[Bug 606382] decodebin: Upstream events (seek,qos) get lost when switching groups/chains

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=606382

--- Comment #51 from Duncan Palmer <dpalmer at digisoft.tv> ---
Created attachment 315338
  --> https://bugzilla.gnome.org/attachment.cgi?id=315338&action=edit
logs showing hang on seek

I'm currently seeing a problem where gst_element_seek_simple(decodebin) is not
returning. 

log.2 shows that hlsdemux receives the seek event fine, but that flush-start is
not propogated to any groups within the decodebin. This is because hlsdemux has
just switched bitrates, and the src pad probe has yet to be put in place by
hlsdemux. 

A quick hack to workaround this is for hlsdemux to ensure the streaming thread
is stopped before sending flush-start downstream.

With the quick hack in place, I can the produce the scenario in log.3. Search
for WARNING and CRITICAL. In this case, flush-start is making it's way into a
group which is in the process of being created, and causing grief.

I'm not sure why either scenario causes a hang, but in any case it's obviously
wrong.

Any suggestions?

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