When using pulseaudio, pidgin crashes randomly with error message "Assertion '!in_worker(m)' failed at pulse/thread-mainloop.c:162, function pa_threaded_mainloop_stop(). Aborting." Pulseaudio Bug Report: http://pulseaudio.org/ticket/838 Gstreamer Bug Report: https://bugzilla.gnome.org/show_bug.cgi?id=618874 fixed in gst-plugins-pulse gst-plugins-good/0.10.23 (http://gstreamer.freedesktop.org/releases/gst-plugins-good/0.10.23.html) I'm requesting a version bump to .23 or .24 to solve this problem. Reproducible: Always Steps to Reproduce: 1. start pidgin 2. wait some time (~1 hour) 3. pidgin crashes, console output: "Assertion '!in_worker(m)' failed at pulse/thread-mainloop.c:162, function pa_threaded_mainloop_stop(). Aborting."
Do you happen to know if gst-plugins-pulse-0.10.21 triggers these crashes as well?
could be, because not such a long time ago, pidgin worked without crashes... i'll revert gst-plugins-pulse to v21 and try to reproduce... or not reproduce ;-)
The reason I ask is to be able to make an informed decision on whether to stabilize gst-plugins-pulse-0.10.22 together with all the other very-soon-to-be-stabilized-0.10.22's, or skip it for just that one if it's a regression in 0.10.22.
I'll keep pidgin open some time to test it with gst-plugins-pulse-0.10.21. Seems to me as if the bug is introduced in version 0.10.22. What about providing gst-plugins-pulse with a patch fixing this problem? (something like http://meego.gitorious.org/maemo-multimedia/gst-plugins-good/commit/a8103facc59b9dde07ee95d026c3fd2323e278bb.patch , it's the commit fixing the bug and included in 0.10.23)
I thought of a revbump after posting my comment and peaked at the changes a bit. I'm concerned with there being many multi-threading fixes inbetween 0.10.22 and 0.10.23 and not knowing how they affect eachother. For example there's commits 5332287e2d and 4a492f2c8a
I'm no expert in gstreamer and don't know what maybe half-way-commited multithreading changes will do. Better mask gst-plugins-pulse-0.10.22 (so that 0.10.21 will be used) or revbump to 0.10.23 (or 0.10.24?). But when bumping to 0.10.23, all gst-plugins-* ebuild must be updated, not only gst-plugins-pulse, right?
Right. Working on that, ETA is tomorrow/Friday late night...
ok. thanks :-)
Bumped to 0.10.23 now in ~arch finally. Looks like x86 and amd64 decided to stabilize 0.10.22, I gave them the option to decide per this bug here in the stabilization bug.
Florian, could you test pidgin with media-plugins/gst-plugins-pulse-0.10.22-r1? If it works for you we'll be able to fast stabilize it and finally fix pidgin in stable tree. Thank you.
I'm running pidgin now with media-plugins/gst-plugins-pulse-0.10.22-r1. Will report crashes when they may occur, but I hope they wont!
No crashes so far with media-plugins/gst-plugins-pulse-0.10.22-r1. I think you can stabilize it :-)