GNOME Bugzilla – Bug 473097
[baseaudiosink] gstreamer-properties hangs when testing sound capture
Last modified: 2007-10-28 13:17:17 UTC
The bug has been opened on https://bugs.launchpad.net/bugs/69592 "A crash happens while testing the ich6 intel- mic adc in system-preference-audio-device-sound caputure. ... gaves this message Failed to construct test pipeline for 'gconfaudiosrc ! audioconvert ! audioresample ! gconfaudiosink profile = chat! (this text was hand copied...not obtained with copy and paste) then i pressed close in the pop up window....nothing happened....finally the upper right cross of the aplication window....and obtained finally a forced quit of the aplication. ... In my Acer TravelMate 4020 with Feisty the following steps: 1- System 2- Preference 3- Sound 4- Sound Capture 5- Intel ICH6 - MIC2 ADC 6- Test reproduce the bug There are other ways...."
Can you please try to use gdb to find out where it hangs (I believe we're talking about a hang here, not a crash, right?)? I don't have the hardware, I can't reproduce, and the code looks ok from a cursory glance.
At https://bugs.launchpad.net/bugs/69592 gdb and valgrind reports where uploaded. Let me know if there is the information you want (where it hangs...). If I should generate something different please let me know with great detail: what packgages should i install, what report should i generate. that was my first time i used those tools.
Just to make sure there are no misunderstandings: Are we talking about a crash or a hang? It's not obvious from the launchpad bug, either. The stack you posted seems to be taken after the application quit. To get a useful one (if we're talking about a frozen app), you need to a) install debugging symbols (for at least control-center, gstreamer, gtk, glib, possibly ALSA) b) gdb gnome-sound-properties c) "run" d) do whatever you need to do to cause the hang e) get back to gdb and press Ctrl-C f) "thread apply all bt"
Created attachment 96767 [details] gdb of the hang of gnome-sound-properties with debugging symbols
Comment on attachment 96767 [details] gdb of the hang of gnome-sound-properties with debugging symbols debugging symobls installed :alsa-utils-dbgsym (1.0.13-1ubuntu5) libesd-alsa0-dbgsym (0.2.36-3ubuntu4) libpt-plugins-alsa-dbgsym (1.10.3-0ubuntu1) libsdl1.2debian-alsa-dbgsym (1.2.11-7ubuntu1) libdbus-glib-1-2-dbgsym (0.73-1) libglib1.2-dbgsym (1.2.10-17build1) libglib2.0-0-dbgsym (2.12.11-0ubuntu1) libglib2.0-cil-dbgsym (2.10.0-0ubuntu4) gtk2-engines-dbgsym (1:2.10.1-0ubuntu1) gtk2-engines-pixbuf-dbgsym (2.10.11-0ubuntu3) gtk2-engines-ubuntulooks-dbgsym (0.9.12-4) gtkhtml3.14-dbgsym (3.14.1-0ubuntu2) libgtk2.0-0-dbgsym (2.10.11-0ubuntu3) libgtk2.0-bin-dbgsym (2.10.11-0ubuntu3) libgtk2.0-cil-dbgsym (2.10.0-0ubuntu4) gstreamer0.10-alsa-dbgsym (0.10.12-0ubuntu1) gstreamer0.10-esd-dbgsym (0.10.5-1ubuntu2) gstreamer0.10-ffmpeg-dbgsym (0.10.2-0ubuntu4) gstreamer0.10-gnomevfs-dbgsym (0.10.12-0ubuntu1) gstreamer0.10-tools-dbgsym (0.10.12-0ubuntu2) gstreamer0.10-x-dbgsym (0.10.12-0ubuntu1) libgstreamer-plugins-base0.10-0-dbgsym (0.10.12-0ubuntu1) gnome-control-center-dbgsym (1:2.18.1-0ubuntu2.1) libgstreamer0.10-0-dbgsym (0.10.12-0ubuntu2)
That looks like a problem in GStreamer or ALSA to me. Moving to gstreamer, maybe they can tell more.
I think this is caused by the audiosink pull activation mode that shipped in core/base 0.10.14. It's been disabled again in CVS: 2007-09-13 Wim Taymans <wim.taymans@gmail.com> * gst-libs/gst/audio/gstbaseaudiosink.c: (gst_base_audio_sink_init): Disable pull mode scheduling, we're not ready for it yet and it subtly breaks a lot of things. This is the patch/fix: http://webcvs.freedesktop.org/gstreamer/gst-plugins-base/gst-libs/gst/audio/gstbaseaudiosink.c?r1=1.87&r2=1.88 Distros might want to ship that patch. FWIW, the problem is unlikely to appear in any 'real' pipelines, but it causes problems with very short and simple pipelines along the lines of src ! audioconvert ! sink.