From scoder at codespeak.net  Fri Nov  6 09:20:55 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  6 Nov 2009 09:20:55 +0100 (CET)
Subject: [Lxml-checkins] r69020 - in lxml/trunk: . src/lxml
Message-ID: <20091106082055.61D81168435@codespeak.net>

Author: scoder
Date: Fri Nov  6 09:20:54 2009
New Revision: 69020

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/lxml.etree.pyx
Log:
 r5308 at delle:  sbehnel | 2009-11-02 16:25:20 +0100
 minor cleanup


Modified: lxml/trunk/src/lxml/lxml.etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/lxml.etree.pyx	(original)
+++ lxml/trunk/src/lxml/lxml.etree.pyx	Fri Nov  6 09:20:54 2009
@@ -265,12 +265,7 @@
         # if there are no more references to the document, it is safe
         # to clean the whole thing up, as all nodes have a reference to
         # the document
-        #print "freeing document:", <int>self._c_doc
-        #displayNode(<xmlNode*>self._c_doc, 0)
-        #print <long>self._c_doc, self._c_doc.dict is __GLOBAL_PARSER_CONTEXT._c_dict
-        #print <long>self._c_doc, canDeallocateChildNodes(<xmlNode*>self._c_doc)
         tree.xmlFreeDoc(self._c_doc)
-        #_deallocDocument(self._c_doc)
 
     cdef getroot(self):
         # return an element proxy for the document root
@@ -397,7 +392,7 @@
 cdef __initPrefixCache():
     cdef int i
     return tuple([ python.PyString_FromFormat("ns%d", i)
-                   for i from 0 <= i < 30 ])
+                   for i in range(30) ])
 
 cdef object _PREFIX_CACHE
 _PREFIX_CACHE = __initPrefixCache()

From scoder at codespeak.net  Fri Nov  6 09:21:31 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  6 Nov 2009 09:21:31 +0100 (CET)
Subject: [Lxml-checkins] r69021 - lxml/trunk
Message-ID: <20091106082131.7CD61168435@codespeak.net>

Author: scoder
Date: Fri Nov  6 09:21:31 2009
New Revision: 69021

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/INSTALL.txt
Log:
 r5310 at delle:  sbehnel | 2009-11-06 09:21:26 +0100
 clarification in installation instructions


Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Fri Nov  6 09:21:31 2009
@@ -1,12 +1,17 @@
 Installing lxml
 ===============
 
+For special installation instructions regarding MS Windows and
+MacOS-X, see below.
+
 Requirements
 ------------
 
 You need Python 2.3 or later.
 
-You need libxml2 and libxslt, in particular:
+Unless you are using a static binary distribution (e.g. a Windows
+binary egg from PyPI), you need to install libxml2 and libxslt, in
+particular:
 
 * libxml 2.6.21 or later. It can be found here:
   http://xmlsoft.org/downloads.html
@@ -25,16 +30,35 @@
 Installation
 ------------
 
-If you have easy_install_, you can run the following as super-user (or
+Get the `easy_install`_ tool and run the following as super-user (or
 administrator)::
 
   easy_install lxml
 
 .. _easy_install: http://peak.telecommunity.com/DevCenter/EasyInstall
 
-This has been reported to work on Linux, MacOS-X 10.4 and Windows, as long as
-libxml2 and libxslt are properly installed (including development packages,
-i.e. header files, etc.).
+* On **MS Windows**, the above will install the binary builds that we
+  provide.  If there is no binary build of the latest release yet,
+  please search PyPI_ for the last release that has them and pass that
+  version to ``easy_install`` like this:
+
+    easy_install lxml==2.2.2
+
+* On **Linux** (and most other well-behaved operating systems),
+  ``easy_install`` will manage to build the source distribution as
+  long as libxml2 and libxslt are properly installed, including
+  development packages, i.e. header files, etc.  Use your package
+  management tool to look for packages like ``libxml2-dev`` or
+  ``libxslt-devel`` if the build fails, and make sure they are
+  installed.
+
+* On **MacOS-X**, use the following to build the source distribution,
+  and make sure you have a working Internet connection, as this will
+  download libxml2 and libxslt in order to build them::
+
+   STATIC_DEPS=true easy_install lxml
+
+.. _PyPI: http://cheeseshop.python.org/pypi/lxml
 
 
 Building lxml from sources
@@ -65,9 +89,14 @@
 no need to install the external libraries if you use an official lxml
 build from PyPI.
 
-If you want to upgrade the libraries and/or compile lxml from sources, you
-should install a `binary distribution`_ of libxml2 and libxslt.  You need both
-libxml2 and libxslt, as well as iconv and zlib.
+Unless you know what you are doing, this means: *do not install
+libxml2 or libxslt if you use a binary build of lxml*.  Just use
+``easy_install`` by following the installation instructions above.
+
+*Only* if you want to upgrade the libraries and/or compile lxml from
+sources, you should install a `binary distribution`_ of libxml2 and
+libxslt.  You need both libxml2 and libxslt, as well as iconv and
+zlib.
 
 .. _`binary distribution`: http://www.zlatkovic.com/libxml.en.html
 .. _`binary egg distribution of lxml`: http://cheeseshop.python.org/pypi/lxml
@@ -76,17 +105,25 @@
 MacOS-X
 -------
 
-The system libraries of libxml2 and libxslt installed under MacOS-X
-tend to be rather outdated.  In any case, they are older than the
-required versions for lxml 2.x, so you will have a hard time getting
-lxml to work without installing newer libraries.
-
-A number of users reported success with updated libraries (e.g. using
-fink_ or macports), but needed to set the runtime environment variable
-``DYLD_LIBRARY_PATH`` to the directory where fink keeps the libraries.
-See the `FAQ entry on MacOS-X`_ for more information.
+A macport of lxml is available.  Try ``port install py25-lxml``.
+
+If you want to use a more recent lxml release, you may have to build
+it yourself.  Apple doesn't help here, as the system libraries of
+libxml2 and libxslt installed under MacOS-X are horribly outdated, and
+updating them is everything but easy.  In any case, you cannot run
+lxml 2.x with the system provided libraries, so you have to use newer
+libraries.
+
+Luckily, lxml's ``setup.py`` script has built-in support for building
+and integrating these libraries statically during the build.  Please
+read the `MacOS-X build instructions`_.
 
+.. _`MacOS-X build instructions`: build.html#building-lxml-on-macos-x
 .. _fink: http://finkproject.org/
-.. _`FAQ entry on MacOS-X`: FAQ.html#my-application-crashes-on-macos-x
 
-A macport of lxml is available.  Try ``port install py25-lxml``.
+A number of users also reported success with updated libraries
+(e.g. using fink_ or macports), but needed to set the runtime
+environment variable ``DYLD_LIBRARY_PATH`` to the directory where fink
+keeps the libraries.  In any case, this method is easy to get wrong
+and everything but safe.  Unless you know what you are doing, follow
+the static build instructions above.

From scoder at codespeak.net  Fri Nov  6 09:22:56 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  6 Nov 2009 09:22:56 +0100 (CET)
Subject: [Lxml-checkins] r69022 - lxml/branch/lxml-2.2
Message-ID: <20091106082256.13A81168435@codespeak.net>

Author: scoder
Date: Fri Nov  6 09:22:55 2009
New Revision: 69022

Modified:
   lxml/branch/lxml-2.2/INSTALL.txt   (contents, props changed)
Log:
trunk merge: docs

Modified: lxml/branch/lxml-2.2/INSTALL.txt
==============================================================================
--- lxml/branch/lxml-2.2/INSTALL.txt	(original)
+++ lxml/branch/lxml-2.2/INSTALL.txt	Fri Nov  6 09:22:55 2009
@@ -1,12 +1,17 @@
 Installing lxml
 ===============
 
+For special installation instructions regarding MS Windows and
+MacOS-X, see below.
+
 Requirements
 ------------
 
 You need Python 2.3 or later.
 
-You need libxml2 and libxslt, in particular:
+Unless you are using a static binary distribution (e.g. a Windows
+binary egg from PyPI), you need to install libxml2 and libxslt, in
+particular:
 
 * libxml 2.6.21 or later. It can be found here:
   http://xmlsoft.org/downloads.html
@@ -25,16 +30,35 @@
 Installation
 ------------
 
-If you have easy_install_, you can run the following as super-user (or
+Get the `easy_install`_ tool and run the following as super-user (or
 administrator)::
 
   easy_install lxml
 
 .. _easy_install: http://peak.telecommunity.com/DevCenter/EasyInstall
 
-This has been reported to work on Linux, MacOS-X 10.4 and Windows, as long as
-libxml2 and libxslt are properly installed (including development packages,
-i.e. header files, etc.).
+* On **MS Windows**, the above will install the binary builds that we
+  provide.  If there is no binary build of the latest release yet,
+  please search PyPI_ for the last release that has them and pass that
+  version to ``easy_install`` like this:
+
+    easy_install lxml==2.2.2
+
+* On **Linux** (and most other well-behaved operating systems),
+  ``easy_install`` will manage to build the source distribution as
+  long as libxml2 and libxslt are properly installed, including
+  development packages, i.e. header files, etc.  Use your package
+  management tool to look for packages like ``libxml2-dev`` or
+  ``libxslt-devel`` if the build fails, and make sure they are
+  installed.
+
+* On **MacOS-X**, use the following to build the source distribution,
+  and make sure you have a working Internet connection, as this will
+  download libxml2 and libxslt in order to build them::
+
+   STATIC_DEPS=true easy_install lxml
+
+.. _PyPI: http://cheeseshop.python.org/pypi/lxml
 
 
 Building lxml from sources
@@ -65,9 +89,14 @@
 no need to install the external libraries if you use an official lxml
 build from PyPI.
 
-If you want to upgrade the libraries and/or compile lxml from sources, you
-should install a `binary distribution`_ of libxml2 and libxslt.  You need both
-libxml2 and libxslt, as well as iconv and zlib.
+Unless you know what you are doing, this means: *do not install
+libxml2 or libxslt if you use a binary build of lxml*.  Just use
+``easy_install`` by following the installation instructions above.
+
+*Only* if you want to upgrade the libraries and/or compile lxml from
+sources, you should install a `binary distribution`_ of libxml2 and
+libxslt.  You need both libxml2 and libxslt, as well as iconv and
+zlib.
 
 .. _`binary distribution`: http://www.zlatkovic.com/libxml.en.html
 .. _`binary egg distribution of lxml`: http://cheeseshop.python.org/pypi/lxml
@@ -76,17 +105,25 @@
 MacOS-X
 -------
 
-The system libraries of libxml2 and libxslt installed under MacOS-X
-tend to be rather outdated.  In any case, they are older than the
-required versions for lxml 2.x, so you will have a hard time getting
-lxml to work without installing newer libraries.
-
-A number of users reported success with updated libraries (e.g. using
-fink_ or macports), but needed to set the runtime environment variable
-``DYLD_LIBRARY_PATH`` to the directory where fink keeps the libraries.
-See the `FAQ entry on MacOS-X`_ for more information.
+A macport of lxml is available.  Try ``port install py25-lxml``.
+
+If you want to use a more recent lxml release, you may have to build
+it yourself.  Apple doesn't help here, as the system libraries of
+libxml2 and libxslt installed under MacOS-X are horribly outdated, and
+updating them is everything but easy.  In any case, you cannot run
+lxml 2.x with the system provided libraries, so you have to use newer
+libraries.
+
+Luckily, lxml's ``setup.py`` script has built-in support for building
+and integrating these libraries statically during the build.  Please
+read the `MacOS-X build instructions`_.
 
+.. _`MacOS-X build instructions`: build.html#building-lxml-on-macos-x
 .. _fink: http://finkproject.org/
-.. _`FAQ entry on MacOS-X`: FAQ.html#my-application-crashes-on-macos-x
 
-A macport of lxml is available.  Try ``port install py25-lxml``.
+A number of users also reported success with updated libraries
+(e.g. using fink_ or macports), but needed to set the runtime
+environment variable ``DYLD_LIBRARY_PATH`` to the directory where fink
+keeps the libraries.  In any case, this method is easy to get wrong
+and everything but safe.  Unless you know what you are doing, follow
+the static build instructions above.

From scoder at codespeak.net  Fri Nov  6 09:27:46 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  6 Nov 2009 09:27:46 +0100 (CET)
Subject: [Lxml-checkins] r69023 - lxml/trunk
Message-ID: <20091106082746.8B3E0168435@codespeak.net>

Author: scoder
Date: Fri Nov  6 09:27:46 2009
New Revision: 69023

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/INSTALL.txt
Log:
 r5312 at delle:  sbehnel | 2009-11-06 09:27:41 +0100
 typo


Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Fri Nov  6 09:27:46 2009
@@ -40,7 +40,7 @@
 * On **MS Windows**, the above will install the binary builds that we
   provide.  If there is no binary build of the latest release yet,
   please search PyPI_ for the last release that has them and pass that
-  version to ``easy_install`` like this:
+  version to ``easy_install`` like this::
 
     easy_install lxml==2.2.2
 

From scoder at codespeak.net  Fri Nov  6 09:28:08 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri,  6 Nov 2009 09:28:08 +0100 (CET)
Subject: [Lxml-checkins] r69024 - lxml/branch/lxml-2.2
Message-ID: <20091106082808.29B11168435@codespeak.net>

Author: scoder
Date: Fri Nov  6 09:28:07 2009
New Revision: 69024

Modified:
   lxml/branch/lxml-2.2/INSTALL.txt   (contents, props changed)
Log:
trunk merge: docs

Modified: lxml/branch/lxml-2.2/INSTALL.txt
==============================================================================
--- lxml/branch/lxml-2.2/INSTALL.txt	(original)
+++ lxml/branch/lxml-2.2/INSTALL.txt	Fri Nov  6 09:28:07 2009
@@ -40,7 +40,7 @@
 * On **MS Windows**, the above will install the binary builds that we
   provide.  If there is no binary build of the latest release yet,
   please search PyPI_ for the last release that has them and pass that
-  version to ``easy_install`` like this:
+  version to ``easy_install`` like this::
 
     easy_install lxml==2.2.2
 

From scoder at codespeak.net  Wed Nov 11 16:30:05 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 16:30:05 +0100 (CET)
Subject: [Lxml-checkins] r69175 - lxml/branch/lxml-2.2
Message-ID: <20091111153005.0BEBE1680FE@codespeak.net>

Author: scoder
Date: Wed Nov 11 16:30:04 2009
New Revision: 69175

Modified:
   lxml/branch/lxml-2.2/setupinfo.py
Log:
fix static build

Modified: lxml/branch/lxml-2.2/setupinfo.py
==============================================================================
--- lxml/branch/lxml-2.2/setupinfo.py	(original)
+++ lxml/branch/lxml-2.2/setupinfo.py	Wed Nov 11 16:30:04 2009
@@ -46,7 +46,6 @@
             'libs', 'build/tmp',
             static_include_dirs, static_library_dirs,
             static_cflags, static_binaries,
-            libiconv_version=OPTION_LIBICONV_VERSION,
             libxml2_version=OPTION_LIBXML2_VERSION,
             libxslt_version=OPTION_LIBXSLT_VERSION)
     if CYTHON_INSTALLED:
@@ -338,4 +337,3 @@
     OPTION_STATIC = True
 OPTION_LIBXML2_VERSION = option_value('libxml2-version')
 OPTION_LIBXSLT_VERSION = option_value('libxslt-version')
-OPTION_LIBICONV_VERSION = option_value('libiconv-version')

From scoder at codespeak.net  Wed Nov 11 16:30:26 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 16:30:26 +0100 (CET)
Subject: [Lxml-checkins] r69176 - in lxml/branch/lxml-2.2: . doc
Message-ID: <20091111153026.A3A381680FE@codespeak.net>

Author: scoder
Date: Wed Nov 11 16:30:26 2009
New Revision: 69176

Modified:
   lxml/branch/lxml-2.2/CHANGES.txt
   lxml/branch/lxml-2.2/doc/main.txt
   lxml/branch/lxml-2.2/version.txt
Log:
prepare release of 2.2.4

Modified: lxml/branch/lxml-2.2/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.2/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.2/CHANGES.txt	Wed Nov 11 16:30:26 2009
@@ -2,6 +2,15 @@
 lxml changelog
 ==============
 
+2.2.4 (2009-11-11)
+==================
+
+Bugs fixed
+----------
+
+* Static build of libxml2/libxslt was broken.
+
+
 2.2.3 (2009-10-30)
 ==================
 

Modified: lxml/branch/lxml-2.2/doc/main.txt
==============================================================================
--- lxml/branch/lxml-2.2/doc/main.txt	(original)
+++ lxml/branch/lxml-2.2/doc/main.txt	Wed Nov 11 16:30:26 2009
@@ -147,8 +147,8 @@
 source release.  If you can't wait, consider trying a less recent
 release version first.
 
-The latest version is `lxml 2.2.3`_, released 2009-10-30
-(`changes for 2.2.3`_).  `Older versions`_ are listed below.
+The latest version is `lxml 2.2.4`_, released 2009-10-30
+(`changes for 2.2.4`_).  `Older versions`_ are listed below.
 
 Please take a look at the `installation instructions`_!
 
@@ -221,7 +221,9 @@
 <http://codespeak.net/lxml/2.1/>`_ and the `current in-development
 version <http://codespeak.net/lxml/dev/>`_.
 
-.. _`PDF documentation`: lxmldoc-2.2.3.pdf
+.. _`PDF documentation`: lxmldoc-2.2.4.pdf
+
+* `lxml 2.2.3`_, released 2009-10-30 (`changes for 2.2.3`_)
 
 * `lxml 2.2.2`_, released 2009-06-21 (`changes for 2.2.2`_)
 
@@ -325,6 +327,7 @@
 
 * `lxml 0.5`_,   released 2005-04-08
 
+.. _`lxml 2.2.4`: lxml-2.2.4.tgz
 .. _`lxml 2.2.3`: lxml-2.2.3.tgz
 .. _`lxml 2.2.2`: lxml-2.2.2.tgz
 .. _`lxml 2.2.1`: lxml-2.2.1.tgz
@@ -377,6 +380,7 @@
 .. _`lxml 0.5.1`: lxml-0.5.1.tgz
 .. _`lxml 0.5`: lxml-0.5.tgz
 
+.. _`changes for 2.2.4`: changes-2.2.4.html
 .. _`changes for 2.2.3`: changes-2.2.3.html
 .. _`changes for 2.2.2`: changes-2.2.2.html
 .. _`changes for 2.2.1`: changes-2.2.1.html

Modified: lxml/branch/lxml-2.2/version.txt
==============================================================================
--- lxml/branch/lxml-2.2/version.txt	(original)
+++ lxml/branch/lxml-2.2/version.txt	Wed Nov 11 16:30:26 2009
@@ -1 +1 @@
-2.2.3
+2.2.4

From scoder at codespeak.net  Wed Nov 11 16:31:33 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 16:31:33 +0100 (CET)
Subject: [Lxml-checkins] r69177 - lxml/tag/lxml-2.2.4
Message-ID: <20091111153133.6D9051680FE@codespeak.net>

Author: scoder
Date: Wed Nov 11 16:31:32 2009
New Revision: 69177

Added:
   lxml/tag/lxml-2.2.4/
      - copied from r69176, lxml/branch/lxml-2.2/
Log:
tag for 2.2.4


From scoder at codespeak.net  Wed Nov 11 21:42:54 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 21:42:54 +0100 (CET)
Subject: [Lxml-checkins] r69203 - lxml/branch/lxml-2.2
Message-ID: <20091111204254.4A03516802D@codespeak.net>

Author: scoder
Date: Wed Nov 11 21:42:52 2009
New Revision: 69203

Modified:
   lxml/branch/lxml-2.2/INSTALL.txt
Log:
docs: install in ActivePython

Modified: lxml/branch/lxml-2.2/INSTALL.txt
==============================================================================
--- lxml/branch/lxml-2.2/INSTALL.txt	(original)
+++ lxml/branch/lxml-2.2/INSTALL.txt	Wed Nov 11 21:42:52 2009
@@ -4,6 +4,15 @@
 For special installation instructions regarding MS Windows and
 MacOS-X, see below.
 
+.. contents::
+   1  Requirements
+   2  Installation
+   3  Installation in ActivePython
+   4  Building lxml from sources
+   5  MS Windows
+   6  MacOS-X
+
+
 Requirements
 ------------
 
@@ -61,6 +70,30 @@
 .. _PyPI: http://cheeseshop.python.org/pypi/lxml
 
 
+Installation in ActivePython
+----------------------------
+
+ActiveState_ provides ready-made lxml builds for different platforms
+in its `package repository`_ for the PyPM_ package manager.  PyPM is
+similar to apt-get in that there is a repository of automaticaly
+pre-built packages for Windows, Mac and Linux.
+
+To install lxml in ActivePython, type the following on one of these
+operating systems::
+
+  $ pypm install lxml
+
+To test the installation, try::
+
+  $ python -c "import lxml; print lxml.__file__"
+
+This should show you the directory where the package was installed.
+
+.. _ActiveState: http://www.activestate.com/
+.. _PyPM: http://docs.activestate.com/activepython/2.6/pypm.html
+.. _`package repository`: http://pypm.activestate.com/
+
+
 Building lxml from sources
 --------------------------
 

From scoder at codespeak.net  Wed Nov 11 21:44:56 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 21:44:56 +0100 (CET)
Subject: [Lxml-checkins] r69204 - in lxml/branch/lxml-2.2: . src/lxml
Message-ID: <20091111204456.9A5AE16802C@codespeak.net>

Author: scoder
Date: Wed Nov 11 21:44:56 2009
New Revision: 69204

Modified:
   lxml/branch/lxml-2.2/CHANGES.txt
   lxml/branch/lxml-2.2/src/lxml/xsltext.pxi
Log:
fix XSLT extension elements when running in a non-element XSLT context

Modified: lxml/branch/lxml-2.2/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.2/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.2/CHANGES.txt	Wed Nov 11 21:44:56 2009
@@ -2,6 +2,22 @@
 lxml changelog
 ==============
 
+2.2.5 (?)
+==================
+
+Features added
+--------------
+
+* Support for running XSLT extension elements on the input root node
+  (e.g. in a template matching on "/").
+
+Bugs fixed
+----------
+
+* Crash in XSLT extension elements when the XSLT context node is not
+  an element.
+
+
 2.2.4 (2009-11-11)
 ==================
 

Modified: lxml/branch/lxml-2.2/src/lxml/xsltext.pxi
==============================================================================
--- lxml/branch/lxml-2.2/src/lxml/xsltext.pxi	(original)
+++ lxml/branch/lxml-2.2/src/lxml/xsltext.pxi	Wed Nov 11 21:44:56 2009
@@ -81,6 +81,7 @@
     cdef XSLTExtension extension
     cdef python.PyObject* dict_result
     cdef char* c_uri
+    cdef xmlNode* c_node
     cdef _ReadOnlyElementProxy context_node, self_node, output_parent
     c_uri = _getNs(c_inst_node)
     if c_uri is NULL:
@@ -100,9 +101,21 @@
 
         try:
             self_node     = _newReadOnlyProxy(None, c_inst_node)
-            context_node  = _newReadOnlyProxy(self_node, c_context_node)
             output_parent = _newAppendOnlyProxy(self_node, c_ctxt.insert)
 
+            if c_context_node.type == tree.XML_ELEMENT_NODE:
+                context_node  = _newReadOnlyProxy(self_node, c_context_node)
+            elif c_context_node.type in (tree.XML_DOCUMENT_NODE,
+                                         tree.XML_HTML_DOCUMENT_NODE):
+                c_node = tree.xmlDocGetRootElement(<xmlDoc*>c_context_node)
+                if c_node is not NULL:
+                    context_node = _newReadOnlyProxy(self_node, c_node)
+                else:
+                    context_node = None
+            else:
+                raise TypeError, \
+                      u"Unsupported XSLT context node type %d, extension elements require an element" % c_context_node.type
+
             context._extension_element_proxy = self_node
             extension.execute(context, self_node, context_node, output_parent)
         finally:

From scoder at codespeak.net  Wed Nov 11 21:50:04 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 21:50:04 +0100 (CET)
Subject: [Lxml-checkins] r69205 - lxml/branch/lxml-2.2/doc
Message-ID: <20091111205004.8D70116802C@codespeak.net>

Author: scoder
Date: Wed Nov 11 21:50:03 2009
New Revision: 69205

Modified:
   lxml/branch/lxml-2.2/doc/main.txt
Log:
fix release date

Modified: lxml/branch/lxml-2.2/doc/main.txt
==============================================================================
--- lxml/branch/lxml-2.2/doc/main.txt	(original)
+++ lxml/branch/lxml-2.2/doc/main.txt	Wed Nov 11 21:50:03 2009
@@ -147,7 +147,7 @@
 source release.  If you can't wait, consider trying a less recent
 release version first.
 
-The latest version is `lxml 2.2.4`_, released 2009-10-30
+The latest version is `lxml 2.2.4`_, released 2009-11-11
 (`changes for 2.2.4`_).  `Older versions`_ are listed below.
 
 Please take a look at the `installation instructions`_!

From scoder at codespeak.net  Wed Nov 11 21:51:50 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 21:51:50 +0100 (CET)
Subject: [Lxml-checkins] r69206 - lxml/trunk
Message-ID: <20091111205150.3FB6E168032@codespeak.net>

Author: scoder
Date: Wed Nov 11 21:51:49 2009
New Revision: 69206

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/INSTALL.txt
Log:
doc merge from 2.2 branch

Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Wed Nov 11 21:51:49 2009
@@ -4,6 +4,15 @@
 For special installation instructions regarding MS Windows and
 MacOS-X, see below.
 
+.. contents::
+   1  Requirements
+   2  Installation
+   3  Installation in ActivePython
+   4  Building lxml from sources
+   5  MS Windows
+   6  MacOS-X
+
+
 Requirements
 ------------
 
@@ -61,6 +70,30 @@
 .. _PyPI: http://cheeseshop.python.org/pypi/lxml
 
 
+Installation in ActivePython
+----------------------------
+
+ActiveState_ provides ready-made lxml builds for different platforms
+in its `package repository`_ for the PyPM_ package manager.  PyPM is
+similar to apt-get in that there is a repository of automaticaly
+pre-built packages for Windows, Mac and Linux.
+
+To install lxml in ActivePython, type the following on one of these
+operating systems::
+
+  $ pypm install lxml
+
+To test the installation, try::
+
+  $ python -c "import lxml; print lxml.__file__"
+
+This should show you the directory where the package was installed.
+
+.. _ActiveState: http://www.activestate.com/
+.. _PyPM: http://docs.activestate.com/activepython/2.6/pypm.html
+.. _`package repository`: http://pypm.activestate.com/
+
+
 Building lxml from sources
 --------------------------
 

From scoder at codespeak.net  Wed Nov 11 22:12:23 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Wed, 11 Nov 2009 22:12:23 +0100 (CET)
Subject: [Lxml-checkins] r69207 - lxml/branch/lxml-2.2/src/lxml/tests
Message-ID: <20091111211223.E71C716802D@codespeak.net>

Author: scoder
Date: Wed Nov 11 22:12:22 2009
New Revision: 69207

Modified:
   lxml/branch/lxml-2.2/src/lxml/tests/test_xslt.py
Log:
additional tests for XSLT extension elements running in non-element contexts

Modified: lxml/branch/lxml-2.2/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/branch/lxml-2.2/src/lxml/tests/test_xslt.py	(original)
+++ lxml/branch/lxml-2.2/src/lxml/tests/test_xslt.py	Wed Nov 11 22:12:22 2009
@@ -721,6 +721,54 @@
         self.assertEquals(self._rootstring(result),
                           _bytes('<A><b>X</b></A>'))
 
+    def test_extension_element_doc_context(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="/">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        tags = []
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                tags.append(input_node.tag)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(tags, ['a'])
+
+    def test_extension_element_unsupported_context(self):
+        tree = self.parse('<a><!--a comment--></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="comment()">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        tags = []
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                tags.append(input_node.tag)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        self.assertRaises(TypeError, tree.xslt, style, extensions=extensions)
+        self.assertEquals(tags, [])
+
     def test_extension_element_content(self):
         tree = self.parse('<a><b>B</b></a>')
         style = self.parse('''\

From scoder at codespeak.net  Fri Nov 13 08:33:54 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Fri, 13 Nov 2009 08:33:54 +0100 (CET)
Subject: [Lxml-checkins] r69245 - lxml/branch/lxml-2.2
Message-ID: <20091113073354.4E7B7168026@codespeak.net>

Author: scoder
Date: Fri Nov 13 08:33:53 2009
New Revision: 69245

Modified:
   lxml/branch/lxml-2.2/INSTALL.txt
Log:
doc fix

Modified: lxml/branch/lxml-2.2/INSTALL.txt
==============================================================================
--- lxml/branch/lxml-2.2/INSTALL.txt	(original)
+++ lxml/branch/lxml-2.2/INSTALL.txt	Fri Nov 13 08:33:53 2009
@@ -5,6 +5,7 @@
 MacOS-X, see below.
 
 .. contents::
+..
    1  Requirements
    2  Installation
    3  Installation in ActivePython

From scoder at codespeak.net  Tue Nov 17 15:20:46 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 17 Nov 2009 15:20:46 +0100 (CET)
Subject: [Lxml-checkins] r69351 - lxml/branch/lxml-2.2
Message-ID: <20091117142046.9FE2D318139@codespeak.net>

Author: scoder
Date: Tue Nov 17 15:20:46 2009
New Revision: 69351

Modified:
   lxml/branch/lxml-2.2/buildlibxml.py
Log:
static build fix for MacOS 10.6

Modified: lxml/branch/lxml-2.2/buildlibxml.py
==============================================================================
--- lxml/branch/lxml-2.2/buildlibxml.py	(original)
+++ lxml/branch/lxml-2.2/buildlibxml.py	Tue Nov 17 15:20:46 2009
@@ -175,15 +175,23 @@
     call_setup = {}
     env_setup = None
     if sys.platform in ('darwin',):
+        import platform
         # We compile Universal if we are on a machine > 10.3
-        major_version = int(os.uname()[2].split('.')[0])
+        major_version, minor_version = map(int, platform.mac_ver()[0].split('.')[:2])
         if major_version > 7:
             env = os.environ.copy()
-            env.update({
-                'CFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk -O2",
-                'LDFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk",
-                'MACOSX_DEPLOYMENT_TARGET' : "10.3"
-                })
+            if minor_version < 6:
+                env.update({
+                    'CFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk -O2",
+                    'LDFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk",
+                    'MACOSX_DEPLOYMENT_TARGET' : "10.3"
+                    })
+            else:
+                env.update({
+                    'CFLAGS' : "-arch ppc -arch i386 -arch x86_64 -O2",
+                    'LDFLAGS' : "-arch ppc -arch i386 -arch x86_64",
+                    'MACOSX_DEPLOYMENT_TARGET' : "10.6"
+                    })
             call_setup['env'] = env
 
         # We may loose the link to iconv, so make sure it's there

From scoder at codespeak.net  Tue Nov 17 20:15:01 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 17 Nov 2009 20:15:01 +0100 (CET)
Subject: [Lxml-checkins] r69357 - lxml/branch/lxml-2.2/doc
Message-ID: <20091117191501.4E0C131813D@codespeak.net>

Author: scoder
Date: Tue Nov 17 20:14:59 2009
New Revision: 69357

Modified:
   lxml/branch/lxml-2.2/doc/build.txt
Log:
mac build docs: show how to install

Modified: lxml/branch/lxml-2.2/doc/build.txt
==============================================================================
--- lxml/branch/lxml-2.2/doc/build.txt	(original)
+++ lxml/branch/lxml-2.2/doc/build.txt	Tue Nov 17 20:14:59 2009
@@ -209,6 +209,8 @@
          --libxml2-version=2.7.3 \
          --libxslt-version=1.1.24 \
 
+  sudo python setup.py install
+
 Instead of ``build``, you can use any target, like ``bdist_egg`` if
 you want to use setuptools to build an installable egg.
 

From scoder at codespeak.net  Sun Nov 22 04:37:55 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Sun, 22 Nov 2009 04:37:55 +0100 (CET)
Subject: [Lxml-checkins] r69500 - lxml/branch/lxml-2.2
Message-ID: <20091122033755.59A3E168026@codespeak.net>

Author: scoder
Date: Sun Nov 22 04:37:53 2009
New Revision: 69500

Modified:
   lxml/branch/lxml-2.2/test.py
Log:
test runner fix for Py2.7/3.2

Modified: lxml/branch/lxml-2.2/test.py
==============================================================================
--- lxml/branch/lxml-2.2/test.py	(original)
+++ lxml/branch/lxml-2.2/test.py	Sun Nov 22 04:37:53 2009
@@ -71,11 +71,18 @@
 import getopt
 import unittest
 import traceback
+
 try:
     set
 except NameError:
     from sets import Set as set
 
+try:
+    # Python >=2.7 and >=3.2
+    from unittest.runner import _TextTestResult
+except ImportError:
+    from unittest import _TextTestResult
+
 __metaclass__ = type
 
 def stderr(text):
@@ -302,14 +309,14 @@
     return results
 
 
-class CustomTestResult(unittest._TextTestResult):
+class CustomTestResult(_TextTestResult):
     """Customised TestResult.
 
     It can show a progress bar, and displays tracebacks for errors and failures
     as soon as they happen, in addition to listing them all at the end.
     """
 
-    __super = unittest._TextTestResult
+    __super = _TextTestResult
     __super_init = __super.__init__
     __super_startTest = __super.startTest
     __super_stopTest = __super.stopTest

From scoder at codespeak.net  Tue Nov 24 18:58:44 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:58:44 +0100 (CET)
Subject: [Lxml-checkins] r69597 - lxml/trunk
Message-ID: <20091124175844.BF4D1168006@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:58:42 2009
New Revision: 69597

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/INSTALL.txt
Log:
 r5315 at delle:  sbehnel | 2009-11-11 15:41:12 +0100
 doc clarification


Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Tue Nov 24 18:58:42 2009
@@ -2,7 +2,7 @@
 ===============
 
 For special installation instructions regarding MS Windows and
-MacOS-X, see below.
+MacOS-X, see the specific sections below.
 
 .. contents::
    1  Requirements
@@ -65,7 +65,7 @@
   and make sure you have a working Internet connection, as this will
   download libxml2 and libxslt in order to build them::
 
-   STATIC_DEPS=true easy_install lxml
+   STATIC_DEPS=true sudo easy_install lxml
 
 .. _PyPI: http://cheeseshop.python.org/pypi/lxml
 

From scoder at codespeak.net  Tue Nov 24 18:58:53 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:58:53 +0100 (CET)
Subject: [Lxml-checkins] r69598 - in lxml/trunk: . src/lxml
Message-ID: <20091124175853.9376D168006@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:58:52 2009
New Revision: 69598

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/apihelpers.pxi
   lxml/trunk/src/lxml/etree_defs.h
   lxml/trunk/src/lxml/iterparse.pxi
   lxml/trunk/src/lxml/lxml.objectify.pyx
   lxml/trunk/src/lxml/objectpath.pxi
   lxml/trunk/src/lxml/parser.pxi
   lxml/trunk/src/lxml/python.pxd
   lxml/trunk/src/lxml/serializer.pxi
   lxml/trunk/src/lxml/xpath.pxi
   lxml/trunk/src/lxml/xslt.pxi
Log:
 r5316 at delle:  sbehnel | 2009-11-11 17:26:12 +0100
 use PyBytes_*() instead of PyString_*()


Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Tue Nov 24 18:58:52 2009
@@ -652,7 +652,7 @@
     elif isinstance(value, CDATA):
         c_text_node = tree.xmlNewCDataBlock(
             c_node.doc, _cstr((<CDATA>value)._utf8_data),
-            python.PyString_GET_SIZE((<CDATA>value)._utf8_data))
+            python.PyBytes_GET_SIZE((<CDATA>value)._utf8_data))
     else:
         # this will raise the right error
        _utf8(value)
@@ -683,7 +683,7 @@
     else:
         c_ns = element._doc._findOrBuildNodeNs(
             element._c_node, _cstr(ns), NULL)
-        return python.PyString_FromFormat('%s:%s', c_ns.prefix, _cstr(tag))
+        return python.PyBytes_FromFormat('%s:%s', c_ns.prefix, _cstr(tag))
 
 cdef inline bint _hasChild(xmlNode* c_node):
     return c_node is not NULL and _findChildForwards(c_node, 0) is not NULL
@@ -1197,7 +1197,7 @@
     cdef char c
     cdef bint is_non_ascii
     s = _cstr(pystring)
-    c_end = s + python.PyString_GET_SIZE(pystring)
+    c_end = s + python.PyBytes_GET_SIZE(pystring)
     is_non_ascii = 0
     while s < c_end:
         if s[0] & 0x80:
@@ -1229,7 +1229,7 @@
     slen = spos - s
     if is_non_ascii:
         return python.PyUnicode_DecodeUTF8(s, slen, NULL)
-    return python.PyString_FromStringAndSize(s, slen)
+    return python.PyBytes_FromStringAndSize(s, slen)
 
 cdef object _utf8(object s):
     cdef int invalid
@@ -1321,13 +1321,13 @@
         try:
             # try to decode with default encoding
             filename = python.PyUnicode_Decode(
-                c_filename, python.PyString_GET_SIZE(filename),
+                c_filename, python.PyBytes_GET_SIZE(filename),
                 _C_FILENAME_ENCODING, NULL)
         except UnicodeDecodeError, decode_exc:
             try:
                 # try if it's UTF-8
                 filename = python.PyUnicode_DecodeUTF8(
-                    c_filename, python.PyString_GET_SIZE(filename), NULL)
+                    c_filename, python.PyBytes_GET_SIZE(filename), NULL)
             except UnicodeDecodeError:
                 raise decode_exc # otherwise re-raise original exception
     if python.PyUnicode_Check(filename):
@@ -1354,13 +1354,13 @@
         if c_ns_end is NULL:
             raise ValueError, u"Invalid tag name"
         nslen  = c_ns_end - c_tag
-        taglen = python.PyString_GET_SIZE(tag) - nslen - 2
+        taglen = python.PyBytes_GET_SIZE(tag) - nslen - 2
         if taglen == 0:
             raise ValueError, u"Empty tag name"
         if nslen > 0:
-            ns = python.PyString_FromStringAndSize(c_tag,   nslen)
-        tag = python.PyString_FromStringAndSize(c_ns_end+1, taglen)
-    elif python.PyString_GET_SIZE(tag) == 0:
+            ns = python.PyBytes_FromStringAndSize(c_tag,   nslen)
+        tag = python.PyBytes_FromStringAndSize(c_ns_end+1, taglen)
+    elif python.PyBytes_GET_SIZE(tag) == 0:
         raise ValueError, u"Empty tag name"
     return ns, tag
 
@@ -1445,7 +1445,7 @@
     elif python.IS_PYTHON3:
         return python.PyUnicode_FromFormat("{%s}%s", href, name)
     else:
-        s = python.PyString_FromFormat("{%s}%s", href, name)
+        s = python.PyBytes_FromFormat("{%s}%s", href, name)
         if isutf8(href) or isutf8(name):
             return python.PyUnicode_FromEncodedObject(s, 'UTF-8', NULL)
         else:

Modified: lxml/trunk/src/lxml/etree_defs.h
==============================================================================
--- lxml/trunk/src/lxml/etree_defs.h	(original)
+++ lxml/trunk/src/lxml/etree_defs.h	Tue Nov 24 18:58:52 2009
@@ -36,6 +36,12 @@
 #  define PyString_GET_SIZE(s)               PyBytes_GET_SIZE(s)
 #  define PyString_AS_STRING(s)              PyBytes_AS_STRING(s)
 #else
+#  define PyBytes_CheckExact(o)              PyString_CheckExact(o)
+#  define PyBytes_Check(o)                   PyString_Check(o)
+#  define PyBytes_FromStringAndSize(s, len)  PyString_FromStringAndSize(s, len)
+#  define PyBytes_FromFormat                 PyString_FromFormat
+#  define PyBytes_GET_SIZE(s)                PyString_GET_SIZE(s)
+#  define PyBytes_AS_STRING(s)               PyString_AS_STRING(s)
 #if PY_VERSION_HEX < 0x02060000
 /* we currently only use three parameters - MSVC can't compile (s, ...) */
 #  define PyUnicode_FromFormat(s, a, b) (NULL)

Modified: lxml/trunk/src/lxml/iterparse.pxi
==============================================================================
--- lxml/trunk/src/lxml/iterparse.pxi	(original)
+++ lxml/trunk/src/lxml/iterparse.pxi	Tue Nov 24 18:58:52 2009
@@ -482,13 +482,13 @@
                 if not python.PyBytes_Check(data):
                     self._source = None
                     raise TypeError, u"reading file objects must return plain strings"
-                c_data_len = python.PyString_GET_SIZE(data)
+                c_data_len = python.PyBytes_GET_SIZE(data)
                 c_data = _cstr(data)
                 done = (c_data_len == 0)
                 error = self._parse_chunk(pctxt, c_data, c_data_len, done)
             else:
                 if self._buffer is None:
-                    self._buffer = python.PyString_FromStringAndSize(
+                    self._buffer = python.PyBytes_FromStringAndSize(
                         NULL, __ITERPARSE_CHUNK_SIZE)
                 c_data = _cstr(self._buffer)
                 with nogil:

Modified: lxml/trunk/src/lxml/lxml.objectify.pyx
==============================================================================
--- lxml/trunk/src/lxml/lxml.objectify.pyx	(original)
+++ lxml/trunk/src/lxml/lxml.objectify.pyx	Tue Nov 24 18:58:52 2009
@@ -466,7 +466,7 @@
     c_node = parent._c_node
     ns, tag = cetree.getNsTag(tag)
     c_tag = tree.xmlDictExists(
-        c_node.doc.dict, _cstr(tag), python.PyString_GET_SIZE(tag))
+        c_node.doc.dict, _cstr(tag), python.PyBytes_GET_SIZE(tag))
     if c_tag is NULL:
         return None
     if ns is None:

Modified: lxml/trunk/src/lxml/objectpath.pxi
==============================================================================
--- lxml/trunk/src/lxml/objectpath.pxi	(original)
+++ lxml/trunk/src/lxml/objectpath.pxi	Tue Nov 24 18:58:52 2009
@@ -153,11 +153,11 @@
                 if index_end is NULL:
                     raise ValueError, u"index must be enclosed in []"
                 index = int(
-                    python.PyString_FromStringAndSize(
+                    python.PyBytes_FromStringAndSize(
                         index_pos + 1, <Py_ssize_t>(index_end - index_pos - 1)))
                 if python.PyList_GET_SIZE(new_path) == 0 and index != 0:
                     raise ValueError, u"index not allowed on root node"
-                name = python.PyString_FromStringAndSize(
+                name = python.PyBytes_FromStringAndSize(
                     c_name, <Py_ssize_t>(index_pos - c_name))
         new_path.append( (ns, name, index) )
     if python.PyList_GET_SIZE(new_path) == 0:

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Tue Nov 24 18:58:52 2009
@@ -347,7 +347,7 @@
             return 0
         try:
             c_byte_count = 0
-            byte_count = python.PyString_GET_SIZE(self._bytes)
+            byte_count = python.PyBytes_GET_SIZE(self._bytes)
             remaining  = byte_count - self._bytes_read
             while c_requested > remaining:
                 c_start = _cstr(self._bytes) + self._bytes_read
@@ -368,7 +368,7 @@
                         raise TypeError, \
                             u"reading from file-like objects must return byte strings or unicode strings"
 
-                remaining = python.PyString_GET_SIZE(self._bytes)
+                remaining = python.PyBytes_GET_SIZE(self._bytes)
                 if remaining == 0:
                     self._bytes_read = -1
                     return c_byte_count
@@ -435,7 +435,7 @@
             c_input = xmlparser.xmlNewInputStream(c_context)
             if c_input is not NULL:
                 c_input.base = _cstr(data)
-                c_input.length = python.PyString_GET_SIZE(data)
+                c_input.length = python.PyBytes_GET_SIZE(data)
                 c_input.cur = c_input.base
                 c_input.end = &c_input.base[c_input.length]
         elif doc_ref._type == PARSER_DATA_FILENAME:
@@ -870,7 +870,7 @@
         py_buffer_len = python.PyUnicode_GET_DATA_SIZE(utext)
         if py_buffer_len > python.INT_MAX or _UNICODE_ENCODING is NULL:
             text_utf = python.PyUnicode_AsUTF8String(utext)
-            py_buffer_len = python.PyString_GET_SIZE(text_utf)
+            py_buffer_len = python.PyBytes_GET_SIZE(text_utf)
             return self._parseDoc(_cstr(text_utf), py_buffer_len, c_filename)
         buffer_len = py_buffer_len
 
@@ -1049,7 +1049,7 @@
             else:
                 c_encoding = self._default_encoding
             c_data = _cstr(data)
-            py_buffer_len = python.PyString_GET_SIZE(data)
+            py_buffer_len = python.PyBytes_GET_SIZE(data)
         elif python.PyUnicode_Check(data):
             if _UNICODE_ENCODING is NULL:
                 raise ParserError, \
@@ -1416,7 +1416,7 @@
                 StringIO(text), filename)
         return (<_BaseParser>parser)._parseUnicodeDoc(text, c_filename)
     else:
-        c_len = python.PyString_GET_SIZE(text)
+        c_len = python.PyBytes_GET_SIZE(text)
         if c_len > python.INT_MAX:
             return (<_BaseParser>parser)._parseDocFromFilelike(
                 BytesIO(text), filename)

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Tue Nov 24 18:58:52 2009
@@ -40,9 +40,10 @@
     cdef char* PyUnicode_AS_DATA(object ustring)
     cdef Py_ssize_t PyUnicode_GET_DATA_SIZE(object ustring)
     cdef Py_ssize_t PyUnicode_GET_SIZE(object ustring)
-    cdef object PyString_FromStringAndSize(char* s, Py_ssize_t size)
-    cdef object PyString_FromFormat(char* format, ...)
-    cdef Py_ssize_t PyString_GET_SIZE(object s)
+    cdef object PyBytes_FromStringAndSize(char* s, Py_ssize_t size)
+    cdef object PyBytes_FromFormat(char* format, ...)
+    cdef object PyString_FromFormat(char* format, ...) # to be deleted!
+    cdef Py_ssize_t PyBytes_GET_SIZE(object s)
 
     cdef object PyNumber_Int(object value)
     cdef Py_ssize_t PyInt_AsSsize_t(object value)
@@ -107,7 +108,7 @@
 
     # some handy functions
     cdef int callable "PyCallable_Check" (object obj)
-    cdef char* _cstr "PyString_AS_STRING" (object s)
+    cdef char* _cstr "PyBytes_AS_STRING" (object s)
 
     # Py_buffer related flags
     cdef int PyBUF_SIMPLE

Modified: lxml/trunk/src/lxml/serializer.pxi
==============================================================================
--- lxml/trunk/src/lxml/serializer.pxi	(original)
+++ lxml/trunk/src/lxml/serializer.pxi	Tue Nov 24 18:58:52 2009
@@ -130,7 +130,7 @@
                 tree.xmlBufferLength(c_result_buffer),
                 'strict')
         else:
-            result = python.PyString_FromStringAndSize(
+            result = python.PyBytes_FromStringAndSize(
                 tree.xmlBufferContent(c_result_buffer),
                 tree.xmlBufferLength(c_result_buffer))
     finally:
@@ -332,7 +332,7 @@
         try:
             if self._filelike is None:
                 raise IOError, u"File is already closed"
-            py_buffer = python.PyString_FromStringAndSize(c_buffer, size)
+            py_buffer = python.PyBytes_FromStringAndSize(c_buffer, size)
             self._filelike.write(py_buffer)
             return size
         except:

Modified: lxml/trunk/src/lxml/xpath.pxi
==============================================================================
--- lxml/trunk/src/lxml/xpath.pxi	(original)
+++ lxml/trunk/src/lxml/xpath.pxi	Tue Nov 24 18:58:52 2009
@@ -473,7 +473,7 @@
         i = 1
         for namespace_def in _find_namespaces(stripped_path):
             if namespace_def not in namespace_defs:
-                prefix = python.PyString_FromFormat("__xpp%02d", i)
+                prefix = python.PyBytes_FromFormat("__xpp%02d", i)
                 i += 1
                 namespace_defs.append(namespace_def)
                 namespace = namespace_def[1:-1] # remove '{}'

Modified: lxml/trunk/src/lxml/xslt.pxi
==============================================================================
--- lxml/trunk/src/lxml/xslt.pxi	(original)
+++ lxml/trunk/src/lxml/xslt.pxi	Tue Nov 24 18:58:52 2009
@@ -689,7 +689,7 @@
             return ''
         # we must not use 'funicode' here as this is not always UTF-8
         try:
-            result = python.PyString_FromStringAndSize(s, l)
+            result = python.PyBytes_FromStringAndSize(s, l)
         finally:
             tree.xmlFree(s)
         return result

From scoder at codespeak.net  Tue Nov 24 18:58:57 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:58:57 +0100 (CET)
Subject: [Lxml-checkins] r69599 - in lxml/trunk: . src/lxml
Message-ID: <20091124175857.DD8AF168008@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:58:57 2009
New Revision: 69599

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/apihelpers.pxi
Log:
 r5317 at delle:  sbehnel | 2009-11-11 20:44:21 +0100
 type fix


Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Tue Nov 24 18:58:57 2009
@@ -675,7 +675,7 @@
     tree.xmlAddNextSibling(c_node, c_text_node)
     return 0
 
-cdef _resolveQNameText(_Element element, value):
+cdef bytes _resolveQNameText(_Element element, value):
     cdef xmlNs* c_ns
     ns, tag = _getNsTag(value)
     if ns is None:

From scoder at codespeak.net  Tue Nov 24 18:59:11 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:11 +0100 (CET)
Subject: [Lxml-checkins] r69600 - in lxml/trunk: . src/lxml
Message-ID: <20091124175911.2773D168006@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:10 2009
New Revision: 69600

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/lxml.etree.pyx
Log:
 r5327 at delle:  sbehnel | 2009-11-12 09:13:43 +0100
 more PyString_ -> PyBytes_ changes


Modified: lxml/trunk/src/lxml/lxml.etree.pyx
==============================================================================
--- lxml/trunk/src/lxml/lxml.etree.pyx	(original)
+++ lxml/trunk/src/lxml/lxml.etree.pyx	Tue Nov 24 18:59:10 2009
@@ -332,7 +332,7 @@
             ns = python.PyTuple_GET_ITEM(_PREFIX_CACHE, self._ns_counter)
             python.Py_INCREF(ns)
         else:
-            ns = python.PyString_FromFormat("ns%d", self._ns_counter)
+            ns = python.PyBytes_FromFormat("ns%d", self._ns_counter)
         if self._prefix_tail is not None:
             ns += self._prefix_tail
         self._ns_counter += 1
@@ -391,7 +391,7 @@
 
 cdef __initPrefixCache():
     cdef int i
-    return tuple([ python.PyString_FromFormat("ns%d", i)
+    return tuple([ python.PyBytes_FromFormat("ns%d", i)
                    for i in range(30) ])
 
 cdef object _PREFIX_CACHE
@@ -1467,11 +1467,10 @@
             return funicode(self._c_node.name)
 
         def __set__(self, value):
-            value = _utf8(value)
+            value_utf = _utf8(value)
             assert u'&' not in value and u';' not in value, \
                 u"Invalid entity name '%s'" % value
-            c_text = _cstr(value)
-            tree.xmlNodeSetName(self._c_node, c_text)
+            tree.xmlNodeSetName(self._c_node, _cstr(value_utf))
 
     property text:
         # FIXME: should this be None or '&[VALUE];' or the resolved

From scoder at codespeak.net  Tue Nov 24 18:59:14 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:14 +0100 (CET)
Subject: [Lxml-checkins] r69601 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20091124175914.0726F168008@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:14 2009
New Revision: 69601

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/cssselect.py
   lxml/trunk/src/lxml/tests/test_css.txt
Log:
 r5328 at delle:  sbehnel | 2009-11-13 14:07:16 +0100
 fix ticket 481641: support treating CSS attribute selectors independent of element selectors


Modified: lxml/trunk/src/lxml/cssselect.py
==============================================================================
--- lxml/trunk/src/lxml/cssselect.py	(original)
+++ lxml/trunk/src/lxml/cssselect.py	Tue Nov 24 18:59:14 2009
@@ -865,7 +865,7 @@
             pos += 2
             continue
         if c in '>+~,.*=[]()|:#':
-            if c in '.#' and preceding_whitespace_pos > 0:
+            if c in '.#[' and preceding_whitespace_pos > 0:
                 yield Token(' ', preceding_whitespace_pos)
             yield Token(c, pos)
             pos += 1

Modified: lxml/trunk/src/lxml/tests/test_css.txt
==============================================================================
--- lxml/trunk/src/lxml/tests/test_css.txt	(original)
+++ lxml/trunk/src/lxml/tests/test_css.txt	Tue Nov 24 18:59:14 2009
@@ -26,6 +26,8 @@
     Pseudo[Element[td]:first]
     >>> parse('a[name]')
     Attrib[Element[a][name]]
+    >>> parse('a [name]')
+    CombinedSelector[Element[a] <followed> Attrib[Element[*][name]]]
     >>> repr(parse('a[rel="include"]')).replace("u'", "'")
     "Attrib[Element[a][rel = String('include', 6)]]"
     >>> repr(parse('a[hreflang |= \'en\']')).replace("u'", "'")

From scoder at codespeak.net  Tue Nov 24 18:59:18 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:18 +0100 (CET)
Subject: [Lxml-checkins] r69602 - in lxml/trunk: . doc
Message-ID: <20091124175918.DC51816800B@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:18 2009
New Revision: 69602

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/INSTALL.txt
   lxml/trunk/doc/main.txt
Log:
 r5329 at delle:  sbehnel | 2009-11-13 14:59:55 +0100
 2.2 branch doc merge


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Tue Nov 24 18:59:18 2009
@@ -8,6 +8,9 @@
 Features added
 --------------
 
+* Support for XSLT context nodes other than elements: document root,
+  comments, processing instructions.
+
 * Support for strings (in addition to Elements) in node-sets returned
   by extension functions.
 
@@ -52,6 +55,15 @@
 * Static builds include libiconv, in addition to libxml2 and libxslt.
 
 
+2.2.4 (2009-11-11)
+==================
+
+Bugs fixed
+----------
+
+* Static build of libxml2/libxslt was broken.
+
+
 2.2.3 (2009-10-30)
 ==================
 

Modified: lxml/trunk/INSTALL.txt
==============================================================================
--- lxml/trunk/INSTALL.txt	(original)
+++ lxml/trunk/INSTALL.txt	Tue Nov 24 18:59:18 2009
@@ -5,6 +5,7 @@
 MacOS-X, see the specific sections below.
 
 .. contents::
+..
    1  Requirements
    2  Installation
    3  Installation in ActivePython

Modified: lxml/trunk/doc/main.txt
==============================================================================
--- lxml/trunk/doc/main.txt	(original)
+++ lxml/trunk/doc/main.txt	Tue Nov 24 18:59:18 2009
@@ -147,8 +147,8 @@
 source release.  If you can't wait, consider trying a less recent
 release version first.
 
-The latest version is `lxml 2.2.2`_, released 2009-06-21
-(`changes for 2.2.2`_).  `Older versions`_ are listed below.
+The latest version is `lxml 2.2.4`_, released 2009-11-11
+(`changes for 2.2.4`_).  `Older versions`_ are listed below.
 
 Please take a look at the `installation instructions`_!
 
@@ -221,7 +221,7 @@
 <http://codespeak.net/lxml/2.1/>`_ and the `current in-development
 version <http://codespeak.net/lxml/dev/>`_.
 
-.. _`PDF documentation`: lxmldoc-2.2.3.pdf
+.. _`PDF documentation`: lxmldoc-2.2.4.pdf
 
 * `lxml 2.2.3`_, released 2009-10-30 (`changes for 2.2.3`_)
 
@@ -327,6 +327,7 @@
 
 * `lxml 0.5`_,   released 2005-04-08
 
+.. _`lxml 2.2.4`: lxml-2.2.4.tgz
 .. _`lxml 2.2.3`: lxml-2.2.3.tgz
 .. _`lxml 2.2.2`: lxml-2.2.2.tgz
 .. _`lxml 2.2.1`: lxml-2.2.1.tgz
@@ -379,7 +380,8 @@
 .. _`lxml 0.5.1`: lxml-0.5.1.tgz
 .. _`lxml 0.5`: lxml-0.5.tgz
 
-.. _`changes for 2.2.3`: changes-2.2.2.html
+.. _`changes for 2.2.4`: changes-2.2.4.html
+.. _`changes for 2.2.3`: changes-2.2.3.html
 .. _`changes for 2.2.2`: changes-2.2.2.html
 .. _`changes for 2.2.1`: changes-2.2.1.html
 .. _`changes for 2.2`: changes-2.2.html

From scoder at codespeak.net  Tue Nov 24 18:59:22 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:22 +0100 (CET)
Subject: [Lxml-checkins] r69603 - lxml/trunk
Message-ID: <20091124175922.13AFB16800B@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:21 2009
New Revision: 69603

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
Log:
 r5330 at delle:  sbehnel | 2009-11-13 17:30:53 +0100
 changelog


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Tue Nov 24 18:59:21 2009
@@ -28,6 +28,10 @@
 Bugs fixed
 ----------
 
+* ``lxml.cssselect`` did not distinguish between ``x[attr="val"]`` and
+  ``x [attr="val"]`` (with a space).  The latter now matches the
+  attribute independent of the element.
+
 * Rewriting multiple links inside of HTML text content could end up
   replacing unrelated content as replacements could impact the
   reported position of subsequent matches.  Modifications are now

From scoder at codespeak.net  Tue Nov 24 18:59:25 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:25 +0100 (CET)
Subject: [Lxml-checkins] r69604 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20091124175925.DD032168008@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:25 2009
New Revision: 69604

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/CHANGES.txt
   lxml/trunk/src/lxml/python.pxd
   lxml/trunk/src/lxml/readonlytree.pxi
   lxml/trunk/src/lxml/tests/test_xslt.py
   lxml/trunk/src/lxml/xsltext.pxi
Log:
 r5331 at delle:  sbehnel | 2009-11-15 11:44:03 +0100
 support for several different node types as XSLT extension element context nodes


Modified: lxml/trunk/CHANGES.txt
==============================================================================
--- lxml/trunk/CHANGES.txt	(original)
+++ lxml/trunk/CHANGES.txt	Tue Nov 24 18:59:25 2009
@@ -8,8 +8,8 @@
 Features added
 --------------
 
-* Support for XSLT context nodes other than elements: document root,
-  comments, processing instructions.
+* XSLT extension elements: support for XSLT context nodes other than
+  elements: document root, comments, processing instructions.
 
 * Support for strings (in addition to Elements) in node-sets returned
   by extension functions.

Modified: lxml/trunk/src/lxml/python.pxd
==============================================================================
--- lxml/trunk/src/lxml/python.pxd	(original)
+++ lxml/trunk/src/lxml/python.pxd	Tue Nov 24 18:59:25 2009
@@ -26,23 +26,22 @@
     cdef bint PyBytes_Check(object obj)
     cdef bint PyBytes_CheckExact(object obj)
 
-    cdef object PyUnicode_FromEncodedObject(object s, char* encoding,
+    cdef unicode PyUnicode_FromEncodedObject(object s, char* encoding,
                                             char* errors)
-    cdef object PyUnicode_AsEncodedString(object u, char* encoding,
+    cdef bytes PyUnicode_AsEncodedString(object u, char* encoding,
                                           char* errors)
-    cdef object PyUnicode_FromFormat(char* format, ...) # Python 3
-    cdef object PyUnicode_Decode(char* s, Py_ssize_t size,
-                                 char* encoding, char* errors)
-    cdef object PyUnicode_DecodeUTF8(char* s, Py_ssize_t size, char* errors)
-    cdef object PyUnicode_DecodeLatin1(char* s, Py_ssize_t size, char* errors)
-    cdef object PyUnicode_AsUTF8String(object ustring)
-    cdef object PyUnicode_AsASCIIString(object ustring)
+    cdef unicode PyUnicode_FromFormat(char* format, ...) # Python 3
+    cdef unicode PyUnicode_Decode(char* s, Py_ssize_t size,
+                                  char* encoding, char* errors)
+    cdef unicode PyUnicode_DecodeUTF8(char* s, Py_ssize_t size, char* errors)
+    cdef unicode PyUnicode_DecodeLatin1(char* s, Py_ssize_t size, char* errors)
+    cdef bytes PyUnicode_AsUTF8String(object ustring)
+    cdef bytes PyUnicode_AsASCIIString(object ustring)
     cdef char* PyUnicode_AS_DATA(object ustring)
     cdef Py_ssize_t PyUnicode_GET_DATA_SIZE(object ustring)
     cdef Py_ssize_t PyUnicode_GET_SIZE(object ustring)
-    cdef object PyBytes_FromStringAndSize(char* s, Py_ssize_t size)
-    cdef object PyBytes_FromFormat(char* format, ...)
-    cdef object PyString_FromFormat(char* format, ...) # to be deleted!
+    cdef bytes PyBytes_FromStringAndSize(char* s, Py_ssize_t size)
+    cdef bytes PyBytes_FromFormat(char* format, ...)
     cdef Py_ssize_t PyBytes_GET_SIZE(object s)
 
     cdef object PyNumber_Int(object value)

Modified: lxml/trunk/src/lxml/readonlytree.pxi
==============================================================================
--- lxml/trunk/src/lxml/readonlytree.pxi	(original)
+++ lxml/trunk/src/lxml/readonlytree.pxi	Tue Nov 24 18:59:25 2009
@@ -1,10 +1,10 @@
 # read-only tree implementation
 
-cdef class _ReadOnlyElementProxy:
-    u"The main read-only Element proxy class (for internal use only!)."
+cdef class _ReadOnlyProxy:
+    u"A read-only proxy class suitable for PIs/Comments (for internal use only!)."
     cdef bint _free_after_use
     cdef xmlNode* _c_node
-    cdef object _source_proxy
+    cdef _ReadOnlyProxy _source_proxy
     cdef list _dependent_proxies
 
     cdef int _assertNode(self) except -1:
@@ -13,6 +13,9 @@
         assert self._c_node is not NULL, u"Proxy invalidated!"
         return 0
 
+    cdef int _raise_unsupported_type(self):
+        raise TypeError("Unsupported node type: %d" % self._c_node.type)
+
     cdef void free_after_use(self):
         u"""Should the xmlNode* be freed when releasing the proxy?
         """
@@ -23,7 +26,16 @@
         """
         def __get__(self):
             self._assertNode()
-            return _namespacedName(self._c_node)
+            if self._c_node.type == tree.XML_ELEMENT_NODE:
+                return _namespacedName(self._c_node)
+            elif self._c_node.type == tree.XML_PI_NODE:
+                return ProcessingInstruction
+            elif self._c_node.type == tree.XML_COMMENT_NODE:
+                return Comment
+            elif self._c_node.type == tree.XML_ENTITY_REF_NODE:
+                return Entity
+            else:
+                self._raise_unsupported_type()
 
     property text:
         u"""Text before the first subelement. This is either a string or 
@@ -31,7 +43,18 @@
         """
         def __get__(self):
             self._assertNode()
-            return _collectText(self._c_node.children)
+            if self._c_node.type == tree.XML_ELEMENT_NODE:
+                return _collectText(self._c_node.children)
+            elif self._c_node.type in (tree.XML_PI_NODE,
+                                       tree.XML_COMMENT_NODE):
+                if self._c_node.content is NULL:
+                    return ''
+                else:
+                    return funicode(self._c_node.content)
+            elif self._c_node.type == tree.XML_ENTITY_REF_NODE:
+                return u'&%s;' % funicode(self._c_node.name)
+            else:
+                self._raise_unsupported_type()
         
     property tail:
         u"""Text after this element's end tag, but before the next sibling
@@ -42,21 +65,6 @@
             self._assertNode()
             return _collectText(self._c_node.next)
 
-    property attrib:
-        def __get__(self):
-            self._assertNode()
-            return dict(_collectAttributes(self._c_node, 3))
-
-    property prefix:
-        u"""Namespace prefix or None.
-        """
-        def __get__(self):
-            self._assertNode()
-            if self._c_node.ns is not NULL:
-                if self._c_node.ns.prefix is not NULL:
-                    return funicode(self._c_node.ns.prefix)
-            return None
-
     property sourceline:
         u"""Original line number as found by the parser or None if unknown.
         """
@@ -70,7 +78,21 @@
                 return None
 
     def __repr__(self):
-        return u"<Element %s at %x>" % (self.tag, id(self))
+        self._assertNode()
+        if self._c_node.type == tree.XML_ELEMENT_NODE:
+            return u"<Element %s at %x>" % (self.tag, id(self))
+        elif self._c_node.type == tree.XML_COMMENT_NODE:
+            return u"<!--%s-->" % self.text
+        elif self._c_node.type == tree.XML_ENTITY_NODE:
+            return u"&%s;" % funicode(self._c_node.name)
+        elif self._c_node.type == tree.XML_PI_NODE:
+            text = self.text
+            if text:
+                return u"<?%s %s?>" % (self.target, text)
+            else:
+                return u"<?%s?>" % self.target
+        else:
+            self._raise_unsupported_type()
 
     def __getitem__(self, x):
         u"""Returns the subelement at the given position or the requested
@@ -81,6 +103,7 @@
         cdef Py_ssize_t c, i
         cdef _node_to_node_function next_element
         cdef list result
+        self._assertNode()
         if python.PySlice_Check(x):
             # slicing
             if _isFullSlice(<python.slice>x):
@@ -133,11 +156,13 @@
         u"__deepcopy__(self, memo)"
         return self.__copy__()
         
-    def __copy__(self):
+    cpdef __copy__(self):
         u"__copy__(self)"
         cdef xmlDoc* c_doc
         cdef xmlNode* c_node
         cdef _Document new_doc
+        if self._c_node is NULL:
+            return self
         c_doc = _copyDocRoot(self._c_node.doc, self._c_node) # recursive
         new_doc = _documentFactory(c_doc, None)
         root = new_doc.getroot()
@@ -166,33 +191,6 @@
             children = children[::-1]
         return iter(children)
 
-    def get(self, key, default=None):
-        u"""Gets an element attribute.
-        """
-        self._assertNode()
-        return _getNodeAttributeValue(self._c_node, key, default)
-
-    def keys(self):
-        u"""Gets a list of attribute names. The names are returned in an
-        arbitrary order (just like for an ordinary Python dictionary).
-        """
-        self._assertNode()
-        return _collectAttributes(self._c_node, 1)
-
-    def values(self):
-        u"""Gets element attributes, as a sequence. The attributes are returned
-        in an arbitrary order.
-        """
-        self._assertNode()
-        return _collectAttributes(self._c_node, 2)
-
-    def items(self):
-        u"""Gets element attributes, as a sequence. The attributes are returned
-        in an arbitrary order.
-        """
-        self._assertNode()
-        return _collectAttributes(self._c_node, 3)
-
     cpdef getchildren(self):
         u"""Returns all subelements. The elements are returned in document
         order.
@@ -240,20 +238,98 @@
         return None
 
 
+cdef class _ReadOnlyPIProxy(_ReadOnlyProxy):
+    u"A read-only proxy for processing instructions (for internal use only!)"
+    property target:
+        def __get__(self):
+            self._assertNode()
+            return funicode(self._c_node.name)
+
+cdef class _ReadOnlyEntityProxy(_ReadOnlyProxy):
+    u"A read-only proxy for entity references (for internal use only!)"
+    property name:
+        def __get__(self):
+            return funicode(self._c_node.name)
+
+        def __set__(self, value):
+            value_utf = _utf8(value)
+            assert u'&' not in value and u';' not in value, \
+                u"Invalid entity name '%s'" % value
+            tree.xmlNodeSetName(self._c_node, _cstr(value_utf))
+
+    property text:
+        def __get__(self):
+            return u'&%s;' % funicode(self._c_node.name)
+
+
+cdef class _ReadOnlyElementProxy(_ReadOnlyProxy):
+    u"The main read-only Element proxy class (for internal use only!)."
+
+    property attrib:
+        def __get__(self):
+            self._assertNode()
+            return dict(_collectAttributes(self._c_node, 3))
+
+    property prefix:
+        u"""Namespace prefix or None.
+        """
+        def __get__(self):
+            self._assertNode()
+            if self._c_node.ns is not NULL:
+                if self._c_node.ns.prefix is not NULL:
+                    return funicode(self._c_node.ns.prefix)
+            return None
+
+    def get(self, key, default=None):
+        u"""Gets an element attribute.
+        """
+        self._assertNode()
+        return _getNodeAttributeValue(self._c_node, key, default)
+
+    def keys(self):
+        u"""Gets a list of attribute names. The names are returned in an
+        arbitrary order (just like for an ordinary Python dictionary).
+        """
+        self._assertNode()
+        return _collectAttributes(self._c_node, 1)
+
+    def values(self):
+        u"""Gets element attributes, as a sequence. The attributes are returned
+        in an arbitrary order.
+        """
+        self._assertNode()
+        return _collectAttributes(self._c_node, 2)
+
+    def items(self):
+        u"""Gets element attributes, as a sequence. The attributes are returned
+        in an arbitrary order.
+        """
+        self._assertNode()
+        return _collectAttributes(self._c_node, 3)
+
+
 cdef extern from "etree_defs.h":
     # macro call to 't->tp_new()' for fast instantiation
-    cdef _ReadOnlyElementProxy NEW_RO_PROXY "PY_NEW" (object t)
+    cdef _ReadOnlyProxy NEW_RO_PROXY "PY_NEW" (object t)
 
-cdef _ReadOnlyElementProxy _newReadOnlyProxy(
-    _ReadOnlyElementProxy source_proxy, xmlNode* c_node):
-    cdef _ReadOnlyElementProxy el
-    el = NEW_RO_PROXY(_ReadOnlyElementProxy)
+cdef _ReadOnlyProxy _newReadOnlyProxy(
+    _ReadOnlyProxy source_proxy, xmlNode* c_node):
+    cdef _ReadOnlyProxy el
+    if c_node.type == tree.XML_ELEMENT_NODE:
+        el = NEW_RO_PROXY(_ReadOnlyElementProxy)
+    elif c_node.type == tree.XML_PI_NODE:
+        el = NEW_RO_PROXY(_ReadOnlyPIProxy)
+    elif c_node.type in (tree.XML_COMMENT_NODE,
+                         tree.XML_ENTITY_REF_NODE):
+        el = NEW_RO_PROXY(_ReadOnlyProxy)
+    else:
+        raise TypeError("Unsupported element type: %d" % c_node.type)
     el._c_node = c_node
     _initReadOnlyProxy(el, source_proxy)
     return el
 
-cdef inline _initReadOnlyProxy(_ReadOnlyElementProxy el,
-                               _ReadOnlyElementProxy source_proxy):
+cdef inline _initReadOnlyProxy(_ReadOnlyProxy el,
+                               _ReadOnlyProxy source_proxy):
     el._free_after_use = 0
     if source_proxy is None:
         el._source_proxy = el
@@ -262,9 +338,9 @@
         el._source_proxy = source_proxy
         source_proxy._dependent_proxies.append(el)
 
-cdef _freeReadOnlyProxies(_ReadOnlyElementProxy sourceProxy):
+cdef _freeReadOnlyProxies(_ReadOnlyProxy sourceProxy):
     cdef xmlNode* c_node
-    cdef _ReadOnlyElementProxy el
+    cdef _ReadOnlyProxy el
     if sourceProxy is None:
         return
     if sourceProxy._dependent_proxies is None:
@@ -276,6 +352,58 @@
             tree.xmlFreeNode(c_node)
     del sourceProxy._dependent_proxies[:]
 
+
+cdef class _ModifyContentOnlyProxy(_ReadOnlyProxy):
+    u"""A read-only proxy that allows changing the text content.
+    """
+    property text:
+        def __get__(self):
+            self._assertNode()
+            if self._c_node.content is NULL:
+                return ''
+            else:
+                return funicode(self._c_node.content)
+
+        def __set__(self, value):
+            cdef tree.xmlDict* c_dict
+            cdef char* c_text
+            self._assertNode()
+            if value is None:
+                c_text = NULL
+            else:
+                value = _utf8(value)
+                c_text = _cstr(value)
+            tree.xmlNodeSetContent(self._c_node, c_text)
+
+cdef class _ModifyContentOnlyPIProxy(_ModifyContentOnlyProxy):
+    u"""A read-only proxy that allows changing the text/target content of a
+    processing instruction.
+    """
+    property target:
+        def __get__(self):
+            self._assertNode()
+            return funicode(self._c_node.name)
+
+        def __set__(self, value):
+            self._assertNode()
+            value = _utf8(value)
+            c_text = _cstr(value)
+            tree.xmlNodeSetName(self._c_node, c_text)
+
+cdef class _ModifyContentOnlyEntityProxy(_ModifyContentOnlyProxy):
+    u"A read-only proxy for entity references (for internal use only!)"
+    property name:
+        def __get__(self):
+            return funicode(self._c_node.name)
+
+        def __set__(self, value):
+            value = _utf8(value)
+            assert u'&' not in value and u';' not in value, \
+                u"Invalid entity name '%s'" % value
+            c_text = _cstr(value)
+            tree.xmlNodeSetName(self._c_node, c_text)
+
+
 cdef class _AppendOnlyElementProxy(_ReadOnlyElementProxy):
     u"""A read-only element that allows adding children and changing the
     text content (i.e. everything that adds to the subtree).
@@ -315,10 +443,16 @@
                     _resolveQNameText(self, value), 'UTF-8', 'strict')
             _setNodeText(self._c_node, value)
 
-cdef _AppendOnlyElementProxy _newAppendOnlyProxy(
-    _ReadOnlyElementProxy source_proxy, xmlNode* c_node):
-    cdef _AppendOnlyElementProxy el
-    el = <_AppendOnlyElementProxy>NEW_RO_PROXY(_AppendOnlyElementProxy)
+
+cdef _ReadOnlyProxy _newAppendOnlyProxy(
+    _ReadOnlyProxy source_proxy, xmlNode* c_node):
+    cdef _ReadOnlyProxy el
+    if c_node.type == tree.XML_ELEMENT_NODE:
+        el = NEW_RO_PROXY(_AppendOnlyElementProxy)
+    elif c_node.type == tree.XML_PI_NODE:
+        el = NEW_RO_PROXY(_ModifyContentOnlyPIProxy)
+    elif c_node.type == tree.XML_COMMENT_NODE:
+        el = NEW_RO_PROXY(_ModifyContentOnlyProxy)
     el._c_node = c_node
     _initReadOnlyProxy(el, source_proxy)
     return el
@@ -327,8 +461,8 @@
     cdef xmlNode* c_node
     if isinstance(element, _Element):
         c_node = (<_Element>element)._c_node
-    elif isinstance(element, _ReadOnlyElementProxy):
-        c_node = (<_ReadOnlyElementProxy>element)._c_node
+    elif isinstance(element, _ReadOnlyProxy):
+        c_node = (<_ReadOnlyProxy>element)._c_node
     else:
         raise TypeError, u"invalid value to append()"
 

Modified: lxml/trunk/src/lxml/tests/test_xslt.py
==============================================================================
--- lxml/trunk/src/lxml/tests/test_xslt.py	(original)
+++ lxml/trunk/src/lxml/tests/test_xslt.py	Tue Nov 24 18:59:25 2009
@@ -628,175 +628,6 @@
         self.assertEquals(self._rootstring(result),
                           _bytes('<C>C</C>'))
 
-    def test_extensions1(self):
-        tree = self.parse('<a><b>B</b></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    exclude-result-prefixes="myns">
-  <xsl:template match="a"><A><xsl:value-of select="myns:mytext(b)"/></A></xsl:template>
-</xsl:stylesheet>''')
-
-        def mytext(ctxt, values):
-            return 'X' * len(values)
-
-        result = tree.xslt(style, {('testns', 'mytext') : mytext})
-        self.assertEquals(self._rootstring(result),
-                          _bytes('<A>X</A>'))
-
-    def test_extensions2(self):
-        tree = self.parse('<a><b>B</b></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    exclude-result-prefixes="myns">
-  <xsl:template match="a"><A><xsl:value-of select="myns:mytext(b)"/></A></xsl:template>
-</xsl:stylesheet>''')
-
-        def mytext(ctxt, values):
-            return 'X' * len(values)
-
-        namespace = etree.FunctionNamespace('testns')
-        namespace['mytext'] = mytext
-
-        result = tree.xslt(style)
-        self.assertEquals(self._rootstring(result),
-                          _bytes('<A>X</A>'))
-
-    def test_variable_result_tree_fragment(self):
-        tree = self.parse('<a><b>B</b><b/></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    exclude-result-prefixes="myns">
-  <xsl:template match="a">
-    <xsl:variable name="content">
-       <xsl:apply-templates/>
-    </xsl:variable>
-    <A><xsl:value-of select="myns:mytext($content)"/></A>
-  </xsl:template>
-  <xsl:template match="b"><xsl:copy>BBB</xsl:copy></xsl:template>
-</xsl:stylesheet>''')
-
-        def mytext(ctxt, values):
-            for value in values:
-                self.assert_(hasattr(value, 'tag'),
-                             "%s is not an Element" % type(value))
-                self.assertEquals(value.tag, 'b')
-                self.assertEquals(value.text, 'BBB')
-            return 'X'.join([el.tag for el in values])
-
-        namespace = etree.FunctionNamespace('testns')
-        namespace['mytext'] = mytext
-
-        result = tree.xslt(style)
-        self.assertEquals(self._rootstring(result),
-                          _bytes('<A>bXb</A>'))
-
-    def test_extension_element(self):
-        tree = self.parse('<a><b>B</b></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    extension-element-prefixes="myns"
-    exclude-result-prefixes="myns">
-  <xsl:template match="a">
-    <A><myns:myext>b</myns:myext></A>
-  </xsl:template>
-</xsl:stylesheet>''')
-
-        class MyExt(etree.XSLTExtension):
-            def execute(self, context, self_node, input_node, output_parent):
-                child = etree.Element(self_node.text)
-                child.text = 'X'
-                output_parent.append(child)
-
-        extensions = { ('testns', 'myext') : MyExt() }
-
-        result = tree.xslt(style, extensions=extensions)
-        self.assertEquals(self._rootstring(result),
-                          _bytes('<A><b>X</b></A>'))
-
-    def test_extension_element_content(self):
-        tree = self.parse('<a><b>B</b></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    extension-element-prefixes="myns">
-  <xsl:template match="a">
-    <A><myns:myext><x>X</x><y>Y</y><z/></myns:myext></A>
-  </xsl:template>
-</xsl:stylesheet>''')
-
-        class MyExt(etree.XSLTExtension):
-            def execute(self, context, self_node, input_node, output_parent):
-                output_parent.extend(list(self_node)[1:])
-
-        extensions = { ('testns', 'myext') : MyExt() }
-
-        result = tree.xslt(style, extensions=extensions)
-        self.assertEquals(self._rootstring(result),
-                          _bytes('<A><y>Y</y><z/></A>'))
-
-    def test_extension_element_apply_templates(self):
-        tree = self.parse('<a><b>B</b></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    extension-element-prefixes="myns">
-  <xsl:template match="a">
-    <A><myns:myext><x>X</x><y>Y</y><z/></myns:myext></A>
-  </xsl:template>
-  <xsl:template match="x" />
-  <xsl:template match="z">XYZ</xsl:template>
-</xsl:stylesheet>''')
-
-        class MyExt(etree.XSLTExtension):
-            def execute(self, context, self_node, input_node, output_parent):
-                for child in self_node:
-                    for result in self.apply_templates(context, child):
-                        if isinstance(result, basestring):
-                            el = etree.Element("T")
-                            el.text = result
-                        else:
-                            el = result
-                        output_parent.append(el)
-
-        extensions = { ('testns', 'myext') : MyExt() }
-
-        result = tree.xslt(style, extensions=extensions)
-        self.assertEquals(self._rootstring(result),
-                          _bytes('<A><T>Y</T><T>XYZ</T></A>'))
-
-    def test_extension_element_raise(self):
-        tree = self.parse('<a><b>B</b></a>')
-        style = self.parse('''\
-<xsl:stylesheet version="1.0"
-    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
-    xmlns:myns="testns"
-    extension-element-prefixes="myns"
-    exclude-result-prefixes="myns">
-  <xsl:template match="a">
-    <A><myns:myext>b</myns:myext></A>
-  </xsl:template>
-</xsl:stylesheet>''')
-
-        class MyError(Exception):
-            pass
-
-        class MyExt(etree.XSLTExtension):
-            def execute(self, context, self_node, input_node, output_parent):
-                raise MyError("expected!")
-
-        extensions = { ('testns', 'myext') : MyExt() }
-        self.assertRaises(MyError, tree.xslt, style, extensions=extensions)
-
     def test_xslt_document_XML(self):
         # make sure document('') works from parsed strings
         xslt = etree.XSLT(etree.XML("""\
@@ -1389,6 +1220,265 @@
         self.assertEquals(root[3].text, "test")
 
 
+class ETreeXSLTExtFuncTestCase(HelperTestCase):
+    """Tests for XPath extension functions in XSLT."""
+
+    def test_extensions1(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="a"><A><xsl:value-of select="myns:mytext(b)"/></A></xsl:template>
+</xsl:stylesheet>''')
+
+        def mytext(ctxt, values):
+            return 'X' * len(values)
+
+        result = tree.xslt(style, {('testns', 'mytext') : mytext})
+        self.assertEquals(self._rootstring(result),
+                          _bytes('<A>X</A>'))
+
+    def test_extensions2(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="a"><A><xsl:value-of select="myns:mytext(b)"/></A></xsl:template>
+</xsl:stylesheet>''')
+
+        def mytext(ctxt, values):
+            return 'X' * len(values)
+
+        namespace = etree.FunctionNamespace('testns')
+        namespace['mytext'] = mytext
+
+        result = tree.xslt(style)
+        self.assertEquals(self._rootstring(result),
+                          _bytes('<A>X</A>'))
+
+    def test_variable_result_tree_fragment(self):
+        tree = self.parse('<a><b>B</b><b/></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="a">
+    <xsl:variable name="content">
+       <xsl:apply-templates/>
+    </xsl:variable>
+    <A><xsl:value-of select="myns:mytext($content)"/></A>
+  </xsl:template>
+  <xsl:template match="b"><xsl:copy>BBB</xsl:copy></xsl:template>
+</xsl:stylesheet>''')
+
+        def mytext(ctxt, values):
+            for value in values:
+                self.assert_(hasattr(value, 'tag'),
+                             "%s is not an Element" % type(value))
+                self.assertEquals(value.tag, 'b')
+                self.assertEquals(value.text, 'BBB')
+            return 'X'.join([el.tag for el in values])
+
+        namespace = etree.FunctionNamespace('testns')
+        namespace['mytext'] = mytext
+
+        result = tree.xslt(style)
+        self.assertEquals(self._rootstring(result),
+                          _bytes('<A>bXb</A>'))
+
+
+class ETreeXSLTExtElementTestCase(HelperTestCase):
+    """Tests for extension elements in XSLT."""
+
+    def test_extension_element(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="a">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                child = etree.Element(self_node.text)
+                child.text = 'X'
+                output_parent.append(child)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(self._rootstring(result),
+                          _bytes('<A><b>X</b></A>'))
+
+    def test_extension_element_doc_context(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="/">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        tags = []
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                tags.append(input_node.tag)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(tags, ['a'])
+
+    def test_extension_element_comment_pi_context(self):
+        tree = self.parse('<?test toast?><a><!--a comment--><?another pi?></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="/">
+    <ROOT><xsl:apply-templates /></ROOT>
+  </xsl:template>
+  <xsl:template match="comment()">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+  <xsl:template match="processing-instruction()">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        text = []
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                text.append(input_node.text)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(text, ['toast', 'a comment', 'pi'])
+
+    def _test_extension_element_attribute_context(self):
+        # currently not supported
+        tree = self.parse('<a test="A"><b attr="B"/></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="@test">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+  <xsl:template match="@attr">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        text = []
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, attr_value, output_parent):
+                text.append(attr_value)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(text, ['A', 'B'])
+
+    def test_extension_element_content(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns">
+  <xsl:template match="a">
+    <A><myns:myext><x>X</x><y>Y</y><z/></myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                output_parent.extend(list(self_node)[1:])
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(self._rootstring(result),
+                          _bytes('<A><y>Y</y><z/></A>'))
+
+    def test_extension_element_apply_templates(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns">
+  <xsl:template match="a">
+    <A><myns:myext><x>X</x><y>Y</y><z/></myns:myext></A>
+  </xsl:template>
+  <xsl:template match="x" />
+  <xsl:template match="z">XYZ</xsl:template>
+</xsl:stylesheet>''')
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                for child in self_node:
+                    for result in self.apply_templates(context, child):
+                        if isinstance(result, basestring):
+                            el = etree.Element("T")
+                            el.text = result
+                        else:
+                            el = result
+                        output_parent.append(el)
+
+        extensions = { ('testns', 'myext') : MyExt() }
+
+        result = tree.xslt(style, extensions=extensions)
+        self.assertEquals(self._rootstring(result),
+                          _bytes('<A><T>Y</T><T>XYZ</T></A>'))
+
+    def test_extension_element_raise(self):
+        tree = self.parse('<a><b>B</b></a>')
+        style = self.parse('''\
+<xsl:stylesheet version="1.0"
+    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:myns="testns"
+    extension-element-prefixes="myns"
+    exclude-result-prefixes="myns">
+  <xsl:template match="a">
+    <A><myns:myext>b</myns:myext></A>
+  </xsl:template>
+</xsl:stylesheet>''')
+
+        class MyError(Exception):
+            pass
+
+        class MyExt(etree.XSLTExtension):
+            def execute(self, context, self_node, input_node, output_parent):
+                raise MyError("expected!")
+
+        extensions = { ('testns', 'myext') : MyExt() }
+        self.assertRaises(MyError, tree.xslt, style, extensions=extensions)
+
+
 class Py3XSLTTestCase(HelperTestCase):
     """XSLT tests for etree under Python 3"""
     def test_xslt_result_bytes(self):
@@ -1452,6 +1542,8 @@
 def test_suite():
     suite = unittest.TestSuite()
     suite.addTests([unittest.makeSuite(ETreeXSLTTestCase)])
+    suite.addTests([unittest.makeSuite(ETreeXSLTExtFuncTestCase)])
+    suite.addTests([unittest.makeSuite(ETreeXSLTExtElementTestCase)])
     if is_python3:
         suite.addTests([unittest.makeSuite(Py3XSLTTestCase)])
     suite.addTests(

Modified: lxml/trunk/src/lxml/xsltext.pxi
==============================================================================
--- lxml/trunk/src/lxml/xsltext.pxi	(original)
+++ lxml/trunk/src/lxml/xsltext.pxi	Tue Nov 24 18:59:25 2009
@@ -28,7 +28,7 @@
         cdef xmlNode* c_node
         cdef xmlNode* c_next
         cdef xmlNode* c_context_node
-        cdef _ReadOnlyElementProxy proxy
+        cdef _ReadOnlyProxy proxy
         cdef list results
         c_context_node = _roNodeOf(node)
         #assert c_context_node.doc is context._xsltContext.node.doc, \
@@ -81,7 +81,8 @@
     cdef XSLTExtension extension
     cdef python.PyObject* dict_result
     cdef char* c_uri
-    cdef _ReadOnlyElementProxy context_node, self_node, output_parent
+    cdef xmlNode* c_node
+    cdef _ReadOnlyProxy context_node = None, self_node = None, output_parent
     c_uri = _getNs(c_inst_node)
     if c_uri is NULL:
         # not allowed, and should never happen
@@ -91,34 +92,60 @@
         return
     context = <_XSLTContext>c_ctxt.xpathCtxt.userData
     try:
-        dict_result = python.PyDict_GetItem(
-            context._extension_elements, (c_uri, c_inst_node.name))
-        if dict_result is NULL:
-            raise KeyError, \
-                u"extension element %s not found" % funicode(c_inst_node.name)
-        extension = <object>dict_result
-
         try:
-            self_node     = _newReadOnlyProxy(None, c_inst_node)
-            context_node  = _newReadOnlyProxy(self_node, c_context_node)
-            output_parent = _newAppendOnlyProxy(self_node, c_ctxt.insert)
+            dict_result = python.PyDict_GetItem(
+                context._extension_elements, (c_uri, c_inst_node.name))
+            if dict_result is NULL:
+                raise KeyError, \
+                    u"extension element %s not found" % funicode(c_inst_node.name)
+            extension = <object>dict_result
+
+            try:
+                # build the context proxy nodes
+                self_node = _newReadOnlyProxy(None, c_inst_node)
+                output_parent = _newAppendOnlyProxy(self_node, c_ctxt.insert)
+                if c_context_node.type in (tree.XML_DOCUMENT_NODE,
+                                           tree.XML_HTML_DOCUMENT_NODE):
+                    c_node = tree.xmlDocGetRootElement(<xmlDoc*>c_context_node)
+                    if c_node is not NULL:
+                        context_node = _newReadOnlyProxy(self_node, c_node)
+                    else:
+                        context_node = None
+                elif c_context_node.type in (tree.XML_ATTRIBUTE_NODE,
+                                             tree.XML_TEXT_NODE,
+                                             tree.XML_CDATA_SECTION_NODE):
+                    # this isn't easy to support using read-only
+                    # nodes, as the smart-string factory must
+                    # instantiate the parent proxy somehow...
+                    raise TypeError("Unsupported element type: %d" % c_context_node.type)
+                else:
+                    context_node  = _newReadOnlyProxy(self_node, c_context_node)
 
-            context._extension_element_proxy = self_node
-            extension.execute(context, self_node, context_node, output_parent)
-        finally:
-            context._extension_element_proxy = None
-            if self_node is not None:
-                _freeReadOnlyProxies(self_node)
-    except Exception, e:
-        e = unicode(e).encode(u"UTF-8")
-        message = python.PyString_FromFormat(
-            "Error executing extension element '%s': %s",
-            c_inst_node.name, _cstr(e))
-        xslt.xsltTransformError(c_ctxt, NULL, c_inst_node, message)
-        context._exc._store_raised()
+                # run the XSLT extension
+                context._extension_element_proxy = self_node
+                extension.execute(context, self_node, context_node, output_parent)
+            finally:
+                context._extension_element_proxy = None
+                if self_node is not None:
+                    _freeReadOnlyProxies(self_node)
+        except Exception, e:
+            try:
+                e = unicode(e).encode(u"UTF-8")
+            except:
+                e = repr(e).encode(u"UTF-8")
+            message = python.PyBytes_FromFormat(
+                "Error executing extension element '%s': %s",
+                c_inst_node.name, _cstr(e))
+            xslt.xsltTransformError(c_ctxt, NULL, c_inst_node, message)
+            context._exc._store_raised()
+        except:
+            # just in case
+            message = python.PyBytes_FromFormat(
+                "Error executing extension element '%s'", c_inst_node.name)
+            xslt.xsltTransformError(c_ctxt, NULL, c_inst_node, message)
+            context._exc._store_raised()
     except:
-        # just in case
-        message = python.PyString_FromFormat(
-            "Error executing extension element '%s'", c_inst_node.name)
-        xslt.xsltTransformError(c_ctxt, NULL, c_inst_node, message)
+        # no Python functions here - everything can fail...
+        xslt.xsltTransformError(c_ctxt, NULL, c_inst_node,
+                                "Error during XSLT extension element evaluation")
         context._exc._store_raised()

From scoder at codespeak.net  Tue Nov 24 18:59:30 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:30 +0100 (CET)
Subject: [Lxml-checkins] r69605 - in lxml/trunk: . doc
Message-ID: <20091124175930.B2046168008@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:30 2009
New Revision: 69605

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/doc/build.txt
Log:
 r5332 at delle:  sbehnel | 2009-11-17 20:13:02 +0100
 mac build docs: show how to install


Modified: lxml/trunk/doc/build.txt
==============================================================================
--- lxml/trunk/doc/build.txt	(original)
+++ lxml/trunk/doc/build.txt	Tue Nov 24 18:59:30 2009
@@ -209,6 +209,8 @@
          --libxml2-version=2.7.3 \
          --libxslt-version=1.1.24 \
 
+  sudo python setup.py install
+
 Instead of ``build``, you can use any target, like ``bdist_egg`` if
 you want to use setuptools to build an installable egg.
 

From scoder at codespeak.net  Tue Nov 24 18:59:34 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:34 +0100 (CET)
Subject: [Lxml-checkins] r69606 - lxml/trunk
Message-ID: <20091124175934.10563168006@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:34 2009
New Revision: 69606

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/buildlibxml.py
Log:
 r5333 at delle:  sbehnel | 2009-11-24 18:55:53 +0100
 build support for MacOS-X 10.6


Modified: lxml/trunk/buildlibxml.py
==============================================================================
--- lxml/trunk/buildlibxml.py	(original)
+++ lxml/trunk/buildlibxml.py	Tue Nov 24 18:59:34 2009
@@ -197,15 +197,23 @@
     call_setup = {}
     env_setup = None
     if sys.platform in ('darwin',):
+        import platform
         # We compile Universal if we are on a machine > 10.3
-        major_version = int(os.uname()[2].split('.')[0])
+        major_version, minor_version = map(int, platform.mac_ver()[0].split('.')[:2])
         if major_version > 7:
             env = os.environ.copy()
-            env.update({
-                'CFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk -O2",
-                'LDFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk",
-                'MACOSX_DEPLOYMENT_TARGET' : "10.3"
-                })
+            if minor_version < 6:
+                env.update({
+                    'CFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk -O2",
+                    'LDFLAGS' : "-arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk",
+                    'MACOSX_DEPLOYMENT_TARGET' : "10.3"
+                    })
+            else:
+                env.update({
+                    'CFLAGS' : "-arch ppc -arch i386 -arch x86_64 -O2",
+                    'LDFLAGS' : "-arch ppc -arch i386 -arch x86_64",
+                    'MACOSX_DEPLOYMENT_TARGET' : "10.6"
+                    })
             call_setup['env'] = env
 
     configure_cmd = ['./configure',

From scoder at codespeak.net  Tue Nov 24 18:59:38 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 18:59:38 +0100 (CET)
Subject: [Lxml-checkins] r69607 - in lxml/trunk: . src/lxml src/lxml/tests
Message-ID: <20091124175938.58F0F168008@codespeak.net>

Author: scoder
Date: Tue Nov 24 18:59:37 2009
New Revision: 69607

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/apihelpers.pxi
   lxml/trunk/src/lxml/cleanup.pxi
   lxml/trunk/src/lxml/tests/test_etree.py
Log:
 r5334 at delle:  sbehnel | 2009-11-24 18:58:17 +0100
 major rewrite in cleanup.pxi: faster and less buggy


Modified: lxml/trunk/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/trunk/src/lxml/apihelpers.pxi	(original)
+++ lxml/trunk/src/lxml/apihelpers.pxi	Tue Nov 24 18:59:37 2009
@@ -400,6 +400,7 @@
     return 0
 
 cdef int _replaceNodeByChildren(_Document doc, xmlNode* c_node) except -1:
+    # NOTE: this does not deallocate the node, just unlink it!
     cdef xmlNode* c_parent
     cdef xmlNode* c_child
     if c_node.children is NULL:
@@ -883,6 +884,35 @@
     else:
         return 0
 
+cdef inline bint _tagMatchesExactly(xmlNode* c_node, char* c_href, char* c_name):
+    u"""Tests if the node matches namespace URI and tag name.
+
+    This differs from _tagMatches() in that it does not consider a
+    NULL value in c_href a wildcard, and that it expects the c_name to
+    be taken from the doc dict, i.e. it only compares the names by
+    address.
+
+    A node matches if it matches both c_href and c_name.
+
+    A node matches c_href if any of the following is true:
+    * its namespace is NULL and c_href is the empty string
+    * its namespace string equals the c_href string
+
+    A node matches c_name if any of the following is true:
+    * c_name is NULL
+    * its name string points to the same address (!) as c_name
+    """
+    cdef char* c_node_href
+    if c_name is not NULL and c_name is not c_node.name:
+        return 0
+    c_node_href = _getNs(c_node)
+    if c_href is NULL:
+        return c_node_href is NULL or c_node_href[0] == '\0'
+    elif c_node_href is NULL:
+        return 0
+    else:
+        return cstd.strcmp(c_href, c_node_href) == 0
+
 cdef int _removeNode(_Document doc, xmlNode* c_node) except -1:
     u"""Unlink and free a node and subnodes if possible.  Otherwise, make sure
     it's self-contained.
@@ -896,6 +926,29 @@
         moveNodeToDocument(doc, c_node.doc, c_node)
     return 0
 
+cdef int _removeSiblings(xmlNode* c_element, int node_type, bint with_tail) except -1:
+    cdef xmlNode* c_node
+    cdef xmlNode* c_next
+    c_node = c_element.next
+    while c_node is not NULL:
+        c_next = _nextElement(c_node)
+        if c_node.type == node_type:
+            if with_tail:
+                _removeText(c_node.next)
+            tree.xmlUnlinkNode(c_node)
+            attemptDeallocation(c_node)
+        c_node = c_next
+    c_node = c_element.prev
+    while c_node is not NULL:
+        c_next = _previousElement(c_node)
+        if c_node.type == node_type:
+            if with_tail:
+                _removeText(c_node.next)
+            tree.xmlUnlinkNode(c_node)
+            attemptDeallocation(c_node)
+        c_node = c_next
+    return 0
+
 cdef void _moveTail(xmlNode* c_tail, xmlNode* c_target):
     cdef xmlNode* c_next
     # tail support: look for any text nodes trailing this node and 

Modified: lxml/trunk/src/lxml/cleanup.pxi
==============================================================================
--- lxml/trunk/src/lxml/cleanup.pxi	(original)
+++ lxml/trunk/src/lxml/cleanup.pxi	Tue Nov 24 18:59:37 2009
@@ -22,40 +22,56 @@
                          'simpleattr',
                          '{http://some/ns}attrname')
     """
-    cdef xmlNode* c_node
-    cdef xmlAttr* c_attr
     cdef _Element element
     cdef list ns_tags
-    cdef char* c_name
+    cdef char** c_ns_tags
+    cdef Py_ssize_t c_tag_count
 
     element = _rootNodeOrRaise(tree_or_element)
     if not attribute_names: return
 
     ns_tags = _sortedTagList([ _getNsTag(attr)
                                for attr in <tuple>attribute_names ])
-    ns_tags = [ (ns, tag if tag != '*' else None)
+    ns_tags = [ (ns, tag if tag != b'*' else None)
                 for ns, tag in ns_tags ]
 
-    c_node = element._c_node
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
+    try:
+        c_tag_count = _mapTagsToCharArray(element._doc._c_doc, ns_tags, c_ns_tags)
+        if c_tag_count > 0:
+            _strip_attributes(element._c_node, c_ns_tags, c_tag_count)
+    finally:
+        cstd.free(c_ns_tags)
+
+cdef _strip_attributes(xmlNode* c_node, char** c_ns_tags, Py_ssize_t c_tag_count):
+    cdef xmlAttr* c_attr
+    cdef Py_ssize_t i
+    cdef char* c_href
+    cdef char* c_name
+
     tree.BEGIN_FOR_EACH_ELEMENT_FROM(c_node, c_node, 1)
     if c_node.type == tree.XML_ELEMENT_NODE:
         if c_node.properties is not NULL:
-            for ns, tag in ns_tags:
-                # must search attributes manually to make sure we only
-                # match on blank tag names if there is no namespace
-                c_name = NULL if tag is None else _cstr(tag)
+            for i in xrange(c_tag_count):
+                c_href = c_ns_tags[2*i]
+                c_name = c_ns_tags[2*i+1]
+                # must compare attributes manually to make sure we
+                # only match on wildcard tag names if the attribute
+                # has no namespace
                 c_attr = c_node.properties
                 while c_attr is not NULL:
-                    if ns is None:
-                        if c_attr.ns is NULL or c_attr.ns.href is NULL:
-                            if c_name is NULL or \
-                                   cstd.strcmp(c_attr.name, c_name) == 0:
+                    if c_name is NULL or c_attr.name == c_name:
+                        if c_href is NULL:
+                            if c_attr.ns is NULL or c_attr.ns.href is NULL:
                                 tree.xmlRemoveProp(c_attr)
                                 break
-                    elif c_attr.ns is not NULL and c_attr.ns.href is not NULL:
-                        if cstd.strcmp(c_attr.ns.href, _cstr(ns)) == 0:
-                            if c_name is NULL or \
-                                   cstd.strcmp(c_attr.name, c_name) == 0:
+                        elif c_attr.ns is not NULL and c_attr.ns.href is not NULL:
+                            if cstd.strcmp(c_attr.ns.href, c_href) == 0:
                                 tree.xmlRemoveProp(c_attr)
                                 break
                     c_attr = c_attr.next
@@ -84,14 +100,11 @@
             Comment                      # comments
             )
     """
-    cdef xmlNode* c_node
-    cdef xmlNode* c_child
-    cdef xmlNode* c_next
-    cdef char* c_href
-    cdef char* c_name
     cdef _Element element
     cdef _Document doc
     cdef list ns_tags
+    cdef char** c_ns_tags
+    cdef Py_ssize_t c_tag_count
     cdef bint strip_comments, strip_pis, strip_entities
 
     doc = _documentOrRaise(tree_or_element)
@@ -101,7 +114,35 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    c_node = element._c_node
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
+    if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
+        # include PIs and comments next to the root node
+        if strip_comments:
+            _removeSiblings(element._c_node, tree.XML_COMMENT_NODE, with_tail)
+        if strip_pis:
+            _removeSiblings(element._c_node, tree.XML_PI_NODE, with_tail)
+
+    try:
+        c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
+        if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:
+            _strip_elements(doc, element._c_node, c_ns_tags, c_tag_count,
+                            strip_comments, strip_pis, strip_entities, with_tail)
+    finally:
+        cstd.free(c_ns_tags)
+
+cdef _strip_elements(_Document doc, xmlNode* c_node,
+                     char** c_ns_tags, Py_ssize_t c_tag_count,
+                     bint strip_comments, bint strip_pis, bint strip_entities,
+                     bint with_tail):
+    cdef xmlNode* c_child
+    cdef xmlNode* c_next
+    cdef Py_ssize_t i
+
     tree.BEGIN_FOR_EACH_ELEMENT_FROM(c_node, c_node, 1)
     if c_node.type == tree.XML_ELEMENT_NODE:
         # we run through the children here to prevent any problems
@@ -109,39 +150,25 @@
         # c_node itself
         c_child = _findChildForwards(c_node, 0)
         while c_child is not NULL:
+            c_next = _nextElement(c_child)
             if c_child.type == tree.XML_ELEMENT_NODE:
-                for ns, tag in ns_tags:
-                    if ns is None:
-                        # _tagMatches() considers NULL a wildcard
-                        # match but we don't
-                        if c_child.ns is not NULL and c_child.ns.href is not NULL:
-                            continue
-                        c_href = NULL
-                    else:
-                        c_href = _cstr(ns)
-                    c_name = NULL if tag is None else _cstr(tag)
-                    if _tagMatches(c_child, c_href, c_name):
-                        c_next = _nextElement(c_child)
+                for i in xrange(c_tag_count):
+                    if _tagMatchesExactly(c_child, c_ns_tags[2*i], c_ns_tags[2*i+1]):
                         if not with_tail:
                             tree.xmlUnlinkNode(c_child)
                         _removeNode(doc, c_child)
-                        c_child = c_next
                         break
-                else:
-                    c_child = _nextElement(c_child)
-            elif strip_comments and c_child.type == tree.XML_COMMENT_NODE or \
-                     strip_pis and c_child.type == tree.XML_PI_NODE or \
-                     strip_entities and c_child.type == tree.XML_ENTITY_REF_NODE:
-                c_next = _nextElement(c_child)
+            elif c_child.type == tree.XML_COMMENT_NODE and strip_comments \
+                     or c_child.type == tree.XML_PI_NODE and strip_pis \
+                     or c_child.type == tree.XML_ENTITY_REF_NODE and strip_entities:
                 if with_tail:
-                    _removeText(c_next)
+                    _removeText(c_child.next)
                 tree.xmlUnlinkNode(c_child)
                 attemptDeallocation(c_child)
-                c_child = c_next
-            else:
-                c_child = _nextElement(c_child)
+            c_child = c_next
     tree.END_FOR_EACH_ELEMENT_FROM(c_node)
 
+
 def strip_tags(tree_or_element, *tag_names):
     u"""strip_tags(tree_or_element, *tag_names)
 
@@ -164,15 +191,12 @@
             Comment                      # comments (including their text!)
             )
     """
-    cdef xmlNode* c_node
-    cdef xmlNode* c_child
-    cdef xmlNode* c_next
-    cdef char* c_href
-    cdef char* c_name
     cdef _Element element
     cdef _Document doc
     cdef list ns_tags
     cdef bint strip_comments, strip_pis, strip_entities
+    cdef char** c_ns_tags
+    cdef Py_ssize_t c_tag_count
 
     doc = _documentOrRaise(tree_or_element)
     element = _rootNodeOrRaise(tree_or_element)
@@ -181,7 +205,34 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    c_node = element._c_node
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
+    if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
+        # include PIs and comments next to the root node
+        if strip_comments:
+            _removeSiblings(element._c_node, tree.XML_COMMENT_NODE, 0)
+        if strip_pis:
+            _removeSiblings(element._c_node, tree.XML_PI_NODE, 0)
+
+    try:
+        c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
+        if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:
+            _strip_tags(doc, element._c_node, c_ns_tags, c_tag_count,
+                        strip_comments, strip_pis, strip_entities)
+    finally:
+        cstd.free(c_ns_tags)
+
+cdef _strip_tags(_Document doc, xmlNode* c_node,
+                 char** c_ns_tags, Py_ssize_t c_tag_count,
+                 bint strip_comments, bint strip_pis, bint strip_entities):
+    cdef xmlNode* c_child
+    cdef xmlNode* c_next
+    cdef Py_ssize_t i
+
     tree.BEGIN_FOR_EACH_ELEMENT_FROM(c_node, c_node, 1)
     if c_node.type == tree.XML_ELEMENT_NODE:
         # we run through the children here to prevent any problems
@@ -190,39 +241,26 @@
         c_child = _findChildForwards(c_node, 0)
         while c_child is not NULL:
             if c_child.type == tree.XML_ELEMENT_NODE:
-                for ns, tag in ns_tags:
-                    if ns is None:
-                        # _tagMatches() considers NULL a wildcard
-                        # match but we don't
-                        if c_child.ns is not NULL and c_child.ns.href is not NULL:
-                            continue
-                        c_href = NULL
-                    else:
-                        c_href = _cstr(ns)
-                    c_name = NULL if tag is None else _cstr(tag)
-                    if _tagMatches(c_child, c_href, c_name):
-                        if c_child.children is not NULL:
-                            c_next = _findChildForwards(c_child, 0)
-                        else:
-                            c_next = _nextElement(c_child)
+                for i in xrange(c_tag_count):
+                    if _tagMatchesExactly(c_child, c_ns_tags[2*i], c_ns_tags[2*i+1]):
+                        c_next = _findChildForwards(c_child, 0) or _nextElement(c_child)
                         _replaceNodeByChildren(doc, c_child)
                         if not attemptDeallocation(c_child):
-                            if c_child.ns is not NULL:
+                            if c_child.nsDef is not NULL:
                                 # make namespaces absolute
                                 moveNodeToDocument(doc, doc._c_doc, c_child)
                         c_child = c_next
                         break
                 else:
-                    c_child = c_child.next
-            elif strip_comments and c_child.type == tree.XML_COMMENT_NODE or \
-                     strip_pis and c_child.type == tree.XML_PI_NODE or \
-                     strip_entities and c_child.type == tree.XML_ENTITY_REF_NODE:
+                    c_child = _nextElement(c_child)
+            else:
                 c_next = _nextElement(c_child)
-                tree.xmlUnlinkNode(c_child)
-                attemptDeallocation(c_child)
+                if c_child.type == tree.XML_COMMENT_NODE and strip_comments \
+                       or c_child.type == tree.XML_PI_NODE and strip_pis \
+                       or c_child.type == tree.XML_ENTITY_REF_NODE and strip_entities:
+                    tree.xmlUnlinkNode(c_child)
+                    attemptDeallocation(c_child)
                 c_child = c_next
-            else:
-                c_child = _nextElement(c_child)
     tree.END_FOR_EACH_ELEMENT_FROM(c_node)
 
 
@@ -235,7 +273,7 @@
     cdef list decorated_list
     cdef tuple ns_tag
     cdef Py_ssize_t i
-    decorated_list = [ (ns_tag[0] or '', ns_tag[1], i, ns_tag)
+    decorated_list = [ (ns_tag[0] or b'', ns_tag[1], i, ns_tag)
                        for i, ns_tag in enumerate(l) ]
     decorated_list.sort()
     return [ item[-1] for item in decorated_list ]
@@ -246,18 +284,38 @@
     pis[0] = 0
     entities[0] = 0
 
-    if Comment in tag_names:
-        comments[0] = 1
-        tag_names = [ tag for tag in tag_names
-                      if tag is not Comment ]
-    if ProcessingInstruction in tag_names:
-        pis[0] = 1
-        tag_names = [ tag for tag in tag_names
-                      if tag is not ProcessingInstruction ]
-    if Entity in tag_names:
-        entities[0] = 1
-        tag_names = [ tag for tag in tag_names
-                      if tag is not Entity ]
-    ns_tags = _sortedTagList([ _getNsTag(tag) for tag in tag_names ])
+    ns_tags = []
+    for tag in tag_names:
+        if tag is Comment:
+            comments[0] = 1
+        elif tag is ProcessingInstruction:
+            pis[0] = 1
+        elif tag is Entity:
+            entities[0] = 1
+        else:
+            ns_tags.append(_getNsTag(tag))
+
     return [ (ns, tag if tag != '*' else None)
-             for ns, tag in ns_tags ]
+             for ns, tag in _sortedTagList(ns_tags) ]
+
+cdef Py_ssize_t _mapTagsToCharArray(xmlDoc* c_doc, list ns_tags,
+                                    char** c_ns_tags) except -1:
+    cdef Py_ssize_t count = 0
+    cdef char* c_tag
+    for ns, tag in ns_tags:
+        if ns is None:
+            c_ns_tags[0] = NULL
+        else:
+            c_ns_tags[0] = _cstr(ns)
+        if tag is None:
+            c_ns_tags[1] = NULL
+        else:
+            c_tag = _cstr(tag)
+            c_ns_tags[1] = tree.xmlDictExists(
+                c_doc.dict, c_tag, cstd.strlen(c_tag))
+            if c_ns_tags[1] == NULL:
+                # not in the dict => not in the document
+                continue
+        c_ns_tags += 2
+        count += 1
+    return count

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 Nov 24 18:59:37 2009
@@ -7,7 +7,14 @@
 test_elementtree
 """
 
-import os.path, unittest, copy, sys, operator, tempfile, gzip
+import os.path
+import unittest
+import copy
+import sys
+import re
+import operator
+import tempfile
+import gzip
 
 this_dir = os.path.dirname(__file__)
 if this_dir not in sys.path:
@@ -312,6 +319,85 @@
         self.assertEquals(_bytes('<test>TEST<a>A<b>BCT</b>BT</a>AT<x>X<a>A<b></b>BTCT</a>AT</x>XT</test>'),
                           self._writeElement(root))
 
+    def test_strip_tags_pi_comment(self):
+        XML = self.etree.XML
+        PI = self.etree.ProcessingInstruction
+        Comment = self.etree.Comment
+        xml = _bytes('<!--comment1-->\n<?PI1?>\n<test>TEST<!--comment2-->XT<?PI2?></test>\n<!--comment3-->\n<?PI1?>')
+
+        root = XML(xml)
+        self.etree.strip_tags(root, PI)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TEST<!--comment2-->XT</test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, Comment)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TESTXT<?PI2?></test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, PI, Comment)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TESTXT</test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, Comment, PI)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TESTXT</test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+    def test_strip_tags_pi_comment_all(self):
+        XML = self.etree.XML
+        ElementTree = self.etree.ElementTree
+        PI = self.etree.ProcessingInstruction
+        Comment = self.etree.Comment
+        xml = _bytes('<!--comment1-->\n<?PI1?>\n<test>TEST<!--comment2-->XT<?PI2?></test>\n<!--comment3-->\n<?PI1?>')
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), PI)
+        self.assertEquals(_bytes('<!--comment1-->\n<test>TEST<!--comment2-->XT</test>\n<!--comment3-->'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), Comment)
+        self.assertEquals(_bytes('<?PI1?>\n<test>TESTXT<?PI2?></test>\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), PI, Comment)
+        self.assertEquals(_bytes('<test>TESTXT</test>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), Comment, PI)
+        self.assertEquals(_bytes('<test>TESTXT</test>'),
+                          self._writeElement(root))
+
+    def test_strip_tags_doc_style(self):
+        XML = self.etree.XML
+        xml = _bytes('''
+        <div>
+	    <div>
+		I like <strong>sheep</strong>.
+		<br/>
+		I like lots of <strong>sheep</strong>.
+		<br/>
+		Click <a href="http://www.sheep.com">here</a> for <a href="http://www.sheep.com">those</a> sheep.
+		<br/>
+	    </div>
+        </div>
+        '''.strip())
+
+        root = XML(xml)
+        self.etree.strip_tags(root, 'a')
+        self.assertEquals(re.sub(_bytes('</?a[^>]*>'), '', xml).replace('<br/>', '<br></br>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, 'a', 'br')
+        self.assertEquals(re.sub(_bytes('</?a[^>]*>'), '',
+                                 re.sub(_bytes('<br[^>]*>'), '', xml)),
+                          self._writeElement(root))
+
     def test_strip_tags_ns(self):
         XML = self.etree.XML
         xml = _bytes('<test>TEST<n:a xmlns:n="urn:a">A<b>B<c xmlns="urn:c"/>CT</b>BT</n:a>AT<x>X<a>A<b xmlns="urn:a"/>BT<c xmlns="urn:x"/>CT</a>AT</x>XT</test>')

From scoder at codespeak.net  Tue Nov 24 19:09:44 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 19:09:44 +0100 (CET)
Subject: [Lxml-checkins] r69609 - in lxml/branch/lxml-2.2: . src/lxml
	src/lxml/tests
Message-ID: <20091124180944.0B723168008@codespeak.net>

Author: scoder
Date: Tue Nov 24 19:09:43 2009
New Revision: 69609

Modified:
   lxml/branch/lxml-2.2/   (props changed)
   lxml/branch/lxml-2.2/CHANGES.txt
   lxml/branch/lxml-2.2/INSTALL.txt   (props changed)
   lxml/branch/lxml-2.2/src/lxml/apihelpers.pxi
   lxml/branch/lxml-2.2/src/lxml/cleanup.pxi
   lxml/branch/lxml-2.2/src/lxml/tests/test_etree.py
Log:
trunk merge: fix strip_tags() and strip_elements()

Modified: lxml/branch/lxml-2.2/CHANGES.txt
==============================================================================
--- lxml/branch/lxml-2.2/CHANGES.txt	(original)
+++ lxml/branch/lxml-2.2/CHANGES.txt	Tue Nov 24 19:09:43 2009
@@ -14,6 +14,9 @@
 Bugs fixed
 ----------
 
+* The functions ``strip_tags()`` and ``strip_elements()`` in
+  ``lxml.etree`` did not remove all occurrences of a tag in all cases.
+
 * Crash in XSLT extension elements when the XSLT context node is not
   an element.
 

Modified: lxml/branch/lxml-2.2/src/lxml/apihelpers.pxi
==============================================================================
--- lxml/branch/lxml-2.2/src/lxml/apihelpers.pxi	(original)
+++ lxml/branch/lxml-2.2/src/lxml/apihelpers.pxi	Tue Nov 24 19:09:43 2009
@@ -395,6 +395,7 @@
     return 0
 
 cdef int _replaceNodeByChildren(_Document doc, xmlNode* c_node) except -1:
+    # NOTE: this does not deallocate the node, just unlink it!
     cdef xmlNode* c_parent
     cdef xmlNode* c_child
     if c_node.children is NULL:
@@ -879,6 +880,35 @@
     else:
         return 0
 
+cdef inline bint _tagMatchesExactly(xmlNode* c_node, char* c_href, char* c_name):
+    u"""Tests if the node matches namespace URI and tag name.
+
+    This differs from _tagMatches() in that it does not consider a
+    NULL value in c_href a wildcard, and that it expects the c_name to
+    be taken from the doc dict, i.e. it only compares the names by
+    address.
+
+    A node matches if it matches both c_href and c_name.
+
+    A node matches c_href if any of the following is true:
+    * its namespace is NULL and c_href is the empty string
+    * its namespace string equals the c_href string
+
+    A node matches c_name if any of the following is true:
+    * c_name is NULL
+    * its name string points to the same address (!) as c_name
+    """
+    cdef char* c_node_href
+    if c_name is not NULL and c_name is not c_node.name:
+        return 0
+    c_node_href = _getNs(c_node)
+    if c_href is NULL:
+        return c_node_href is NULL or c_node_href[0] == '\0'
+    elif c_node_href is NULL:
+        return 0
+    else:
+        return cstd.strcmp(c_href, c_node_href) == 0
+
 cdef int _removeNode(_Document doc, xmlNode* c_node) except -1:
     u"""Unlink and free a node and subnodes if possible.  Otherwise, make sure
     it's self-contained.
@@ -892,6 +922,29 @@
         moveNodeToDocument(doc, c_node.doc, c_node)
     return 0
 
+cdef int _removeSiblings(xmlNode* c_element, int node_type, bint with_tail) except -1:
+    cdef xmlNode* c_node
+    cdef xmlNode* c_next
+    c_node = c_element.next
+    while c_node is not NULL:
+        c_next = _nextElement(c_node)
+        if c_node.type == node_type:
+            if with_tail:
+                _removeText(c_node.next)
+            tree.xmlUnlinkNode(c_node)
+            attemptDeallocation(c_node)
+        c_node = c_next
+    c_node = c_element.prev
+    while c_node is not NULL:
+        c_next = _previousElement(c_node)
+        if c_node.type == node_type:
+            if with_tail:
+                _removeText(c_node.next)
+            tree.xmlUnlinkNode(c_node)
+            attemptDeallocation(c_node)
+        c_node = c_next
+    return 0
+
 cdef void _moveTail(xmlNode* c_tail, xmlNode* c_target):
     cdef xmlNode* c_next
     # tail support: look for any text nodes trailing this node and 

Modified: lxml/branch/lxml-2.2/src/lxml/cleanup.pxi
==============================================================================
--- lxml/branch/lxml-2.2/src/lxml/cleanup.pxi	(original)
+++ lxml/branch/lxml-2.2/src/lxml/cleanup.pxi	Tue Nov 24 19:09:43 2009
@@ -22,40 +22,56 @@
                          'simpleattr',
                          '{http://some/ns}attrname')
     """
-    cdef xmlNode* c_node
-    cdef xmlAttr* c_attr
     cdef _Element element
     cdef list ns_tags
-    cdef char* c_name
+    cdef char** c_ns_tags
+    cdef Py_ssize_t c_tag_count
 
     element = _rootNodeOrRaise(tree_or_element)
     if not attribute_names: return
 
     ns_tags = _sortedTagList([ _getNsTag(attr)
                                for attr in <tuple>attribute_names ])
-    ns_tags = [ (ns, tag if tag != '*' else None)
+    ns_tags = [ (ns, tag if tag != b'*' else None)
                 for ns, tag in ns_tags ]
 
-    c_node = element._c_node
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
+    try:
+        c_tag_count = _mapTagsToCharArray(element._doc._c_doc, ns_tags, c_ns_tags)
+        if c_tag_count > 0:
+            _strip_attributes(element._c_node, c_ns_tags, c_tag_count)
+    finally:
+        cstd.free(c_ns_tags)
+
+cdef _strip_attributes(xmlNode* c_node, char** c_ns_tags, Py_ssize_t c_tag_count):
+    cdef xmlAttr* c_attr
+    cdef Py_ssize_t i
+    cdef char* c_href
+    cdef char* c_name
+
     tree.BEGIN_FOR_EACH_ELEMENT_FROM(c_node, c_node, 1)
     if c_node.type == tree.XML_ELEMENT_NODE:
         if c_node.properties is not NULL:
-            for ns, tag in ns_tags:
-                # must search attributes manually to make sure we only
-                # match on blank tag names if there is no namespace
-                c_name = NULL if tag is None else _cstr(tag)
+            for i in xrange(c_tag_count):
+                c_href = c_ns_tags[2*i]
+                c_name = c_ns_tags[2*i+1]
+                # must compare attributes manually to make sure we
+                # only match on wildcard tag names if the attribute
+                # has no namespace
                 c_attr = c_node.properties
                 while c_attr is not NULL:
-                    if ns is None:
-                        if c_attr.ns is NULL or c_attr.ns.href is NULL:
-                            if c_name is NULL or \
-                                   cstd.strcmp(c_attr.name, c_name) == 0:
+                    if c_name is NULL or c_attr.name == c_name:
+                        if c_href is NULL:
+                            if c_attr.ns is NULL or c_attr.ns.href is NULL:
                                 tree.xmlRemoveProp(c_attr)
                                 break
-                    elif c_attr.ns is not NULL and c_attr.ns.href is not NULL:
-                        if cstd.strcmp(c_attr.ns.href, _cstr(ns)) == 0:
-                            if c_name is NULL or \
-                                   cstd.strcmp(c_attr.name, c_name) == 0:
+                        elif c_attr.ns is not NULL and c_attr.ns.href is not NULL:
+                            if cstd.strcmp(c_attr.ns.href, c_href) == 0:
                                 tree.xmlRemoveProp(c_attr)
                                 break
                     c_attr = c_attr.next
@@ -84,14 +100,11 @@
             Comment                      # comments
             )
     """
-    cdef xmlNode* c_node
-    cdef xmlNode* c_child
-    cdef xmlNode* c_next
-    cdef char* c_href
-    cdef char* c_name
     cdef _Element element
     cdef _Document doc
     cdef list ns_tags
+    cdef char** c_ns_tags
+    cdef Py_ssize_t c_tag_count
     cdef bint strip_comments, strip_pis, strip_entities
 
     doc = _documentOrRaise(tree_or_element)
@@ -101,7 +114,35 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    c_node = element._c_node
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
+    if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
+        # include PIs and comments next to the root node
+        if strip_comments:
+            _removeSiblings(element._c_node, tree.XML_COMMENT_NODE, with_tail)
+        if strip_pis:
+            _removeSiblings(element._c_node, tree.XML_PI_NODE, with_tail)
+
+    try:
+        c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
+        if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:
+            _strip_elements(doc, element._c_node, c_ns_tags, c_tag_count,
+                            strip_comments, strip_pis, strip_entities, with_tail)
+    finally:
+        cstd.free(c_ns_tags)
+
+cdef _strip_elements(_Document doc, xmlNode* c_node,
+                     char** c_ns_tags, Py_ssize_t c_tag_count,
+                     bint strip_comments, bint strip_pis, bint strip_entities,
+                     bint with_tail):
+    cdef xmlNode* c_child
+    cdef xmlNode* c_next
+    cdef Py_ssize_t i
+
     tree.BEGIN_FOR_EACH_ELEMENT_FROM(c_node, c_node, 1)
     if c_node.type == tree.XML_ELEMENT_NODE:
         # we run through the children here to prevent any problems
@@ -109,39 +150,25 @@
         # c_node itself
         c_child = _findChildForwards(c_node, 0)
         while c_child is not NULL:
+            c_next = _nextElement(c_child)
             if c_child.type == tree.XML_ELEMENT_NODE:
-                for ns, tag in ns_tags:
-                    if ns is None:
-                        # _tagMatches() considers NULL a wildcard
-                        # match but we don't
-                        if c_child.ns is not NULL and c_child.ns.href is not NULL:
-                            continue
-                        c_href = NULL
-                    else:
-                        c_href = _cstr(ns)
-                    c_name = NULL if tag is None else _cstr(tag)
-                    if _tagMatches(c_child, c_href, c_name):
-                        c_next = _nextElement(c_child)
+                for i in xrange(c_tag_count):
+                    if _tagMatchesExactly(c_child, c_ns_tags[2*i], c_ns_tags[2*i+1]):
                         if not with_tail:
                             tree.xmlUnlinkNode(c_child)
                         _removeNode(doc, c_child)
-                        c_child = c_next
                         break
-                else:
-                    c_child = _nextElement(c_child)
-            elif strip_comments and c_child.type == tree.XML_COMMENT_NODE or \
-                     strip_pis and c_child.type == tree.XML_PI_NODE or \
-                     strip_entities and c_child.type == tree.XML_ENTITY_REF_NODE:
-                c_next = _nextElement(c_child)
+            elif c_child.type == tree.XML_COMMENT_NODE and strip_comments \
+                     or c_child.type == tree.XML_PI_NODE and strip_pis \
+                     or c_child.type == tree.XML_ENTITY_REF_NODE and strip_entities:
                 if with_tail:
-                    _removeText(c_next)
+                    _removeText(c_child.next)
                 tree.xmlUnlinkNode(c_child)
                 attemptDeallocation(c_child)
-                c_child = c_next
-            else:
-                c_child = _nextElement(c_child)
+            c_child = c_next
     tree.END_FOR_EACH_ELEMENT_FROM(c_node)
 
+
 def strip_tags(tree_or_element, *tag_names):
     u"""strip_tags(tree_or_element, *tag_names)
 
@@ -164,15 +191,12 @@
             Comment                      # comments (including their text!)
             )
     """
-    cdef xmlNode* c_node
-    cdef xmlNode* c_child
-    cdef xmlNode* c_next
-    cdef char* c_href
-    cdef char* c_name
     cdef _Element element
     cdef _Document doc
     cdef list ns_tags
     cdef bint strip_comments, strip_pis, strip_entities
+    cdef char** c_ns_tags
+    cdef Py_ssize_t c_tag_count
 
     doc = _documentOrRaise(tree_or_element)
     element = _rootNodeOrRaise(tree_or_element)
@@ -181,7 +205,34 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    c_node = element._c_node
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
+    if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
+        # include PIs and comments next to the root node
+        if strip_comments:
+            _removeSiblings(element._c_node, tree.XML_COMMENT_NODE, 0)
+        if strip_pis:
+            _removeSiblings(element._c_node, tree.XML_PI_NODE, 0)
+
+    try:
+        c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
+        if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:
+            _strip_tags(doc, element._c_node, c_ns_tags, c_tag_count,
+                        strip_comments, strip_pis, strip_entities)
+    finally:
+        cstd.free(c_ns_tags)
+
+cdef _strip_tags(_Document doc, xmlNode* c_node,
+                 char** c_ns_tags, Py_ssize_t c_tag_count,
+                 bint strip_comments, bint strip_pis, bint strip_entities):
+    cdef xmlNode* c_child
+    cdef xmlNode* c_next
+    cdef Py_ssize_t i
+
     tree.BEGIN_FOR_EACH_ELEMENT_FROM(c_node, c_node, 1)
     if c_node.type == tree.XML_ELEMENT_NODE:
         # we run through the children here to prevent any problems
@@ -190,39 +241,26 @@
         c_child = _findChildForwards(c_node, 0)
         while c_child is not NULL:
             if c_child.type == tree.XML_ELEMENT_NODE:
-                for ns, tag in ns_tags:
-                    if ns is None:
-                        # _tagMatches() considers NULL a wildcard
-                        # match but we don't
-                        if c_child.ns is not NULL and c_child.ns.href is not NULL:
-                            continue
-                        c_href = NULL
-                    else:
-                        c_href = _cstr(ns)
-                    c_name = NULL if tag is None else _cstr(tag)
-                    if _tagMatches(c_child, c_href, c_name):
-                        if c_child.children is not NULL:
-                            c_next = _findChildForwards(c_child, 0)
-                        else:
-                            c_next = _nextElement(c_child)
+                for i in xrange(c_tag_count):
+                    if _tagMatchesExactly(c_child, c_ns_tags[2*i], c_ns_tags[2*i+1]):
+                        c_next = _findChildForwards(c_child, 0) or _nextElement(c_child)
                         _replaceNodeByChildren(doc, c_child)
                         if not attemptDeallocation(c_child):
-                            if c_child.ns is not NULL:
+                            if c_child.nsDef is not NULL:
                                 # make namespaces absolute
                                 moveNodeToDocument(doc, doc._c_doc, c_child)
                         c_child = c_next
                         break
                 else:
-                    c_child = c_child.next
-            elif strip_comments and c_child.type == tree.XML_COMMENT_NODE or \
-                     strip_pis and c_child.type == tree.XML_PI_NODE or \
-                     strip_entities and c_child.type == tree.XML_ENTITY_REF_NODE:
+                    c_child = _nextElement(c_child)
+            else:
                 c_next = _nextElement(c_child)
-                tree.xmlUnlinkNode(c_child)
-                attemptDeallocation(c_child)
+                if c_child.type == tree.XML_COMMENT_NODE and strip_comments \
+                       or c_child.type == tree.XML_PI_NODE and strip_pis \
+                       or c_child.type == tree.XML_ENTITY_REF_NODE and strip_entities:
+                    tree.xmlUnlinkNode(c_child)
+                    attemptDeallocation(c_child)
                 c_child = c_next
-            else:
-                c_child = _nextElement(c_child)
     tree.END_FOR_EACH_ELEMENT_FROM(c_node)
 
 
@@ -235,7 +273,7 @@
     cdef list decorated_list
     cdef tuple ns_tag
     cdef Py_ssize_t i
-    decorated_list = [ (ns_tag[0] or '', ns_tag[1], i, ns_tag)
+    decorated_list = [ (ns_tag[0] or b'', ns_tag[1], i, ns_tag)
                        for i, ns_tag in enumerate(l) ]
     decorated_list.sort()
     return [ item[-1] for item in decorated_list ]
@@ -246,18 +284,38 @@
     pis[0] = 0
     entities[0] = 0
 
-    if Comment in tag_names:
-        comments[0] = 1
-        tag_names = [ tag for tag in tag_names
-                      if tag is not Comment ]
-    if ProcessingInstruction in tag_names:
-        pis[0] = 1
-        tag_names = [ tag for tag in tag_names
-                      if tag is not ProcessingInstruction ]
-    if Entity in tag_names:
-        entities[0] = 1
-        tag_names = [ tag for tag in tag_names
-                      if tag is not Entity ]
-    ns_tags = _sortedTagList([ _getNsTag(tag) for tag in tag_names ])
+    ns_tags = []
+    for tag in tag_names:
+        if tag is Comment:
+            comments[0] = 1
+        elif tag is ProcessingInstruction:
+            pis[0] = 1
+        elif tag is Entity:
+            entities[0] = 1
+        else:
+            ns_tags.append(_getNsTag(tag))
+
     return [ (ns, tag if tag != '*' else None)
-             for ns, tag in ns_tags ]
+             for ns, tag in _sortedTagList(ns_tags) ]
+
+cdef Py_ssize_t _mapTagsToCharArray(xmlDoc* c_doc, list ns_tags,
+                                    char** c_ns_tags) except -1:
+    cdef Py_ssize_t count = 0
+    cdef char* c_tag
+    for ns, tag in ns_tags:
+        if ns is None:
+            c_ns_tags[0] = NULL
+        else:
+            c_ns_tags[0] = _cstr(ns)
+        if tag is None:
+            c_ns_tags[1] = NULL
+        else:
+            c_tag = _cstr(tag)
+            c_ns_tags[1] = tree.xmlDictExists(
+                c_doc.dict, c_tag, cstd.strlen(c_tag))
+            if c_ns_tags[1] == NULL:
+                # not in the dict => not in the document
+                continue
+        c_ns_tags += 2
+        count += 1
+    return count

Modified: lxml/branch/lxml-2.2/src/lxml/tests/test_etree.py
==============================================================================
--- lxml/branch/lxml-2.2/src/lxml/tests/test_etree.py	(original)
+++ lxml/branch/lxml-2.2/src/lxml/tests/test_etree.py	Tue Nov 24 19:09:43 2009
@@ -7,7 +7,14 @@
 test_elementtree
 """
 
-import os.path, unittest, copy, sys, operator, tempfile, gzip
+import os.path
+import unittest
+import copy
+import sys
+import re
+import operator
+import tempfile
+import gzip
 
 this_dir = os.path.dirname(__file__)
 if this_dir not in sys.path:
@@ -312,6 +319,85 @@
         self.assertEquals(_bytes('<test>TEST<a>A<b>BCT</b>BT</a>AT<x>X<a>A<b></b>BTCT</a>AT</x>XT</test>'),
                           self._writeElement(root))
 
+    def test_strip_tags_pi_comment(self):
+        XML = self.etree.XML
+        PI = self.etree.ProcessingInstruction
+        Comment = self.etree.Comment
+        xml = _bytes('<!--comment1-->\n<?PI1?>\n<test>TEST<!--comment2-->XT<?PI2?></test>\n<!--comment3-->\n<?PI1?>')
+
+        root = XML(xml)
+        self.etree.strip_tags(root, PI)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TEST<!--comment2-->XT</test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, Comment)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TESTXT<?PI2?></test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, PI, Comment)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TESTXT</test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, Comment, PI)
+        self.assertEquals(_bytes('<!--comment1-->\n<?PI1?>\n<test>TESTXT</test>\n<!--comment3-->\n<?PI1?>'),
+                          self._writeElement(root))
+
+    def test_strip_tags_pi_comment_all(self):
+        XML = self.etree.XML
+        ElementTree = self.etree.ElementTree
+        PI = self.etree.ProcessingInstruction
+        Comment = self.etree.Comment
+        xml = _bytes('<!--comment1-->\n<?PI1?>\n<test>TEST<!--comment2-->XT<?PI2?></test>\n<!--comment3-->\n<?PI1?>')
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), PI)
+        self.assertEquals(_bytes('<!--comment1-->\n<test>TEST<!--comment2-->XT</test>\n<!--comment3-->'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), Comment)
+        self.assertEquals(_bytes('<?PI1?>\n<test>TESTXT<?PI2?></test>\n<?PI1?>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), PI, Comment)
+        self.assertEquals(_bytes('<test>TESTXT</test>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(ElementTree(root), Comment, PI)
+        self.assertEquals(_bytes('<test>TESTXT</test>'),
+                          self._writeElement(root))
+
+    def test_strip_tags_doc_style(self):
+        XML = self.etree.XML
+        xml = _bytes('''
+        <div>
+	    <div>
+		I like <strong>sheep</strong>.
+		<br/>
+		I like lots of <strong>sheep</strong>.
+		<br/>
+		Click <a href="http://www.sheep.com">here</a> for <a href="http://www.sheep.com">those</a> sheep.
+		<br/>
+	    </div>
+        </div>
+        '''.strip())
+
+        root = XML(xml)
+        self.etree.strip_tags(root, 'a')
+        self.assertEquals(re.sub(_bytes('</?a[^>]*>'), '', xml).replace('<br/>', '<br></br>'),
+                          self._writeElement(root))
+
+        root = XML(xml)
+        self.etree.strip_tags(root, 'a', 'br')
+        self.assertEquals(re.sub(_bytes('</?a[^>]*>'), '',
+                                 re.sub(_bytes('<br[^>]*>'), '', xml)),
+                          self._writeElement(root))
+
     def test_strip_tags_ns(self):
         XML = self.etree.XML
         xml = _bytes('<test>TEST<n:a xmlns:n="urn:a">A<b>B<c xmlns="urn:c"/>CT</b>BT</n:a>AT<x>X<a>A<b xmlns="urn:a"/>BT<c xmlns="urn:x"/>CT</a>AT</x>XT</test>')

From scoder at codespeak.net  Tue Nov 24 21:39:02 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 21:39:02 +0100 (CET)
Subject: [Lxml-checkins] r69616 - in lxml/trunk: . src/lxml
Message-ID: <20091124203902.53EE4168021@codespeak.net>

Author: scoder
Date: Tue Nov 24 21:39:01 2009
New Revision: 69616

Modified:
   lxml/trunk/   (props changed)
   lxml/trunk/src/lxml/cleanup.pxi
Log:
 r5351 at delle:  sbehnel | 2009-11-24 21:38:56 +0100
 prevent memory leaks on errors


Modified: lxml/trunk/src/lxml/cleanup.pxi
==============================================================================
--- lxml/trunk/src/lxml/cleanup.pxi	(original)
+++ lxml/trunk/src/lxml/cleanup.pxi	Tue Nov 24 21:39:01 2009
@@ -114,12 +114,6 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    # tag names are passes as C pointers as this allows us to take
-    # them from the doc dict and do pointer comparisons
-    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
-    if c_ns_tags is NULL:
-        python.PyErr_NoMemory()
-
     if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
         # include PIs and comments next to the root node
         if strip_comments:
@@ -127,6 +121,12 @@
         if strip_pis:
             _removeSiblings(element._c_node, tree.XML_PI_NODE, with_tail)
 
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
     try:
         c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
         if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:
@@ -205,12 +205,6 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    # tag names are passes as C pointers as this allows us to take
-    # them from the doc dict and do pointer comparisons
-    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
-    if c_ns_tags is NULL:
-        python.PyErr_NoMemory()
-
     if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
         # include PIs and comments next to the root node
         if strip_comments:
@@ -218,6 +212,12 @@
         if strip_pis:
             _removeSiblings(element._c_node, tree.XML_PI_NODE, 0)
 
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
     try:
         c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
         if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:

From scoder at codespeak.net  Tue Nov 24 21:40:44 2009
From: scoder at codespeak.net (scoder at codespeak.net)
Date: Tue, 24 Nov 2009 21:40:44 +0100 (CET)
Subject: [Lxml-checkins] r69617 - in lxml/branch/lxml-2.2: . src/lxml
Message-ID: <20091124204044.83562168021@codespeak.net>

Author: scoder
Date: Tue Nov 24 21:40:43 2009
New Revision: 69617

Modified:
   lxml/branch/lxml-2.2/   (props changed)
   lxml/branch/lxml-2.2/INSTALL.txt   (props changed)
   lxml/branch/lxml-2.2/src/lxml/cleanup.pxi
Log:
trunk merge: fix memory leak in error case

Modified: lxml/branch/lxml-2.2/src/lxml/cleanup.pxi
==============================================================================
--- lxml/branch/lxml-2.2/src/lxml/cleanup.pxi	(original)
+++ lxml/branch/lxml-2.2/src/lxml/cleanup.pxi	Tue Nov 24 21:40:43 2009
@@ -114,12 +114,6 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    # tag names are passes as C pointers as this allows us to take
-    # them from the doc dict and do pointer comparisons
-    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
-    if c_ns_tags is NULL:
-        python.PyErr_NoMemory()
-
     if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
         # include PIs and comments next to the root node
         if strip_comments:
@@ -127,6 +121,12 @@
         if strip_pis:
             _removeSiblings(element._c_node, tree.XML_PI_NODE, with_tail)
 
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
     try:
         c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
         if c_tag_count > 0 or strip_comments or strip_pis or strip_entities:
@@ -205,12 +205,6 @@
     ns_tags = _filterSpecialTagNames(
         tag_names, &strip_comments, &strip_pis, &strip_entities)
 
-    # tag names are passes as C pointers as this allows us to take
-    # them from the doc dict and do pointer comparisons
-    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
-    if c_ns_tags is NULL:
-        python.PyErr_NoMemory()
-
     if (strip_comments or strip_pis) and isinstance(tree_or_element, _ElementTree):
         # include PIs and comments next to the root node
         if strip_comments:
@@ -218,6 +212,12 @@
         if strip_pis:
             _removeSiblings(element._c_node, tree.XML_PI_NODE, 0)
 
+    # tag names are passes as C pointers as this allows us to take
+    # them from the doc dict and do pointer comparisons
+    c_ns_tags = <char**> cstd.malloc(sizeof(char*) * len(ns_tags) * 2 + 2)
+    if c_ns_tags is NULL:
+        python.PyErr_NoMemory()
+
     try:
         c_tag_count = _mapTagsToCharArray(doc._c_doc, ns_tags, c_ns_tags)
         if c_tag_count > 0 or strip_comments or strip_pis or strip_entities: