[Bug 666174] [ffdec] segfault because gst_pad_alloc_buffer_and_set_caps return a buffer with a size of 0

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=666174
  GStreamer | gst-ffmpeg | git

--- Comment #7 from Mark Nauwelaerts <mnauw at users.sourceforge.net> 2011-12-19 11:20:30 UTC ---
Created an attachment (id=203853)
 View: https://bugzilla.gnome.org/attachment.cgi?id=203853
 Review: https://bugzilla.gnome.org/review?bug=666174&attachment=203853

basetransform: suggestion compatible with upstream is not much of a suggestion

Problem with current suggest mechanics in basetransform seems to be that it
bypasses the core test in a silly way.  That is, where core tests for caps ==
newcaps as a check for renegotiation, it is well possible to have newcaps
equivalent/compatible (but obviously different pointer) and as such to slip by.

Attached patch modifies some suggestion mechanics in basetransform to ignore
the suggestion if suggested caps are compatible with (i.e. can intersect) with
upstream requested caps (rather than having to be exactly equal).  Not sure
about original semantics, but seems to make sense that the suggestion
could/should hardly coax upstream into doing something "new" if it is already
doing stuff that is compatible with the suggestion.

So, for a typical fixed src pad scenario that would then mean the suggested
stuff is either compatible and essentially ignored (thereby fixing this), or it
is not compatible and would run into some failure to negotiate (arranged for by
_alloc_and_set_caps).

-- 
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.