[Bug 704507] gstglimagesink: Replace pointer properties with signals

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=704507
  GStreamer | gst-plugins-gl | unspecified

Matthew Waters <ystreet00> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ystreet00 at gmail.com

--- Comment #9 from Matthew Waters <ystreet00 at gmail.com> 2013-07-22 02:58:46 UTC ---
(In reply to comment #7)
> Created an attachment (id=249656)
 View: https://bugzilla.gnome.org/attachment.cgi?id=249656
 Review: https://bugzilla.gnome.org/review?bug=704507&attachment=249656

> Patch to update the recordgraphic example.
> 
> I'll do all this but I've got a problem:
> 
> To remove the client-draw-callback as well in gstglfilterapp, I think it would
> make sense to refactor a little the way the callback is handled.
> 
> Currently it's passed along with 15 other arguments to gst_gldisplay_use_fbo,
> which in turn passes it to gst_glframebuffer_use, and I'm not sure this is
> optimal. On that codepath ther's also an intermediary step which I don't
> understand the purpose of, namely stuff all these arguments in a structure just
> to unpack them and pass them as arguments in another function, I don't see how
> that makes sense. (Actually I think it doesn't :)

The reason it is stuffed into a structure is to perform the
gst_gl_framebuffer_use function in the GL thread which may be different from
the calling thread.  Ideally, that function should be split up into set matrix,
set texture attachements, etc and be moved into gstglframebuffer.c.  Eventually
gstglutils.c should be incinerated, it was and still is a temporary solution
for speeding along the process of using GstContext.

> So yeah I kind of don't want to go and messup with all this too much before
> someone else with more knowledge of the code base comes and does a little
> cleanup, and when done I'd like to rebase my history a little, no point in
> having two different commits for gstglfilterapp IMHO. I think I'll rename the
> commits etc.. during that rebase, except if we decide to switch to using a
> signal for now.

That might take a while :)

> the attachment I added allows me to run the only example using this
> client-draw-callback.

-- 
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.