When seeking near the end of the file and the next file loads, seeking
continues on the next file at the same position and then immediately
the file after that. This patch stops slider seeking when a new file is
loaded, which is the standard behavior of many other players.
The player fully restarts playback when the edition or disk title is
changed. Before this, the player tried to reinitialized playback
partially. For example, it did not print a new "Playing: <file>"
message, and did not send playback end to libmpv users (scripts or
applications).
This playback restart code was a bit messy and could have unforeseen
interactions with various state. There have been bugs before. Since it's
a mostly cosmetic thing for an obscure feature, just change it to a full
restart. This works well, though since it may have consequences for
scripts or client API users, mention it in interface-changes.rst.
When changing video filters during initialization, there was a small
time window where video was initialized, but playback restart was not
complete yet. In this time window, playback_pts is not set. But since
issue_refresh_seek() was using this, it could lead to no refresh being
done _if_ the "video" had only 1 frame (such as cover art).
Fix this by using get_current_time() instead, which is the current time
with corner cases such as ongoing loading or seeks taken into account.
See also the previous commit. Without that, get_current_time() could
return NOPTS during init.
Fixes#5831.
Fixes several issues playing back mpegts with video streams marked
as having "still images". For example, see this video which has
frames only every 6s: https://s3.amazonaws.com/tmm1/music-choice.ts
Changes include:
- start playback right away, without waiting for first video frame
- do not consider the sparse video stream in demuxer underrun detection
- do not require multiple video frames for the VO
- use audio as the master stream for demuxer metadata events
- use audio stream for playback time
Signed-off-by: Aman Gupta <aman@tmm1.net>
Fixes `--ytdl-format="dash-fastly_skyfire-video-363357330+dash-fastly_skyfire_sep-audio-363357330" https://vimeo.com/108650530`
This happened because the video track also had audio available and after
adding it expecting an audio-only track, there were no more tracks with video.
Could make it behave differently (and leak memory) in certain cases.
Basically, m_option_parse() randomly returns 0 or 1, but most time 1,
with the difference due to legacy reasons that don't matter anymore.
The "if (prop.name)" check is redundant, because an assert above it
implies that it never can be NULL.
Deduplicate some code for initializing the "prop" variable.
This gets rid of run_command() and its big switch statement, which was
an idiotically big function of almost 1000 lines.
The switch is replaced with a callback per command, and each command is
now implemented in its own function. Command IDs are not needed anymore,
so the mp_command_type enum disappears.
There should be no functional changes, but since this refactors 64
commands, regressions are possible.
The handler() parameter is void*, because in theory the input code is
supposed to be independent of the player core code. For example, you
should be able to reuse the command parser code for some other part of
mpv. In practice, the variable containing command list is defined in the
player core anyway, so you could say this doesn't work. But I'm still
trying to hold onto this idea, so I went with void*.
The plan is to remove the command ID enum. This will happen by replacing
the big switch statement in command.c with dispatching to per-command
callbacks. As preparation, remove uses of the command IDs outside of the
actual dispatching mechanism.
Also remove some instances of checking cmd->def for NULL. We now require
this always to be set.
Until recently, ao_lavc and vo_lavc started encoding whenever the core
happened to send them data. Since audio and video are not initialized at
the same time, and the muxer was not necessarily opened when the first
encoder started to produce data, the resulting packets were put into a
queue. As soon as the muxer was opened, the queue was flushed.
Change this to make the core wait with sending data until all encoders
are initialized. This has the advantage that we don't need to queue up
the packets.
The video timing code could just decide that EOF was reached before it
was displayed. This is not really a problem for normal playback (if you
use something like --keep-open it'd show the last frame anyway,
otherwise it'd at best flash it on screen before destroying the window).
But in encode mode, it really matters, and makes the difference between
having one frame more or less in the output file.
Fix this by waiting for the VO before starting the real EOF.
vo_is_ready_for_frame() is normally used to determine when the VO frame
queue has enough space to send a new frame. Since the VO frame queue is
currently at most 1 frame, it being signaled means the remaining frame
was consumed and thus sent to the VO driver. If it returns false, it
will wake up the playloop as soon as the state changes.
I also considered using vo_still_displaying(), but it's not reliable,
because it checks the realtime of the frame end display time.
Instead of using an internal counter to keep track of the value that was
set last, attempt to find the current value of the property/option in
the value list, and then set the next value in the list.
There are some potential problems. If a property refuses to accept a
specific value, the cycle-values command will fail, and start from the
same position again. It can't know that it's supposed to skip the next
value. The same can happen to properties which behave "strangely", such
as the "aspect" property, which will return the current aspect if you
write "-1" to it. As a consequence, cycle-values can appear to get
"stuck".
I still think the new behavior is what users expect more, and which is
generally more useful. We won't restore the ability to get the old
behavior, unless we decide to revert this commit entirely.
Fixes#5772, and hopefully other complaints.
If playback has not been initialized yet (decoders not initialized
etc.), or if in idle mode, let the track properties cycle through "no"
and "auto". This should be slightly more helpful than making it simply
exit.
Depending on the stage of loading, more could be done. For example, if
youtube-dl loads additional subtitle files, it can happen that these get
added before the main file, and this could be cycled through to an
extent. This is probably too clever, and also sort of dangerous
(unintended interactions with messy in-loading state), so don't do it.
The main change is that we wait with opening the muxer ("writing
headers") until we have data from all streams. This fixes race
conditions at init due to broken assumptions in the old code.
This also changes a lot of other stuff. I found and fixed a few API
violations (often things for which better mechanisms were invented, and
the old ones are not valid anymore). I try to get away from the public
mutex and shared fields in encode_lavc_context. For now it's still
needed for some timestamp-related fields, but most are gone. It also
removes some bad code duplication between audio and video paths.
1. I want to get away from mp_image_params (maybe).
2. For encoding mode, it's convenient to get the nominal_fps, which is
a mp_image field, and not in mp_image_params.
Also rename stereo3d to stereo_in. The only real change is that the
vo_gpu OSD code now uses the actual stereo 3D mode, instead of the
--video-steroe-mode value. (Why does this vo_gpu code even exist?)
DR (letting the decoder allocate texture memory) requires running the
allocation on the render thread. This is rather hard with the render
API, because the user controls this thread and when it's entered. It was
not possible until now.
This commit adds a bunch of infrastructure to make this possible. We add
a new optional mode (MPV_RENDER_PARAM_ADVANCED_CONTROL) which basically
lets the user's render thread and libmpv agree how this should be done.
Misuse would lead to deadlocks. To make this less likely, strictly
document thread safety/locking issues. In particular, document which
libmpv functions can be called without issues. (The rest has to be
assumed unsafe.)
The worst issue is destruction of the render context while video is
still active. To avoid certain unintended recursive locks (i.e.
deadlocks, unless we'd make the locks recursive), make the update
callback lock separate. Make "killing" the video chain asynchronous, so
we can do extra work while video is being destroyed.
Because losing wakeups is a big deal, setting the update callback now
triggers a wakeup. (It would have been better if the wakeup callback
were a parameter to mpv_render_context_create(), but too late.)
This commit does not add DR yet; the following commit does this.
Fundamentally, scripts are loaded asynchronously, but as a feature,
there was code to wait until a script is loaded (for a certain arbitrary
definition of "loaded"). This was done in scripting.c with the
wait_loaded() function.
This called mp_idle(), and since there are commands to load/unload
scripts, it meant the player core loop could be entered recursively. I
think this is a major complication and has some problems. For example,
if you had a script that does 'os.execute("sleep inf")', then every time
you ran a command to load an instance of the script would add a new
stack frame of mp_idle(). This would lead to some sort of reentrancy
horror that is hard to debug. Also misc/dispatch.c contains a somewhat
tricky mess to support such recursive invocations. There were also some
bugs due to this and due to unforeseen interactions with other messes.
This scripting stuff was the only thing making use of that reentrancy,
and future commands that have "logical" waiting for something should be
implemented differently. So get rid of it.
Change the code to wait only in the player initialization phase: the
only place where it really has to wait is before playback is started,
because scripts might want to set options or hooks that interact with
playback initialization. Unloading of builtin scripts (can happen with
e.g. "set osc no") is left asynchronous; the unloading wasn't too robust
anyway, and this change won't make a difference if someone is trying to
break it intentionally. Note that this is not in mp_initialize(),
because mpv_initialize() uses this by locking the core, which would have
the same problem.
In the future, commands which logically wait should use different
mechanisms. Originally I thought the current approach (that is removed
with this commit) should be used, but it's too much of a mess and can't
even be used in some cases. Examples are:
- "loadfile" should be made blocking (needs to run the normal player
code and manually unblock the thread issuing the command)
- "add-sub" should not freeze the player until the URL is opened (needs
to run opening on a separate thread)
Possibly the current scripting behavior could be restored once new
mechanisms exist, and if it turns out that anyone needs it.
With this commit there should be no further instances of recursive
playloop invocations (other than the case in the following commit),
since all mp_idle()/mp_wait_events() calls are done strictly from the
main thread (and not commands/properties or libmpv client API that
"lock" the main thread).
ffmpeg marks audio tracks which are not meant to be played standalone
as DEPENDENT. these are typically used in DVB broadcasts for audio
descriptions, and are meant to be mixed into the main audio track during
playback.
This was slightly broken: since mp_initialize() did not necessarily
interrupt core_thread() (which is waiting for initialization), it did
not enter mp_play_files(), which would have sent an IDLE event.
I suppose that in some cases (like with mpv-android), the initial IDLE
event was never actually sent, because the first wakeup of the core
thread happens with the "loadfile" command, which will disallow the core
thread an IDLE event.
Although this was never observed to be happening, it seems definitely
possible: we first tell the main thread to exit, and then we ask the
main thread to do some work for us (with mp_dispatch_run()). Obviously
this is racy, and the main thread could exit without doing this, which
would block mp_dispatch_run() forever.
Fix this by changing the order of operation, so that it makes sense.
We could also just store the pthread_t of the main thread in some
variable, but the fact that pthread_create() might set the pthread_t
argument _after_ starting the thread makes this a bit messy (at least it
doesn't seem to be guaranteed on a superficial look at the manpage).
Until recently, the AO was reinitialized strictly only on decoder format
changes. But the commit for simplifying audio format negotiation removed
this. Now the AO is recreated for any format change.
This is sort of annoying if you change playback speed. The
insertion/removal of af_scaletempo can change the sample format. For
example, the acompressor filter will convert output to double, so
toggling scaletempo will force the format back to float. This recreates
the AO under the --gapless-audio=weak default. This likely affects a lot
of other filters too.
Work this around by allowing sample format changes, and keeping the
current AO format in these cases. This is probably not a big problem.
Most audio APIs force the output format to float anyway.
This means you actually have to worry about what the default gapless
mode does to your audio. If you start with a file that uses 8 bit per
sample, and then continue playing a 24 bit FLAC, it will be converted
down to 8 bit per sample. (Assuming they are played in a way that uses
the gapless logic.)
When dump's argument is an array, it was displaying <VISITED> for all
the array's object elements (objects, arrays, etc), regardless if they're
actually visited or not.
The reason is that we try to stringify twice: once normally which may
throw (on cycles), and a second time while excluding visited items which
is indicated by binding the replacer to an empty array - in which we hold
the visited items, where the replacer tests if its 'this' is an array or
not and acts accordingly.
However, its "this" may also be an array even if we don't bind it to one,
because its "normal" this is the main stringified object, so the test of
Array.isArray(this) is true when the top object is an array, and the object
items are indeed are in it - so the replacer considers them visited.
Fix by binding to null on the first attempt such that "this" is an array
only when we want it to test for visited items and not when the argument
itself is an array.
Due to earlier misinterpretation of the Lua docs as if mp.register_idle
registers a one-shot callback, the JS docs suggested to use setTimeout.
But the behavior and Lua docs are such that it's a repeating callback
which fires just before the script thread goes to sleep.
Implement it for JS too.
There was a "generic" function to run a hook and to wait for its
completion, yet there were two duplicated functions doing the same
anyway. Replace them with a single function.
They differed in how stop_play was handled, but it was broken anyway.
stop_play is set when playback is stopped due to quitting or changing
the playlist entry - but we still can't stop hook processing, because
that would mean asynchronously doing something else while the user hook
code is still busy and might still have the expectation that running the
hook stops everything else. So not waiting until the hook ends properly
is against the whole hook idea. That this was done inconsistently is
even worse. (Though it could be argued that when quitting the player,
everything should just be stopped violently. But I still think that's
up to the hook handler.)
process_hooks() does not return anything, since hook processing doesn't
really have a result (it's all about blocking and letting some other
code synchronously do something). Just let the caller check whether
loading was aborted in the meantime.
Also change the potentially misleading name of mp_hook_run().
As it turns out, there are multiple libmpv users who saw a need to
use the hook API. The API is kind of shitty and was never meant to be
actually public (it was mostly a hack for the ytdl script).
Introduce a proper API and deprecate the old one. The old one will
probably continue to work for a few releases, but will be removed
eventually.
There are some slight changes to the old API, but if a user followed
the manual properly, it won't break.
Mostly untested. Appears to work with ytdl_hook.
Move all of this stuff to a common function. This makes the error
messages less specific, but I don't think anyone will miss it.
The OSD flag handling is annoying, but it's nothing that should be
changed with this commit.
I think this will help with reducing code duplication (see following
commit). The error messages loses the multiplication factor, but the
error message will be replaced by a generic one in the following commit
anyway.
Hardware decoding things often need access to additional handles from
the windowing system, such as the X11 or Wayland display when using
vaapi. The opengl-cb had nothing dedicated for this, and used the weird
GL_MP_MPGetNativeDisplay GL extension (which was mpv specific and not
officially registered with OpenGL).
This was awkward, and a pain due to having to emulate GL context
behavior (like needing a TLS variable to store context for the pseudo GL
extension function). In addition (and not inherently due to this), we
could pass only one resource from mpv builtin context backends to
hwdecs. It was also all GL specific.
Replace this with a newer mechanism. It works for all RA backends, not
just GL. the API user can explicitly pass the objects at init time via
mpv_render_context_create(). Multiple resources are naturally possible.
The API uses MPV_RENDER_PARAM_* defines, but internally we use strings.
This is done for 2 reasons: 1. trying to leave libmpv and internal
mechanisms decoupled, 2. not having to add public API for some of the
internal resource types (especially D3D/GL interop stuff).
To remain sane, drop support for obscure half-working opengl-cb things,
like the DRM interop (was missing necessary things), the RPI window
thing (nobody used it), and obscure D3D interop things (not needed with
ANGLE, others were undocumented). In order not to break ABI and the C
API, we don't remove the associated structs from opengl_cb.h.
The parts which are still needed (in particular DRM interop) needs to be
ported to the render API.
Julian
fatalis
wm4
Aman Gupta
sfan5
Ricardo Constantino
Avi Halachmi (:avih)