[Webkit-unassigned] [Bug 79795] Unreleased gst_object_reference to audio sink in MediaPlayerPrivateGStreamer

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=79795





--- Comment #8 from Philippe Normand <pnormand at igalia.com>  2012-02-29 01:11:36 PST ---
(In reply to comment #7)
> (In reply to comment #4)
> > (From update of attachment 129267 [details] [details])
> > Attachment 129267 [details] [details] did not pass chromium-ews (chromium-xvfb):
> > Output: http://queues.webkit.org/results/11662224
> > 
> > New failing tests:
> > css3/filters/effect-invert-hw.html
> 
> I don't see the relationship between the patch and this failing test: is Chrome even using gstreamer ? What am I supposed to do ?

You can ignore this warning. Chromium indeed doesn't use the gstreamer backend.

(In reply to comment #6)
> (In reply to comment #5)
> > Looks like a good candidate for a RefPtr!
> 
> Not really, I think, as this is purely related to the specific gstreamer refcounting (it could be if a specific type of RfPtr was added for gst_objects, to explicitly call gst_object_unref upon object destruction).
> What happens here is that the audio-sink is neither "disposed" nor "finalized" (in the gstreamer terminology), because its gstreamer refcount stays > 0.

Well Martin is right indeed. We have implemented exactly what you describe in GRefPtrGStreamer.cpp. So if you can update the patch it would be awesome :)

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.