GNOME Bugzilla – Bug 611721
Assertion fails in pygobject_new_full()
Last modified: 2012-04-25 05:47:57 UTC
Actually this is not a new bug, it already has been posted as bug 580022 in Rhythmbox. I'm posting it here again because it looks like a PyGTK or PyGObject issue. It's about this assertion failure: >>> snip ERROR:/build/buildd/pygobject-2.18.0/gobject/pygobject.c:924:pygobject_new_full: assertion failed: (tp != NULL) Aborted <<< snip For steps to reproduce and stack trace information, please have a look in 580022. The strange thing about this bug is that it occurs only on some systems. For instance users reported this bug on Ubuntu Karmic while others don't see it in Karmic. However, recently I managed it to reliably reproduce this bug using VMs. The following is a copy of comment 20 in bug 580022: >>> snip Some news on this. After a user reported this issue when using the Remuco plugin in Ubuntu Studio 9.10 I tried again to reproduce this in virtual machines using VirtualBox. Indeed I could reproduce this bug in "Ubuntu Studio 9.10" using this ISO image: http://cdimage.ubuntu.com/ubuntustudio/releases/9.10/release/ubuntustudio-9.10-alternate-i386.iso. I did the test right after a fresh installation as well as after getting all updates (as of 17 Feb 2010). OTH, the same test did not reproduce the bug in a regular "Ubuntu 9.10" (http://cdimage.ubuntu.com/releases/9.10/release/ubuntu-9.10-dvd-i386.iso), so here everything works fine. <<< snip Possible related bugs: bug 555384, bug 538401 Apparently this bug is not directly related to a specific version of pygobject (it has been seen from 2.16 to 2.20). What other parameters could cause this issue? Or am I completely wrong and this isn't a pygobject problem at all?
We had a similar issue with the pygi hooks we put into pygobject a while ago. Those hooks could be disabled, so some distributions did not ship them. Have you experienced those issues with more recent releases?
Thanks for some feedback. I did not checked it for a while now. Personally I did not experience the bug (except in the VMs), but some Rhythmbox users did. There were no related bug reports in the last time, so maybe the problem does not occur in recent distributions. If I'm wrong, I'll check this again. BTW .. is there a way to check if these hooks are en/disabled in a system?
This report and the affected versions are quite old, and this does not have a reproducer. Is this still an issue? Did you get any duplicates with newer versions?
Now news here. I did not receive any reports of this issue in the last year, so I guess it is okay to close this issue.
Thanks for the feedback. Closing.