- tree.c result/ result/noent/: trying to get 52712 solved, this
also made me clean up the fact that XML output in general should
not add formating blanks by default, this changed the output of
a few tests
Daniel
Fixes in the plane:
- xpath.c: fixed a [] evaluation problem reported
- test/XPath/tests/simpleaddr: extended test
- result/XPath/simpleaddr: updated result
Daniel
Small fixes:
- xpath.c: fixed a memleak when comparing nodesets
- HTMLtree.c: don't invent the HTML doctype if not available (XSLT)
- tree.c: added a TODO
Daniel
Spring cleanup ...:
- configure.in Makefile.am config.h.in xmlversion.h.in: detect if
we need string functions
- trio.[ch] strio.[ch]: embedded the Trio-0.23 string functions
to be able to use them where needed. Applied some changes
to reduce name linking pollution and compile in only what's
needed.
- HTMLtree.c debugXML.c entities.c error.c nanoftp.c valid.c
xlink.c xmlversion.h.in xpath.c: got rid of the #ifdef
for the string manipulation functions
- xmlmemory.[ch]: removed DEBUG_MEMORY_FREED and added it automatically
to the free() function of xmlmemory.c
- entities.c HTMLtree.c parserInternals.c tree.c uri.c valid.c
xinclude.c xmlIO.c xpath.c xpointer.c: removed the MEM_CLEANUP
usage.
Daniel
testHTML.c testSAX.c tree.c valid.c xmlIO.c xmlmemory.c
xmlversion.h.in xpointer.c: of course the way I defined
UNUSED breaks on old gcc version. Try to be smart and
also define it directly in xmlversion.h
- configure.in: removed -ansi flag from the pedantic set
Daniel
- parser.c: fixing bug 52299 strange condition leading
to a parser crash due to a buffer overflow
- result/noent/attrib.xml result/attrib.xml test/attrib.xml:
added the specific test case
Daniel
- xpath.[ch]: still a lot of cleanup based on XSLT, added
xmlXPathConvert{String,Number,Boolean} to be able to make
type casts without a context stack, fixed some implementation
problems related to the absence of context at parse-time,
added xmlXPathEvalPredicate() and xmlXPathFreeCompExpr()
in the public API too
- xpointer.c xpathInternals.h: we need to know at parse time
whether we are compiling an XPointer
Daniel
- xpath.[ch] xpointer.c: restaured the Binary and API compatibility
cleaned up the parser internals, refactored XPath code, added
new compilation based APIs and cleanly separated public and
private APIs.
Daniel
stuff like libxslt right now:
- Makefile.am: detect XPath memleaks in regreson tests
- error.c: fixed and error w.r.t. error reporting still using
stderr
- hash.c: added new line at end of file
- tree.h: minor cleanup
- xpath.[ch] xpointer.[ch]: Major changes ! Separated XPath
expression parsing from evaluation, resulted in a number of
changes internally, and in XPointer. Likely to break stuff
using xpathInternals.h but should remain binary compatible,
new interfaces will be added.
Daniel
- configure.in: bumped to 2.3.4
- error.c: fixed bug #51860
- tree.c: fixed bug #51861
- valid.c: cleanup, more debug, failed to fix one bug crap ...
- tree.[ch] : added xmlDefaultBufferSize
- nanoftp.c: typo in function name header block
- doc/xml.html : updated, added link to XML::LibXSLT
- doc/html/* : rebuilt the docs
Daniel
before real 0m2.483s (2.3.2 release yesterday)
current real 0m1.763s
when parsing (with tree build/freeing) db10000.xml from XSLTMark:
- xmlmemory.h HTMLparser.c HTMLtree.c entities.c parser.c
xpath.c xpointer.c tree.c uri.c valid.c xinclude.c xmlIO.c:
avoiding memcpy in production builds MEM_CLEANUP macro use
- parser.[ch] parserInternals.c: optimizations of the tightest
internal loops inside the parser. Better checking of I/O
flushing/loading conditions
- xmllint.c : added --timing
Daniel
