* libxslt/attributes.c: removed a warning
* libxslt/transform.c libxslt/transform.h win32/libxslt.def.src:
added xsltRunStylesheetUser() API needed to fix#78546
* xsltproc/xsltproc.c: second part of the fix#78546
Daniel
* tests/docs/Makefile.am tests/docs/bug-83.xml
tests/general/Makefile.am tests/general/bug-83.*: added a
specific example for bug #78662 in the regression tests
* tests/docbook/: this also changed a couple of DocBook results
Daniel
* configure.in: trying to kill #77827 IEEE conformance on alphas
* libxslt/imports.c libxslt/transform.c libxslt/xslt.c: fixing
bug #78211
* tests/docs/Makefile.am tests/docs/bug-82.xml
tests/general/Makefile.am tests/general/bug-82.*: added a
specific example for bug #78211 in the regression tests
Daniel
* tests/docs/Makefile.am tests/docs/bug-81.xml
tests/general/Makefile.am tests/general/bug-81.*: added a
specific example for bug #76927 in the regression tests
Daniel
* AUTHORS HACKING: Added Igor Zlatkovic as official maintainer
* python/Makefile.am python/tests/Makefile.am: Albert Chin pointed
that $(datadir) should be used for docs
Daniel
* libxslt/attributes.c libxslt/attributes.h libxslt/pattern.c
libxslt/xslt.c: Fix bug #76043 about cascading attribute sets
* tests/docs/Makefile.am tests/docs/bug-80.xml
tests/general/Makefile.am tests/general/bug-80.*: added a
specific example for bug #76043 in the regression tests
Daniel
* libxslt/pattern.c: Fixing bug #75902 error with @foo[..]
steps which were not compiled
* tests/docs/Makefile.am tests/docs/bug-79.xml
tests/general/Makefile.am tests/general/bug-79.*: added a
specific example for bug #75902 in the regression tests
Daniel
* libxslt/pattern.c: Fixing bug #75777 error with namespaced
attribute match rules evaluation
* tests/docs/Makefile.am tests/docs/bug-78.xml
tests/general/Makefile.am tests/general/bug-78.*: added a
specific example for bug #75777 in the regression tests
Daniel
* libxslt/transform.c: found another stupid bug by step by
step processing of the code
* libxslt/pattern.c: idem, except that once stupid mistake
ELEM vs NODE forced the templates to be stored in a list
instead of a hash table, fixing this stupidity should
again lead to a substantive improvement of processing speed.
Like divide by 2 processing time for DocBook stylesheets.
Daniel
* libxslt/transform.c: fixing bug #75603
* tests/docs/Makefile.am tests/docs/bug-77.xml
tests/general/Makefile.am tests/general/bug-77.*: added a
specific example for bug #75603 in the regression tests
Daniel
* Makefile.am tests/Makefile.am tests/*/Makefile.am
tests/*/*/Makefile.am : added "make valgrind" targets
to run the test suite under the debugger control
* transform.c: valgrind spotted 2 bugs, one related to
the ordering of the deallocation of the data associated to
a transofrmation, the second in xsltCopyTree when the new
node may have been coalesced with an adjacent text node.
The regression tests now pass cleanly under testgrind.
Daniel
* configure.in: preparing 1.0.14
* doc/*: updated rebuilt
* libxslt/*.c libexslt/*.c libxslt/libxslt.h libexslt/libexslt.h:
implemented the IN_LIBXSLT and IN_LIBEXSLT mechanism discussed
with the Windows maintainers
Daniel
* python/Makefile.am python/generator.py python/libxslt.c
python/types.c python/tests/Makefile.am: applied the same kind of
fixes to the Python Makefiels than to libxml2 ones. Updates
and cleanups too.
Daniel
* libexslt/common.c libxslt/pattern.c libxslt/transform.c
libxslt/variables.c: chaing result tree values which may
be deallocated and must not be kept in the template
pattern lookup cache. Thanks to Valgrin to allow finding
the real problem in bug #74857
Daniel
* tests/docs/Makefile.am tests/docs/bug-76.xml
tests/general/Makefile.am tests/general/bug-76.*: added a
home brewed test for path computation elmininating duplicate
in result sets.
Daniel
2002-03-08 jacob berkman <jacob@ximian.com>
* python/Makefile.am (libxsltmodule_la_SOURCES): remove $(srcdir)
as make will automatically look there for these files
* configure.in libxslt/xsltwin32config.h: preparing release
1.0.13
* doc/*: updated and rebuilt the docs
* python/libxslt.c: fixed a possible reentrancy problem
Daniel
* libxslt/pattern.c: Fixes the problems exposed by #73880
those ought to be computed at stylesheet compile time, not
at run-time, and the computation was wrong.
* libxslt/transform.c: get rid of fake nodes coming from node-set
transformations. At least if they are still produced they will
become easy to spot as resulting document won't be well-formed.
Daniel
* libxslt/extensions.c: fixed bug #73791 related to extension
function declared in included stylesheets
* tests/exslt/functions/function.7.*: added specific test
Daniel
* tests/docs/Makefile.am tests/docs/bug-75.xml
tests/general/Makefile.am tests/general/bug-75.*: added a
specific example for bug #72150 in the regression tests
Daniel
* libexslt/date.c: patch from Charles Bozeman for the exslt date
extension
* configure.in tests/exslt/Makefile.am tests/exslt/date/*: added
the associated regression test provided by Charles
Daniel
* xsltproc/xsltproc.c: similar to #72663 and #72658, don't memdump
unless compiled explicitely with memory debugging switched on
* TODO: refreshed a bit
Daniel
* python/generator.py python/libxslt.c: changes for the 'usual'
setup.py to allow building a libxml2-python
module based on the same code. The initialization is however
different the 2 .so files fo libxml2 and libxslt are identical and
they entry point initialize both libraries. this is done to avoid
some possible nasty problem since the Python don't merge the maps
of all shared modules.
* python/libxsl.py: attempt to cope with the shared library loading
problem when both modules are not merged.
Daniel
