Hi,

upon trying to play an MMS stream added to the play list, I got this:

    mpd: /tmp/mpd/./src/input_stream.c:85: input_stream_open: Assertion `is->plugin->open == ((void *)0) || is->plugin == plugin' failed.

With the following patch applied, it works perfectly.

Thanks for having implemented MMS support :-).

Best regards,
Peter

The configure.ac script does not care about the host architecture, it
only cares about the OS.  Use ${host_os} instead of ${host} to
simplify the matching expressions.

Without AC_CANONICAL_HOST, ${host_os} etc. are not defined.

Added an inline assembly function for the 64 bit multiplication.

Benchmark results on a Pentium II 266 MHz, 512 MB of 24 bit PCM data:

 dd if=/dev/zero bs=64k count=8k |
     time ./test/software_volume 48000:24:2 >/dev/null

Before this patch 22.94s, after this patch 7.24s.

This program is a simple utility for debugging and benchmarking the
software volume code.

The switch syntax is "--enable-test".  There are no test programs yet.

When mp4ff_read_sample() returns a value bigger than zero, it
guarantees that the buffer is set.  Remove the check.

Don't waste any precious memory when the seek_table cannot be used.

Don't include limits.h, use GLib constants instead.

Use faacDecInit2() instead of AudioSpecificConfig() to detect the AAC
track in the MP4 file.  This has a great advantage: it initializes the
libfaad decoder, which the caller would normally do anyway - but now
we can go without the AudioSpecificConfig() call.  When decoder==NULL
(called from mp4_tag_dup()), fall back to a mp4ff_get_track_type()==1
check, like other audio players do.

Moved the libfaad decoder initialization to mp4_faad_new(), and also
fill the audio_format struct there.  This eliminates a little bit of
complexity in mp4_decode().

Don't wait for the first frame to be decoded.  We already have the
sample rate and the channel count from faacDecInit2().

The function mp4_load_tag() is used only once, and mp4_tag_dup() is a
one-liner.  Merge them.

When a file is not seekable, MPD dropped the audio buffers before even
attempting to seek.  This caused noticable sound corruption.  Fix:
first attempt to seek, and only if that succeeds, call
audio_output_all_cancel().

Make some variables more local, and eliminate superfluous ones.

Instead of returning the sample rate and channel count as separate
values, fill an audio_format struct.

Don't wait for the first frame to be decoded.  We already have the
sample rate and the channel count from faacDecInit().

The MPD core will never send a SEEK command to a decoder which has
declared to be not seekable.

Define G_LOG_DOMAIN.

Replace this plugin's own buffer library with the new decoder_buffer
library.

Instead of checking if the buffer is empty after adts_find_frame(),
check adts_find_frame()'s return value.  This is more robust.

Moved libfaad API quirks to the wrapper functions faad_decoder_init()
and faad_decoder_decode().

Instead of writing the song duration into a float pointer, return it
from the function.

There are no callers which pass NULL here.

All callers of adts_find_frame() use faad_buffer_fill() before that.
Move that faad_buffer_fill() call into adts_find_frame() instead.
adts_find_frame() will get its own logic for on-demand filling.

The decoder buffer library may be used by decoder libraries such as
"faad".

adts_check_frame() must not be called with a buffer length smaller
than 8.  We can eliminate that duplicate check, and convert it into an
assertion.

It's not valid to use the buffer's data without ensuring that the
buffer contains enough data.

"aac" -> "faad"

The function is unused, since we added the pcm_buffer library.

When I implemented the pcm_buffer library, I forgot to set the new
buffer size.  This caused a new allocation in each pcm_buffer_get(),
fortunately no memory was leaked.

Fixed the log domains of the renamed decoders.  Added G_LOG_DOMAIN
macros in decoders which don't have one already.

This plugin is based on "libmpcdec".

This plugin is based on "libmp4ff".

Renamed functions and variables.

The element fileOffset is only written, but never read.  It can be
removed safely.

This plugin uses libvorbis.