From scoder at codespeak.net  Tue Sep  2 22:10:00 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue,  2 Sep 2008 22:10:00 +0200 (CEST)
Subject: [Lxml-checkins] r57776 - in lxml/branch/lxml-2.1: . src/lxml
	src/lxml/tests
Message-ID: <20080902201000.DB553169F97@codespeak.net>

Author: scoder
Date: Tue Sep  2 22:09:53 2008
New Revision: 57776

Modified:
   lxml/branch/lxml-2.1/CHANGES.txt
   lxml/branch/lxml-2.1/src/lxml/parsertarget.pxi
   lxml/branch/lxml-2.1/src/lxml/tests/test_etree.py
Log:
let target parser honour recover option

Modified: lxml/branch/lxml-2.1/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.1/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.1/CHANGES.txt	Tue Sep  2 22:09:53 2008
@@ -16,6 +16,9 @@
 Bugs fixed
 ----------
 
+* Target parser did not honour the ``recover`` option and raised an
+  exception instead of calling ``.close()`` on the target.
+
 Other changes
 -------------
 

Modified: lxml/branch/lxml-2.1/src/lxml/parsertarget.pxi
==============================================================================
--- lxml/branch/lxml-2.1/src/lxml/parsertarget.pxi	(original)
+++ lxml/branch/lxml-2.1/src/lxml/parsertarget.pxi	Tue Sep  2 22:09:53 2008
@@ -110,13 +110,17 @@
 
     cdef object _handleParseResult(self, _BaseParser parser, xmlDoc* result,
                                    filename):
-        if not self._c_ctxt.wellFormed:
+        cdef bint recover
+        recover = parser._parse_options & xmlparser.XML_PARSE_RECOVER
+        if not self._c_ctxt.wellFormed and not recover:
             _raiseParseError(self._c_ctxt, filename, self._error_log)
         self._raise_if_stored()
         return self._python_target.close()
 
     cdef xmlDoc* _handleParseResultDoc(self, _BaseParser parser,
                                        xmlDoc* result, filename) except NULL:
+        cdef bint recover
+        recover = parser._parse_options & xmlparser.XML_PARSE_RECOVER
         if result is not NULL and result._private is NULL:
             # no _Document proxy => orphen
             tree.xmlFreeDoc(result)
@@ -126,7 +130,7 @@
                 # no _Document proxy => orphen
                 tree.xmlFreeDoc(self._c_ctxt.myDoc)
             self._c_ctxt.myDoc = NULL
-        if not self._c_ctxt.wellFormed:
+        if not self._c_ctxt.wellFormed and not recover:
             _raiseParseError(self._c_ctxt, filename, self._error_log)
         self._raise_if_stored()
         raise _TargetParserResult(self._python_target.close())

Modified: lxml/branch/lxml-2.1/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/lxml-2.1/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/lxml-2.1/src/lxml/tests/test_etree.py	Tue Sep  2 22:09:53 2008
@@ -525,6 +525,31 @@
                            "data-ca", "end-a", "data-B", "end-root"],
                           events)
 
+    def test_parser_target_recover(self):
+        events = []
+        class Target(object):
+            def start(self, tag, attrib):
+                events.append("start-" + tag)
+            def end(self, tag):
+                events.append("end-" + tag)
+            def data(self, data):
+                events.append("data-" + data)
+            def close(self):
+                events.append("close")
+                return "DONE"
+
+        parser = self.etree.XMLParser(target=Target(),
+                                      recover=True)
+
+        parser.feed(_bytes('<root>A<a>ca</a>B</not-root>'))
+        done = parser.close()
+
+        self.assertEquals("DONE", done)
+        self.assertEquals(["start-root", "data-A", "start-a",
+                           "data-ca", "end-a", "data-B",
+                           "end-root", "close"],
+                          events)
+
     def test_iterwalk_tag(self):
         iterwalk = self.etree.iterwalk
         root = self.etree.XML(_bytes('<a><b><d/></b><c/></a>'))

From scoder at codespeak.net  Tue Sep  2 22:17:03 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue,  2 Sep 2008 22:17:03 +0200 (CEST)
Subject: [Lxml-checkins] r57778 - in lxml/branch/lxml-2.0: . src/lxml
	src/lxml/tests
Message-ID: <20080902201703.89A67169F97@codespeak.net>

Author: scoder
Date: Tue Sep  2 22:16:59 2008
New Revision: 57778

Modified:
   lxml/branch/lxml-2.0/CHANGES.txt
   lxml/branch/lxml-2.0/src/lxml/parsertarget.pxi
   lxml/branch/lxml-2.0/src/lxml/tests/test_etree.py
Log:
let target parser honour recover option

Modified: lxml/branch/lxml-2.0/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.0/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.0/CHANGES.txt	Tue Sep  2 22:16:59 2008
@@ -2,6 +2,22 @@
 lxml changelog
 ==============
 
+Under development
+=================
+
+Features added
+--------------
+
+Bugs fixed
+----------
+
+* Target parser did not honour the ``recover`` option and raised an
+  exception instead of calling ``.close()`` on the target.
+
+Other changes
+-------------
+
+
 2.0.8 (2008-07-24)
 ==================
 

Modified: lxml/branch/lxml-2.0/src/lxml/parsertarget.pxi
==============================================================================
--- lxml/branch/lxml-2.0/src/lxml/parsertarget.pxi	(original)
+++ lxml/branch/lxml-2.0/src/lxml/parsertarget.pxi	Tue Sep  2 22:16:59 2008
@@ -110,13 +110,17 @@
 
     cdef object _handleParseResult(self, _BaseParser parser, xmlDoc* result,
                                    filename):
-        if not self._c_ctxt.wellFormed:
+        cdef bint recover
+        recover = parser._parse_options & xmlparser.XML_PARSE_RECOVER
+        if not self._c_ctxt.wellFormed and not recover:
             _raiseParseError(self._c_ctxt, filename, self._error_log)
         self._raise_if_stored()
         return self._python_target.close()
 
     cdef xmlDoc* _handleParseResultDoc(self, _BaseParser parser,
                                        xmlDoc* result, filename) except NULL:
+        cdef bint recover
+        recover = parser._parse_options & xmlparser.XML_PARSE_RECOVER
         if result is not NULL and result._private is NULL:
             # no _Document proxy => orphen
             tree.xmlFreeDoc(result)
@@ -126,7 +130,7 @@
                 # no _Document proxy => orphen
                 tree.xmlFreeDoc(self._c_ctxt.myDoc)
             self._c_ctxt.myDoc = NULL
-        if not self._c_ctxt.wellFormed:
+        if not self._c_ctxt.wellFormed and not recover:
             _raiseParseError(self._c_ctxt, filename, self._error_log)
         self._raise_if_stored()
         raise _TargetParserResult(self._python_target.close())

Modified: lxml/branch/lxml-2.0/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/lxml-2.0/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/lxml-2.0/src/lxml/tests/test_etree.py	Tue Sep  2 22:16:59 2008
@@ -439,6 +439,31 @@
                           tree.parse, StringIO("<TAG/>"), parser=parser)
         self.assertEquals(["start", "end"], events)
 
+    def test_parser_target_recover(self):
+        events = []
+        class Target(object):
+            def start(self, tag, attrib):
+                events.append("start-" + tag)
+            def end(self, tag):
+                events.append("end-" + tag)
+            def data(self, data):
+                events.append("data-" + data)
+            def close(self):
+                events.append("close")
+                return "DONE"
+
+        parser = self.etree.XMLParser(target=Target(),
+                                      recover=True)
+
+        parser.feed('<root>A<a>ca</a>B</not-root>')
+        done = parser.close()
+
+        self.assertEquals("DONE", done)
+        self.assertEquals(["start-root", "data-A", "start-a",
+                           "data-ca", "end-a", "data-B",
+                           "end-root", "close"],
+                          events)
+
     def test_iterwalk_tag(self):
         iterwalk = self.etree.iterwalk
         root = self.etree.XML('<a><b><d/></b><c/></a>')

From scoder at codespeak.net  Tue Sep  2 22:18:55 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue,  2 Sep 2008 22:18:55 +0200 (CEST)
Subject: [Lxml-checkins] r57779 - in lxml/trunk: . src/lxml
Message-ID: <20080902201855.9228C169F95@codespeak.net>

Author: scoder
Date: Tue Sep  2 22:18:46 2008
New Revision: 57779

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/lxml.etree.pyx
Log:
 r4754 at delle:  sbehnel | 2008-09-01 15:51:55 +0200
 small fixes


Modified: lxml/trunk/src/lxml/lxml.etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/lxml.etree.pyx	(original)
+++ lxml/trunk/src/lxml/lxml.etree.pyx	Tue Sep  2 22:18:46 2008
@@ -38,7 +38,7 @@
 cdef object BytesIO, StringIO
 try:
     from io import BytesIO, StringIO
-except ImportError:
+except (ImportError, AttributeError):
     from StringIO import StringIO, StringIO as BytesIO
 
 cdef object _elementpath
@@ -103,7 +103,7 @@
     # xml schema
     "http://www.w3.org/2001/XMLSchema": "xs",
     "http://www.w3.org/2001/XMLSchema-instance": "xsi",
-    # dublic core
+    # dublin core
     "http://purl.org/dc/elements/1.1/": "dc",
     # objectify
     "http://codespeak.net/lxml/objectify/pytype" : "py",

From scoder at codespeak.net  Tue Sep  2 22:19:01 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue,  2 Sep 2008 22:19:01 +0200 (CEST)
Subject: [Lxml-checkins] r57780 - in lxml/trunk: . src/lxml/tests
Message-ID: <20080902201901.283B2169F95@codespeak.net>

Author: scoder
Date: Tue Sep  2 22:18:59 2008
New Revision: 57780

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/tests/test_xpathevaluator.py
Log:
 r4755 at delle:  sbehnel | 2008-09-01 15:52:26 +0200
 disabled broken test that uses non-ASCII characters in URIs


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	Tue Sep  2 22:18:59 2008
@@ -555,19 +555,21 @@
         self.assertEquals(1, len(r))
         self.assertEquals('{nsb}b', r[0].tag)
 
-    def test_xpath_compile_unicode(self):
-        x = self.parse(_bytes('<a><b xmlns="nsa\\uf8d2"/><b xmlns="nsb\\uf8d1"/></a>'
+    # disabled this test as non-ASCII characters in namespace URIs are
+    # not acceptable
+    def _test_xpath_compile_unicode(self):
+        x = self.parse(_bytes('<a><b xmlns="http://nsa/\\uf8d2"/><b xmlns="http://nsb/\\uf8d1"/></a>'
                               ).decode("unicode_escape"))
 
-        expr = etree.ETXPath(_bytes("/a/{nsa\\uf8d2}b").decode("unicode_escape"))
+        expr = etree.ETXPath(_bytes("/a/{http://nsa/\\uf8d2}b").decode("unicode_escape"))
         r = expr(x)
         self.assertEquals(1, len(r))
-        self.assertEquals(_bytes('{nsa\\uf8d2}b').decode("unicode_escape"), r[0].tag)
+        self.assertEquals(_bytes('{http://nsa/\\uf8d2}b').decode("unicode_escape"), r[0].tag)
 
-        expr = etree.ETXPath(_bytes("/a/{nsb\\uf8d1}b").decode("unicode_escape"))
+        expr = etree.ETXPath(_bytes("/a/{http://nsb/\\uf8d1}b").decode("unicode_escape"))
         r = expr(x)
         self.assertEquals(1, len(r))
-        self.assertEquals(_bytes('{nsb\\uf8d1}b').decode("unicode_escape"), r[0].tag)
+        self.assertEquals(_bytes('{http://nsb/\\uf8d1}b').decode("unicode_escape"), r[0].tag)
 
 SAMPLE_XML = etree.parse(BytesIO("""
 <body>

From scoder at codespeak.net  Tue Sep  2 22:19:15 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue,  2 Sep 2008 22:19:15 +0200 (CEST)
Subject: [Lxml-checkins] r57781 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20080902201915.B39A9169F95@codespeak.net>

Author: scoder
Date: Tue Sep  2 22:19:08 2008
New Revision: 57781

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/parsertarget.pxi
   lxml/trunk/src/lxml/tests/test_etree.py
Log:
 r4759 at delle:  sbehnel | 2008-09-02 22:07:11 +0200
 let target parser honour recover option


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Tue Sep  2 22:19:08 2008
@@ -20,6 +20,9 @@
 Bugs fixed
 ----------
 
+* Target parser did not honour the ``recover`` option and raised an
+  exception instead of calling ``.close()`` on the target.
+
 Other changes
 -------------
 

Modified: lxml/trunk/src/lxml/parsertarget.pxi
==============================================================================
--- lxml/trunk/src/lxml/parsertarget.pxi	(original)
+++ lxml/trunk/src/lxml/parsertarget.pxi	Tue Sep  2 22:19:08 2008
@@ -110,13 +110,17 @@
 
     cdef object _handleParseResult(self, _BaseParser parser, xmlDoc* result,
                                    filename):
-        if not self._c_ctxt.wellFormed:
+        cdef bint recover
+        recover = parser._parse_options & xmlparser.XML_PARSE_RECOVER
+        if not self._c_ctxt.wellFormed and not recover:
             _raiseParseError(self._c_ctxt, filename, self._error_log)
         self._raise_if_stored()
         return self._python_target.close()
 
     cdef xmlDoc* _handleParseResultDoc(self, _BaseParser parser,
                                        xmlDoc* result, filename) except NULL:
+        cdef bint recover
+        recover = parser._parse_options & xmlparser.XML_PARSE_RECOVER
         if result is not NULL and result._private is NULL:
             # no _Document proxy => orphen
             tree.xmlFreeDoc(result)
@@ -126,7 +130,7 @@
                 # no _Document proxy => orphen
                 tree.xmlFreeDoc(self._c_ctxt.myDoc)
             self._c_ctxt.myDoc = NULL
-        if not self._c_ctxt.wellFormed:
+        if not self._c_ctxt.wellFormed and not recover:
             _raiseParseError(self._c_ctxt, filename, self._error_log)
         self._raise_if_stored()
         raise _TargetParserResult(self._python_target.close())

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	Tue Sep  2 22:19:08 2008
@@ -525,6 +525,31 @@
                            "data-ca", "end-a", "data-B", "end-root"],
                           events)
 
+    def test_parser_target_recover(self):
+        events = []
+        class Target(object):
+            def start(self, tag, attrib):
+                events.append("start-" + tag)
+            def end(self, tag):
+                events.append("end-" + tag)
+            def data(self, data):
+                events.append("data-" + data)
+            def close(self):
+                events.append("close")
+                return "DONE"
+
+        parser = self.etree.XMLParser(target=Target(),
+                                      recover=True)
+
+        parser.feed(_bytes('<root>A<a>ca</a>B</not-root>'))
+        done = parser.close()
+
+        self.assertEquals("DONE", done)
+        self.assertEquals(["start-root", "data-A", "start-a",
+                           "data-ca", "end-a", "data-B",
+                           "end-root", "close"],
+                          events)
+
     def test_iterwalk_tag(self):
         iterwalk = self.etree.iterwalk
         root = self.etree.XML(_bytes('<a><b><d/></b><c/></a>'))

From scoder at codespeak.net  Fri Sep  5 13:50:01 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 13:50:01 +0200 (CEST)
Subject: [Lxml-checkins] r57837 - in lxml/branch/lxml-2.1: . src/lxml
Message-ID: <20080905115001.21531169F13@codespeak.net>

Author: scoder
Date: Fri Sep  5 13:50:00 2008
New Revision: 57837

Modified:
   lxml/branch/lxml-2.1/CHANGES.txt
   lxml/branch/lxml-2.1/src/lxml/proxy.pxi
Log:
fix another threading-related memory problem

Modified: lxml/branch/lxml-2.1/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.1/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.1/CHANGES.txt	Fri Sep  5 13:50:00 2008
@@ -16,6 +16,8 @@
 Bugs fixed
 ----------
 
+* Memory problem when passing documents between threads.
+
 * Target parser did not honour the ``recover`` option and raised an
   exception instead of calling ``.close()`` on the target.
 

Modified: lxml/branch/lxml-2.1/src/lxml/proxy.pxi
==============================================================================
--- lxml/branch/lxml-2.1/src/lxml/proxy.pxi	(original)
+++ lxml/branch/lxml-2.1/src/lxml/proxy.pxi	Fri Sep  5 13:50:00 2008
@@ -46,12 +46,12 @@
     python.Py_XDECREF(proxy._gc_doc)
     proxy._gc_doc = NULL
 
-cdef inline void _updateProxyDocument(xmlNode* c_node, _Document doc):
+cdef inline _Document _updateProxyDocument(xmlNode* c_node, _Document doc):
     u"""Replace the document reference of a proxy.
 
     This may deallocate the original document of the proxy!
     """
-    cdef _Document old_doc
+    cdef _Document old_doc = None
     cdef _Element element = <_Element>c_node._private
     if element._doc is not doc:
         old_doc = element._doc
@@ -59,6 +59,7 @@
         python.Py_INCREF(doc)
         element._gc_doc = <python.PyObject*>doc
         python.Py_DECREF(old_doc)
+    return old_doc
 
 ################################################################################
 # temporarily make a node the root node of its document
@@ -315,6 +316,7 @@
     cdef xmlNs* c_nsdef
     cdef xmlNs* c_del_ns_list
     cdef cstd.size_t i
+    cdef list old_docs = []
 
     if not tree._isElementOrXInclude(c_element):
         return 0
@@ -365,9 +367,11 @@
         if c_node is NULL:
             # no children => back off and continue with siblings and parents
 
-            # 4) fix _Document reference (may dealloc the original document!)
+            # 4) fix _Document reference
             if c_element._private is not NULL:
-                _updateProxyDocument(c_element, doc)
+                old_doc = _updateProxyDocument(c_element, doc)
+                if old_doc not in old_docs:
+                    old_docs.append(old_doc)
 
             if c_element is c_start_node:
                 break # all done
@@ -383,9 +387,11 @@
                 if c_element is NULL or not tree._isElementOrXInclude(c_element):
                     break
 
-                # 4) fix _Document reference (may dealloc the original document!)
+                # 4) fix _Document reference
                 if c_element._private is not NULL:
-                    _updateProxyDocument(c_element, doc)
+                    old_doc = _updateProxyDocument(c_element, doc)
+                    if old_doc not in old_docs:
+                        old_docs.append(old_doc)
 
                 if c_element is c_start_node:
                     break
@@ -402,6 +408,9 @@
     if doc._c_doc.dict is not c_source_doc.dict:
         fixThreadDictNames(c_start_node, c_source_doc.dict, doc._c_doc.dict)
 
+    # *now* free all _Documents the original tree referred to
+    old_docs = None
+
     # free now unused namespace declarations
     if c_del_ns_list is not NULL:
         tree.xmlFreeNsList(c_del_ns_list)

From scoder at codespeak.net  Fri Sep  5 13:50:10 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 13:50:10 +0200 (CEST)
Subject: [Lxml-checkins] r57838 - in lxml/branch/lxml-2.0: . src/lxml
Message-ID: <20080905115010.A02A9169F15@codespeak.net>

Author: scoder
Date: Fri Sep  5 13:50:09 2008
New Revision: 57838

Modified:
   lxml/branch/lxml-2.0/CHANGES.txt
   lxml/branch/lxml-2.0/src/lxml/proxy.pxi
Log:
fix another threading-related memory problem

Modified: lxml/branch/lxml-2.0/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.0/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.0/CHANGES.txt	Fri Sep  5 13:50:09 2008
@@ -11,6 +11,8 @@
 Bugs fixed
 ----------
 
+* Memory problem when passing documents between threads.
+
 * Target parser did not honour the ``recover`` option and raised an
   exception instead of calling ``.close()`` on the target.
 

Modified: lxml/branch/lxml-2.0/src/lxml/proxy.pxi
==============================================================================
--- lxml/branch/lxml-2.0/src/lxml/proxy.pxi	(original)
+++ lxml/branch/lxml-2.0/src/lxml/proxy.pxi	Fri Sep  5 13:50:09 2008
@@ -46,12 +46,12 @@
     python.Py_XDECREF(proxy._gc_doc)
     proxy._gc_doc = NULL
 
-cdef inline void _updateProxyDocument(xmlNode* c_node, _Document doc):
+cdef inline _Document _updateProxyDocument(xmlNode* c_node, _Document doc):
     u"""Replace the document reference of a proxy.
 
     This may deallocate the original document of the proxy!
     """
-    cdef _Document old_doc
+    cdef _Document old_doc = None
     cdef _Element element = <_Element>c_node._private
     if element._doc is not doc:
         old_doc = element._doc
@@ -59,6 +59,7 @@
         python.Py_INCREF(doc)
         element._gc_doc = <python.PyObject*>doc
         python.Py_DECREF(old_doc)
+    return old_doc
 
 ################################################################################
 # temporarily make a node the root node of its document
@@ -315,6 +316,7 @@
     cdef xmlNs* c_nsdef
     cdef xmlNs* c_del_ns_list
     cdef cstd.size_t i
+    cdef list old_docs = []
 
     if not tree._isElementOrXInclude(c_element):
         return 0
@@ -365,9 +367,11 @@
         if c_node is NULL:
             # no children => back off and continue with siblings and parents
 
-            # 4) fix _Document reference (may dealloc the original document!)
+            # 4) fix _Document reference
             if c_element._private is not NULL:
-                _updateProxyDocument(c_element, doc)
+                old_doc = _updateProxyDocument(c_element, doc)
+                if old_doc not in old_docs:
+                    old_docs.append(old_doc)
 
             if c_element is c_start_node:
                 break # all done
@@ -383,9 +387,11 @@
                 if c_element is NULL or not tree._isElementOrXInclude(c_element):
                     break
 
-                # 4) fix _Document reference (may dealloc the original document!)
+                # 4) fix _Document reference
                 if c_element._private is not NULL:
-                    _updateProxyDocument(c_element, doc)
+                    old_doc = _updateProxyDocument(c_element, doc)
+                    if old_doc not in old_docs:
+                        old_docs.append(old_doc)
 
                 if c_element is c_start_node:
                     break
@@ -402,6 +408,9 @@
     if doc._c_doc.dict is not c_source_doc.dict:
         fixThreadDictNames(c_start_node, c_source_doc.dict, doc._c_doc.dict)
 
+    # *now* free all _Documents the original tree referred to
+    old_docs = None
+
     # free now unused namespace declarations
     if c_del_ns_list is not NULL:
         tree.xmlFreeNsList(c_del_ns_list)

From scoder at codespeak.net  Fri Sep  5 13:50:18 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 13:50:18 +0200 (CEST)
Subject: [Lxml-checkins] r57839 - in lxml/trunk: . src/lxml
Message-ID: <20080905115018.0B017169F1D@codespeak.net>

Author: scoder
Date: Fri Sep  5 13:50:17 2008
New Revision: 57839

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/proxy.pxi
Log:
 r4765 at delle:  sbehnel | 2008-09-05 13:47:15 +0200
 fix another threading-related memory problem


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Fri Sep  5 13:50:17 2008
@@ -20,6 +20,8 @@
 Bugs fixed
 ----------
 
+* Memory problem when passing documents between threads.
+
 * Target parser did not honour the ``recover`` option and raised an
   exception instead of calling ``.close()`` on the target.
 

Modified: lxml/trunk/src/lxml/proxy.pxi
==============================================================================
--- lxml/trunk/src/lxml/proxy.pxi	(original)
+++ lxml/trunk/src/lxml/proxy.pxi	Fri Sep  5 13:50:17 2008
@@ -44,12 +44,12 @@
     python.Py_XDECREF(proxy._gc_doc)
     proxy._gc_doc = NULL
 
-cdef inline void _updateProxyDocument(xmlNode* c_node, _Document doc):
+cdef inline _Document _updateProxyDocument(xmlNode* c_node, _Document doc):
     u"""Replace the document reference of a proxy.
 
     This may deallocate the original document of the proxy!
     """
-    cdef _Document old_doc
+    cdef _Document old_doc = None
     cdef _Element element = <_Element>c_node._private
     if element._doc is not doc:
         old_doc = element._doc
@@ -57,6 +57,7 @@
         python.Py_INCREF(doc)
         element._gc_doc = <python.PyObject*>doc
         python.Py_DECREF(old_doc)
+    return old_doc
 
 ################################################################################
 # temporarily make a node the root node of its document
@@ -313,6 +314,7 @@
     cdef xmlNs* c_nsdef
     cdef xmlNs* c_del_ns_list
     cdef cstd.size_t i
+    cdef list old_docs = []
 
     if not tree._isElementOrXInclude(c_element):
         return 0
@@ -363,9 +365,11 @@
         if c_node is NULL:
             # no children => back off and continue with siblings and parents
 
-            # 4) fix _Document reference (may dealloc the original document!)
+            # 4) fix _Document reference
             if c_element._private is not NULL:
-                _updateProxyDocument(c_element, doc)
+                old_doc = _updateProxyDocument(c_element, doc)
+                if old_doc not in old_docs:
+                    old_docs.append(old_doc)
 
             if c_element is c_start_node:
                 break # all done
@@ -381,9 +385,11 @@
                 if c_element is NULL or not tree._isElementOrXInclude(c_element):
                     break
 
-                # 4) fix _Document reference (may dealloc the original document!)
+                # 4) fix _Document reference
                 if c_element._private is not NULL:
-                    _updateProxyDocument(c_element, doc)
+                    old_doc = _updateProxyDocument(c_element, doc)
+                    if old_doc not in old_docs:
+                        old_docs.append(old_doc)
 
                 if c_element is c_start_node:
                     break
@@ -400,6 +406,9 @@
     if doc._c_doc.dict is not c_source_doc.dict:
         fixThreadDictNames(c_start_node, c_source_doc.dict, doc._c_doc.dict)
 
+    # *now* free all _Documents the original tree referred to
+    old_docs = None
+
     # free now unused namespace declarations
     if c_del_ns_list is not NULL:
         tree.xmlFreeNsList(c_del_ns_list)

From scoder at codespeak.net  Fri Sep  5 14:31:42 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:31:42 +0200 (CEST)
Subject: [Lxml-checkins] r57845 - in lxml/branch/lxml-2.1: . doc
Message-ID: <20080905123142.76A4F16A0AD@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:31:38 2008
New Revision: 57845

Modified:
   lxml/branch/lxml-2.1/CHANGES.txt
   lxml/branch/lxml-2.1/doc/main.txt
   lxml/branch/lxml-2.1/version.txt
Log:
release version 2.1.2

Modified: lxml/branch/lxml-2.1/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.1/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.1/CHANGES.txt	Fri Sep  5 14:31:38 2008
@@ -2,8 +2,8 @@
 lxml changelog
 ==============
 
-Under development
-=================
+2.1.2 (2008-09-05)
+==================
 
 Features added
 --------------

Modified: lxml/branch/lxml-2.1/doc/main.txt
==============================================================================
--- lxml/branch/lxml-2.1/doc/main.txt	(original)
+++ lxml/branch/lxml-2.1/doc/main.txt	Fri Sep  5 14:31:38 2008
@@ -147,8 +147,8 @@
 source release.  If you can't wait, consider trying a less recent
 release version first.
 
-The latest version is `lxml 2.1.1`_, released 2008-07-24
-(`changes for 2.1.1`_).  `Older versions`_ are listed below.
+The latest version is `lxml 2.1.2`_, released 2008-09-05
+(`changes for 2.1.2`_).  `Older versions`_ are listed below.
 
 Please take a look at the `installation instructions`_!
 
@@ -220,7 +220,9 @@
 `2.0 <http://codespeak.net/lxml/2.0/>`_ and the `current
 in-development version <http://codespeak.net/lxml/dev/>`_.
 
-.. _`PDF documentation`: lxmldoc-2.1.1.pdf
+.. _`PDF documentation`: lxmldoc-2.1.2.pdf
+
+* `lxml 2.1.1`_, released 2008-07-24 (`changes for 2.1.1`_)
 
 * `lxml 2.1`_, released 2008-07-09 (`changes for 2.1`_)
 
@@ -300,6 +302,7 @@
 
 * `lxml 0.5`_,   released 2005-04-08
 
+.. _`lxml 2.1.2`: lxml-2.1.2.tgz
 .. _`lxml 2.1.1`: lxml-2.1.1.tgz
 .. _`lxml 2.1`: lxml-2.1.tgz
 .. _`lxml 2.1beta3`: lxml-2.1beta3.tgz
@@ -340,6 +343,7 @@
 .. _`lxml 0.5.1`: lxml-0.5.1.tgz
 .. _`lxml 0.5`: lxml-0.5.tgz
 
+.. _`changes for 2.1.2`: changes-2.1.2.html
 .. _`changes for 2.1.1`: changes-2.1.1.html
 .. _`changes for 2.1`: changes-2.1.html
 .. _`changes for 2.1beta3`: changes-2.1beta3.html

Modified: lxml/branch/lxml-2.1/version.txt
==============================================================================
--- lxml/branch/lxml-2.1/version.txt	(original)
+++ lxml/branch/lxml-2.1/version.txt	Fri Sep  5 14:31:38 2008
@@ -1 +1 @@
-2.1.1
+2.1.2

From scoder at codespeak.net  Fri Sep  5 14:32:21 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:32:21 +0200 (CEST)
Subject: [Lxml-checkins] r57846 - lxml/tag/lxml-2.1.2
Message-ID: <20080905123221.4251916A099@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:32:19 2008
New Revision: 57846

Added:
   lxml/tag/lxml-2.1.2/
      - copied from r57845, lxml/branch/lxml-2.1/
Log:
tag for 2.1.2

From scoder at codespeak.net  Fri Sep  5 14:32:40 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:32:40 +0200 (CEST)
Subject: [Lxml-checkins] r57847 - in lxml/branch/lxml-2.0: . doc
Message-ID: <20080905123240.BEA2116A099@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:32:39 2008
New Revision: 57847

Modified:
   lxml/branch/lxml-2.0/CHANGES.txt
   lxml/branch/lxml-2.0/doc/main.txt
   lxml/branch/lxml-2.0/version.txt
Log:
release version 2.0.9

Modified: lxml/branch/lxml-2.0/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.0/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.0/CHANGES.txt	Fri Sep  5 14:32:39 2008
@@ -2,11 +2,8 @@
 lxml changelog
 ==============
 
-Under development
-=================
-
-Features added
---------------
+2.0.9 (2008-09-05)
+==================
 
 Bugs fixed
 ----------
@@ -16,9 +13,6 @@
 * Target parser did not honour the ``recover`` option and raised an
   exception instead of calling ``.close()`` on the target.
 
-Other changes
--------------
-
 
 2.0.8 (2008-07-24)
 ==================

Modified: lxml/branch/lxml-2.0/doc/main.txt
==============================================================================
--- lxml/branch/lxml-2.0/doc/main.txt	(original)
+++ lxml/branch/lxml-2.0/doc/main.txt	Fri Sep  5 14:32:39 2008
@@ -146,8 +146,8 @@
 source release.  If you can't wait, consider trying a less recent
 release version first.
 
-The latest version is `lxml 2.0.8`_, released 2008-07-24
-(`changes for 2.0.8`_).  `Older versions`_ are listed below.
+The latest version is `lxml 2.0.9`_, released 2008-09-05
+(`changes for 2.0.9`_).  `Older versions`_ are listed below.
 
 Please take a look at the `installation instructions`_!
 
@@ -215,7 +215,9 @@
 Old Versions
 ------------
 
-.. _`PDF documentation`: lxmldoc-2.0.8.pdf
+.. _`PDF documentation`: lxmldoc-2.0.9.pdf
+
+* `lxml 2.0.8`_, released 2008-07-24 (`changes for 2.0.8`_)
 
 * `lxml 2.0.7`_, released 2008-06-20 (`changes for 2.0.7`_)
 
@@ -283,6 +285,7 @@
 
 * `lxml 0.5`_,   released 2005-04-08
 
+.. _`lxml 2.0.9`: lxml-2.0.9.tgz
 .. _`lxml 2.0.8`: lxml-2.0.8.tgz
 .. _`lxml 2.0.7`: lxml-2.0.7.tgz
 .. _`lxml 2.0.6`: lxml-2.0.6.tgz
@@ -317,6 +320,7 @@
 .. _`lxml 0.5.1`: lxml-0.5.1.tgz
 .. _`lxml 0.5`: lxml-0.5.tgz
 
+.. _`changes for 2.0.9`: changes-2.0.9.html
 .. _`changes for 2.0.8`: changes-2.0.8.html
 .. _`changes for 2.0.7`: changes-2.0.7.html
 .. _`changes for 2.0.6`: changes-2.0.6.html

Modified: lxml/branch/lxml-2.0/version.txt
==============================================================================
--- lxml/branch/lxml-2.0/version.txt	(original)
+++ lxml/branch/lxml-2.0/version.txt	Fri Sep  5 14:32:39 2008
@@ -1 +1 @@
-2.0.8
+2.0.9

From scoder at codespeak.net  Fri Sep  5 14:33:13 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:33:13 +0200 (CEST)
Subject: [Lxml-checkins] r57848 - lxml/tag/lxml-2.0.9
Message-ID: <20080905123313.4ED4816A099@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:33:12 2008
New Revision: 57848

Added:
   lxml/tag/lxml-2.0.9/
      - copied from r57847, lxml/branch/lxml-2.0/
Log:
tag for 2.0.9

From scoder at codespeak.net  Fri Sep  5 14:34:34 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:34:34 +0200 (CEST)
Subject: [Lxml-checkins] r57849 - lxml/tag/lxml-2.0.8
Message-ID: <20080905123434.57BCB16A09E@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:34:33 2008
New Revision: 57849

Added:
   lxml/tag/lxml-2.0.8/
      - copied from r56762, lxml/branch/lxml-2.0/
Log:
tag for 2.0.8

From scoder at codespeak.net  Fri Sep  5 14:35:11 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:35:11 +0200 (CEST)
Subject: [Lxml-checkins] r57850 - lxml/tag/lxml-2.1.1
Message-ID: <20080905123511.9738316A0A1@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:35:10 2008
New Revision: 57850

Added:
   lxml/tag/lxml-2.1.1/
      - copied from r56760, lxml/branch/lxml-2.1/
Log:
tag for 2.1.1

From scoder at codespeak.net  Fri Sep  5 14:36:45 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:36:45 +0200 (CEST)
Subject: [Lxml-checkins] r57851 - lxml/tag/lxml-2.1
Message-ID: <20080905123645.7397016A0A1@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:36:44 2008
New Revision: 57851

Added:
   lxml/tag/lxml-2.1/
      - copied from r56392, lxml/trunk/
Log:
tag for 2.1

From scoder at codespeak.net  Fri Sep  5 14:38:13 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:38:13 +0200 (CEST)
Subject: [Lxml-checkins] r57852 - lxml/tag/lxml-2.1
Message-ID: <20080905123813.2B23516A05D@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:38:12 2008
New Revision: 57852

Removed:
   lxml/tag/lxml-2.1/
Log:
wrong tag

From scoder at codespeak.net  Fri Sep  5 14:38:19 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:38:19 +0200 (CEST)
Subject: [Lxml-checkins] r57853 - lxml/tag/lxml-2.1
Message-ID: <20080905123819.C572B16A09C@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:38:18 2008
New Revision: 57853

Added:
   lxml/tag/lxml-2.1/
      - copied from r56386, lxml/trunk/
Log:
tag for 2.1

From scoder at codespeak.net  Fri Sep  5 14:55:14 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 14:55:14 +0200 (CEST)
Subject: [Lxml-checkins] r57855 - lxml/branch/lxml-2.1/doc
Message-ID: <20080905125514.F0CEB169EA5@codespeak.net>

Author: scoder
Date: Fri Sep  5 14:55:14 2008
New Revision: 57855

Modified:
   lxml/branch/lxml-2.1/doc/main.txt
Log:
fixed version list in docs

Modified: lxml/branch/lxml-2.1/doc/main.txt
==============================================================================
--- lxml/branch/lxml-2.1/doc/main.txt	(original)
+++ lxml/branch/lxml-2.1/doc/main.txt	Fri Sep  5 14:55:14 2008
@@ -234,7 +234,9 @@
 
 * `lxml 2.1alpha1`_, released 2008-03-27 (`changes for 2.1alpha1`_)
 
-* `lxml 2.0.8`_, released 2008-07-24 (`changes for 2.0.7`_)
+* `lxml 2.0.9`_, released 2008-09-05 (`changes for 2.0.9`_)
+
+* `lxml 2.0.8`_, released 2008-07-24 (`changes for 2.0.8`_)
 
 * `lxml 2.0.7`_, released 2008-06-20 (`changes for 2.0.7`_)
 
@@ -309,6 +311,7 @@
 .. _`lxml 2.1beta2`: lxml-2.1beta2.tgz
 .. _`lxml 2.1beta1`: lxml-2.1beta1.tgz
 .. _`lxml 2.1alpha1`: lxml-2.1alpha1.tgz
+.. _`lxml 2.0.9`: lxml-2.0.9.tgz
 .. _`lxml 2.0.8`: lxml-2.0.8.tgz
 .. _`lxml 2.0.7`: lxml-2.0.7.tgz
 .. _`lxml 2.0.6`: lxml-2.0.6.tgz
@@ -350,6 +353,7 @@
 .. _`changes for 2.1beta2`: changes-2.1beta2.html
 .. _`changes for 2.1beta1`: changes-2.1beta1.html
 .. _`changes for 2.1alpha1`: changes-2.1alpha1.html
+.. _`changes for 2.0.9`: changes-2.0.9.html
 .. _`changes for 2.0.8`: changes-2.0.8.html
 .. _`changes for 2.0.7`: changes-2.0.7.html
 .. _`changes for 2.0.6`: changes-2.0.6.html

From scoder at codespeak.net  Fri Sep  5 15:05:09 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 15:05:09 +0200 (CEST)
Subject: [Lxml-checkins] r57857 - lxml/branch/lxml-2.0
Message-ID: <20080905130509.F3F0216A086@codespeak.net>

Author: scoder
Date: Fri Sep  5 15:05:08 2008
New Revision: 57857

Modified:
   lxml/branch/lxml-2.0/setup.py
Log:
declare 2.0.x mature

Modified: lxml/branch/lxml-2.0/setup.py
==============================================================================
--- lxml/branch/lxml-2.0/setup.py	(original)
+++ lxml/branch/lxml-2.0/setup.py	Fri Sep  5 15:05:08 2008
@@ -88,7 +88,7 @@
 """ + branch_link) % { "branch_version" : versioninfo.branch_version() }) +
                       versioninfo.changes()),
     classifiers = [
-    versioninfo.dev_status(),
+    'Development Status :: 6 - Mature', # versioninfo.dev_status(),
     'Intended Audience :: Developers',
     'Intended Audience :: Information Technology',
     'License :: OSI Approved :: BSD License',

From scoder at codespeak.net  Fri Sep  5 21:07:53 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 21:07:53 +0200 (CEST)
Subject: [Lxml-checkins] r57866 - in lxml/trunk: . doc
Message-ID: <20080905190753.5145F16A104@codespeak.net>

Author: scoder
Date: Fri Sep  5 21:07:49 2008
New Revision: 57866

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/doc/main.txt
Log:
 r4769 at delle:  sbehnel | 2008-09-05 14:49:27 +0200
 post-release updates


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Fri Sep  5 21:07:49 2008
@@ -8,15 +8,28 @@
 Features added
 --------------
 
+* New options for exclusive C14N and C14N without comments.
+
+* Instantiating a custom Element classes creates a new Element.
+
+Bugs fixed
+----------
+
+Other changes
+-------------
+
+
+2.1.2 (2008-09-05)
+==================
+
+Features added
+--------------
+
 * lxml.etree now tries to find the absolute path name of files when
   parsing from a file-like object.  This helps custom resolvers when
   resolving relative URLs, as lixbml2 can prepend them with the path
   of the source document.
 
-* New options for exclusive C14N and C14N without comments.
-
-* Instantiating a custom Element classes creates a new Element.
-
 Bugs fixed
 ----------
 
@@ -29,6 +42,18 @@
 -------------
 
 
+2.0.9 (2008-09-05)
+==================
+
+Bugs fixed
+----------
+
+* Memory problem when passing documents between threads.
+
+* Target parser did not honour the ``recover`` option and raised an
+  exception instead of calling ``.close()`` on the target.
+
+
 2.1.1 (2008-07-24)
 ==================
 

Modified: lxml/trunk/doc/main.txt
==============================================================================
--- lxml/trunk/doc/main.txt	(original)
+++ lxml/trunk/doc/main.txt	Fri Sep  5 21:07:49 2008
@@ -147,8 +147,8 @@
 source release.  If you can't wait, consider trying a less recent
 release version first.
 
-The latest version is `lxml 2.1.1`_, released 2008-07-24
-(`changes for 2.1.1`_).  `Older versions`_ are listed below.
+The latest version is `lxml 2.1.2`_, released 2008-09-05
+(`changes for 2.1.2`_).  `Older versions`_ are listed below.
 
 Please take a look at the `installation instructions`_!
 
@@ -220,12 +220,16 @@
 `2.0 <http://codespeak.net/lxml/2.0/>`_ and the `current
 in-development version <http://codespeak.net/lxml/dev/>`_.
 
-.. _`PDF documentation`: lxmldoc-2.1.1.pdf
+.. _`PDF documentation`: lxmldoc-2.1.2.pdf
+
+* `lxml 2.1.2`_, released 2008-09-05 (`changes for 2.1.2`_)
 
 * `lxml 2.1.1`_, released 2008-07-24 (`changes for 2.1.1`_)
 
 * `lxml 2.1`_, released 2008-07-09 (`changes for 2.1`_)
 
+* `lxml 2.0.9`_, released 2008-09-05 (`changes for 2.0.9`_)
+
 * `lxml 2.0.8`_, released 2008-07-24 (`changes for 2.0.8`_)
 
 * `lxml 2.0.7`_, released 2008-06-20 (`changes for 2.0.7`_)
@@ -294,8 +298,10 @@
 
 * `lxml 0.5`_,   released 2005-04-08
 
+.. _`lxml 2.1.2`: lxml-2.1.2.tgz
 .. _`lxml 2.1.1`: lxml-2.1.1.tgz
 .. _`lxml 2.1`: lxml-2.1.tgz
+.. _`lxml 2.0.9`: lxml-2.0.9.tgz
 .. _`lxml 2.0.8`: lxml-2.0.8.tgz
 .. _`lxml 2.0.7`: lxml-2.0.7.tgz
 .. _`lxml 2.0.6`: lxml-2.0.6.tgz
@@ -330,8 +336,10 @@
 .. _`lxml 0.5.1`: lxml-0.5.1.tgz
 .. _`lxml 0.5`: lxml-0.5.tgz
 
+.. _`changes for 2.1.2`: changes-2.1.2.html
 .. _`changes for 2.1.1`: changes-2.1.1.html
 .. _`changes for 2.1`: changes-2.1.html
+.. _`changes for 2.0.9`: changes-2.0.9.html
 .. _`changes for 2.0.8`: changes-2.0.8.html
 .. _`changes for 2.0.7`: changes-2.0.7.html
 .. _`changes for 2.0.6`: changes-2.0.6.html

From scoder at codespeak.net  Fri Sep  5 21:08:05 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  5 Sep 2008 21:08:05 +0200 (CEST)
Subject: [Lxml-checkins] r57867 - in lxml/trunk: . src/lxml
Message-ID: <20080905190805.E63D116A104@codespeak.net>

Author: scoder
Date: Fri Sep  5 21:08:04 2008
New Revision: 57867

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/proxy.pxi
Log:
 r4770 at delle:  sbehnel | 2008-09-05 21:07:40 +0200
 some cleanup after last fix: one document reference should be enough


Modified: lxml/trunk/src/lxml/proxy.pxi
==============================================================================
--- lxml/trunk/src/lxml/proxy.pxi	(original)
+++ lxml/trunk/src/lxml/proxy.pxi	Fri Sep  5 21:08:04 2008
@@ -314,7 +314,6 @@
     cdef xmlNs* c_nsdef
     cdef xmlNs* c_del_ns_list
     cdef cstd.size_t i
-    cdef list old_docs = []
 
     if not tree._isElementOrXInclude(c_element):
         return 0
@@ -368,8 +367,6 @@
             # 4) fix _Document reference
             if c_element._private is not NULL:
                 old_doc = _updateProxyDocument(c_element, doc)
-                if old_doc not in old_docs:
-                    old_docs.append(old_doc)
 
             if c_element is c_start_node:
                 break # all done
@@ -388,8 +385,6 @@
                 # 4) fix _Document reference
                 if c_element._private is not NULL:
                     old_doc = _updateProxyDocument(c_element, doc)
-                    if old_doc not in old_docs:
-                        old_docs.append(old_doc)
 
                 if c_element is c_start_node:
                     break
@@ -406,8 +401,8 @@
     if doc._c_doc.dict is not c_source_doc.dict:
         fixThreadDictNames(c_start_node, c_source_doc.dict, doc._c_doc.dict)
 
-    # *now* free all _Documents the original tree referred to
-    old_docs = None
+    # *now* allow the original _Document to be deleted
+    old_doc = None
 
     # free now unused namespace declarations
     if c_del_ns_list is not NULL:

From scoder at codespeak.net  Sun Sep  7 15:56:08 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun,  7 Sep 2008 15:56:08 +0200 (CEST)
Subject: [Lxml-checkins] r57946 - in lxml/trunk: . src/lxml
Message-ID: <20080907135608.0FC7A16A19B@codespeak.net>

Author: scoder
Date: Sun Sep  7 15:56:07 2008
New Revision: 57946

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/apihelpers.pxi
Log:
 r4784 at delle:  sbehnel | 2008-09-07 15:48:36 +0200
 code cleanup and minor speed-up


Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Sun Sep  7 15:56:07 2008
@@ -446,11 +446,11 @@
         value = _utf8(value)
     c_value = _cstr(value)
     if ns is None:
-        tree.xmlSetProp(element._c_node, c_tag, c_value)
+        c_ns = NULL
     else:
         c_ns = element._doc._findOrBuildNodeNs(element._c_node,
                                                _cstr(ns), NULL)
-        tree.xmlSetNsProp(element._c_node, c_ns, c_tag, c_value)
+    tree.xmlSetNsProp(element._c_node, c_ns, c_tag, c_value)
     return 0
 
 cdef int _delAttribute(_Element element, key) except -1:

From scoder at codespeak.net  Sun Sep  7 15:56:14 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun,  7 Sep 2008 15:56:14 +0200 (CEST)
Subject: [Lxml-checkins] r57947 - in lxml/trunk: . src/lxml
Message-ID: <20080907135614.9F68D16A19F@codespeak.net>

Author: scoder
Date: Sun Sep  7 15:56:13 2008
New Revision: 57947

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/proxy.pxi
Log:
 r4785 at delle:  sbehnel | 2008-09-07 15:55:32 +0200
 another restructuring in moveNodeToDocument(): use separate steps for namespaces, dict cleanup and doc references (last\!)


Modified: lxml/trunk/src/lxml/proxy.pxi
==============================================================================
--- lxml/trunk/src/lxml/proxy.pxi	(original)
+++ lxml/trunk/src/lxml/proxy.pxi	Sun Sep  7 15:56:13 2008
@@ -44,12 +44,11 @@
     python.Py_XDECREF(proxy._gc_doc)
     proxy._gc_doc = NULL
 
-cdef inline _Document _updateProxyDocument(xmlNode* c_node, _Document doc):
-    u"""Replace the document reference of a proxy.
-
-    This may deallocate the original document of the proxy!
+cdef inline void _updateProxyDocument(xmlNode* c_node, _Document doc):
+    u"""Replace the document reference of a proxy and return the old one
+    iff it was replaced (None otherwise).
     """
-    cdef _Document old_doc = None
+    cdef _Document old_doc
     cdef _Element element = <_Element>c_node._private
     if element._doc is not doc:
         old_doc = element._doc
@@ -57,7 +56,6 @@
         python.Py_INCREF(doc)
         element._gc_doc = <python.PyObject*>doc
         python.Py_DECREF(old_doc)
-    return old_doc
 
 ################################################################################
 # temporarily make a node the root node of its document
@@ -313,7 +311,7 @@
     cdef xmlNs* c_ns_next
     cdef xmlNs* c_nsdef
     cdef xmlNs* c_del_ns_list
-    cdef cstd.size_t i
+    cdef cstd.size_t i, proxy_count = 0
 
     if not tree._isElementOrXInclude(c_element):
         return 0
@@ -326,7 +324,11 @@
     c_ns_cache.size = 0
     c_ns_cache.last = 0
 
-    while c_element is not NULL:
+    tree.BEGIN_FOR_EACH_FROM(c_element, c_element, 1)
+    if tree._isElementOrXInclude(c_element):
+        if c_element._private is not NULL:
+            proxy_count += 1
+
         # 1) cut out namespaces defined here that are already known by
         #    the ancestors
         if c_element.nsDef is not NULL:
@@ -354,55 +356,7 @@
                 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
-
-            # 4) fix _Document reference
-            if c_element._private is not NULL:
-                old_doc = _updateProxyDocument(c_element, doc)
-
-            if c_element is c_start_node:
-                break # all done
-
-            # 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
-
-                # 4) fix _Document reference
-                if c_element._private is not NULL:
-                    old_doc = _updateProxyDocument(c_element, 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 # all done
-        c_element = c_node
-
-    # 3) fix the names in the tree in case we moved it to a different thread
-    if doc._c_doc.dict is not c_source_doc.dict:
-        fixThreadDictNames(c_start_node, c_source_doc.dict, doc._c_doc.dict)
-
-    # *now* allow the original _Document to be deleted
-    old_doc = None
+    tree.END_FOR_EACH_FROM(c_element)
 
     # free now unused namespace declarations
     if c_del_ns_list is not NULL:
@@ -414,9 +368,31 @@
     if c_ns_cache.old is not NULL:
         cstd.free(c_ns_cache.old)
 
+    # 3) fix the names in the tree if we moved it from a different thread
+    if doc._c_doc.dict is not c_source_doc.dict:
+        fixThreadDictNames(c_start_node, c_source_doc.dict, doc._c_doc.dict)
+
+    # 4) fix _Document references
+    #    (and potentially deallocate the source document)
+    if proxy_count > 0:
+        if proxy_count == 1 and c_start_node._private is not NULL:
+            _updateProxyDocument(c_start_node, doc)
+        else:
+            fixElementDocument(c_start_node, doc, proxy_count)
+
     return 0
 
 
+cdef void fixElementDocument(xmlNode* c_element, _Document doc,
+                             cstd.size_t proxy_count):
+    tree.BEGIN_FOR_EACH_FROM(c_element, c_element, 1)
+    if c_element._private is not NULL:
+        _updateProxyDocument(c_element, doc)
+        proxy_count -= 1
+        if proxy_count == 0:
+            return
+    tree.END_FOR_EACH_FROM(c_element)
+
 cdef void fixThreadDictNames(xmlNode* c_element,
                              tree.xmlDict* c_src_dict,
                              tree.xmlDict* c_dict) nogil:

From scoder at codespeak.net  Sun Sep  7 22:23:27 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun,  7 Sep 2008 22:23:27 +0200 (CEST)
Subject: [Lxml-checkins] r57956 - in lxml/trunk: . benchmark src/lxml
Message-ID: <20080907202327.B585316A1FD@codespeak.net>

Author: scoder
Date: Sun Sep  7 22:23:26 2008
New Revision: 57956

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/benchmark/bench_etree.py
   lxml/trunk/src/lxml/apihelpers.pxi
   lxml/trunk/src/lxml/python.pxd
Log:
 r4788 at delle:  sbehnel | 2008-09-07 19:55:29 +0200
 streamlined the _utf8() function (a very common helper function)


Modified: lxml/trunk/benchmark/bench_etree.py
==============================================================================
--- lxml/trunk/benchmark/bench_etree.py	(original)
+++ lxml/trunk/benchmark/bench_etree.py	Sun Sep  7 22:23:26 2008
@@ -228,6 +228,7 @@
 
     @with_attributes(True)
     @children
+    @nochange
     def bench_get_attributes(self, children):
         for child in children:
             child.get('bla1')

Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Sun Sep  7 22:23:26 2008
@@ -1130,13 +1130,17 @@
     is_non_ascii = 0
     while s < c_end:
         c = s[0]
-        if c & 0x80:
-            is_non_ascii = 1
-        elif c == c'\0':
-            return -1 # invalid!
-        elif is_non_ascii == 0 and not tree.xmlIsChar_ch(c):
+        if not tree.xmlIsChar_ch(c):
             return -1 # invalid!
+        elif c & 0x80:
+            is_non_ascii = 1
+            break
         s += 1
+    if is_non_ascii:
+        while s < c_end:
+            if not tree.xmlIsChar_ch(s[0]):
+                return -1 # invalid!
+            s += 1
     return is_non_ascii
 
 cdef object funicode(char* s):
@@ -1161,17 +1165,19 @@
     return python.PyString_FromStringAndSize(s, slen)
 
 cdef object _utf8(object s):
-    if python.PyString_Check(s):
-        if isutf8py(s):
-            raise ValueError, \
-                u"All strings must be XML compatible: Unicode or ASCII, no NULL bytes"
-    elif python.PyUnicode_Check(s):
+    cdef bint invalid
+    if python.PyString_CheckExact(s):
+        invalid = isutf8py(s)
+    elif python.PyUnicode_CheckExact(s) or python.PyUnicode_Check(s):
         s = python.PyUnicode_AsUTF8String(s)
-        if isutf8py(s) == -1:
-            raise ValueError, \
-                u"All strings must be XML compatible: Unicode or ASCII, no NULL bytes"
+        invalid = isutf8py(s) == -1
+    elif python.PyString_Check(s):
+        invalid = isutf8py(s)
     else:
         raise TypeError, u"Argument must be string or unicode."
+    if invalid:
+        raise ValueError, \
+            u"All strings must be XML compatible: Unicode or ASCII, no NULL bytes"
     return s
 
 cdef bint _isFilePath(char* c_path):

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Sun Sep  7 22:23:26 2008
@@ -22,9 +22,10 @@
 
     cdef FILE* PyFile_AsFile(object p)
 
-    cdef int PyUnicode_Check(object obj)
-    cdef int PyString_Check(object obj)
-    cdef int PyString_CheckExact(object obj)
+    cdef bint PyUnicode_Check(object obj)
+    cdef bint PyUnicode_CheckExact(object obj)
+    cdef bint PyString_Check(object obj)
+    cdef bint PyString_CheckExact(object obj)
 
     cdef object PyUnicode_FromEncodedObject(object s, char* encoding,
                                             char* errors)

From scoder at codespeak.net  Sun Sep  7 22:23:33 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun,  7 Sep 2008 22:23:33 +0200 (CEST)
Subject: [Lxml-checkins] r57957 - in lxml/trunk: . src/lxml
Message-ID: <20080907202333.8B00216A213@codespeak.net>

Author: scoder
Date: Sun Sep  7 22:23:30 2008
New Revision: 57957

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/apihelpers.pxi
Log:
 r4789 at delle:  sbehnel | 2008-09-07 22:18:12 +0200
 fix API string input checking, misc. fixes and switch-statement optimisations


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Sun Sep  7 22:23:30 2008
@@ -15,6 +15,9 @@
 Bugs fixed
 ----------
 
+* 0-bytes could slip through the API when used inside of Unicode
+  strings.
+
 Other changes
 -------------
 

Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Sun Sep  7 22:23:30 2008
@@ -1120,7 +1120,11 @@
         c = s[0]
     return 0
 
-cdef int isutf8py(pystring):
+cdef int check_string_utf8(pystring):
+    u"""Check if a string looks like valid UTF-8 XML content.  Returns 0
+    for ASCII, 1 for UTF-8 and -1 in the case of errors, such as NULL
+    bytes or ASCII control characters.
+    """
     cdef char* s
     cdef char* c_end
     cdef char c
@@ -1130,17 +1134,15 @@
     is_non_ascii = 0
     while s < c_end:
         c = s[0]
-        if not tree.xmlIsChar_ch(c):
-            return -1 # invalid!
-        elif c & 0x80:
+        if c & 0x80:
+            # skip the entire multi byte sequence
+            while c & 0x80:
+                s += 1
+                c = s[0]
             is_non_ascii = 1
-            break
+        elif not tree.xmlIsChar_ch(c):
+            return -1 # invalid!
         s += 1
-    if is_non_ascii:
-        while s < c_end:
-            if not tree.xmlIsChar_ch(s[0]):
-                return -1 # invalid!
-            s += 1
     return is_non_ascii
 
 cdef object funicode(char* s):
@@ -1165,14 +1167,14 @@
     return python.PyString_FromStringAndSize(s, slen)
 
 cdef object _utf8(object s):
-    cdef bint invalid
+    cdef int invalid
     if python.PyString_CheckExact(s):
-        invalid = isutf8py(s)
+        invalid = check_string_utf8(s)
     elif python.PyUnicode_CheckExact(s) or python.PyUnicode_Check(s):
         s = python.PyUnicode_AsUTF8String(s)
-        invalid = isutf8py(s) == -1
+        invalid = check_string_utf8(s) == -1
     elif python.PyString_Check(s):
-        invalid = isutf8py(s)
+        invalid = check_string_utf8(s)
     else:
         raise TypeError, u"Argument must be string or unicode."
     if invalid:
@@ -1182,6 +1184,7 @@
 
 cdef bint _isFilePath(char* c_path):
     u"simple heuristic to see if a path is a filename"
+    cdef char c
     # test if it looks like an absolute Unix path or a Windows network path
     if c_path[0] == c'/':
         return 1
@@ -1192,11 +1195,12 @@
             return 1
     # test if it looks like a relative path
     while c_path[0] != c'\0':
-        if c_path[0] == c':':
+        c = c_path[0]
+        if c == c':':
             return 0
-        if c_path[0] == c'/':
+        elif c == c'/':
             return 1
-        if c_path[0] == c'\\':
+        elif c == c'\\':
             return 1
         c_path += 1
     return 1
@@ -1245,7 +1249,7 @@
     if filename is None:
         return None
     elif python.PyString_Check(filename):
-        if not isutf8py(filename):
+        if not check_string_utf8(filename):
             # plain ASCII!
             return filename
         c_filename = _cstr(filename)
@@ -1305,20 +1309,13 @@
     return tree.xmlValidateNCName(c_name, 0) == 0
 
 cdef int _htmlNameIsValid(char* c_name):
+    cdef char c
     if c_name is NULL or c_name[0] == c'\0':
         return 0
     while c_name[0] != c'\0':
-        if c_name[0] == c'&' or \
-                c_name[0] == c'<' or \
-                c_name[0] == c'>' or \
-                c_name[0] == c'/' or \
-                c_name[0] == c'"' or \
-                c_name[0] == c"'" or \
-                c_name[0] == c'\x09' or \
-                c_name[0] == c'\x0A' or \
-                c_name[0] == c'\x0B' or \
-                c_name[0] == c'\x0C' or \
-                c_name[0] == c'\x20':
+        c = c_name[0]
+        if c in (c'&', c'<', c'>', c'/', c'"', c"'",
+                 c'\t', c'\n', c'\x0B', c'\x0C', c'\r', c' '):
             return 0
         c_name = c_name + 1
     return 1

From scoder at codespeak.net  Sun Sep  7 22:23:38 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun,  7 Sep 2008 22:23:38 +0200 (CEST)
Subject: [Lxml-checkins] r57958 - in lxml/trunk: . src/lxml
Message-ID: <20080907202338.06D9B16A214@codespeak.net>

Author: scoder
Date: Sun Sep  7 22:23:38 2008
New Revision: 57958

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/apihelpers.pxi
Log:
 r4790 at delle:  sbehnel | 2008-09-07 22:22:36 +0200
 cleanup


Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Sun Sep  7 22:23:38 2008
@@ -1133,14 +1133,12 @@
     c_end = s + python.PyString_GET_SIZE(pystring)
     is_non_ascii = 0
     while s < c_end:
-        c = s[0]
-        if c & 0x80:
+        if s[0] & 0x80:
             # skip the entire multi byte sequence
-            while c & 0x80:
+            while s[0] & 0x80:
                 s += 1
-                c = s[0]
             is_non_ascii = 1
-        elif not tree.xmlIsChar_ch(c):
+        elif not tree.xmlIsChar_ch(s[0]):
             return -1 # invalid!
         s += 1
     return is_non_ascii

From lxml-checkins at codespeak.net  Mon Sep 15 05:22:11 2008
From: lxml-checkins at codespeak.net (VIAGRA INC)
Date: Mon, 15 Sep 2008 05:22:11 +0200 (CEST)
Subject: [Lxml-checkins] SALE 89% OFF
Message-ID: <20080915042004.3172.qmail@cpc2-mfld11-0-0-cust369.nott.cable.ntl.com>

An HTML attachment was scrubbed...
URL: http://codespeak.net/pipermail/lxml-checkins/attachments/20080915/3d81a058/attachment.htm 

From scoder at codespeak.net  Sun Sep 21 16:15:12 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:12 +0200 (CEST)
Subject: [Lxml-checkins] r58301 - in lxml/trunk: . src/lxml/html
Message-ID: <20080921141512.AFF81169F28@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:12 2008
New Revision: 58301

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/html/__init__.py
Log:
 r4794 at delle:  sbehnel | 2008-09-09 08:48:03 +0200
 regexp cleanup


Modified: lxml/trunk/src/lxml/html/__init__.py
==============================================================================
--- lxml/trunk/src/lxml/html/__init__.py	(original)
+++ lxml/trunk/src/lxml/html/__init__.py	Sun Sep 21 16:15:12 2008
@@ -1370,7 +1370,7 @@
 # This isn't a general match, but it's a match for what libxml2
 # specifically serialises:
 __replace_meta_content_type = re.compile(
-    r'<meta http-equiv="Content-Type".*?>').sub
+    r'<meta http-equiv="Content-Type"[^>]*>').sub
 
 def tostring(doc, pretty_print=False, include_meta_content_type=False,
              encoding=None, method="html"):

From scoder at codespeak.net  Sun Sep 21 16:15:17 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:17 +0200 (CEST)
Subject: [Lxml-checkins] r58302 - in lxml/trunk: . src/lxml/tests
Message-ID: <20080921141517.02BA6169F2A@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:17 2008
New Revision: 58302

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/tests/test_elementtree.py
Log:
 r4795 at delle:  sbehnel | 2008-09-09 09:11:50 +0200
 fix namespaces in tests


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	Sun Sep 21 16:15:17 2008
@@ -2650,15 +2650,15 @@
 
     def test_iterparse_attrib_ns(self):
         iterparse = self.etree.iterparse
-        f = BytesIO('<a xmlns="ns1"><b><c xmlns="ns2"/></b></a>')
+        f = BytesIO('<a xmlns="http://ns1/"><b><c xmlns="http://ns2/"/></b></a>')
 
-        attr_name = '{testns}bla'
+        attr_name = '{http://testns/}bla'
         events = []
         iterator = iterparse(f, events=('start','end','start-ns','end-ns'))
         for event, elem in iterator:
             events.append(event)
             if event == 'start':
-                if elem.tag != '{ns1}a':
+                if elem.tag != '{http://ns1/}a':
                     elem.set(attr_name, 'value')
 
         self.assertEquals(

From scoder at codespeak.net  Sun Sep 21 16:15:22 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:22 +0200 (CEST)
Subject: [Lxml-checkins] r58303 - in lxml/trunk: . doc
Message-ID: <20080921141522.8F689169F2C@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:22 2008
New Revision: 58303

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/doc/parsing.txt
Log:
 r4796 at delle:  sbehnel | 2008-09-09 09:13:51 +0200
 fix namespaces in tests, some test setup fixes


Modified: lxml/trunk/doc/parsing.txt
==============================================================================
--- lxml/trunk/doc/parsing.txt	(original)
+++ lxml/trunk/doc/parsing.txt	Sun Sep 21 16:15:22 2008
@@ -189,7 +189,8 @@
   >>> parser = etree.HTMLParser()
   >>> tree   = etree.parse(StringIO(broken_html), parser)
 
-  >>> result = etree.tostring(tree.getroot(), pretty_print=True)
+  >>> result = etree.tostring(tree.getroot(),
+  ...                         pretty_print=True, method="html")
   >>> print(result)
   <html>
     <head>
@@ -206,7 +207,7 @@
 .. sourcecode:: pycon
 
   >>> html = etree.HTML(broken_html)
-  >>> result = etree.tostring(html, pretty_print=True)
+  >>> result = etree.tostring(html, pretty_print=True, method="html")
   >>> print(result)
   <html>
     <head>
@@ -426,7 +427,7 @@
   ... <root>
   ...   <element key='value'>text</element>
   ...   <element>text</element>tail
-  ...   <empty-element xmlns="testns" />
+  ...   <empty-element xmlns="http://testns/" />
   ... </root>
   ... '''
 
@@ -435,7 +436,7 @@
   ...     print("%s: %s" % (action, elem.tag))
   end: element
   end: element
-  end: {testns}empty-element
+  end: {http://testns/}empty-element
   end: root
 
 The resulting tree is available through the ``root`` property of the iterator:
@@ -458,8 +459,8 @@
   end: element
   start: element
   end: element
-  start: {testns}empty-element
-  end: {testns}empty-element
+  start: {http://testns/}empty-element
+  end: {http://testns/}empty-element
   end: root
 
 The 'start-ns' and 'end-ns' events notify about namespace
@@ -476,7 +477,7 @@
   <root>
     <element key='value'>text</element>
     <element>text</element>tail
-    <empty-element xmlns="testns" />
+    <empty-element xmlns="http://testns/" />
   </root>
 
   >>> events = ("start", "end", "start-ns", "end-ns")
@@ -493,9 +494,9 @@
   end: element
   start: element
   end: element
-  start-ns: ('', 'testns')
-  start: {testns}empty-element
-  end: {testns}empty-element
+  start-ns: ('', 'http://testns/')
+  start: {http://testns/}empty-element
+  end: {http://testns/}empty-element
   end-ns
   end: root
 
@@ -517,11 +518,11 @@
 
   >>> events = ("start", "end")
   >>> context = etree.iterparse(
-  ...             StringIO(xml), events=events, tag="{testns}*")
+  ...             StringIO(xml), events=events, tag="{http://testns/}*")
   >>> for action, elem in context:
   ...     print("%s: %s" % (action, elem.tag))
-  start: {testns}empty-element
-  end: {testns}empty-element
+  start: {http://testns/}empty-element
+  end: {http://testns/}empty-element
 
 
 Comments and PIs
@@ -540,7 +541,7 @@
   ...   <element key='value'>text</element>
   ...   <!-- another comment -->
   ...   <element>text</element>tail
-  ...   <empty-element xmlns="testns" />
+  ...   <empty-element xmlns="http://testns/" />
   ... </root>
   ... '''
 
@@ -561,8 +562,8 @@
   comment: - another comment -
   start: element
   end: element
-  start: {testns}empty-element
-  end: {testns}empty-element
+  start: {http://testns/}empty-element
+  end: {http://testns/}empty-element
   end: root
 
   >>> print(context.root.tag)

From scoder at codespeak.net  Sun Sep 21 16:15:26 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:26 +0200 (CEST)
Subject: [Lxml-checkins] r58304 - lxml/trunk
Message-ID: <20080921141526.DF131169F72@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:26 2008
New Revision: 58304

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/IDEAS.txt
Log:
 r4797 at delle:  sbehnel | 2008-09-17 15:20:13 +0200
 idea: reimplement iterparse()


Modified: lxml/trunk/IDEAS.txt
==============================================================================
--- lxml/trunk/IDEAS.txt	(original)
+++ lxml/trunk/IDEAS.txt	Sun Sep 21 16:15:26 2008
@@ -22,3 +22,12 @@
   access check methods
 
   * maybe custom resolvers are enough, or can be combined with this?
+
+* reimplement iterparse() using the libxml2 xmlReader API
+
+  * Advantage: the implementation can be made safer than the current
+    SAX implementation, as the parser would not interact with the
+    Python-level tree.
+
+  * Disadvantage: the tree has to be built manually. In the current
+    SAX based implementation, libxml2 does it for us.

From scoder at codespeak.net  Sun Sep 21 16:15:33 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:33 +0200 (CEST)
Subject: [Lxml-checkins] r58305 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20080921141533.3F63A16A0AC@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:32 2008
New Revision: 58305

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/parser.pxi
   lxml/trunk/src/lxml/tests/test_etree.py
Log:
 r4798 at delle:  sbehnel | 2008-09-17 15:21:11 +0200
 test case and fix for file(-like) resolver


Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Sun Sep 21 16:15:32 2008
@@ -441,7 +441,7 @@
             c_input = xmlparser.xmlNewInputFromFile(
                 c_context, _cstr(doc_ref._filename))
         elif doc_ref._type == PARSER_DATA_FILE:
-            file_context = _FileReaderContext(doc_ref._file, context, url)
+            file_context = _FileReaderContext(doc_ref._file, context, url, None)
             c_input = file_context._createParserInput(c_context)
             data = file_context
         else:

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	Sun Sep 21 16:15:32 2008
@@ -14,7 +14,7 @@
     sys.path.insert(0, this_dir) # needed for Py3
 
 from common_imports import etree, StringIO, BytesIO, HelperTestCase, fileInTestDir
-from common_imports import LargeFileLikeUnicode, doctest, make_doctest
+from common_imports import SillyFileLike, LargeFileLikeUnicode, doctest, make_doctest
 from common_imports import canonicalize, sorted, _str, _bytes
 
 print("")
@@ -667,6 +667,27 @@
         root = tree.getroot()
         self.assertEquals(root.text, test_url)
 
+    def test_resolve_filelike_dtd(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(dtd_validation=True)
+        assertEqual = self.assertEqual
+        test_url = _str("__nosuch.dtd")
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                assertEqual(url, test_url)
+                return self.resolve_file(
+                    SillyFileLike(
+                        _str('''<!ENTITY myentity "%s">
+                        <!ELEMENT doc ANY>''') % url), context)
+
+        parser.resolvers.add(MyResolver())
+
+        xml = _str('<!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_filename_dtd(self):
         parse = self.etree.parse
         parser = self.etree.XMLParser(attribute_defaults=True)
@@ -712,6 +733,28 @@
         self.assertEquals(
             root[0].attrib, {'default': 'valueB'})
 
+    def test_resolve_file_dtd(self):
+        parse = self.etree.parse
+        parser = self.etree.XMLParser(attribute_defaults=True)
+        assertEqual = self.assertEqual
+        test_url = _str("__nosuch.dtd")
+
+        class MyResolver(self.etree.Resolver):
+            def resolve(self, url, id, context):
+                assertEqual(url, test_url)
+                return self.resolve_file(
+                    open(fileInTestDir('test.dtd'), 'rb'), context)
+
+        parser.resolvers.add(MyResolver())
+
+        xml = _str('<!DOCTYPE a SYSTEM "%s"><a><b/></a>') % test_url
+        tree = parse(StringIO(xml), parser)
+        root = tree.getroot()
+        self.assertEquals(
+            root.attrib,    {'default': 'valueA'})
+        self.assertEquals(
+            root[0].attrib, {'default': 'valueB'})
+
     def test_resolve_empty(self):
         parse = self.etree.parse
         parser = self.etree.XMLParser(load_dtd=True)

From scoder at codespeak.net  Sun Sep 21 16:15:37 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:37 +0200 (CEST)
Subject: [Lxml-checkins] r58306 - in lxml/trunk: . src/lxml
Message-ID: <20080921141537.A6DCB169F2A@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:37 2008
New Revision: 58306

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/parser.pxi
Log:
 r4799 at delle:  sbehnel | 2008-09-17 21:30:57 +0200
 fix for parser encoding keyword


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Sun Sep 21 16:15:37 2008
@@ -15,6 +15,8 @@
 Bugs fixed
 ----------
 
+* Overriding the parser encoding didn't work for many encodings.
+
 * 0-bytes could slip through the API when used inside of Unicode
   strings.
 

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Sun Sep 21 16:15:37 2008
@@ -683,6 +683,7 @@
     def __init__(self, int parse_options, bint for_html, XMLSchema schema,
                  remove_comments, remove_pis, strip_cdata, target,
                  filename, encoding):
+        cdef tree.xmlCharEncodingHandler* enchandler
         cdef int c_encoding
         if not isinstance(self, HTMLParser) and \
                 not isinstance(self, XMLParser) and \
@@ -705,10 +706,10 @@
             self._default_encoding_int = tree.XML_CHAR_ENCODING_NONE
         else:
             encoding = _utf8(encoding)
-            c_encoding = tree.xmlParseCharEncoding(_cstr(encoding))
-            if c_encoding == tree.XML_CHAR_ENCODING_ERROR or \
-                   c_encoding == tree.XML_CHAR_ENCODING_NONE:
+            enchandler = tree.xmlFindCharEncodingHandler(_cstr(encoding))
+            if enchandler is NULL:
                 raise LookupError, u"unknown encoding: '%s'" % encoding
+            tree.xmlCharEncCloseFunc(enchandler)
             self._default_encoding = encoding
             self._default_encoding_int = c_encoding
 

From scoder at codespeak.net  Sun Sep 21 16:15:42 2008
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 21 Sep 2008 16:15:42 +0200 (CEST)
Subject: [Lxml-checkins] r58307 - in lxml/trunk: . src/lxml
Message-ID: <20080921141542.0140A169F72@codespeak.net>

Author: scoder
Date: Sun Sep 21 16:15:42 2008
New Revision: 58307

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/iterparse.pxi
   lxml/trunk/src/lxml/parser.pxi
Log:
 r4800 at delle:  sbehnel | 2008-09-17 22:13:58 +0200
 fix encoding override also for the feed parser


Modified: lxml/trunk/src/lxml/iterparse.pxi
==============================================================================
--- lxml/trunk/src/lxml/iterparse.pxi	(original)
+++ lxml/trunk/src/lxml/iterparse.pxi	Sun Sep 21 16:15:42 2008
@@ -415,6 +415,19 @@
 
         context = <_IterparseContext>self._getPushParserContext()
         __GLOBAL_PARSER_CONTEXT.initParserDict(context._c_ctxt)
+
+        if self._default_encoding is not None:
+            if self._for_html:
+                error = _htmlCtxtResetPush(
+                    context._c_ctxt, NULL, 0,
+                    _cstr(self._default_encoding), self._parse_options)
+            else:
+                xmlparser.xmlCtxtUseOptions(
+                    context._c_ctxt, self._parse_options)
+                error = xmlparser.xmlCtxtResetPush(
+                    context._c_ctxt, NULL, 0, NULL,
+                    _cstr(self._default_encoding))
+
         context.prepare()
         # parser will not be unlocked - no other methods supported
 

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Sun Sep 21 16:15:42 2008
@@ -678,7 +678,6 @@
     cdef object _filename
     cdef object _target
     cdef object _default_encoding
-    cdef int _default_encoding_int
 
     def __init__(self, int parse_options, bint for_html, XMLSchema schema,
                  remove_comments, remove_pis, strip_cdata, target,
@@ -703,7 +702,6 @@
 
         if encoding is None:
             self._default_encoding = None
-            self._default_encoding_int = tree.XML_CHAR_ENCODING_NONE
         else:
             encoding = _utf8(encoding)
             enchandler = tree.xmlFindCharEncodingHandler(_cstr(encoding))
@@ -711,7 +709,6 @@
                 raise LookupError, u"unknown encoding: '%s'" % encoding
             tree.xmlCharEncCloseFunc(enchandler)
             self._default_encoding = encoding
-            self._default_encoding_int = c_encoding
 
     cdef _ParserContext _getParserContext(self):
         cdef xmlparser.xmlParserCtxt* pctxt
@@ -777,7 +774,7 @@
             c_filename = NULL
         if self._for_html:
             c_ctxt = htmlparser.htmlCreatePushParserCtxt(
-                NULL, NULL, NULL, 0, c_filename, self._default_encoding_int)
+                NULL, NULL, NULL, 0, c_filename, tree.XML_CHAR_ENCODING_NONE)
             if c_ctxt is not NULL:
                 htmlparser.htmlCtxtUseOptions(c_ctxt, self._parse_options)
         else:
@@ -785,9 +782,6 @@
                 NULL, NULL, NULL, 0, c_filename)
             if c_ctxt is not NULL:
                 xmlparser.xmlCtxtUseOptions(c_ctxt, self._parse_options)
-                if self._default_encoding_int != tree.XML_CHAR_ENCODING_NONE:
-                    xmlparser.xmlSwitchEncoding(
-                        c_ctxt, self._default_encoding_int)
         return c_ctxt
 
     property error_log: