Probable regression from 95a18cc9
(9 months old, I guess no developers use S/PDIF)

git failed to apply the entire patch, sorry for the noise
Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

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

Note that this code seems more likely to trigger race-induced crashes
in the audio mixer core.

Some weird hardware only support weird formats that VLC does not know
of (e.g. unsigned 18-bits padded to 3 bytes). The ALSA plug plugin will
do the conversion if needed.

VLC resamples all the time anyway. We don't need two resamplers.

ALSA-lib will convert the sample format if needed, notably on weird
cheap hardware. It will also remix the channels based on the selected
device.

In any case, ALSA version 1.0.24 is recommended as any older versions
may cause spurious failures when used in VLC.

With this change, the ALSA device failure dialog gets a fair chance to
appear in case of error. Previously, Probe() would silently make the
ALSA plugin fail, bypassing the error dialogs.

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

This commits crashes saving preferences
This reverts commit 3d3cacbf.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Its only supported codec (of sorts) cannot occur anymore, as the fake
access_demux was superseded by the image demux.

That codec cannot occur anymore.

Sometimes it didn't work, when stopping, as pointed out by Rémi on IRC.

This keeps the VLC value in sync with the PulseAudio one. For some
reason, the Qt4 interface does not pick up the correct default though.

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

Notes:

- The variable value represents the PulseAudio sink index.
  It is not a member of the AOUT_VAR_ enumeration.

- There is no configuration item. PulseAudio normally remembers the
  last sink VLC used. So I think there is no need value in writing
  code to duplicate that effort.

- I am not sure if we need more locking inside the variable callback.
  This needs to be clarified with PulseAudio guys.

Fixes the int64_t issue.