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 627342 - [tests] gnloperation:test_complex_operations timeout
[tests] gnloperation:test_complex_operations timeout
Status: RESOLVED INCOMPLETE
Product: GStreamer
Classification: Platform
Component: gnonlin
git master
Other Linux
: Normal major
: git master
Assigned To: GStreamer Maintainers
Edward Hervey
Depends on:
Blocks:
 
 
Reported: 2010-08-19 08:32 UTC by Edward Hervey
Modified: 2014-09-01 09:37 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Edward Hervey 2010-08-19 08:32:13 UTC
make gnl/gnloperation.torture

gnl/common.h:117:E:gnloperation:test_complex_operations:0: (after this point) Test timeout expired

The problem is due to a race in videomixer with:
* one incoming stream EOS'd (i.e. it's not pushing any data) : stream1
* one incoming stream starting to push data (because it was just activated/linked) : stream2
* sending a seek at the same time

The summary is that:
* the newsegment from stream2 arrives *after* the videomixer seek handler has removed the flushing state of collectpads and *before* the seek event was sent on that stream
* videomixer pushes the seek to stream1 (the one that was EOS'd)
 * flush start comes in that other pad
 * flush stop comes in that other pad
 * collectpad thinks the two streams have been properly setup and forwards FLUSH_STOP downstream
 * data from that stream1 will arrive and block in collectpads (waiting for data from the other stream)
* videomixer pushes the seek to stream2
 * flush start comes in that other pad
   * collectpad thinks it's a new flush and forwards it
   * stream1 stops itself ... and never restarts
Comment 1 Edward Hervey 2013-07-17 09:21:03 UTC
Still happens with git master as of today.
Comment 2 Tim-Philipp Müller 2014-03-06 18:00:12 UTC
Edward, do you still get this? I can't reproduce this any more, with .forever
Comment 3 Edward Hervey 2014-09-01 09:37:01 UTC
Can't manage to trigger it either. Going to close this for the time being.