Use this struct for db_walk().

Improve error reporting and handling.  command.c gets the new function
print_error(), which sends a GError to the client.

Reduce header dependencies.

Only used there.

First draft, not really pluggable currently - hard-coded to use the
"simple" plugin, and calls several of its internal functions.

The API is very simple currently, all searches are still performed
over the root "directory" object.  Future changes to the API will move
those search implementations into the plugin, to allow more efficient
implementations, or implementations that don't have the whole tree in
memory all the time.

For TAG_NUM_OF_ITEM_TYPES.

No writers.

Better error messages.

Remove the slash hack, allocate memory for the user name.

Just advance the "path" pointer.

config_get_path() was somewhat flawed, because it pretended to be a
function, when it really had a side effect.  The second flaw was that
it did not return the parser error, instead it aborted the whole
process, which is bad style.  The new function returns a duplicated
(modified) string that must be freed by the caller, and returns a
GError on failure.

No caller needs to write.

config_add_block_param() cannot fail, which makes it easier to use.

Reduce indent.

Allow calling findAddIn() without a client object.

Conflicts:
	NEWS
	configure.ac

libOggFLAC has been deprecated for 5 years now, and we havn't been
testing it for a long time.  Let's delete it.

Conflicts:
	configure.ac
	src/output_control.c

.. and a minor tweak for libFLAC+libogg detection.

Eliminate an unnecessary source of deadlocks.

The pkg-config file was added by the Debian package maintainers, and
unfortunately, the rest of the world doesn't have it.

Add M4 function MPD_AUTO_PKG_LIB for pkg-config with AC_CHECK_LIB
fallback.

Don't abort the configure script when avahi could not be
auto-detected.  It previously did, because there was no custom "fail"
action for PKG_CHECK_MODULES.

Don't add those to MPD_CFLAGS / MPD_LIBS.

Better name, better documentation.

The output thread could hang indefinitely after finishing CANCEL,
because it could have missed the signal while the output was not
unlocked in ao_command_finished().

This patch removes the wait() call after CANCEL, and adds the flag
"allow_play" instead.  While this flag is set, playback is skipped.
With this flag, there will not be any excess wait() call after the
pipe has been cleared.

This patch fixes a bug that causes mpd to discontinue playback after
seeking, due to the race condition described above.