Re: [libxml-devel] Code snippet to crash libxml (segfault)

2008-02-07 Thread Saurabh Nanda
> C sure is lovely, isn't it. I will try to reproduce this. If you would > not mind, please post this as a bug so I can track it at: > > http://rubyforge.org/tracker/?atid=1971&group_id=494&func=browse Done - http://rubyforge.org/tracker/index.php?func=detail&aid=17885&group_id=494&atid=1971 Wer

[libxml-devel] Release and Windows Gem

2008-02-07 Thread Charlie Savage
And talking about releases, has a windows build system been setup yet? If not, I'm happy to volunteer and create a windows gem. Charlie Dan Janowski wrote: Glad you found that. I would like to make a release, but I am not sure of the finality of the build system changes. As soon as there is a

Re: [libxml-devel] Crash in ruby_xml_xpath_object_mark

2008-02-07 Thread Dan Janowski
Glad you found that. I would like to make a release, but I am not sure of the finality of the build system changes. As soon as there is a read on that, I will bundle and release. Dan On Feb 7, 2008, at 12:48, Luc Heinrich wrote: > On 7 févr. 08, at 17:27, Luc Heinrich wrote: > >> So it seems

Re: [libxml-devel] Code snippet to crash libxml (segfault)

2008-02-07 Thread Dan Janowski
C sure is lovely, isn't it. I will try to reproduce this. If you would not mind, please post this as a bug so I can track it at: http://rubyforge.org/tracker/?atid=1971&group_id=494&func=browse Dan On Feb 7, 2008, at 03:36, Saurabh Nanda wrote: >> Are you sure it is the doc reference and not

[libxml-devel] Does libxml-ruby install on windows or not

2008-02-07 Thread Sun Tea
Hi, I am having the same issues trying to install the libxml-ruby gem c:/ruby/bin/ruby.exe extconf.rb install libxml-ruby checking for socket() in socket.lib... no checking for gethostbyname() in nsl.lib... no checking for atan() in m.lib... yes checking for inflate() in z.lib... no checking for

Re: [libxml-devel] Crash in ruby_xml_xpath_object_mark

2008-02-07 Thread Luc Heinrich
On 7 févr. 08, at 17:27, Luc Heinrich wrote: > So it seems that simply checking for a NULL 'nodesetval' in > 'ruby_xml_xpath_object_mark' actually makes sense after all. Er, nevermind, just ignore me. It seems that this is already fixed in the subversion repository. Sorry for the noise. Any pla

Re: [libxml-devel] Crash in ruby_xml_xpath_object_mark

2008-02-07 Thread Luc Heinrich
On 7 févr. 08, at 13:17, Luc Heinrich wrote: > I don't know the libxml-ruby code well enough yet to propose a clean > fix (I'm not sure that simply testing for NULL on nodesetval would be > enough or appropriate), so if anyone could chime in... :) After a bit of debugging, it seems that having th

[libxml-devel] Crash in ruby_xml_xpath_object_mark

2008-02-07 Thread Luc Heinrich
Greetings, I have recently switched from REXML to libxml-ruby (0.5.2) because I really needed a major speed boost for a web application and libxml- ruby really delivered. Some parts of the code which were taking more than 30 seconds are now running under 0.2 seconds, I'll take a 150x speed

Re: [libxml-devel] Code snippet to crash libxml (segfault)

2008-02-07 Thread Saurabh Nanda
> Are you sure it is the doc reference and not the GC finalizers > running? Put a print statement after it and see what happens. Probably what you're saying is right. I don't fully understand what effect the GC finalizers have, but I've tried the following script (same script without the doc refer