[Bug 668471] [h264parse] Negotiation regression for autoplugging

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=668471
  GStreamer | gst-plugins-bad | 0.11.x

Mark Nauwelaerts <mnauw> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mnauw at users.sourceforge.net

--- Comment #1 from Mark Nauwelaerts <mnauw at users.sourceforge.net> 2012-01-23 10:31:22 UTC ---
That sounds well, but what real meaning has h264parse's selection in such case
?

It will then happen to pick whatever caps happened to end up first in those
filter caps, which may then very well come down to non-passthrough, although
passthrough would be possible, and that happening is what prompted the above
commit.

Also, if the filter caps are set to all *possible* caps and h264parse finds
that passthrough is among these and selects that, then it should not go wrong
afterwards since they were included in possible caps (so there should have been
a decoder capable of handling that selected passthrough case).

Problem might be due to decodebin2 (afaics) always adding the (non-fixed)
parser src pad caps to the filter_caps which renders mute the rest of the
*possible* caps as well as any sort of nifty negotiation a parser might be
trying to do (like avoiding useless conversion if possible).

-- 
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.