After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 114406 - upstream events are not forwarded by default
upstream events are not forwarded by default
Status: RESOLVED FIXED
Product: GStreamer
Classification: Platform
Component: gstreamer (core)
git master
Other other
: Normal normal
: 0.6.x
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2003-06-04 17:18 UTC by Joshua N Pritikin
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
forward upstream events by default (476 bytes, patch)
2003-06-04 17:19 UTC, Joshua N Pritikin
none Details | Review
What about that one for tee (1.35 KB, patch)
2003-06-04 17:56 UTC, Julien MOUTTE
none Details | Review
The first one was sending the event to emitting pad which is wrong (1.38 KB, patch)
2003-06-04 18:04 UTC, Julien MOUTTE
none Details | Review

Description Joshua N Pritikin 2003-06-04 17:18:26 UTC
Here's one possible fix.  Dunno if this is "correct" but it works for me.
Comment 1 Joshua N Pritikin 2003-06-04 17:19:22 UTC
Created attachment 17139 [details] [review]
forward upstream events by default
Comment 2 Joshua N Pritikin 2003-06-04 17:38:52 UTC
Why is this needed?  For example, if you insert a tee or identity
element in your pipeline then you want upstream events forwarded
through these elements.
Comment 3 Julien MOUTTE 2003-06-04 17:56:56 UTC
Created attachment 17143 [details] [review]
What about that one for tee
Comment 4 Julien MOUTTE 2003-06-04 18:04:56 UTC
Created attachment 17144 [details] [review]
The first one was sending the event to emitting pad which is wrong
Comment 5 Julien MOUTTE 2003-06-04 18:05:47 UTC
Maybe that's totally stupid but my idea was that event should go
downstream in the other branches of tee ..

Does that make sense ?
Comment 6 Joshua N Pritikin 2003-06-04 18:09:35 UTC
There are two patches here.  i don't know about extra forwarding in
tee, but the default upstream event behavior needs to change.
Comment 7 Joshua N Pritikin 2003-06-04 18:41:02 UTC
Also beware of reference count problems.  i really don't understand
when to incr ref counts.
Comment 8 Benjamin Otte (Company) 2003-06-07 13:13:05 UTC
This is not the right fix.

If a pad does not have an event handler, it's a scheduler bug. (We
should probably emit a g_warning there.) Schedulers set the event
handler (which just calls the event function normally). And the event
function is set in gst_real_pad_init, so there should always be one
present.

And Julien, you may not change the direction of events so that event
handler is wrong.
Comment 9 David Schleef 2003-09-18 22:36:41 UTC
I changed it to a g_warning().  The patches in this bug are incorrect.