|
7c693dad
|
2012-07-25T16:32:18
|
|
Cleanups and new limit APIs for dictionaries
* include/libxml/dict.h dict.c: adding 2 new functions xmlDictGetUsage
and xmlDictSetLimit allowing to review the amount of memory allocated
for dictionary strings. Aslo cleanup of various signed int used as
size values in the code.
|
|
|
379ebc1d
|
2012-05-18T15:41:31
|
|
Cleanup on randomization
tsan reported that rand() is not thread safe, so create
a thread safe wrapper, use rand_r() if available.
Consolidate the function, initialization and cleanup in
dict.c and make sure it is initialized in xmlInitParser()
|
|
1441251f
|
2005-01-21T23:53:26
|
|
a single lock version mostly avoid the cost penalty of the lock in case of
* dict.c parser.c include/libxml/dict.h: a single lock version
mostly avoid the cost penalty of the lock in case of low
parallelism, so applying that version instead.
Daniel
|
|
|
6bb3e86d
|
2004-11-24T12:39:00
|
|
added xmlDictExists() to the dictionnary interface. applying
* dict.c include/libxml/dict.h: added xmlDictExists() to the
dictionnary interface.
* xmlreader.c: applying xmlTextReaderHasAttributes fix for namespaces
from Rob Richards
Daniel
|
|
|
4773df2a
|
2004-01-23T13:15:13
|
|
added io1.c an example ox xmlIO usage and io1.res test result, fixed a
* doc/examples/*: added io1.c an example ox xmlIO usage and io1.res
test result, fixed a awful lot of memory leaks showing up in
testWriter.c, changed the examples and the Makefiles to test
memory leaks.
* xmlwriter.c: fixed a memory leak
* Makefile.am: run the doc/examples regression tests as part of
make tests
* xpath.c include/libxml/xpath.h: added xmlXPathCtxtCompile() to
compile an XPath expression within a context, currently the goal
is to be able to reuse the XSLT stylesheet dictionnary, but this
opens the door to others possible optimizations.
* dict.c include/libxml/dict.h: added xmlDictCreateSub() which allows
to build a new dictionnary based on another read-only dictionnary.
This is needed for XSLT to keep the stylesheet dictionnary read-only
while being able to reuse the strings for the transformation
dictionnary.
* xinclude.c: fixed a dictionnar reference counting problem occuring
when document parsing failed.
* testSAX.c: adding option --repeat for timing 100times the parsing
* doc/* : rebuilt all the docs
Daniel
|
|
|
be586972
|
2003-11-18T20:56:51
|
|
modified the file header to add more informations, painful... updated to
* include/libxml/*.h include/libxml/*.h.in: modified the file
header to add more informations, painful...
* genChRanges.py genUnicode.py: updated to generate said changes
in headers
* doc/apibuild.py: extract headers, add them to libxml2-api.xml
* *.html *.xsl *.xml: updated the stylesheets to flag geprecated
APIs modules. Updated the stylesheets, some cleanups, regenerated
* doc/html/*.html: regenerated added back book1 and libxml-lib.html
Daniel
|
|
|
e96a2a4b
|
2003-09-24T21:23:56
|
|
adding repeated parsing and validating tests make the new DOM tree
* Makefile.am: adding repeated parsing and validating tests
* SAX2.c parser.c tree.c include/libxml/parser.h: make the new
DOM tree building interfaces use the dictionary from the
parsing context to build the element and attributes names
as well as formatting spaces and short text nodes
* include/libxml/dict.h dict.c: added some reference counting
for xmlDictPtr because they can be shared by documents and
a parser context.
* xmlreader.c: a bit of cleanup, remove the specific tree freeing
functions and use the standard ones now.
* xmllint.c: add --nodict
* python/libxml.c: fix a stupid bug so that ns() works on
attribute nodes.
Daniel
|
|
|
e72c508c
|
2003-09-19T12:44:05
|
|
preparing a 2.6.0-beta2 release avoid a warning avoid duplicate code in
* configure.in: preparing a 2.6.0-beta2 release
* xmlIO.c: avoid a warning
* tree.c: avoid duplicate code in xmlReplaceNode as pointed out
by Chris Ryland
* include/libxml/dict.h: add a QName access lookup to the
dictionary.
* xmlreader.c include/libxml/xmlreader.h: adding const access
based on the dictionary interface for string read from the
reader, the node content access is still TODO, it's too different
Daniel
|
|
|
81514ba4
|
2003-09-16T23:17:26
|
|
do string allocations in large pools, allowing to find if a string pertain
* dict.c include/libxml/dict.h: do string allocations in large
pools, allowing to find if a string pertain to a dict quickly
* xmllint.c: fix --stream --repeat --timing
* Makefile.am: the testThreads run output should be seen.
Daniel
|
|
76874e45
|
2003-08-25T09:05:12
|
|
Exportability taint of the headers
|
|
|
2fdbd32d
|
2003-08-18T12:15:38
|
|
new dictionary module to keep a single instance of the names used by the
* dict.c include/libxml/dict.h Makefile.am include/libxml/Makefile.am:
new dictionary module to keep a single instance of the names used
by the parser
* DOCBparser.c HTMLparser.c parser.c parserInternals.c valid.c:
switched all parsers to use the dictionary internally
* include/libxml/HTMLparser.h include/libxml/parser.h
include/libxml/parserInternals.h include/libxml/valid.h:
Some of the interfaces changed as a result to receive or return
"const xmlChar *" instead of "xmlChar *", this is either
insignificant from an user point of view or when the returning
value changed, those function are really parser internal methods
that no user code should really change
* doc/libxml2-api.xml doc/html/*: the API interface changed and
the docs were regenerated
Daniel
|