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 114976 - [0.6.3/patch] videotestsrc and xvideosink do not reset their state when going to NULL
[0.6.3/patch] videotestsrc and xvideosink do not reset their state when going...
Status: RESOLVED FIXED
Product: GStreamer
Classification: Platform
Component: gst-plugins
0.6.x
Other Linux
: Normal normal
: 0.6.3
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2003-06-11 22:17 UTC by Thomas Vander Stichele
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
fix against 0.6 (5.37 KB, patch)
2003-06-11 22:18 UTC, Thomas Vander Stichele
none Details | Review
patch (2.50 KB, patch)
2003-08-19 08:29 UTC, David Schleef
none Details | Review

Description Thomas Vander Stichele 2003-06-11 22:17:16 UTC
noticable by running videotestsrc ! colorspace ! xvideosink in the editor
and setting to play, pause, play.  Watch xvideosink not output anything until 
the same time as the previous play time has elapsed
Comment 1 Thomas Vander Stichele 2003-06-11 22:18:08 UTC
Created attachment 17469 [details] [review]
fix against 0.6
Comment 2 Thomas Vander Stichele 2003-06-11 22:37:48 UTC
commited to HEAD and 0.6.2.1, closed
Comment 3 Ronald Bultje 2003-06-12 05:53:52 UTC
Shouldn't they reset in READY instead of NULL? NULL->READY is the
resource allocation phase, READY->PAUSE is the pipeline intialization
phase (and on PAUSE->READY, the pipeline data is therefore reset,
including time and so on).
Comment 4 David Schleef 2003-08-19 08:28:48 UTC
yes, the timing information should reset in PAUSED->READY.  Also,
element parameters should never be reset to their default value, even
in READY->NULL transitions.
Comment 5 David Schleef 2003-08-19 08:29:26 UTC
Created attachment 19339 [details] [review]
patch
Comment 6 Thomas Vander Stichele 2003-08-21 09:25:25 UTC
why shouldn't element data be reset even from READY to NULL ?

AFAIK, all plugins going to NULL should behave as if they were freshly
created, why is this not the case ?
Comment 7 David Schleef 2003-08-22 00:43:47 UTC
> why shouldn't element data be reset even from READY to NULL ?

Element data, yes.  But (writable) parameters should (ideally) never
be changed except under the control of the application.

Reseting parameters to the default in READY->NULL makes it much more
difficult for elements to use the NULL state to free their resources.
 A good example is osssink -- I don't think that it's a good idea to
reset the device parameter every time we want osssink to detach from
the device (i.e., every time the user presses pause on a GUI).

If an application wants a pristine element, it can create one.  I
don't see a good reason to support restoring elements to a pristine
state.  One could probably approximate it by going through all the
parameters and calling g_param_value_set_default(), though.
Comment 8 Thomas Vander Stichele 2003-08-26 14:28:35 UTC
commited to 0.6.2.1, closing.