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 589779 - GStreamer-CRITICAL **: gst_segment_set_seek: assertion `start <= stop' failed
GStreamer-CRITICAL **: gst_segment_set_seek: assertion `start <= stop' failed
Status: RESOLVED OBSOLETE
Product: GStreamer
Classification: Platform
Component: gnonlin
git master
Other Linux
: Normal normal
: NONE
Assigned To: GStreamer Maintainers
Edward Hervey
Depends on:
Blocks:
 
 
Reported: 2009-07-26 16:54 UTC by Jean-François Fortin Tam
Modified: 2016-02-21 22:25 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
GST_DEBUG=3,gnl*:5,python:5 PITIVI_DEBUG=5 (264.27 KB, application/x-bzip)
2009-08-03 22:56 UTC, Jean-François Fortin Tam
Details
GST_DEBUG=3,gnl*:5,python:5 PITIVI_DEBUG=5 while pulseaudio is uninstalled and not running (293.49 KB, application/x-bzip)
2009-08-05 02:58 UTC, Jean-François Fortin Tam
Details
screenshot (12.28 KB, image/png)
2009-08-05 03:00 UTC, Jean-François Fortin Tam
Details

Description Jean-François Fortin Tam 2009-07-26 16:54:24 UTC
I keep getting this when having clip A and B on 2 different layers, with clip B overlapping just a little, when the playhead arrives at the place where it needs to switch from A to B.

Looks like this, with "v" being the playhead that stops/freezes at that position:

                    v
[----- clip A ----- ]
          [-------- clip B -------]
Comment 1 Edward Hervey 2009-08-03 13:58:17 UTC
Jeff, can't reproduce it anymore with git gnonlin/pitivi, can you confirm ?
Comment 2 Jean-François Fortin Tam 2009-08-03 22:52:36 UTC
Yes, I still see this (simply by playing a bit with 2 theora files -- not files created by recordmydesktop, for the record). What flavor of debug log would you like? :)
Comment 3 Jean-François Fortin Tam 2009-08-03 22:56:27 UTC
Created attachment 139830 [details]
GST_DEBUG=3,gnl*:5,python:5 PITIVI_DEBUG=5

Actually, I don't see that particular error (the one in the bug title) in the pitivi terminal, however I see the same kind of symptoms (either the playback stops between the two clips, or it lags when playing).
Comment 4 Edward Hervey 2009-08-04 07:54:21 UTC
Could you try again with pulseaudio disabled please :)

I can see in the logs than pulsesink is bitching like mad.
Comment 5 Jean-François Fortin Tam 2009-08-05 02:56:39 UTC
Ok, still see this.

jeff@kusanagi:~/trunks/pitivi/pitivi$ export GST_PLUGIN_PATH="../../gnonlin/"jeff@kusanagi:~/trunks/pitivi/pitivi$ bin/pitivi 
AttributeError: 'module' object has no attribute 'Element'
Traceback (most recent call last):
  • File "pitivi/elements/singledecodebin.py", line 314 in _removeUnusedElements
    self._removeUnusedElements(peer)
  • File "pitivi/elements/singledecodebin.py", line 310 in _removeUnusedElements
    for pad in element.src_pads():
TypeError: Resync

(pitivi:9231): GStreamer-CRITICAL **: gst_event_new_qos: assertion `diff >= 0 || -diff <= timestamp' failed


# Then, I tried with suspending pulseaudio:

jeff@kusanagi:~/trunks/pitivi/pitivi$ pasuspender -- bin/pitivi 
AttributeError: 'module' object has no attribute 'Element'
# Same problem, but no errors


# Then, I removed the pulseaudio package and killed the process with gnome-system-monitor, and did:

jeff@kusanagi:~/trunks/pitivi/pitivi$ bin/pitivi 
AttributeError: 'module' object has no attribute 'Element'
0:00:06.609280590  9604 0xb193a4d0 ERROR         gnlcomposition gnlcomposition.c:1640:no_more_pads_object_cb:<gnlcomposition1> Sending seek event failed!

(pitivi:9604): GStreamer-CRITICAL **: gst_event_new_qos: assertion `diff >= 0 || -diff <= timestamp' failed


So it seems to happen no matter what I do.
Comment 6 Jean-François Fortin Tam 2009-08-05 02:58:53 UTC
Created attachment 139918 [details]
GST_DEBUG=3,gnl*:5,python:5 PITIVI_DEBUG=5 while pulseaudio is uninstalled and not running
Comment 7 Jean-François Fortin Tam 2009-08-05 03:00:54 UTC
Created attachment 139919 [details]
screenshot

This is how I actually set my clips to overlap. It allows me to reproduce the problem very easily.
Comment 8 Edward Hervey 2009-09-04 09:32:51 UTC
The QoS issue is gone.

But the seek bug is in fact in GNonLin, moving the bug there.
Comment 9 Tim-Philipp Müller 2011-12-09 16:56:03 UTC
This one has been a "blocker" for more than 2 years now :)
Comment 10 Jean-François Fortin Tam 2011-12-09 21:02:49 UTC
Le sigh, I do still see this failed assertion regularly in pitivi's terminal, so frequently in fact that I've become accustomed to it and don't know anymore if it's supposed to be considered a critical failure or insignificant...
Comment 11 Nicolas Dufresne (ndufresne) 2013-05-26 14:03:45 UTC
Jeff, is Pitivi master in good enough condition to retest this today?
Comment 12 Mathieu Duponchelle 2013-05-28 20:47:55 UTC
I confirm that this bug still exists with the audio_mixing GES branch. It doesn't seem to prevent playback and the playhead doesn't freeze, I would suggest removing the BLOCKER tag.
Comment 13 Nicolas Dufresne (ndufresne) 2014-01-13 16:02:51 UTC
Can this still be reproduced ?
Comment 14 Bassam Kurdali 2014-01-15 23:06:59 UTC
getting occasionally this same error using pitivi master during an edit, (pitivi:4720): GStreamer-CRITICAL **: gst_segment_do_seek: assertion 'start <= stop' failed
however it does not seem to impact editing.
Comment 15 Jean-François Fortin Tam 2014-01-16 00:31:18 UTC
Yeah, still happens, though I'm not sure how to trigger/reproduce it (Bassam, any hints on that front?)
Comment 16 Tim-Philipp Müller 2016-02-21 22:25:43 UTC
gnonlin is dead, long live ..