The "mode" of the second ReplayGainFilter was never set, and thus
replay gain was never applied to the new song during cross-fade.

Wake up the OutputThread only if it hasn't already been woken up and
if it isn't already in the playback loop.

Prepare to migrate away from GLib.  Currently, we're still using GLib
as a backend.

Replaces GLib's GError.

Push the new mode to the filter instead of accessing global variables
through replay_gain_get_real_mode().

Pass audio_output objects around instead of void pointers.  This will
give some more control to the plugin, and prepares for non-blocking
audio outputs.

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.

.. and destruct the output object properly in test/run_output.