Taming the directory.c monster, part II: move the database management
stuff to database. directory.c should only contain code which works
on directory objects.
Provide separate constructors for creating a remote song, a local
song, and one for loading data from a song file. This way, we can add
more assertions.
We already know if a song is a URL or not based on whether it
has parentDir defined or not. Hopefully one day in the future
we can drop HTTP support from MPD entirely when an HTTP
filesystem comes along and we can access streams via open(2).
Since all callers of song_id_exists() will map it to a song position
after the check, introduce a new function called song_id_to_position()
which performs both the check and the map lookup, including nice
assertions.
The shared code in showPlaylist() isn't worth it, because we aim to
remove fdprintf(). Duplicate this small function, and enable stdio
buffering for saved playlists.
The function loadPlaylist() wants to report incremental errors to the
client, for this reason we cannot remove its protocol dependency right
now. Instead, make it use the client struct instead of the raw file
descriptor.
Don't pass the raw file descriptor around. This migration patch is
rather large, because all of the sources have inter dependencies - we
have to change all of them at the same time.
Continuing the effort of removing protocol specific calls from the
core libraries: let the command.c code call commandError() based on
PlaylistInfo's return value.
The playlist library shouldn't talk to the client if possible.
Introduce the "enum playlist_result" type which the caller
(i.e. command.c) may use to generate an error message.
Client's input values should be validated by the command
implementation, and the core libraries shouldn't talk to the client
directly if possible. Thus, setPlaylistRepeatStatus() and
setPlaylistRandomStatus() don't get the file descriptor, and cannot
fail (return void).
The function valid_playlist_name() checks the name, but it insists on
reporting an eventual error to the client. The new function
is_valid_playlist_name() is more generic: it just returns a boolean,
and does not care what the caller will use it for. The old function
valid_playlist_name() will be removed later.
Move everything which dumps song information (via tag_print.c) to a
separate source file. song_print.c gets code which writes song data
to the client; song_save.c is responsible for serializing songs from
the tag cache.
There is no unlocked caller of clearPlayerQueue(), and the functions
lockPlaylistInteraction() and unlockPlaylistInteraction() are trivial
- merge them.
Since playerPlay() already calls playerStop(), we can remove its
invocation of playerStop() from playPlaylistOrderNumber().
We can also make playerStop a static function.
All (indirect) callers of queueSong() ensure that the queue state is
BLANK, so there is no need to check it in queueSong() again. As a
side effect, queueSong() cannot fail anymore, and can return void.
Also, playlist_queueError and all its error handling can go away.
Also enable -Wunused-parameter - this forces us to add the gcc
"unused" attribute to a lot of parameters (mostly library callback
functions), but it's worth it during code refactorizations.
Try to only include headers which are really needed. We should
particularly check all "headers including other headers". The
long-term goal is to have a manageable, small API for plugins
(decoders, output) without so many mpd internals cluttering the
namespace.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7319 09075e82-0dd4-0310-85a5-a0d7c8717e4f
since clearPlayerQueue() is always called within
lockPlaylistInteraction() / unlockPlaylistInteraction(), it simplifies
the code to add another function which calls these three functions.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7253 09075e82-0dd4-0310-85a5-a0d7c8717e4f
strtok() may return NULL if the input is an empty string. The
playlist parser did not check for that.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7200 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Local variables which are never read before the first assignment don't
need initialization. Saves a few bytes of text. Also don't reset
variables which are never read until function return.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7199 09075e82-0dd4-0310-85a5-a0d7c8717e4f
There is unreachable code at several positions, e.g. after an
#if/#end, or after an endless loop. Remove that.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7197 09075e82-0dd4-0310-85a5-a0d7c8717e4f
The while() loop only checks for interrupted system calls (which woudl
never happen if the signal mask were set up properly), but nobody
checks if the fopen() actually succeeds.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7195 09075e82-0dd4-0310-85a5-a0d7c8717e4f
It's too ugly and broken (both technically and usability-wise)
to be worth supporting in any stable release.
In one sentence: The queue is a very crippled version of the
playlist that takes precedence over the normal playlist.
How is it crippled?
* The "queueid" command only allows the queuing of songs
ALREADY IN THE PLAYLIST! This promotes having the entire mpd
database of songs in the playlist, which is a stupid practice
to begin with.
* It doesn't allow for meaningful rearranging and movement
of songs within the queue. To move a song, you'd need to
dequeue and requeue it (and other songs on the list).
Why? The playlist already allows _all_ these features
and shows everything a client needs to know about the ordering
of songs in a _single_ command!
* Random was a stupid idea to begin with and unfortunately
we're stuck supporting it since we've always had it. Users
should learn to use "shuffle" instead and not look at their
playlists. Implementing queue because we have the problem of
random is just a bandage fix and digging ourselves a new hole.
This protocol addition was never in a stable release of mpd, so
reverting it will only break things for people following trunk;
which I'm not too worried about. I am however worried about
long-term support of this misfeature, so I'm removing it.
Additionally, there are other points:
* It's trivially DoS-able:
(while true; do echo queueid $song_id; done) | nc $MPD_HOST $MPD_PORT
The above commands would cause the queue to become infinitely
expanding, taking up all available memory in the system. The
mpd playlist was implemented as an array with a fixed (but
configurable) size limit for this reason.
* It's not backwards-compatible. All clients would require
upgrades (and additional complexity) to even know what the
next song in the playlist is. mpd is a shared architecture,
and we should not violate the principle of least astonishment
here.
This removes the following commands:
queueid, dequeue, queueinfo
Additionally, the status field of "playlistqueue: " is removed
from the status command.
While this DoS is trivial to fix, the design is simply too
broken to ever support in a real release.
The overloading of the "addid" command and the allowing of
negative numbers to be used as offsets is far more flexible.
This improved "addid" is completely backwards-compatible with
all clients, and does not require clients to have UI changes or
run additional commands to display the queue.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7155 09075e82-0dd4-0310-85a5-a0d7c8717e4f
While mpd has always protected against the infinite expansion of
the main playlist by limiting its size in memory, however the
new storedPlaylist code has never checked for this limit.
Malicious (or clumsy) users could repeatedly append songs to
stored playlists, causing files to grow increasingly large
on disk. Attempting to load extremely large files into memory
will require mpd to slurp that all into memory, and ultimately
the file would be unusable by mpd because of the configurable
playlist size limit.
Now we limit stored playlists to the max_playlist_length
configuration variable set by the user (default is 16384). We
will refuse to append to playlist files if they hit that limit;
and also refuse to load more than the specified amount of songs
into memory.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7154 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This disables moving the bonkered moving of the current song to
a (negative) offset of itself (introduced in the last commit).
This also short circuits no-op moves when (from == to) and
avoid needless increasing of the playlist version and causes
clients to issue pointless no-op plchanges commands.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7153 09075e82-0dd4-0310-85a5-a0d7c8717e4f
If (and only if) there is a current song in the playlist,
(player could be stopped), allow the move destination
argument to be specified as a negative number.
This means moving any song (besides the current one) to the -1
position will allow it to be moved to the next song in the
playlist. Moving any song to position -2 will move it
to the song after the next, and so forth.
Moving a song to -playlist.length will move it to the song
_before_ the current song on the playlist; so this will
work for repeating playlists, too.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7152 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Instead of printing out the Id from playlist.c, instead set
the integer that added_id poitns to if added_id is non-NULL.
This makes the API cleaner and will allow us to use additional
commands to manipulate the newly-added song_id. Callers
(handleAddId) that relied on printId to print it to the given
fd have now been modified to print the ID at a higher-level;
making playlist.c less-dependent on protocol details.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7149 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This will make refactoring features easier, especially now that
pthreads support and larger refactorings are on the horizon.
Hopefully, this will make porting to other platforms (even
non-UNIX-like ones for masochists) easier, too.
os_compat.h will house all the #includes for system headers
considered to be the "core" of MPD. Headers for optional
features will be left to individual source files.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7130 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Abstractions should reduce the amount of code and make things
easier to follow. The StoredPlaylist struct did not do that, so
get rid of it and just use our standard linked list
implementation instead.
diffstat agrees with me that the abstraction is pointless:
3 files changed, 60 insertions(+), 90 deletions(-)
git-svn-id: https://svn.musicpd.org/mpd/trunk@7120 09075e82-0dd4-0310-85a5-a0d7c8717e4f
It was a nice way to double the memory needed to write the
playlist to a file.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7116 09075e82-0dd4-0310-85a5-a0d7c8717e4f
thread-safety work in preparation for rewrite to use pthreads
Expect no regressions against trunk (r7078), possibly minor
performance improvements in update (due to fewer heap
allocations), but increased stack usage.
Applied the following patches:
* maxpath_str for reentrancy (temporary fix, reverted)
* path: start working on thread-safe variants of these methods
* Re-entrancy work on path/character-set conversions
* directory.c: exploreDirectory() use reentrant functions here
* directory/update: more use of reentrant functions + cleanups
* string_toupper: a strdup-less version of strDupToUpper
* get_song_url: a static-variable-free version of getSongUrl()
* Use reentrant/thread-safe get_song_url everywhere
* replace rmp2amp with the reentrant version, rmp2amp_r
* Get rid of the non-reentrant/non-thread-safe rpp2app, too.
* buffer2array: assert strdup() returns a usable value in unit tests
* replace utf8ToFsCharset and fsCharsetToUtf8 with thread-safe variants
* fix storing playlists w/o absolute paths
* parent_path(), a reentrant version of parentPath()
* parentPath => parent_path for reentrancy and thread-safety
* allow "make test" to automatically run embedded unit tests
* remove convStrDup() and maxpath_str()
* use MPD_PATH_MAX everywhere instead of MAXPATHLEN
* path: get rid of appendSlash, pfx_path and just use pfx_dir
* get_song_url: fix the ability to play songs in the top-level music_directory
git-svn-id: https://svn.musicpd.org/mpd/trunk@7106 09075e82-0dd4-0310-85a5-a0d7c8717e4f
this was caused by always having the handleStoredPlaylist function return 0.
git-svn-id: https://svn.musicpd.org/mpd/trunk@7105 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This adds the following commands:
* queueid <id> Add song <id> to the queue.
* dequeue <pos> Remove song from <pos> from the queue
* queueinfo List the queue
To the statusfield it adds the following entry:
playlistqueue: <uid> UID can be used by clients to track changes in the playlist queue.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6927 09075e82-0dd4-0310-85a5-a0d7c8717e4f
the force flag will issue FATAL() if an invalid value is
specified
git-svn-id: https://svn.musicpd.org/mpd/trunk@6857 09075e82-0dd4-0310-85a5-a0d7c8717e4f
completely stopped. Instead, send them SIGSTOP to pause the process until
they're needed again. Then send them SIGCONT instead of re-spawning them.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6485 09075e82-0dd4-0310-85a5-a0d7c8717e4f
playback is stopped completely. This means the player process will no
longer have to wake up 100 times per second to see if it's been told to
start playing (the main process will just spawn a new player process when
it needs to). On the downside, this means an extra pair of forks() and the
re-initializing of the player and decode processes each time playback is
restarted.
git-svn-id: https://svn.musicpd.org/mpd/trunk@6446 09075e82-0dd4-0310-85a5-a0d7c8717e4f
relative paths in the DB or URLs. The main functional difference is that
playlistmove and playlistdelete will rewrite playlists in the correct
encoding and remove invalid lines instead of potentially modifying them.
The specific changes are:
appendSongToStoredPlaylist:
* Don't convert to FS charset
* Don't prepend music_directory if saving absolute paths
writeStoredPlaylistToPath:
* Convert to FS charset
* Prepend music_directory if saving absolute paths
loadStoredPlaylist:
* Make sure each line is either in the DB or a URL
loadPlaylist:
* Don't bother checking paths, since it's done in loadStoredPlaylist now
git-svn-id: https://svn.musicpd.org/mpd/trunk@6266 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Some compilers and linkers aren't smart enough to optimize this,
as global variables are implictly initialized to zero. As a
result, binaries are a bit smaller as more goes in the .bss and
less in the text section.
git-svn-id: https://svn.musicpd.org/mpd/trunk@5254 09075e82-0dd4-0310-85a5-a0d7c8717e4f
I'm checking for zero-size allocations and assert()-ing them,
so we can more easily get backtraces and debug problems, but we'll
also allow -DNDEBUG people to live on the edge if they wish.
We do not rely on errno when checking for OOM errors because
some implementations of malloc do not set it, and malloc
is commonly overridden by userspace wrappers.
I've spent some time looking through the source and didn't find any
obvious places where we would explicitly allocate 0 bytes, so we
shouldn't trip any of those assertions.
We also avoid allocating zero bytes because C libraries don't
handle this consistently (some return NULL, some not); and it's
dangerous either way.
git-svn-id: https://svn.musicpd.org/mpd/trunk@4690 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Add -Wmissing-prototypes if compiling with gcc
Static where possible
git-svn-id: https://svn.musicpd.org/mpd/trunk@4657 09075e82-0dd4-0310-85a5-a0d7c8717e4f
utf8ToFsCharset() and fsCharsetToUtf8() got very broken in r4311, and
resulted in several commits to fix those leaks. Unfortunately, not all
of those newly introduced leaks were fixed, nor was the result pretty.
Also, fixed a double-free in lsPlaylists(). This is very hard
to trigger (and therefore exploit) at the moment because we
check printDirectoryInfo() beforehand.
Intended behavior for utf8ToFsCharset() and fsCharsetToUtf8() as
God^H^H^Hshank originally intended is now documented in path.h
to prevent future errors like this.
mpd could still use some good valgrind testing before the 0.12.0
release.
<plug>In addition to reducing heap fragmentation, malloc
reductions from mpd-ke greatly reduces the chance of leaks from
happening due to programming errors.</plug>
git-svn-id: https://svn.musicpd.org/mpd/trunk@4639 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This shaves another 5-6k because we've removed the paranoid
fflush() calls after every fprintf. Now we only fflush()
when we need to
git-svn-id: https://svn.musicpd.org/mpd/trunk@4493 09075e82-0dd4-0310-85a5-a0d7c8717e4f
stripped binary size reduced by 9k on my machine from making
commandError a function. We'll print out error messages slightly
slower before, but the smaller binary is more than worth it.
git-svn-id: https://svn.musicpd.org/mpd/trunk@4488 09075e82-0dd4-0310-85a5-a0d7c8717e4f
This patch massively reduces the amount of heap allocations at
the interface/command layer. Most commands with minimal output
should not allocate memory from the heap at all. Things like
repeatedly polling status, currentsong, and volume changes
should be faster as a result, and more importantly, not a source
of memory fragmentation.
These changes should be safe in that there's no way for a
remote-client to corrupt memory or otherwise do bad stuff to
MPD, but an extra set of eyes to review would be good. Of
course there's never any warranty :)
No longer do we use FILE * structures in the interface, which means
we don't have to allocate any new memory for most connections.
Now, before you go on about losing the buffering that FILE *
+implies+, remember that myfprintf() never took advantage of
any of the stdio buffering features.
To reduce the diff and make bugs easier to spot in the diff,
I've kept myfprintf in places where we write to files (and not
network interfaces). Expect myfprintf to go away entirely soon
(we'll use fprintf for writing regular files).
git-svn-id: https://svn.musicpd.org/mpd/trunk@4483 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Add a few new options for indent to try to make
things a bit cleaner
git-svn-id: https://svn.musicpd.org/mpd/trunk@4411 09075e82-0dd4-0310-85a5-a0d7c8717e4f
These are just warnings from sparse, but it makes the output
easier to read. I ran this through a quick perl script, but
of course verified the output by looking at the diff and making
sure the thing still compiles.
here's the quick perl script I wrote to generate this patch:
----------- 8< -----------
use Tie::File;
defined(my $pid = open my $fh, '-|') or die $!;
if (!$pid) {
open STDERR, '>&STDOUT' or die $!;
exec 'sparse', @ARGV or die $!;
}
my $na = 'warning: non-ANSI function declaration of function';
while (<$fh>) {
print STDERR $_;
if (/^(.+?\.[ch]):(\d+):(\d+): $na '(\w+)'/o) {
my ($f, $l, $pos, $func) = ($1, $2, $3, $4);
$l--;
tie my @x, 'Tie::File', $f or die "$!: $f";
print '-', $x[$l], "\n";
$x[$l] =~ s/\b($func\s*)\(\s*\)/$1(void)/;
print '+', $x[$l], "\n";
untie @x;
}
}
git-svn-id: https://svn.musicpd.org/mpd/trunk@4378 09075e82-0dd4-0310-85a5-a0d7c8717e4f
Yes, I know about what I said on IRC earlier, but this
is an allocation + free inside a loop.
git-svn-id: https://svn.musicpd.org/mpd/trunk@3953 09075e82-0dd4-0310-85a5-a0d7c8717e4f
#2) Fix qball's bug, basically when we try todo initial buffering in decodeStart(), sleep for a few seconds after each attempt to finish init'ing the connection and begin buffering.
git-svn-id: https://svn.musicpd.org/mpd/trunk@2236 09075e82-0dd4-0310-85a5-a0d7c8717e4f
also some slight optimizations to interfacePrintWithFD() and myfprintf()
git-svn-id: https://svn.musicpd.org/mpd/trunk@1503 09075e82-0dd4-0310-85a5-a0d7c8717e4f
not time played, for previous add a hack where we record the difference
in time between previous presses
git-svn-id: https://svn.musicpd.org/mpd/trunk@1440 09075e82-0dd4-0310-85a5-a0d7c8717e4f
vorbis comments are updated on the fly for streams
need to decode icy metadata
buffering of metadata needs to be hardened, to ensure that player has already read a particular metachunk or passed over it
git-svn-id: https://svn.musicpd.org/mpd/trunk@1358 09075e82-0dd4-0310-85a5-a0d7c8717e4f
