[Bug 755732] rtmpsrc: Slow teardown of connection when pipeline is interrupted

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=755732

--- Comment #4 from John Slade <john at jtes.net> ---
>@@ +661,2 @@
>   if (src->rtmp) {
>+    gst_poll_set_flushing (src->poll, FALSE);
>
> Shouldn't this be flushing in stop(), and non-flushing in start()?

My assumption was that create and stop are in the same thread so we can't
simultaneously by blocked in create and in stop.  Where as unlock was a
separate thread (and can flush the polling).  Is this correct view of how it
works?

I put this non-flushing in stop() to just reset the poll to a good state as
this is where we clear up src->rtmp and close any sockets.  I could also do the
same in start() to ensure the the poll is non-flushing before we start to use
it.

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