On NetBSD, PTHREAD_MUTEX_INITIALIZER and PTHREAD_COND_INITIALIZER are
not compatible with C++11 "constexpr" (see Mantis ticket 0004110). As
a workaround, don't ues "constexpr", and use the functions
pthread_mutex_init(), pthread_mutex_destroy(), pthread_cond_init() and
pthread_cond_destroy() instead. This adds some runtime overhead, but
is portable to POSIX implementations that have awkward initializer
macros.
Casting std::numeric_limits<unsigned>::max() to "long" leads to an
overflow if sizeof(unsigned)==sizeof(long), and the result will be -1.
This happens on some 32 bit architectures, for example ARM and WIN32.
Workaround: use std::numeric_limits<int>::max(), which is the largest
signed integer. Since sizeof(long)>=sizeof(int), this will never
overflow.
Fixes Mantis ticket 0004080.
The IsActive() method returned true even if the timer was not active,
after it completed once. This broke the state file timer, and the
state file was not saved periodically.
Read one block at a time. This discards the last partial block, which
cannot be interleaved anyway. Previously, uninitialised memory was
used to interleave the last block, which generated some noise.
When the data chunk size is not a multiple of the frame size, the last
partial frame lead to an endless loop. We fix this by checking
chunk_sze>=frame instead of chunk_sze>0. This way, the partial frame
is simply skipped.
Previously, MPD tried to slurp the whole song file, count the number
of frames and calculate the song duration from that. That however is
extremely expensive for remote files, and will delay playback for a
long time. Workaround: check only the first 128 frames and try to
extrapolate from here. Fixes Mantis ticket 0004035.
Implement a "bulk" edit mode that postpones both UpdateQueuedSong()
and OnModified(). This way, the playlist version gets incremented
only once. More importantly: when adding multiple songs to a queue
that consists of only one song, the first song that got added will
always be played next. By postponing this choice, all newly added
songs get a chance to become the next song. Fixes the second (and
last) part of Mantis ticket 0004005.
Don't restore the current song after shufflung when MPD is stopped
(but still remembers the current song internally). Fixes the first
part of Mantis ticket 0004005.
Turns out the lock-free code using atomics was not thread-safe. The
given callback could be invoked by GLib before the source_id attribute
was assigned. This commit changes the DeferredMonitor class to use a
Mutex to block the event loop until source_id is assigned. This bug
does not exist in the 0.19 branch because it does not use the GLib
main loop anymore.
This corrects a major mistake from commit 724a59aa - there was one
small thing that commit was supposed to do, and it failed.
AV_TIME_BASE is not a seek flag.
Don't wait for an optimistic write to fail. This is an improved
workaround for the infamous Raspberry Pi bug (see commit af991765).
It works much better and comes without the negative side effects. The
old workaround is now obsolete.
Move code from CreateConfiguredDatabase() and add XDG support. This
implements an automatic Linux fallback for the setting "db_file" if
none was specified.
This commit adds the NeighborPlugin API which can be used to detect
nearby file servers that can be used by input plugins. This list of
servers is exported using the new "listneighbors" command. The idle
even "neighbor" notifies interested clients when a new neighbor is
found or an existing one is lost.
There's a lot missing currently: protocol&user documentation, and a
way to "mount" remote servers into the music database. Obviously,
some code from the UPnP database plugin can be moved to a neighbor
plugin.
The update thread should not affect the rest of the system, therefore
set "idle" priority, and let it only run when nobody else is using the
resources.
This encoding plugin features a fixed-point mp3 encoder,
with faster encoding on architectures without a FPU.
Right now the encoder is limited to stereo and 16 bit depth.
The bitrate and sample rate can be modified in audio_output.
audio_output {
type "httpd"
name "My shine stream"
encoder "shine"
port "8000"
format "44100:16:2"
bitrate "320" # default: 128
}
If we have libyajl 2.0.1 (without a pkg-config file), our configure.ac
would assume this is the libyajl 1.0 API, because the function
yajl_alloc() exists in both. This commit changes the library check to
the function yajl_parse_complete() which was removed in the 2.0 API.
This fixes build failure with libyajl 2.0.1.
I've created an elementary input plugin that plays sound from the
soundcard, so you can use MPD to listen to anything connected to the
line-in jack, or to Video4Linux FM radio cards that send audio through
the soundcard. There has been a small number of posts here in the
past requesting line-in input, so here is a first, simplistic stab at
it.
The patch adds a new sheme, alsa://, which causes mpd to play data
read directly from a souncdard. It defaults to hw:0,0, but you can
pass any ALSA device name in the URI. So, using mpc for example:
mpc add alsa://
mpc play
will play from device hw:0,0.
To use a diffferent device:
mpc add alsa://hw:1,0
Commit 77c63511 caused MPD to become stuck right after a song change.
The problem was that at some point, the MusicBuffer became full, and
the DecoderThread working on the next song waits for the PlayerThread.
However, the PlayerThread was stuck in a loop of g_usleep() calls, and
never bothered to tell the DecoderThread that the MusicBuffer is not
full anymore. This bug is very old, but its chance to occur went from
nearly 0% to nearly 100%.
The fix is to wake up the DecoderThread before waiting for it. As a
side effect, I replaced the g_usleep() call with a Cond::Wait() call.
Allows big-endian users to configure the fallback byte order to
little-endian. Without this setting, MPD assumes native byte order if
the CD drive can't decide.
Add a new CommandResult code called "FINISH" which, unlike "CLOSE",
will attempt to flush the output buffer. This is a one-shot attempt;
it will do one write, and not try again.
Prior to version 0.3, the "length" callback returned a "long" instead
of AFfileoffset. Now that this API bug fix is a few years old, let's
drop 0.2 support for good.
Migrate from the old curl_multi_perform() API to the newer
curl_multi_socket_action() API (since CURL 7.16).
This allows working around a bug with HTTP redirections with epoll:
when CURL closes a socket and the new one happens to have the same
file number, MPD did not have a chance to remove the old one from
epoll and subsequently attempted to use EPOLL_CTL_MOD, which was not
allowed by epoll, because it's a new socket now.
This prevented using the "volume_normalization" feature with some
codecs (e.g. mp3), because the normalization code requires 16 bit
samples. If the codec happens to deliver formats other than S16, the
AutoConvert filter succeeds to initialize the conversion filter, but
the returned input audio format was wrong.
Share the Mutex between the DecoderThread and the PlayerThread. This
simplifies synchronization between the two threads and fixes a freeze
problem: while the PlayerThread waits for the DeocderThread, it cannot
answer requests from the main thread, and the main thread will block
until the DecoderThread finishes.
This command was removed by commit 206392ad (MPD 0.16), even though it
was been proven useful for some very simple clients. On request, I
add it to the protocol again.
The "loop_count" configuration parameter allows the user to set how
many times a module with backward loops shall loop. "0" (the default)
means a module is not allowed to use backward loops at all. "-1"
enables inifinite looping.
This patch allows the user to configure the mikmod decoder plugin to loop
modules. It adds a configuration parameter to the mikmod decoder called
"loop" which can be "no" (the old behaviour, default) or "yes" to allow
modules to use backward loops.
Player_LoadTitle() returns an aligned pointer in libmikmod-3.2 that
cannot be freed with free(). The correct way to do this now is
MikMod_free() which extracts the original pointer from the buffer and
frees that.
Signed-off-by: Christoph Mende <mende.christoph@gmail.com>
This plugin is cumbersome to support, now that MPD is migrating away
from GLib and the GLib event loop. It has no practical advantages
over the CURL plugin. Soup requires the bloated GType library.
Save the state file 2 minutes after the last change. This reduces the
disruptions by an idle MPD, and MPD can be paged out permanently until
it is used.
Implement the "scan_stream" method that can read tags from any
input_stream object. This requires a FLAC__IOCallbacks implementation
based on the input_stream API.
Improves quality by not squeezing 32 bit samples down to 16 bit, and
then back to 32 bit floating point. Reduces CPU usage by skipping a
conversion step.
Internally the vorbis (non-Tremor) decoder is working in floating
point, and it's not really necessary to cut the output back to 16-bit
if the soundcard or OS supports higher resolution.
The decoder can be trivially modified to bypass its internal
quantisation and produce floating-point output, and a separate
quantisation can be used as appropriate to the platform.
libvorbisidec and libvorbis export the same symbols, which is a
dangerous thing. Since libvorbisenc depends on libvorbis, this can
get nasty, so let's disable the Vorbis encoder unless the user
explicitly wants it.
This commit reimplements the core of the "single" mode. Instead of
doing the detection in the playlist code from the outside, it is moved
to the player thread, which gets a new option called "border_pause".
It will now pause playback exactly at the beginning of the new song,
making the feature more reliable.
Now that the player thread knows what will happen, it can suppress
cross-fading.
Fixes mantis tickets 0003055 and 0003166.
That function is not pure, it writes to error.
When marked as pure, the compiler is allowed to assume it does not do
anything to error, so it can remain NULL, which would result in an
invalid read in print_error().
Ignore APE tags that have no usable tags, and use the ID3 tag instead.
This is useful when the APE tag only contains replay gain, and the
real tags are stored as ID3. This implements feature request Mantis
#0003521.
g_file_test is redefined to be g_file_test_utf8 and thus can't handle
non-ASCII characters. This fix adds simple wrapper (taken from glib)
that fixes encoding and calls g_file_test_utf8. All required inclusions
of glib_compat.h are added as well.
This plugin is horrible code, I mean it. Last year, I tried hard to
fix it, but I figured would take less time to do a full rewrite.
Given that I don't even have any device that supports RAOP, I can't do
that properly. After 16 months, nobody volunteered for fixing it.
Hereby, I delete it, because having no RAOP plugin is better than
having this mess. Sorry.
The existing buffer implementation has a major flaw: it is unable to
re-fill the buffer until it has been consumed completely, leading to
many occasions where the render callback needs to generate silence,
just because the play() implementation was unable to append more
data. The fifo_buffer library handles that well.
Requires YAJL to build, and this doesn't include the necessary
automake changes. Can be built using
./configure CFLAGS="-I/usr/include/yajl" LIBS="-lyajl" --enable-soundcloud
Add the following to your config:
playlist_plugin {
name "soundcloud"
enabled "true"
apikey "c4c979fd6f241b5b30431d722af212e8"
}
Then you can stream from soundcloud using calls like:
mpc load soundcloud://track/<track-id>
mpc load soundcloud://playlist/<playlist-id>
mpc load soundcloud://url/http://soundcloud.com/some/track/or/playlist
For the last case, you can leave off the http:// or
http://soundcloud.com/ .
This was disabled when compiled with a new ffmpeg version. Older
ffmpeg versions used it explicitly, while newer ones may pass it
through from the codec.
This fixes a bug when libsamplerate returns an empty buffer for a very
small input buffer. The caller thinks this is an error, bug there is
no GError object.
This finally enables the new embedded CUE sheet code: when a song file
contains a playlist, it is printed in the "lsinfo" output, so clients
get to know about this.
Use libasound's polling functions, implement a bridge to GSource /
GPollFD and send idle events to clients when an external program
changes the ALSA mixer volume.
Moving songs using either 'move' or 'moveid' to position -1 (after the
current song) would fail for a song which is just before the current
song.
This patch corrects the check to see if the current song is in the range
to be moved. Since the range is from `start` up to `end` (exclusive) the
check was incorrect, but is now fixed.
The implementation of cancel() did not work well: you cannot use
alSourceUnqueueBuffers() to unqueue queued buffers, and our function
openal_unqueue_buffers() left the OpenAL library in a rather undefined
state; nothing was supposed to be queued, but the "filled" variable
was not reset.
The local variable was already divided by 1000, and the return value
was being divided by 1000 again - doh! This caused delays in the
httpd output plugin that were too small by three orders of magnitude,
and the buffer was filled too quickly.
WinAPI explicitly declares filesystem encoding.
It can be determined by GetACP().
Use that instead of Glib routine that always "detects" UTF-8 on Win32,
which is incorrect for MPD case.
Ensure that WINVER is defined early enough, so other system headers
won't fall back to their default value. Specifically, this solves a
build failure (-Werror) with mingw-w64 ("WINVER redefined").
When we have an absolute path that's not inside the music directory,
allow loading it anyway, if we're in "secure" mode (i.e. the client is
connected via UNIX socket).
Right now, a playlist with absolute pathnames can only add songs that
are in the same the directory of the playlist or under it.
If uri is an absolute pathname and base_uri is set,
playlist_check_translate_song() will check that base_uri is a prefix
of uri, excluding every other song in the music directory outside
base_uri.
I think in this case base_uri should be completely ignored (and made
NULL) and uri should just be checked against music root directory.
Previously, the condition "defined(play_audio_format)" was used to see
if an output device has been opened, but if the device had failed on
startup, an assertion failure could occur. This patch adds a separate
flag.
The Naim Uniti does not appear to support icecast-style streaming of FLAC
music but does support the codec from a DLNA server. This change looks for
"transferMode.dlna.org: Streaming" in the HTTP request header and responds
with something the Uniti (and hopefully other DLNA clients) accepts.
The only difference in the DLNA streaming mode is the reponse header and
that icecast metadata is disabled. If a client request indicates both modes
are supported, the DLNA mode is preferred (as the Uniti says it supports
both but then rejects a FLAC ICY stream).
Note: This change may be specific to Naim equipment (the only device it was
tested on). E.g. the hardcoding of Content-Length which works but is not a
logically correct value. The change should be backwards-compatible, so
only those clients requesting a DLNA stream will see any difference.
When playing a CUE track, the player thread waited for the decoder to
become ready, and then sent a SEEK command to the beginning of the CUE
track. If that is near the start of the song file, and the track is
short enough, the decoder could have finished decoding already at that
point, and seeking fails.
This commit makes this initial seek more robust: instead of letting
the player thread deal with the difficult timings, let the decoder API
emulate a SEEK command, and return it to the decoder plugin, as soon
as the plugin finishes its initialization.
Add GMutex, GCond attributes which will be used by callers to
conditionally wait on the stream.
Remove the (now-useless) plugin method buffer(), wait on GCond
instead. Lock the input_stream before each method call. Do the same
with the playlist plugins.
D'oh, we were reading 16 bit integers instead of 32 bit integers!
That caused silence when trying to play a 32 bit input file on a 24
bit sound card (e.g. USB sound chips with 24 bit packed samples).
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.
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.
To demonstrate the new I/O thread. libsoup is well-integrated into
the GLib main loop, which made this plugin pretty easy to write.
As a side effect, we have to initialize the I/O thread in all debug
programs that use the input API.
This warning should only be logged when we really received something.
When the client disconnects, G_IO_IN is triggered, and the read
returns G_IO_STATUS_EOF.
In the "vorbis" plugin, this is a copy of the old flush() method,
while flush() gets a lot of code remove, it just sets the "flush" flag
and nothing else. It doesn't start a new stream now, which should fix
a few problems in some players.
This makes FreeBSD detect libogg correctly. The '==' operator is an
undocumented GNU extension to test(1) and cannot be relied upon to
exist and do the right thing. POSIX mandates string comparisons to be
done using "test foo = bar".
From http://bugs.debian.org/513291
"In mpd.conf, the "admin" permission covers updating the db and
killing mpd.
"Since there are quite some usecases in which the user can upload
music to the mpd's directory by means of anonymous FTP or so, it is
desirable that any user may issue a db update, while killing the mpd
should not be possible.
"I'd suggest to remove the db update from the admin group and either
add it to "control" or an own group."
With mono sound, jack_sample_size is smaller than frame_size (4 vs 2
bytes), and "space/jack_sample_size==0". That means mpd_jack_play()
will return 0, although no error has occurred.
Version 1.0.0 of the libao library added a new field to the
ao_sample_format struct. It is a char * named matrix. When
an ao_sample_format is allocated on the stack, this field contains
garbage. The proper course is to insure that is initialized to NULL.
NULL indicates that we do not want any mapping.
The struct is now initialized using a static initializer, and this
technique is compatible with all known versions of libao.
This fixes the following valgrind warning occuring on the first call of
httpd_output_read_page:
==20124== Conditional jump or move depends on uninitialised value(s)
==20124== at 0x425E65: httpd_output_read_page (httpd_output_plugin.c:240)
==20124== by 0x426087: httpd_output_open (httpd_output_plugin.c:279)
==20124== by 0x41D862: ao_open (output_plugin.h:206)
==20124== by 0x41E133: audio_output_task (output_thread.c:590)
I wanted mpd to play a mp3 stream from a music website. The stream is
only available to subscribers, which restriction is enforced through
normal http authentication. However, the URL I get from the website
is not the final URL of the stream, but a generic URL which points to
the real one through a redirect (code 301). Thus, I cannot predict
the final URL, and so I cannot use the username:password hack to force
the authentication, and mpd (libcurl on mpds behalf) fails to grab the
stream.
libcurl allows the option CURLOPT_NETRC to be set and then the
credentials can be stored in the good old .netrc file (in this case it
would be ~mpd/.netrc, of course). But mpd doesn't set this option. I
think it should.
When a music_chunk to be crossfaded consists only of a tag,
cross-fading is not possible, and led to an assertion failure. This
patch just discards those, as if cross-fading was not enabled.
During the whole output thread, the audio_output object is locked, and
it is only unlocked while waiting for the GCond and while running a
plugin method. The error handler in ao_play_chunk() attempted to lock
the object again, which was code from MPD 0.15.x which should have
been removed a long time ago.
Until the decoder plugin has called decoder_initialized(), the player
may not submit seek commands. This however could occur with a slow
decoder and a CUE file with a virtual song offset. This patch adds
another check.
When you don't explicitly set an output sample rate, liblame tries to
guess an output sample rate from the input sample rate. You would
think that this "guessing" consists of just setting both equal, but
that is not the case. For 44.1kHz at 96kbit/s, liblame chooses
32kHz. This patch explicitly configures the output sample rate, to
stop the bad guessing.
This is a MPD 0.16 regression: when playing a 24 bit file, the switch
to 16 bit was made only partially, after mBytesPerPacket and
mBytesPerFrame had already been applied.
That means mBytesPerFrame referred to 24 bit, and mBitsPerChannel
referred to 16 bits. Of course, that cannot work.
Rename the "version" struct, because it seems to be a reserved name on
Solaris:
"src/decoder/mad_decoder_plugin.c", line 550: (enum) tag redeclared: version
cc: acomp failed for src/decoder/mad_decoder_plugin.c
Add new config parameter 'device' to audio_output type "osx":
- if not supplied or set to "default", open default device
- if set to "system", open system device
- otherwise 'device' should be an audio device name: mpd will find and
open the specified audio device, falling back to the default
device if it's not found
After popular demand, I've switched the order of "artist" and "title"
in the stream title. There is no standard, and there is no reliable
way to parse those from the stream title.
When one song is played twice, and the decoder is working on the
second "instance", but the first should be seeked, the check in
player_seek_decoder() may assume that it can reuse the decoder without
exchanging pipes. The last thing was the mistake: the pipe pointer
was different, which led to an assertion failure. This patch adds
another check which exchanges the player pipe.
Change the assertion on "fail_timer==NULL" in OPEN to a runtime check.
This assertion crashed when the output thread failed while the player
thread was calling audio_output_open().
Added support for a new optional configuration setting for the httpd output
named "bind_to_address". Setting it to a specific IP address (v4 or v6) will
cause the httpd output to bind to that address exclusively. Supporting
multiple addresses in parallel is future work.
This implements the feature requests #2998 and #2646.
Clear the notification before finishing the CANCEL command, so the
notify_wait() after that will always wait for the right notification,
sent by audio_output_all_cancel().
Some users reported that MPD crashes when using a new CURL version
with the threaded DNS resolver enabled. It seems that
curl_multi_fdset() returns no file descriptor when the DNS resolver
runs in another thread, so MPD does not have any event to wait for.
On the CURL mailing list, somebody suggested to sleep for a fixed
amount of time. This is not an elegant solution, because daemons
should never have to sleep without waiting for an event. I hope the
CURL developers will review the API and remove the threaded DNS
resolver.
Meanwhile, I'm removing the assertion in question, to allow those
unfortunate users running the latest CURL version to continue using
MPD.
In libwildmidi 0.2.3, the function WildMidi_SampledSeek() was removed,
without changing the SO name. This patch adds an autoconf check for
that function. Fall back to WildMidi_FastSeek() if
WildMidi_SampledSeek() is not available anymore.
libavformat 0.6 does not pass the original URI pointer to the "open"
method, which leads to a crash because MPD was using a dirty hack to
pass a pointer to that method.
This patch switches to av_open_input_stream() with a custom
ByteIOContext class, instead of doing the URI string hack with
av_open_input_file().
Loosely based on a patch from Jasper St. Pierre.
I've attached a patch that will make file URIs work on operating systems
that provide the getpeereid() function call to check the user ID of the
peer connected to a UNIX domain socket.
I took this tag name from a MusePack sample file I got from a user.
It is not documented in the APE specification:
http://wiki.hydrogenaudio.org/index.php?title=APE_key
People seem to be using undocumented extensions to the specification
anyway, and the best we can do is attempt to support them.
I took these tag names from a MusePack sample file I got from a user.
These are not documented in the APE specification:
http://wiki.hydrogenaudio.org/index.php?title=APE_key
People seem to be using undocumented extensions to the specification
anyway, and the best we can do is attempt to support them.
Without libid3tag, we were trying to skip the ID3 frame (since
0.15.2). Its length however was not calculated at all, we were just
dropping everything from the current input buffer. This lead to the
first few seconds of the file being skipped. This patch attempts to
calculate the ID3v2 frame size with the formula from:
http://www.id3.org/id3v2.4.0-structure 3.1 and 6.2
"When playing musepack files with mpd v0.15.8, rg seems to have no effect.
Using sample file below, mpd says 'computing ReplayGain album scale with gain 122.879997, peak 0.549150'.
One thing though, if I build mpd against old libmpcdec-1.2.6, rg works
as expected: 'computing ReplayGain album scale with gain 16.820000,
peak 0.099765'"
Previously, tags of the new song being cross-faded in were sent
immediately. That can cause wrong information being displayed,
because the "previous" song might send its tag at the end again,
overriding the "next" song's tag. This patch saves & merges the tag
of the next song, and sends it when cross-fading is finished, and the
next song really starts.
"There is a bug in fixed-point musepack (musepack_src_r435) playback.
In floating-point audio is OK but in fixed audio is distorted. I have
made a patch for this"
With single+repeat enabled, it is expected that MPD repeats the
current song over andd over. With random mode also enabled, this
didn't work, because the song order was shuffled internally. This
patch adds a special check for this case.
This is a very basic check, which only ensures that the path does not
begin with a slash, doesn't have double slashes and the special names
"." and ".." are forbidden.
Removed the decoder_command_finished() call at the end of
mp3_decode(). This is invalid, because decoder_command_finished() has
already been called in mp3_read().
Add an option for each audio output which enables the use of the
hardware mixer, instead of the software volume code.
This is hardware specific, and assumes linear volume control. This is
not the case for hardware mixers which were tested, making this patch
somewhat useless, but we will use it to experiment with the settings,
to find a good solution.
Apply the replay gain in the output thread. This means a new setting
will be active instantly, without going through the whole music pipe.
And we might have different replay gain settings for each audio output
device.
When all plugins have failed, MPD used to fall back to the "mad"
decoder plugin, to handle those radio streams without a Content-Type
response header. This however leads to unexpected results (garbage
being played) when the stream isn't really mp3. Since we care little
about "bad" streams, we shouldn't have hacks which have bad side
effects.
Let's get rid of this hack now! Only try to "mad" plugin if there was
no match at all (Content-Type, path suffix) and no other plugin has
been tried.
To allow libavformat to detect the format of the input file, append
the suffix of the input file to the URL of the virtual stream. This
specifically enables the "shorten" codec, which is supported by
libavformat/raw.c, detected only by the suffix.