Commit Graph

102 Commits

Author SHA1 Message Date
Niklas Haas 8bd0dee531 osdep: rename MP_UNREACHABLE
It was pointed out on IRC that the name is misleading, since the actual
semantics of the macro is to assert first.
2021-11-03 15:15:20 +01:00
Niklas Haas c704824b45 osdep: add MP_UNREACHABLE
This seems to work on gcc, clang and mingw as-is, but I made it
conditional on __GNUC__ just in case, even though I can't figure out
which compilers we care about that don't export this define.

Also replace all instances of assert(0) in the code by MP_UNREACHABLE(),
which is a strict improvement.
2021-11-03 14:09:27 +01:00
wm4 654eef9b82 ring: remove this
The code is OK, and it could be restored if it's needed again. But it is
unused now, so remove it.
2020-08-29 13:12:32 +02:00
wm4 50177aaa3b dispatch: add strange mechanism for making worker threads responsive
This is probably a sin for the sake of user experience. See a following
commit that wires up f_decoder_wrapper with it.
2020-03-05 22:00:50 +01:00
wm4 ba45b67370 bstr: remove unused bstr_splitlines() function 2019-12-23 11:01:56 +01:00
Aman Gupta 1a006d6f06 misc: add jni helpers 2019-11-19 12:10:26 -08:00
wm4 c10ba5eb8e Use mp_log2() instead of av_log2() 2019-10-31 13:17:18 +01:00
wm4 6d92e55502 Replace uses of FFMIN/MAX with MPMIN/MAX
And remove libavutil includes where possible.
2019-10-31 11:24:20 +01:00
wm4 7ac622bc5f json: write NaN/Infinity float values as strings
JSON doesn't support these for some god-awful reason. (JSON would have
been so much better if it weren't based on JavaScript, the plague of
this world.)

We don't really care whether these specific values "round trip", so we
might as well write them in a standard-compliant way.

Untested. I was too lazy to even run this, but it probably works.

See #6691.
2019-10-25 00:30:04 +02:00
Philip Sequeira a7158ceec0 demux: sort filenames naturally when playing a directory / archive 2019-09-29 01:13:00 +03:00
wm4 5f83b6a5f5 rendezvous: fix a typo 2018-08-31 11:40:33 +02:00
wm4 29a51900c6 player: some further cleanup of the mp_cancel crap
Alway give each demuxer its own mp_cancel instance. This makes
management of the mp_cancel things much easier. Also, instead of having
add/remove functions for mp_cancel slaves, replace them with a simpler
to use set_parent function. Remove cancel_and_free_demuxer(), which had
mpctx as parameter only to check an assumption. With this commit,
demuxers have their own mp_cancel, so add demux_cancel_and_free() which
makes use of it.
2018-05-24 19:56:35 +02:00
wm4 1ad027e8fd thread_pool: add a helper function
The behavior of mp_thread_pool_queue() doesn't or shouldn't change, but
the new helper function requires touching its logic.
2018-05-24 19:56:35 +02:00
wm4 f0cc6ba18d thread_pool: move comments to .h file 2018-05-24 19:56:35 +02:00
wm4 a253c72dbb thread_tools: unify mp_cancel POSIX/win32 paths, add features
The OS specifics are merged because the resulting ifdeffery is not much
worse than the old ifdeffery, but the logic that is now shared is
becoming more complex.

Create all objects lazily. The intention is to make mp_cancel instances
cheaper. POSIX pipes and win32 Events are pretty heavy weight, and are
only needed in special situations.

Add a mechanism to "chain" mp_cancel instances. Needed by the later
commits for whatever reasons.

Untested on win32.
2018-05-24 19:56:35 +02:00
wm4 782e428284 misc: add linked list helpers
This provides macros for managing intrusive doubly linked lists.

There are many ways how to do those in a "generic" way in C. For example
Solaris style lists are pretty nice:

https://github.com/illumos/illumos-gate/blob/master/usr/src/uts/common/sys/list.h
https://github.com/illumos/illumos-gate/blob/master/usr/src/common/list/list.c

I even have an independent implementation of this, which could be ISC
licensed. But I think it's easier to vomit ~100 lines of preprocessor
garbage, which has a lower footprint, and I think it wins slightly on
the side of type safety, simplicity, and ease of use, even if it doesn't
look as magically nice.
2018-05-24 19:56:35 +02:00
wm4 094f6be600 thread_tools: minor simplification 2018-05-24 19:56:35 +02:00
wm4 31b78ad7fa misc: move mp_cancel from stream.c to thread_tools.c
It seems a bit inappropriate to have dumped this into stream.c, even if
it's roughly speaking its main user. At least it made its way somewhat
unfortunately to other components not related to the stream or demuxer
layer at all.

I'm too greedy to give this weird helper its own file, so dump it into
thread_tools.c.

Probably a somewhat pointless change.
2018-05-24 19:56:35 +02:00
wm4 d36b85cfdf json: add some non-standard extensions
Also clarify this and previously existing differences to standard JSON
in ipc.rst.
2018-05-24 19:56:34 +02:00
wm4 76bff1a000 json: format slightly nicer escape sequences
Make use the escape sequences allowed by JSON.

Also update the linked RFC to the newest one.
2018-05-24 19:56:34 +02:00
wm4 cec37e98d5 misc: move some helper code from client.c
(Slightly oddly function names, because I want to avoid starting them
with mpv_*, which is reserved for public API.)
2018-05-24 19:56:34 +02:00
wm4 1157f07c5b node: move a mpv_node helper from ipc.c to shared code
This particular one is needed in a following commit.
2018-05-24 19:56:34 +02:00
wm4 4fd3ad8d63 thread_pool: set thread name 2018-05-24 19:56:34 +02:00
wm4 a1ed1f8be0 thread_pool: make it slightly less dumb
The existing thread pool code is the most primitive thread pool
possible. That's fine, but one annoying thing was that it used a static
number of threads. Make it dynamic, so we don't need to "waste" idle
threads.

This tries to add threads as needed. If threads are idle for some time,
destroy them again until a minimum number of threads is reached.

Also change the license to ISC.
2018-05-24 19:56:34 +02:00
wm4 22c002138d misc: add a synchronization helper
This is almost like rendezvous(), except it allows async wakeup, and
does not require global state. It will be used by a later commit.

struct mp_waiter is intended to be allocated on the stack, and uses an
initializer including PTHREAD_MUTEX_INITIALIZER. This is the first case
in mpv that it uses PTHREAD_MUTEX_INITIALIZER for stack-allocated
mutexes. It seems POSIX still does not allow this formally, but since
POSIX is worth less than used toilet paper, I don't really care. Modern
OSes use futexes, which means you can make _every_ memory location a
lock, and this code tries to make use of it, without using OS specific
code.

The name of the source file is rather generic, because I intend to dump
further small helpers there (or maybe move mp_rendezvous() to it).
2018-05-24 19:56:34 +02:00
wm4 88f9506f42 dispatch: add an assert() 2018-05-24 19:56:33 +02:00
wm4 020730da0b player: remove in_dispatch field
(Not sure if worth the trouble, but it does seem less awkward.)
2018-04-18 01:17:42 +03:00
wm4 435bc003c0 dispatch: simplify, disallow recursive invocation
Recursive invocation was needed up until the previous commit. Drop this
feature, and simplify the code. It's more logical, and easier to detect
miuses of the API.

This partially reverts commit 3878a59e. The original reason for it was
removed.
2018-04-18 01:17:41 +03:00
wm4 ef402a1c8c command: use mpv_node helpers instead of duplicated code
They didn't exist yet when this code was added.

Completely untested.
2018-03-26 19:47:08 +02:00
wm4 694157e024 m_option: pretty print mpv_node for OSD
Somewhat useful for debugging. Unfortunately libass (or something else)
strips leading whitespace, making it look slightly more ugly than
necessary. Still an improvement.
2017-10-30 15:32:24 +01:00
wm4 f08ec22567 command: change demuxer-cache-state property to return multiple ranges
Even if the demuxer cache does not multiple ranges yet. This is to
reduce the pain should caching of multiple ranges ever be implemented.

Also change it from the sub properties stuff to return a mpv_node
directly, which is less roundabout. Sub-property access won't work
anymore, though.

Remove the seekable-start/-end fields as well, as they're redundant with
the ranges.

All this would normally be considered an API change, but since it's been
only a few days with no known users, change it immediately.

This adds some node.c helpers as well, as the code would be too damn
fugly otherwise.
2017-10-26 22:31:04 +02:00
wm4 d2bdb72b69 options: add a thread-safe way to notify option updates
So far, we had a thread-safe way to read options, but no option update
notification mechanism. Everything was funneled though the main thread's
central mp_option_change_callback() function. For example, if the
panscan options were changed, the function called vo_control() with
VOCTRL_SET_PANSCAN to manually notify the VO thread of updates. This
worked, but's pretty inconvenient. Most of these problems come from the
fact that MPlayer was written as a single-threaded program.

This commit works towards a more flexible mechanism. It adds an update
callback to m_config_cache (the thing that is already used for
thread-safe access of global options).

This alone would still be rather inconvenient, at least in context of
VOs. Add another mechanism on top of it that uses mp_dispatch_queue, and
takes care of some annoying synchronization issues. We extend
mp_dispatch_queue itself to make this easier and slightly more
efficient.

As a first application, use this to reimplement certain VO scaling and
renderer options. The update_opts() function translates these to the
"old" VOCTRLs, though.

An annoyingly subtle issue is that m_config_cache's destructor now
releases pending notifications, and must be released before the
associated dispatch queue. Otherwise, it could happen that option
updates during e.g. VO destruction queue or run stale entries, which is
not expected.

Rather untested. The singly-linked list code in dispatch.c is probably
buggy, and I bet some aspects about synchronization are not entirely
sane.
2017-08-22 15:50:33 +02:00
Niklas Haas 779031ad8f bstr: short-circuit bstr_equals on pointer equality
More efficient in cases where we're comparing a bstr against itself,
which can happen in e.g. the ICC profile code.
2017-08-03 21:48:25 +02:00
Niklas Haas 345bb193fe
vo_opengl: support loading custom user textures
Parsing the texture data as raw strings makes the textures the most
portable and self-contained. In order to facilitate different types of
shaders, the parse_user_shader interaction has been changed to instead
have it loop through blocks and call the passed functions for each valid
block parsed. This is more modular and also cleaner, with better code
separation.

Closes #4586.
2017-07-27 23:51:05 +02:00
Kevin Mitchell 0b412bb60b misc/ring: fix comment typo 2017-07-09 13:46:13 -07:00
wm4 66478cff14 bstr: short-circuit bstr_equals() 2017-07-02 12:43:18 +02:00
Raúl Peñacoba d39f3bcd3a
misc/json: avoid redundant condition
Closes #4415
2017-07-01 15:09:21 +01:00
wm4 36fadac750 ring: use 64 bit counters
Apparently, this messes up on wraparound (although it shouldn't). After
2^32 bytes an audio "blip" happens with AOs that use this ringbuffer.

Whatever, we can fix this by switching to a 64 bit counter. There's also
a good chance the ringbuffer will be dropped completely, so don't waste
more time on this.
2017-06-28 18:43:14 +02:00
wm4 ec3dd7164c misc: add a thread pool
To be used by the following commits.
2017-04-01 20:32:01 +02:00
Akemi 8bbdecea83 osx: consistent normalisation when searching for external files
several unicode characters can be encoded in two different ways, either
in a precomposed (NFC) or decomposed (NFD) representation. everywhere
besides on macOS, specifically HFS+, precomposed strings are being used.
furthermore on macOS we can get either precomposed or decomposed
strings, for example when not HFS+ formatted volumes are used. that can
be the case for network mounted devices (SMB, NFS) or optical/removable
devices (UDF). this can lead to an inequality of actual equal strings,
which can happen when comparing strings from different sources, like the
command line or filesystem. this makes it mainly a problem on macOS
systems.

one case that can potential break is the sub-auto option. to prevent
that we convert the search string as well as the string we search in to
the same normalised representation, specifically we use the decomposed
form which is used anywhere else.

this could potentially be a problem on other platforms too, though the
potential of occurring is very minor. for those platforms we don't
convert anything and just fallback to the input.

Fixes #4016
2017-02-02 16:21:04 +01:00
wm4 06c55ac6c3 charset_conv: fallback to interpreting subs as latin1 if iconv fails
For display purposes, it's better to show scrambled text - at least
that's a more actionable failure mode than spamming the terminal with
FFmpeg nonsense error messages.

This avoids the obnoxious and pointless

  "Invalid UTF-8 in decoded subtitles text; maybe missing -sub_charenc option"

FFmpeg error, which will be spammed on every single subtitle event. We
don't even have a -sub-charenc option, fuck FFmpeg.

Did I mention fuck FFmpeg yet? Because fuck FFmpeg.
2017-01-22 13:10:16 +01:00
wm4 310671e91a charset_conv: support minimum compatibility to utf8:... syntax
Because it's the most commonly used one, and trivial to support.
2017-01-22 13:06:36 +01:00
wm4 4adfde5dd1 options: drop deprecated --sub-codepage syntax 2017-01-19 15:46:59 +01:00
wm4 35716b53db charset_conv: fix "auto" fallback with uchardet not compiled
Tried to open iconv with "auto" as source codepage, instead of using
the latin1 fallback. This also neutralizes the libavcodec dumbass
UTF-8 check, which discards subtitles not in UTF-8 and shows an
error message for ffmpeg CLI instead.

Fixes #3954.
2016-12-28 15:31:25 +01:00
wm4 c21c68edd4 bstr: change to LGPL
Was started by Uoti Urpala in commit 5f631d1c. Although it was made part
of demux_mkv.c, it's quite obvious that it's not based on any
pre-existing demux_mkv.c code (or ebml.c/.h for that matter). Anyone
else who has touched this code every since has already agreed to LGPL
relicensing.
2016-12-11 18:09:44 +01:00
wm4 c324bfab59 charset_conv: simplify and change --sub-codepage option
As documented in interface-changes.rst. This makes it much easier to
follow what the heck is going on.

Whether this is adequate for real-world use is unknown.
2016-12-09 19:51:29 +01:00
wm4 0eb87e1baf charset_conv: drop enca and libguess support
Enca is dead. libguess is relatively useless due to not having an
universal detection mode. On the other hand, libuchardet is actively
developed.

Manpages changes in the following commit.
2016-12-09 19:48:59 +01:00
wm4 4395a4f837 player: don't enter playloop for client API requests
This _actually_ does what commit 8716c2e8 promised, and gives a slight
performance improvement for client API users which make a lot of
requests (like reading properties).

The main issue was that mp_dispatch_lock() (which client.c uses to get
exclusive access to the core) still called the wakeup callback, which
made mp_dispatch_queue_process() exit. So the playloop got executed
again, and since it does a lot of stuff, performance could be reduced.
2016-09-16 20:24:52 +02:00
wm4 17e3e800e1 dispatch: fix a race condition triggering an assert()
If we were waiting, and then exiting due to timeout, we still have to
recheck the condition protected by the condition variable/mutex in order
to get back to a consistent state. In this case, the queue was locked
with mp_dispatch_lock(), and mp_dispatch_queue_process() got to return
without waiting for unlock.

Also caused commit 8716c2e8. Probably an argument for replacing the
dispatch queue by a simple mutex.
2016-09-16 16:11:33 +02:00
wm4 8716c2e88f player: use better way to wait for input and dispatching commands
Instead of using input_ctx for waiting, use the dispatch queue directly.
One big change is that the dispatch queue will just process commands
that come in (e.g. from client API) without returning. This should
reduce unnecessary playloop excutions (which is good since the playloop
got a bit fat from rechecking a lot of conditions every iteration).

Since this doesn't force a new playloop iteration on every access, this
has to be enforced manually in some cases.

Normal input (via terminal or VO window) still wakes up the playloop
every time, though that's not too important. It makes testing this
harder, though. If there are missing wakeup calls, it will be noticed
only when using the client API in some form.

At this point we could probably use a normal lock instead of the
dispatch queue stuff.
2016-09-16 14:49:23 +02:00