Bug#485738: reproduction
Lucas Nussbaum wrote: A guy I know replied. It is up to LibXml which is doing wrong. As the error message says, object allocation by LibXml should not be performed during GC. As of ruby 1.8.7, it exactly checks this issue and will report the bug message if it finds. Apparently the cairo bindings have the same problem I hope it's limited to cairo + libxml! OK. Annoying in some ways, but nice to have narrowed the source down somewhat. I'll see if I can get in touch with the rcairo developer. My b.d.o foo is weak, otherwise I'd clone the bug and reassign it to the right packages. Sorry! -- James Healy jimmy-at-deefa-dot-com Thu, 19 Jun 2008 15:56:44 +1000 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#485738: reproduction
Lucas Nussbaum wrote: A guy I know replied. It is up to LibXml which is doing wrong. As the error message says, object allocation by LibXml should not be performed during GC. As of ruby 1.8.7, it exactly checks this issue and will report the bug message if it finds. Apparently the cairo bindings have the same problem I hope it's limited to cairo + libxml! For anyone doing further research, here's the bug on rubyforge that prompted the addition of the check in 1.8.7: http://rubyforge.org/tracker/index.php?func=detailaid=11859group_id=426atid=1698 -- James Healy jimmy-at-deefa-dot-com Thu, 19 Jun 2008 17:28:00 +1000 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#485738: reproduction
On 11/06/08 at 15:23 +0200, Blaž Lorger wrote: Correction, this is simplest script that reproduces bug: #!/usr/bin/env ruby require 'xml/libxml' i = 0; while true XML::Document.new; i += 1; puts * #{i} * if i%100 == 0; end Thanks for the testcase. I reproduced the problem with 1.8.7.17. Arg. I also rebuilt libxml-ruby to make sure it wasn't caused by a subtle ABI change, but the problem is still there with a freshly build libxml-ruby. Akira-san, Daigo-san, could one of you raise the issue on ruby-dev? Or should I raise it on ruby-core? -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#485738: reproduction
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Lucas Nussbaum wrote: Akira-san, Daigo-san, could one of you raise the issue on ruby-dev? Or should I raise it on ruby-core? I have just reported this issue to ruby-dev and wait for a reply. Daigo - -- Daigo Moriwaki daigo at debian dot org -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFIWQ2GNcPj+ukc0lARAgw6AJ9sngCSyTxdrAz8rD4XjpmDb14S9ACfdjVL 4L4aJHdybhNO1+28YkZVU/8= =7uNt -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#485738: reproduction
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Daigo Moriwaki wrote: Lucas Nussbaum wrote: Akira-san, Daigo-san, could one of you raise the issue on ruby-dev? Or should I raise it on ruby-core? I have just reported this issue to ruby-dev and wait for a reply. A guy I know replied. It is up to LibXml which is doing wrong. As the error message says, object allocation by LibXml should not be performed during GC. As of ruby 1.8.7, it exactly checks this issue and will report the bug message if it finds. Daigo - -- Daigo Moriwaki daigo at debian dot org -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFIWRqbNcPj+ukc0lARArGZAJ907oA2JHvpzeUwRQVNTa3k55B+QQCgwU/1 2fh/DpwvUNJJ5e0tglrfqaU= =MXZd -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#485738: reproduction
On 18/06/08 at 23:24 +0900, Daigo Moriwaki wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Daigo Moriwaki wrote: Lucas Nussbaum wrote: Akira-san, Daigo-san, could one of you raise the issue on ruby-dev? Or should I raise it on ruby-core? I have just reported this issue to ruby-dev and wait for a reply. A guy I know replied. It is up to LibXml which is doing wrong. As the error message says, object allocation by LibXml should not be performed during GC. As of ruby 1.8.7, it exactly checks this issue and will report the bug message if it finds. Apparently the cairo bindings have the same problem I hope it's limited to cairo + libxml! -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#485738: reproduction
Apparently it is sufficient to upgrade packages libruby1.8, ruby1.8 and ruby1.8-dev to encounter the bug. It is possible that bug is actually caused only when you use libxml-ruby1.8. Maybe it is worth mentioning that original script used libxml, but didn't fail while using libxml (line in script). Here is script that reproduces the bug: #!/usr/bin/env ruby require 'xml/libxml' i = 0; while true doc = XML::Document.new; doc.root = XML::Node.new('gc-bug'); i += 1; puts * #{i} * if i%100 == 0; end And output on my computer: $ ./gcbug.rb * 100 * * 200 * * 300 * * 400 * * 500 * * 600 * * 700 * * 800 * * 900 * * 1000 * * 1100 * * 1200 * ./gcbug.rb:8: [BUG] object allocation during garbage collection phase ruby 1.8.7 (2008-05-31 patchlevel 0) [x86_64-linux] Aborted gcbug.rb Description: application/ruby
Bug#485738: reproduction
Correction, this is simplest script that reproduces bug: #!/usr/bin/env ruby require 'xml/libxml' i = 0; while true XML::Document.new; i += 1; puts * #{i} * if i%100 == 0; end -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]