When switching to another song manually, the player checks if the
decoder is already decoding that song; if so, it will attempt to reuse
it by seeking it to the new position.  That however fails if the
decoder is not seekable (e.g. a radio stream) which leaves the user
unable to switch to that song with the bogus error message "Not
seekable".

There was a discrepancy between what was written to the buffer and the
size returned by pcm_dsd_to_dop(): the "for" loop uses num_frames/2,
rounding down, while the return value is num_samples which is
num_frames*channels, without rounding.  This could cause undefined
data at the end of the destination buffer if the source buffer size
was not aligned to multiples of 8 bytes (4 DSD bytes per channel).

The latter however can occur in the 0.21 branch after commit
a06bf388

Closes #233

Fixes crash due to "pure virtual method called" in the "mms" input
plugin.  Closes #253

Closes #247

Without setting the "ready" flag, the caller will wait in WaitReady()
forever, locking up MPD.  Closes #252

lockfree library used by ALSA output plugin is part of Boost from version 1.53,
so this can be theoretically the lowest required version, however
there are issues which are resolved from 1.54 onwards.

First commit for issue #69

Instead of stopping playback (due to seek time overflow), reject the
seek command.  Closes #240

Relative negative values (with "seekcur") are still allowed, and MPD
will fix the resulting position if it turns out to be negative.  But
the "seek" and "seekid" commands use an unsigned time stamp which must
not be negative.

With Grand Central Dispatch used in Main.cxx, debug builds on macOS
crash as the IsInside() assertion gets triggered in the event loop. As
a simple fix, usage of GCD is removed. Plugging and unplugging
headphones or changes of the default output device was tested without
issues. Whatever the original commit tried to fix by GCD probably does
not need fixing anymore.

Now that all build failures have been fixed, we can enable the
feature.

Fixes two build failures with libnfs on Windows.

configure.ac sets this, but this wasn't used for compiling third-party
libraries.  This setting however is important for libnfs, which adds
fallback definitions for POLLIN and POLLOUT with bogus values.

It indeed is a macro on Windows.

Instead of setting state=IGNORE_TRACK, ignore only the following
"INDEX" lines.

Correction for commit 8461d71b.  Closes #227

Let FFmpeg do the MP3 decoding.  See commit
a4de9650

Closes #232

libmad has been unmaintained for a long time, and it fails to build on
Windows.  I could go and fix libmad's broken configure script, but I
prefer to just assign MP3 decoding to FFmpeg for now.

Closes #228

Enable the NFS storage plugin on Android.

Closes #226

This addresses two problems:

1. the libFLAC write callback had to send an error status to its
caller when SubmitData() returned a command; this disrupted libFLAC
and the resulting command could not be used for anything;

2. the libFLAC function FLAC__stream_decoder_seek_absolute() also
calls the write callback, but its result cannot be used, because
seeking is still in progress, so we lose all data from one FLAC frame.
By moving the SubmitData() call until after CommandFinished(), we
avoid losing this data.  This fixes another part of #113

See code comment.

Instead of passing whole chunks to the MusicPipe and checking the
end_time after each chunk, truncate the last chunk if it would exceed
the end_time.  This requires keeping track of the absolute PCM frame
number.

This fixes a problem with gapless CUE song transitions: a small part
of the following song was always played twice.

Closes #113

it was uninitialized before