Do error handling with GError instead of aborting with g_error().

Don't hard code the "bits" parameter to 16.  Try to use the input's
sample format, if possible.

The MPD core logs the audio format of all audio outputs.  Remove the
duplicate message from the plugins.

To aid debugging, print the audio format of the output plugin in a
debug message, and print information about PCM conversion.

To aid debugging, print the audio format of the decoder plugin in a
debug message, and print information about PCM conversion.

During the listen_add_host() API transition, the windows code wasn't
tested, and several removed arguments are still in use there.

Windows doesn't have the standard headers sys/socket.h and netdb.h.

Even when --enable-documentation is passed, don't abort if "xmlto" was
not found.

The output path must be ./doc/protocol/, not ./protocol/

Renamed functions and variables.

Use GLIB_CHECK_VERSION() instead of manually checking
GLIB_MAJOR_VERSION, ...

On some hardware, reading the mixer value from hardware is an
expensive operation, and MPD has to do it for every client.  Throttle
access to the hardware, cache the result for one second.

time() is not a monotonic timer, and MPD might get confused by clock
skews.  clock_gettime() provides a monotonic clock, but is not
portable to non-POSIX systems (i.e. Windows).  This patch uses GLib's
GTimer API, which aims to be portable.

The option "enabled" is on by default.  If you specify "enabled no" in
an audio_output section, then this device is disabled by default.

decoder_data() returns a decoder_command, no need to call
decoder_get_command() twice after decoder_command().

If an input_stream is not seekable, libaudiofile fails to play at all:

 Audio File Library: unrecognized audio file format [error 0]

Since we know in advance whether the input_stream is seekable, just
refuse to play on a non-seekable stream.

Renamed several variables and a function.

[mk: adapted to new output plugin API]

Print the list of suffixes supported by each decoder, instead of
prining a list of all suffixes of all decoders with duplicates.

The generic sockaddr struct is too small for some addresses.  For
accept(), we have to allocate a sockaddr_storage struct on the stack,
which is large enough for all addresses.

Removed the sockaddr_to_tmp_string() hack, use
the new function sockaddr_to_string() instead.

Unpack IPv4 addresses which are packed inside an IPv6 address,
i.e. return "127.0.0.1" rather than "::ffff:127.0.0.1".

Create the socket_util.c library, the first function is
sockaddr_to_string(): it converts a sockaddr struct to a string
containing the IP address in a human-readable form.

When checking whether database entries have been deleted, don't check
if an archive file is a directory (G_FILE_TEST_IS_DIR), use
G_FILE_TEST_IS_REGULAR for this case instead.  To determine if a
"struct directory" is an archive, check for device==DEVICE_INARCHIVE.
This is always false after loading the database, so this patch is not
complete yet.

When the mtime of an archive time hasn't changed, don't update it
again.

Remember the modification time of each directory.  This is important
for archives (which are virtual directories right now), but may also
be useful for an automatic update mechanism.

Simplify the rather large function update_regular_file().

Moved some of them to to directory_save.c, and others to database.c.