[Bug 750694] alsasink: Brief gap after unpause if audio clock is not the pipeline clock

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=750694

--- Comment #2 from Nicolas Dufresne (ndufresne) <nicolas at ndufresne.ca> ---
If I understand well, we call reset to unblock snd_pcm_write(). But this will
also discard the data being written (or buffered). From a ringbuffer/clock
perspective, it is like if this data was played, so the clock jumps forward,
resulting in a gap which get filled with silence.

I think the implementation of reset() is correct, but it's usage in pause is
wrong. It should call GstBaseSink::unlock/unlock_stop() instead, and alsasink
should implement this. But for that, we'd need to move away from snd_pcm_wait()
and retrieve a poll descriptor. A bit like what was done for ALSA midi src
recently.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.