[Bug 720435] rtpbasepayload: Prefer downstream SSRC after collision if possible

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=720435
  GStreamer | gst-plugins-good | unspecified

George Kiagiadakis <kiagiadakis.george> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #13 from George Kiagiadakis <kiagiadakis.george at gmail.com> 2013-12-31 15:23:21 UTC ---
(In reply to comment #10)
> I think there are other conditions where you don't want to timeout an internal
> source completely, like if it's the only one left (even if it doesn't match the
> suggested ssrc).

The last source will always match the suggested ssrc, because there is a check
in add_source() that changes the suggested_ssrc to be the same as the ssrc of
the last added source. Together with rtpbasepayload using the suggested ssrc
after a collision, one can assert that the last source left in the session
always uses the suggested ssrc.

> Also, I just though of this... I'm not not sure we want to timeout the
> internal-ssrc from the RTX session even if it is idle for a longish time. Maybe
> internal ssrcs should just BYE having having received a new GstRTPSsrcEOS
> "partial-EOS" event?

I also thought of this. It doesn't necessarily harm to send BYE and
re-establish a connection later, but it sure doesn't look very nice.

I think that this is all unrelated to this bug report, though. This one is
about rtpbasepayload using the suggested ssrc after a collision, which afaict
is fixed now. I will put the rest of the patches in bug 720440, together with a
unit test.

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