[Bug 746010] [oggdemux] doesn't go into pull mode even when queue2 ring-buffer is enabled

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=746010

--- Comment #6 from Thibault Saunier <tsaunier at gnome.org> ---
(In reply to Young Han Lee from comment #4)
> (In reply to Thibault Saunier from comment #3)
> > 
> > How does it restart?
> 
> 
> 0:00:10.551185004 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:00.006177622
> 0:00:10.553194798 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:00.004167656
> 0:00:10.555073647 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:00.002283863
> 0:00:10.556928716 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:00.000428234
> 0:00:10.566048749 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:08.433333328
> 0:00:10.587855374 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:08.233333330
> 0:00:10.615290166 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:07.999999999
> 0:00:10.639490290 15740       0x9cc000 DEBUG               basesink
> gstbasesink.c:4613:gst_base_sink_get_position:<fakesink1> res: 1, POSITION:
> 0:00:07.766666663
> 
> Here are test logs. You can see the position value went to 0:00:08.433333328
> once the reverse-playback was done.
> 
> Note that, even in this case, reverse-playback test is passed with your
> latest patch(http://phabricator.freedesktop.org/differential/diff/52/).
> Because there are no 3 contiguous wrong positions. But, I think it would be
> also good to merge your patch first and write another make-up patch to test
> this if this is a problem.

Right, the impression I have about that is that the position reporting is wrong
rather than new ((wrong) buffers are pushed, which is also a bug. I did not
investigate it in details so I am not sure about that either.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.