ns.next.next
+                        else:
+                            c_new_ns = c_new_ns.next
                     else:
-                        c_ns = c_ns.next
+                        c_new_ns = c_new_ns.next
 
         # make sure the namespace of an element and its attributes is declared
         # in this document

From scoder at codespeak.net  Tue Dec 12 10:44:56 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 12 Dec 2006 10:44:56 +0100 (CET)
Subject: [Lxml-checkins] r35616 - lxml/branch/nscleanup/src/lxml/tests
Message-ID: <20061212094456.27DC010070@code0.codespeak.net>

Author: scoder
Date: Tue Dec 12 10:44:53 2006
New Revision: 35616

Modified:
   lxml/branch/nscleanup/src/lxml/tests/test_etree.py
Log:
cleanup, new test case for namespace serialisation (fails)

Modified: lxml/branch/nscleanup/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/nscleanup/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/nscleanup/src/lxml/tests/test_etree.py	Tue Dec 12 10:44:53 2006
@@ -907,16 +907,13 @@
             e1.prefix)
         self.assertEquals(
             None,
-            e2.prefix)
+            e1[0].prefix)
         self.assertEquals(
             '{http://ns.infrae.com/foo}bar',
             e1.tag)
         self.assertEquals(
             '{http://ns.infrae.com/foo}bar',
-            e2.tag)
-
-        print etree.tostring(e1)
-        print etree.tostring(e2)
+            e1[0].tag)
 
     def test_namespaces_copy_element(self):
         etree = self.etree
@@ -940,13 +937,7 @@
             '{http://ns.infrae.com/foo}bar',
             e2.tag)
 
-        print etree.tostring(e1)
-        print etree.tostring(e2)
-
     def test_namespaces_reuse_after_move(self):
-        Element = self.etree.Element
-        ElementTree = self.etree.ElementTree
-
         ns_href = "http://a.b.c"
         one = self.etree.parse(
             StringIO('<foo><bar xmlns:ns="%s"><ns:baz/></bar></foo>' % ns_href))
@@ -962,6 +953,20 @@
             '<root xmlns:ns="%s"><ns:baz/></root>' % ns_href,
             self.etree.tostring(two))
 
+    def test_namespaces_after_serialize(self):
+        parse = self.etree.parse
+        tostring = self.etree.tostring
+
+        ns_href = "http://a.b.c"
+        one = parse(
+            StringIO('<foo><bar xmlns:ns="%s"><ns:baz/></bar></foo>' % ns_href))
+        baz = one.getroot()[0][0]
+
+        print tostring(baz)
+        parsed = parse(StringIO( tostring(baz) )).getroot()
+
+        self.assertEquals('{%s}baz' % ns_href, parsed.tag)
+
     def test_element_nsmap(self):
         etree = self.etree
 

From scoder at codespeak.net  Tue Dec 12 10:45:24 2006
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 12 Dec 2006 10:45:24 +0100 (CET)
Subject: [Lxml-checkins] r35617 - lxml/branch/nscleanup/src/lxml
Message-ID: <20061212094524.5ACA510070@code0.codespeak.net>

Author: scoder
Date: Tue Dec 12 10:45:21 2006
New Revision: 35617

Modified:
   lxml/branch/nscleanup/src/lxml/parser.pxi
Log:
enable COMPACT parsing

Modified: lxml/branch/nscleanup/src/lxml/parser.pxi
==============================================================================
--- lxml/branch/nscleanup/src/lxml/parser.pxi	(original)
+++ lxml/branch/nscleanup/src/lxml/parser.pxi	Tue Dec 12 10:45:21 2006
@@ -614,7 +614,8 @@
 cdef int _XML_DEFAULT_PARSE_OPTIONS
 _XML_DEFAULT_PARSE_OPTIONS = (
     xmlparser.XML_PARSE_NOENT |
-    xmlparser.XML_PARSE_NOCDATA
+    xmlparser.XML_PARSE_NOCDATA |
+    xmlparser.XML_PARSE_COMPACT
     )
 
 cdef class XMLParser(_BaseParser):
@@ -754,7 +755,8 @@
 ############################################################
 
 cdef int _HTML_DEFAULT_PARSE_OPTIONS
-_HTML_DEFAULT_PARSE_OPTIONS = 0
+_HTML_DEFAULT_PARSE_OPTIONS = \
+    htmlparser.HTML_PARSE_COMPACT
 
 cdef class HTMLParser(_BaseParser):
     """The HTML parser.  This parser allows reading HTML into a normal XML