It is cleaner to not send event from caller threadand will fix a deadlock in the
playlist code.

This is ugly but we can do jack about this (short of forking ffmpeg).

avcodec is used from at least three different plugins: avcodec,
switcher (which seems dead though) and indirectly in avformat. We need
to have a common interlock for all of these.

This fixes a crash from the avformat (de)mux.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

It makes prerolling less precise but it will avoid loosing valid
frame.

Using vlc_mutex_trylock(), we make sure that the video output thread
will not yield because of an incomplete snapshot request. It can still
yield in the very unlikely case that a thread is doing a snapshot
request while the video output thread finishes sorting SPUs _and_ there
already was a pending snapshot earlier.

The macros are indeed a little bit too intrusive, and break calls to
QString::sprintf
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

When the socket is congested, we should simply skip sending.
When there is a soft-error, we should resend and simply ignore.

Instead, we were dropping the socket if the failure remained on the
second try, which was always the case in the congestion scenario.
This should fix the socket dropping problem described at:
http://forum.videolan.org/viewtopic.php?f=4&t=56493