Reduce the overhead.  Most buffers used by MPD are around 2 to 4 kB.
8 kB seems large enough to keep heap fragmentation low.

Additionally, this patch fixes an off-by-one error in the alignment
formula.

Declaring the global variable _CRT_fmode is a documented hack.  What
an ugly platform!

Add the flag "b" to fopen().  Windows compatibility.

Windows compatibility.

On mingw32, snprintf() expects a 64 bit integer instead of a "long
int" for "%li" - this is not consistent with our expectation, so we're
using plain sprintf().

For some unknown reason, read() blocks on WIN32, even though it was
invoked inside the G_IO_IN callback.  By switching to GIOChannel
functions, this problem is solved, and it works on both Linux and
Windows.

On WIN32, use g_io_channel_win32_new_fd() instead of
g_io_channel_unix_new().  There doesn't seem to be a practical
difference, but it seems more correct.

inode numbers don't work on WIN32, and very few WIN32 users actually
use symlinks - seems ok to disable that check.

Not a CPP macro.  This adds the "unused" attribute to "saveptr", which
eliminates one gcc warning.

On WIN32, "%F" and "%T" are not implemented.

In the winsock headers, the setsockopt() argument is declared as
"const char *", not "const void *".

This function is empty on WIN32, but we need to return a value.

input_stream_buffer() has a check for "buffer==NULL", so we don't need
an explicit empty implementation.

Pay attention to the ReplayGain mode "auto" and the
replay_gain_missing_preamp parameter inside of decoder_replay_gain().

Added #ifdefs on the macros AFMT_S24_NE and other OSS4 extensions.

Removed the "supported" / "unsupported" arrays, and auto-detect on
every open() call, just like the ALSA output plugin.

Removed the macro AFMT_S16_MPD.

In mingw32, int16_t is not defined by sys/types.h, but it is by stdint.h,
and it is in the int16_t man page as being defined in stdint.h. Thanks to
mithi for help debugging.

This is portable for win32 and fixes client logins there.

Don't add it to the filter chain, because we need to apply replay gain
before cross-fading with the next song.  Add a second replay_gain
filter which is used for the song being faded in (chunk->other).

Remove cross_fade_apply(), and call pcm_mix() in the output thread,
mixing the chunk and chunk->other together.

Preparation for cross-fading support in the output thread.

Move encoder before audio outputs. Makes sense to keep them after
decoders and some audio outputs need to know wheather encoders are
present.

Add libogg section rather than checking for it every time it's necessary.