aboutsummaryrefslogtreecommitdiffstats
path: root/src (unfollow)
Commit message (Collapse)AuthorFilesLines
2014-10-31Revert "Main: run the OS X native event loop"NanoTech3-50/+0
This reverts commit f0be48ff90503d9ffa5b295fd4454eec753950ee (except for the NEWS entry). If libdispatch (GCD) is used before forking, it can't safely be used again after forking.
2014-10-31PlaylistFile: don't allow empty playlist nameMax Kellermann1-0/+4
2014-10-28decoder/mad: fix negative replay gain valuesMax Kellermann1-1/+1
Negating an unsigned integer does not work.
2014-10-28TextInputStream: don't ignore unterminated last lineMax Kellermann1-4/+16
2014-10-26playlist/m3u: recognize the file suffix ".m3u8"Max Kellermann2-0/+2
2014-10-25AllCommands: "commands" returns playlist commands only if playlist_directory ↵Max Kellermann3-0/+23
configured
2014-10-25SongSticker, playlist/soundcloud: add missing stdlib.h includeMax Kellermann2-0/+2
For free().
2014-10-25OutputThread: fall back to PCM if given DSD sample rate is not supportedMax Kellermann1-0/+30
Works around the "PCM conversion from f to dsd is not implemented" error message that prevents DSD playback.
2014-10-25decoder/faad: remove workaround for ancient libfaad2 ABI bugMax Kellermann2-20/+4
Many years ago, FAAD had a serious ABI bug: the NeAACDecInit() prototype in its header declared the "samplerate" parameter to be "unsigned long *", but internally, the function assumed it was "uint32_t *" instead. On 32 bit machines, that was no difference, but on 64 bit, this left one portion of the return value uninitialized; and worse, on big-endian, the wrong word was filled. This bug had to be worked around in MPD (commit 9c4e97a6). A few months later, the bug was fixed in the FAAD CVS in commit 1.117 on file libfaad/decoder.c; the commit message was: "Use public headers internally to prevent duplicate declarations" The commit message was too brief at best; the problem was not duplicate declarations, but a prototype mismatch. No mention of the bug fix in the ChangeLog. The MPD project never learned about this bug fix, and so MPD would always pass a "uin32_t *" dressed up as a "unsigned long *". Nearly 6 years later, it's about time to fix this second ABI problem. Let's kill the workaround!
2014-10-25SongSticker: don't use GLibMax Kellermann1-4/+3
2014-10-25decoder/wavpack: don't use GLibMax Kellermann1-3/+3
2014-10-25playlist/SoundCloud: don't use GLib for string allocationMax Kellermann1-31/+31
2014-10-25Util/Alloc: add xstrcatdup(), replacing g_strconcat()Max Kellermann2-0/+82
2014-10-25Set pulseaudio channel map to WAVE-EXSteven Newbury1-1/+5
Pulseaudio expects clients to specify their channel-map if the default (ALSA) map does not route the audio to the expected speakers. Many Google results suggest dealing with this by re-routing the audio channels with the appropriate ALSA plugin, but this will then simply break any clients which expect the default ALSA mapping. Virtually all media files and codecs, certainly flac, dca, a52, and of course anything based on Microsoft's WAVEFORMAT_EXTENSIBLE specification, assume the layout in the table here: http://en.wikipedia.org/wiki/Surround_sound#Standard_speaker_channels Fortunately, pulseaudio directly addresses this with a built-in channel map for WAVE-EX which can be set automatically in the stream sample-spec.
2014-10-25decoder/ffmpeg: recognize MIME type audio/aacpSteven OBrien1-0/+1
2014-10-25fs/Charset: work around clang's -Wunused-const-variableMax Kellermann1-1/+2
MPD_PATH_MAX_UTF8 is only used by GLib-specific code currently.
2014-10-24OutputThread: close the output plugin after filter failureMax Kellermann1-0/+1
Fixes memory leak because ao_plugin_close() never gets called.
2014-10-24OutputThread: unlock mutex for CloseFilter()Max Kellermann2-0/+14
Be consistent.
2014-10-23output/roar: remove unnecessary "volatile" keywordMax Kellermann1-1/+1
A mutex acts as a memory barrier, and thus "volatile" is not necessary.
2014-10-23pcm/ChannelsConverter: fix variable used to generate error messageMax Kellermann1-1/+1
Use the "_format" parameter instead of the (uninitialized) "format" attribute.
2014-10-23pcm/FormatConverter: move check to Open()Max Kellermann1-10/+24
Report unsupported format while opening the filter, not later when the first conversion takes place.
2014-10-23pcm/PcmConvert: assign {src,dest}_format at the endMax Kellermann1-12/+13
Fixes assertion failure in destructor by not assigning {src,dest}_format when an error occurs.
2014-10-23pcm/PcmConvert: make AudioFormat parameters "const"Max Kellermann1-1/+1
2014-10-19TextInputStream: manually shift the buffer before readingMax Kellermann1-0/+2
Fixes truncated lines in m3u and cue files (regression by commit 67958f7).
2014-10-19util/StaticFifoBuffer: make Shift() publicMax Kellermann1-2/+0
2014-10-17playlist/extm3u: strip first line for #EXTM3U detectionMax Kellermann1-2/+6
2014-10-11ThreadInputStream: call ThreadRead() inside the thread instead of Read()Max Kellermann1-1/+1
Fixes deadlock bug in the "mms" plugin.
2014-10-11ThreadInputStream: add assertionsMax Kellermann1-0/+8
2014-10-11util/DynamicFifoBuffer: make the "Range" type publicMax Kellermann1-0/+1
Export it from the protected base class. This fixes a build failure on Mac OS X.
2014-10-10db/upnp/Util: move caturl() to util/UriUtil.cxxMax Kellermann6-32/+39
2014-10-10event/IdleMonitor: remove redundant commentMax Kellermann1-2/+1
2014-10-10ZeroconfBonjour: use htons() instead of g_htons()Max Kellermann1-3/+3
2014-10-10TagString: implement fix_utf8() without GLibMax Kellermann1-22/+34
2014-10-10util/UTF8: add SequenceLengthUTF8()Max Kellermann2-0/+96
2014-10-10util/UTF8: new libraryMax Kellermann2-0/+339
2014-10-10TagString: return WritableBuffer<char>Max Kellermann3-24/+26
2014-10-10TagString: use strndup() for unterminated stringMax Kellermann1-1/+1
Fixes buffer overflow bug.
2014-10-10TagString: remove ISO-Latin-1 fallbackMax Kellermann1-14/+1
MPD handles all strings in UTF-8 internally. Those decoders which read Latin-1 tags are supposed to implement the conversion, instead of passing Latin-1 to TagBuilder::AddItem(). FixTagString() is simply the wrong place to do that, and hard-coding Latin-1 is kind of arbitrary.
2014-10-10fs/Traits: use value_typeMax Kellermann1-4/+4
2014-10-10Main: run the OS X native event loopnanotech3-0/+50
2014-10-09storage/nfs: use the libnfs async APIMax Kellermann3-49/+397
Share the NFS connection with the NFS input plugin.
2014-10-09lib/nfs/Connection: reduce Error instance allocationsMax Kellermann1-3/+2
2014-10-09lib/nfs/Glue: add assertionMax Kellermann1-0/+1
2014-10-09StoragePlugin: pass EventLoop to constructorMax Kellermann9-14/+23
2014-10-07storage/nfs: move code to class MemoryStorageDirectoryReaderMax Kellermann3-73/+150
Read all directory entries into memory and close the struct nfsdir before returning the StorageDirectoryReader instance. This is what libnfs does, anyway.
2014-10-07storage/nfs: move code to Copy()Max Kellermann1-18/+30
2014-10-07lib/nfs/Connection: add methods {Open,Read,Close}Directory()Max Kellermann2-0/+53
2014-10-07lib/nfs/Connection: add method Stat(path)Max Kellermann2-0/+33
2014-10-06lib/nfs/Connection: make GetEventLoop() publicMax Kellermann1-4/+4
2014-10-05lib/nfs/Base: kludge to reduce number of NFS mountsMax Kellermann4-7/+127
Creating a NfsStorage sets its own export_name as the "base". Now NfsFileReader can use this information to derive the export_name to be mounted, instead of guessing. This solves the "too many connection" problem on the NFS server while updating the database.