[Bug 661049] matroskademux: support seek with start_type NONE

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=661049
  GStreamer | gst-plugins-good | git

--- Comment #5 from Wim Taymans <wim.taymans at gmail.com> 2011-10-06 09:42:25 UTC ---
(In reply to comment #3)
> Yes, but that does not say what to do with all the downstream sending of
> closing segment and stuff (assuming that "a seek should happen" means a new
> last_stop is at hand).
> 
> For example, if the update field is false (basically meaning last_stop does not
> give a new position), we could forego sending any closing segments and so on,
> but that would not convey any new stop position to downstream.  If we do still
> want to sent stuff there, it would have to act differently depending on update
> or not (and any of this would presumably apply to all demuxer stuff out there,
> not only matroskademux).

You would have to send a segment update in all cases IMO. If start (and thus
also last_stop) was left unchanged because of _NONE, you would send the segment
update anyway because maybe the stop or the rate changed. If the user sent a
seek event that didn't update anything (we could also disallow that..) a
segment update will be sent that does nothing.

So I guess the seek handling would check the update flag, if updated, do
regular seek logic, otherwise send segment update.

I'm sure not many demuxers do this right currently..

-- 
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.