On Wed, 2007-22-08 at 18:21 -0600, Charlie Savage wrote: > Marc, > > > I have a reproducible crash using libxml...
> Can you provide the test data? That would be hugely helpful. I could but it's a bit of a monster, probably timing critical, and likely to run differently on different platforms. For the next day I'd like to bash on on my own. If I get nowhere, I'll package up a test case and make it available. I'll also try on another platform to see if I can make it crash reliably. > Also what OS, platform, etc? Linux bloodnok 2.6.15.6 #4 SMP PREEMPT Sat Aug 5 20:40:37 PDT 2006 i686 GNU/Linux > I've successfully used Valgrind. That's what I've been trying to use. I've also added some instrumentation to the free calls in libxml which seem to be helping. I have discovered that commenting out the call: xmlFreeNode(rxn->node); in ruby_xml_node_free (in ruby_xml_node.c) removes my crash. Consequently, I suspect the logic for determining whether to free the node is faulty. This is by no means definite as almost every minor change I make seems to affect the timing and thereby the crash behaviour, but it is a reasonable working hypothesis. Right now I'm suspicious of the reference counting in the _private field of the xmlNode structures. > Did you try these patches? > http://rubyforge.org/pipermail/libxml-devel/2007-June/000314.html No. I'll give them a look. Thanks. __ Marc
signature.asc
Description: This is a digitally signed message part
_______________________________________________ libxml-devel mailing list libxml-devel@rubyforge.org http://rubyforge.org/mailman/listinfo/libxml-devel
