After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 573889 - [pitivi] Discoverer: Race in filesink causes some thumbnails to not be generated
[pitivi] Discoverer: Race in filesink causes some thumbnails to not be generated
Status: RESOLVED OBSOLETE
Product: GStreamer
Classification: Platform
Component: gst-plugins-base
unspecified
Other Linux
: Normal normal
: NONE
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2009-03-03 09:43 UTC by Edward Hervey
Modified: 2012-10-16 20:16 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Edward Hervey 2009-03-03 09:43:47 UTC
There's a race in the Discoverer causing some thumbnail files to not be generated.

The problem is the following:
* We detect a video stream, we add pngenc+filesink to the file sink
* We wait for the pipeline to change state to PAUSED (i.e. prerolled)
 * This happens when the encoded buffer arrives in the filesink BUT is not processed (it's waiting)
* In the bus message handler we set the state to PLAYING
 => RACE between the moment the prerolled buffer is handled (passed on to filesink_render) and the EOS coming in from pngenc in snapshot mode.
Comment 1 Jean-François Fortin Tam 2012-10-14 16:51:49 UTC
Wondering if that's still the case nowadays. Well, we do have all sorts of fun races like bug #681015, still.
Comment 2 Tim-Philipp Müller 2012-10-15 18:41:44 UTC
I'm not sure this analysis ever made sense. It would've made sense if pngenc had ever posted an EOS message directly (instead of sending an eos event downstream and/or returning FLOW_UNEXPECTED to make upstream push an eos event downstream), but I can't find any trace of that going back as far as 2006.

I would suggest we close this as OBSOLETE and handle specific bugs (e.g. with debug logs) in other reports.

Also, is it possible this was a different discoverer than the one in pbutils, namely the gst-python discoverer?
Comment 3 Tim-Philipp Müller 2012-10-16 20:16:24 UTC
Closing as per last comment.