change the _delay API doc so that it matches its implementation behavior.

This further optimize the performance. 

This is now working properly with a conditional variable bug being fixed (8bbfb5cd).

do not hog system device.

which ensures mpd do bit perfect playback on OS X

Make all the locks exception-safe.

Now that CloseFilter() may be called again, we can reduce code
duplication.

This allows calling CloseFilter() multiple times.

With the configuration "replay_gain_handler none", the Filter pointers
are never initialized, which can crash MPD.

 https://bugs.musicpd.org/view.php?id=4571

Currently the switch statement is invalidated by ss.format being overwritten
with the default value of PA_SIMPLE_S16NE which results in white noise during
playback as my server is expected S16LE (S16NE).
Signed-off-by: Earnestly <zibeon@gmail.com>

This is based on a patch from Ian Scott in 2014. It was never committed,
so I figured I'd fix the outstanding issue and resubmit it.

https://www.mail-archive.com/mpd-devel%40musicpd.org/msg00139.html

Closes https://bugs.musicpd.org/view.php?id=4537 .

Removed the 'cancel' function because it violates 'single producer,
single consumer'.

The user manual specifies "encoder", which is consistent with other
output plugins.  "encoding" should be deprecated.

In e068d62a I added code that zeros the remainder of the output buffer
if there are not enough input frames available. I have now learned
that we can signal to the caller of the render callback how much data
is in the output buffers. In practice, the input buffer is so large
that it does not matter so much how we handle input buffer underruns,
but I suppose that saying 'there is no data' is better than 'here is
some silence for you'.

For easier state management inside filter plugins.