This reverts commit 75dd3ec210.
This broke seeking with ordered chapters in some situations. While
the reverted commit was perfectly fine for playback of normal files,
it overlooked that in the ordered chapters case switching segments
actually reinitialized the audio chain completely, including the
decoder. And decoders still read packets on initialization. We can
restore the original commit as soon as decoders stop doing this.
This AO pretended to support volume operations when in spdif passthrough
mode, but actually did nothing. This is wrong: at least the GET
operations must write their argument. Signal that volume is unsupported
instead.
This was probably a hack to prevent insertion of volume filters or so,
but it didn't work anyway, while recovering after failed volume filter
insertion does work, so this is not needed at all.
It is possible to have ao->reset() called between ao->pause() and
ao->resume() when seeking during the pause. If the underlying PCM
supports pausing, resuming an already reset PCM will produce an error.
Avoid that by explicitly checking PCM state before calling
snd_pcm_pause().
Signed-off-by: wm4 <wm4@nowhere>
The player didn't quit when the end of a file was reached. The reason
for this is that jack reported a constant audio delay even when all
audio was done playing. Whether that was recognized as EOF by the player
depended whether the exact value was higher or lower than the player's
threshhold for what it considers no more audio.
get_delay() should return amount of time it takes until the last sample
written to the audio buffer reaches the speaker. Therefore, we have to
track the estimated time when the last sample is done, and subtract it
from the calculated latency. Basically, the latency is the only amount
of time left in the delay, and it should go towards 0 as audio reaches
ths speakers.
I'm not sure if this is correct, but at least it solves the problem. One
suspicious thing is that we use system time to estimate the end of the
audio time. Maybe using jack_frame_time() would be more correct. But
apart from this, there doesn't seem to be a better way to handle this.
MP_CMD_COMMAND_LIST commands (used to implement key bindings with
multiple commands) were not checked for abort commands. Implement it.
Remove the remarks about multi-commands being special from the manpage.
Seek coalescing is handled differently now, and the issue with abort
commands is fixed with this commit.
When timeline was used, and the --start option was not used, the initial
seek (needed to switch to the first timeline segment) seeked to -1 due
to an oversight.
Conflicts:
player/loadfile.c
Set subtitle resolution to video resolution when avctx->width and
avctx->height are zero.
This can happen with broken vobsubs that have no size set in their
.idx file (or Matroska extradata). At least with the test file provided
in issue #551, using the video resolution as fallback instead of what
guess_resolution() does is better.
Note that these files clearly are broken. It seems this particular
file was created by trying to use ffmpeg to transcode DVB subtitles
to vobsub, and ffmpeg "forgot" to set the subtitle resolution in the
destination file. On the other hand, ffmpeg DVB and PGS decoders set
the resolution on the first subtitle packet (or somewhere close), so
it's not really clear what to do here.
Closes#551.
Signed-off-by: wm4 <wm4@nowhere>
Patch by xylosper, rewritten commit message by wm4.
It seems that it was causing issues with certain perl setups (such as
the one on issue #549). It also turns out that it was not behaving correctly
(not all constants were being promoted to big nums as they should), so we
use explicit objects to derive the constants.
There were also precedence issues. I wonder if this even worked right to
begin with.
The 'double' path (8-byte floats) is untested, as I couldn't easily find
a file with such a field.
Closes#549.
Cygwin's libc (newlib) doesn't obey a lot of unix feature test macros,
including _GNU_SOURCE; as a result, a lot of functions and defines get
masked out -- important defines such as M_PI and strcasecmp. Work around
it by undefining __STRICT_ANSI__ on cygwin systems.
This will still cause compilation issues on any non-cygwin system that
uses newlib, but hopefully nobody does that, or if they do, they will
find this commit message and know to add -U__STRICT_ANSI__ to their
CFLAGS. Hopefully.
The mplayer decoder (spudec.c) actually handled this. There was explicit
code for binary palettes (16 32 bit values), and the subtitle resolution
was handled by video resolution coincidentally matching the subtitle
resolution.
Whoever puts vobsub into mp4 should be punished.
Fixes the sample gundam_sample.mp4, closes github issue #547.
The minimum required version was bumped in the old configure script, but
for the waf build system is was somehow forgotten or overlooked.
Probably happened while the waf build system was developed in a separate
branch.
Closes#546.
This is the same issue as addressed by 257d9f1, except this time for
the :srgb option as well. (257d9f1 only addressed :icc-profile)
The conditions of the srgb_compand mix() call are also flipped to
prevent an off-by-one error.
Use of these is "discouraged", but they're there to select these special
cases with the "aspect" property. They really should use some sort of
choice option type, but since it would be some work to make these work
with float values, the simple and dumb alternative was picked.
This sometimes happened when changing playback speed (= reinitializing
audio) after seeking of playback start. The assertion in audio.c:441 was
triggered, because buffer_playable_samples wasn't reset correctly when
the audio buffer was cleared or shortened. The assertion is correct and
should hold up any time.
I could not see any difference whatsoever, but for usage with a 3DLUT
there's zero performance difference so we might as well follow the spec to
the letter.
Legacy GL context creation (glXCreateContext) explicitly requires a X
visual, while the modern one (glXCreateContextAttribsARB) does not for
some reason. So fail only on the legacy code path if we don't find a
visual. Note that vo_x11_config_vo_window() will select a default visual
if a NULL visual is passed to it.
This fixes issue #504. For some reason, glXChooseFBConfig() will return
a fbconfig with no associated visual. (I'm not sure if this allowed.
They don't always have a visual, but since GLX_X_RENDERABLE is set
and GLX_DRAWABLE_TYPE is (implicitly) set to GLX_WINDOW_BIT, why would
there be no visual?)
Even worse, a test program seems to show that a 16 bit fbconfig is
selected (instead of 24/32 bit), which doesn't sound nice at all. Since
there _are_ better fbconfigs available, glXChooseFBConfig() should
normally sort them by quality, and return the better ones first. It's
worth noting that this function should also prefer GLX_TRUE_COLOR
over anything else, although this comes last in the sort order.
Whatever is going on, requesting GLX_X_VISUAL_TYPE with GLX_TRUE_COLOR
seems to fix it.
Use an arbitrary constant instead, which is as good as PATH_MAX.
This helps us to avoid having to think about pull request #523.
Also fix a case where a potentially signed char was passed to isspace().
This was done incorrectly in the previous commit: the fallback size used
the window size as requested with the first config call, which is the
size of the hidden window in the vo_opengl case. (That damn hidden
window again...)
Conflicts:
video/out/x11_common.c
This code essentially does nothing. As far as I could find out, this
actually used to do something. Then it was removed with commit efe7c39f,
leaving some leftover code that didn't do anything useful. This happened
12 years ago!
Also remove a commented debug printf.
vo_opengl creates a hidden X11 window to probe the OpenGL context. It
must do that before creating a visible window, because VO creation and
VO config are separate phases.
There's a race condition involving the hidden window: when starting with
--fs, and then leaving fullscreen, the unfullscreened window is
sometimes set to the aspect ratio of the hidden window. I'm not sure why
the window size itself uses the correct size (but corrupted by the wrong
aspect), but that's perhaps because the window manager is free to ignore
the size hint while honoring the aspect, or something equally messed up.
It turns out this happens because x11_common.c thinks the size of the
hidden window is the size of the unfullscreened window. This in turn
happens because vo_x11_update_geometry() reads the size of the hidden
window when called in vo_x11_fullscreen() (called from
vo_x11_config_vo_window()) when mapping the fullscreen window. At that
point, the window could be mapped, but not necessarily. If it's not
mapped, it will get the size of the unfullscreened window... I think.
One could fix this by actively waiting until the window is mapped. Try
to pick a less hacky approach instead, and never read the window size
until MapNotify is received.
vo_x11_create_window() needs a hack, because we'd possibly set the VO's
size to 0, resulting e.g. in vdpau to fail initialization. (It'll print
error messages until a proper resize is received.)
Conflicts:
video/out/x11_common.c
This fixes a weird bug with aspect ratio handling. It has to do with
float handling: with -std=gnu99, gcc implicitly enables broken non-
standard semantics giving float variables excess precision. This can for
example make this fail in theory: "float a = 0.1; assert(a == a);"
While standard C allows excess precision _within_ expressions, it
requires truncation when storing float values in variables of types
"float" or "double". The "gnu99" mode breaks this. It can be unbroken by
using "c99", or by specifying -fexcess-precision=standard. The former
seems less likely to break compilers other than modern gcc. Note that
-ffloat-store would also fix this, but also makes float expressions less
efficient and less precise for no reason.
The code that mistakenly fails because of this is dec_video.c line 393.
It caused the container aspect to be ignored in some or all situations,
depending how the compiler optimizes. For example, on gcc-4.6 with -Os,
the aspect is always ignored.
In future, we should probably just get rid of storing aspects as floats.
Note: requires mingw-w64 trunk for math.h fixes (M_PI was hidden).
Note that this still happens in the stream level, so we can't have
nice highlevel behavior restricting seeking. Instead, if a seek leads
to the demuxer requesting data outside of the cached range, the seek
will simply fail. This might confuse the demuxer, and the resulting
behavior is not necessarily useful.
Note that this also doesn't try to skip data on a forward seek. This
would just freeze the stream with slow unseekable streams.
One nice thing is that stream.h has a separate function for merely
skipping data (separate from seeking forward), which is pretty useful
in this case: we want skipping of data to work, even if we reject
seeking forward by skipping data as too expensive. This probably is
or will be useful for demux_mkv.c.
This used to work; I'm not sure when or why it regressed. When setting
AVProbeData.filename to NULL, libavformat will crash in rtp_probe() by
unconditionally accessing the string.
We used to set the filename to NULL to prevent probing by file extension
when we don't deem it as necessary. Using an empty string also works for
this purpose.
Larger sizes can introduce overflows, depending on the image format. In
the worst case, something larger than 16000x16000 with 8 bytes per pixel
will overflow 31 bits.
Maybe there should be a proper failure path instead of a hard crash, but
not yet. I imagine anything that sets a higher image size than a known
working size should be forced to call a function to check the size (much
like in ffmpeg/libavutil).
Not everything in the OSD path handles 0x0 sized sub-bitmaps well. At
least the code implementing --sub-gray had severe problems with it.
Fix this by skipping such bitmaps.
RGB565 is one of the fastest and most supported formats on low end consumer
devices, but ffmpeg spams warning when using it. Make it opt-in instead of
opt-out.
The problem seems to have solved itself. I guess the previous changes to
resizing and commit ba101ab made this possible. Consider me happy for removing
that crap.
Windows applications that use LoadLibrary are vulnerable to DLL
preloading attacks if a malicious DLL with the same name as a system DLL
is placed in the current directory. mpv had some code to avoid this in
ao_wasapi.c. This commit just moves it to main.c, since there's no
reason it can't be used process-wide.
This change can affect how plugins are loaded in AviSynth, but it
shouldn't be a problem since MPC-HC also does this and it's a very
popular AviSynth client.
Windows users expect this when a program crashes. Without it, the
program just disappears. Also change the SetErrorMode call to use macros
instead of a hardcoded constant.
If, for some reason, the subtitle renderer attempts to render a
subtitle before SD_CTRL_SET_VIDEO_PARAMS was called, it passed a
value calculated from invalid values. This can happen with --vf=sub
and --start. The crash happens if 1. there was a subtitle packet that
falls into the timestamp of the rendered video frame, 2. the playloop
hasn't informed the subtitle decoder about the video resolution yet
(normally unneeded, because that is used for weird corner cases only,
so this code is a bit fuzzy), and 3. something actually requests a
frame to be drawn from the subtitle renderer, like with vf_sub.
The actual crash was due to passing NaN as pixel aspect to libass,
which then created glyphs with ridiculous sizes, involving a few
integer overflows and unchecked mallocs.
The sd_lavc.c and sd_spu.c cases probably don't crash, but I'm not
sure, and it's better fix them anyway.
Not bothering with sd_spu.c, this crap is for compatibility and will
be removed soon.
Note that this would have been no problem, had the code checked whether
SD_CTRL_SET_VIDEO_PARAMS was actually called. This commit adds such a
check (although it basically checks after using the parameters).
Regression since 49caa0a7 and 633fde4a.
This generally affects mp3 files that don't have any (or many) mp3
frames in the first 2 MB. 2 MB is the maximum probe size, and
libavformat returns a low probescore even if we give it the full 2 MB.
Trying to probe a larger buffer (or even the full file) doesn't work for
mysterious reasons.
The workaround consists in accepting a very weak probescore if the
format is detected as mp3 and we probed already 2 MB.
Restructure it a bit, so we can use the format hack list even if no mime
type applies. Shouldn't change anything functionally yet. Preparation
for the next commit.
Balance controls as used by mixer.c was broken, because af_pan.c stopped
accepting its arguments. We have to allow 0 channels explicitly. Also,
fix null pointer access if the matrix parameter is not used.
Regression from commit 82983970.
Some mpv builds identify with e.g. "mpv b'0.3.3' ". The version looks
like str() was called on a Python byte string. I couldn't reproduce it
on my machine (I tried with both Python 2 and 3), so I'm not exactly
sure what's going on here, but I'm hoping this commit does fix it.
Signed-off-by: wm4 <wm4@nowhere>
This merges pull request #496. The problem was that at least the
initialization of the distance[] array accessed af_fmtstr_table[]
entries that were out of bounds. Small cosmetic changes applied to
the original pull request.
Doesn't affect the generated code, but avoids confusion
in both humans and newer Coverity versions.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@36623 b3059339-0415-0410-9bf9-f77b7e298cf2
The struct we need to copy is actually a cdrom_msf0, not cdrom_msf.
Even though the kernel for no good reason reads it in as a
cdrom_msf struct, but only uses the part shared with cdrom_msf0 -
this is probably a kernel bug.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@36622 b3059339-0415-0410-9bf9-f77b7e298cf2
Starting a network stream could stall by executing uncacheable stream
control requests (STREAM_CTRL_GET_LANG and STREAM_CTRL_GET_DVD_INFO).
Being uncacheable means the player has to wait until the cache is done
reading the current block of data. These requests can't be cached
because they're too complicated, so the only way to avoid them is
special casing the DVD and Bluray streams (which are the only things
which need these requests), and not doing them in other cases.
(This is kind of inelegant, but so is the rest of the DVD/BD code.)
Matroska makes it pretty hard to resync correctly on broken files:
random data returns "valid" EBML IDs with a high probability, and when
trying to skip them it's likely that you skip a random amount of data
(instead of considering the element length invalid).
Improve upon this by skipping known level 1 elements only. Consider
everything else invalid and call the resync code. This might result in
annoying behavior when Matroska adds new level 1 elements, although it
won't be particularly harmful. Matroska doesn't really allow us to do
better (even mkvtoolnix explicitly checks for known level 1 elements).
Since we now don't always want to combine EBML element skipping and
resyncing, remove ebml_read_skip_or_resync_cluster(), and make
ebml_read_skip() more tolerant against skipping broken elements.
Also, don't resync when reading sub-elements, and instead do resyncing
when reading them results in an error.
Until now, corrupted files were detected if the size of an element (that
should be skipped) was larger than the remaining file. This still could
skip larger regions of the file itself if the broken size happened to be
within the file.
Change it so that it's never allowed to skip outside the parent's
element.
Seeking usually show the status on OSD. In terminal OSD mode, no status
is shown, because there is already a separate status line.
Unfortunately, the mechanism for showing the status was still active,
which forced showing no message while the code for showing seek status
was active.
Insane .ass subtitle scripts can cause severe slowdown (depending on the
speed of the machine, or the insanity of the script), so mention how to
test without subtitles. This is mainly to make the user aware that
subtitle rendering can be a problem. For longwinded explanation, there
isn't enough space.
For a long time the cocoa backend set the xinerama_x/y and used dx/dy from the
VO instance. This somewhat worked with some workarounds but wasn't really
what was supposed to be happening. Moreover 27e4360, which touched this
workaround introduced a regression.
New code doesn't set the xinerama_x/y values so that dx/dy are offsets in the
current screen (not a virtual screen composed of all the screens). The screen
reference detected during VOCTRL_UPDATE_SCREENINFO is also passed down to the
window initialization code.
Fixes#472
MicroDVD files _can_ contain real timestamps instead of frame timestamps
if they declare a FPS. But this seems to be rare, so ignore that if the
FPS happens to match with the libavformat microdvd parser's default FPS.
This might actually break files that declare 23.976 FPS, but the video
file is not 23.976 FPS, but the chance that this happens is probably
very low, and the commit fixes the more common breakage with 25 FPS
video.
Subtitle formats with frame based timing require using the video FPS to
compute proper subtitle timestamps. But it looks like the calculation to
do that was inversed.
On X11, if no wayland compositor is running, wl_list_init() will never
be called. This will cause destroy_display() to segfault when trying to
iterate over the list.
There are still some leaks from wayland-cursor stuff, but there is no way to
free the memory as user of the cursor library.
Conflicts:
video/out/wayland_common.c
The user_data is passed on add_listener and can later be changed with
set_user_data. But because we don't want to change it later and because it is
the same object remove the set_user_data call.
This might be a copy&paste leftover from the initial draft for the wayland
backend.
I added enough logic to never set ontop or fullscreen twitce.
This commit keeps also the size of the video if multiple videos are played.
If the ratio differs the width will be kept at the same size and only the
height changes.
libwayland-client contains the following code [1]:
runtime_dir = getenv("XDG_RUNTIME_DIR");
if (!runtime_dir) {
fprintf(stderr,
"error: XDG_RUNTIME_DIR not set in the environment.\n");
This means this message will unconditionally and unavoidably be printed
if XDG_RUNTIME_DIR is not set. Since mpv is a terminal program, and we
want to avoid unnecessary output, work it around by not attempting to
use wayland if this environment variable is not set.
[1] http://cgit.freedesktop.org/wayland/wayland/tree/src/wayland-client.c#n636
(cd0dccd01e16fa404e03974d30ded3aebdb1c4bc)
This commonly happens when initializing vo_opengl on a X11-only system.
Unfortunately, most wl_*_destroy() functions appear not to accept NULL
pointers, making partial deinitialization a pain: you have to add your
own NULL checks everywhere to avoid crashes.
xkb.context is uninitialized separately, because you can initialize it
just fine, even if the rest of input initialization fails.
When both libavresample and libswresample were detected, the script
enabled both at the same time. This is not supported; although nothing
bad happened apparently. Make the dependencies both mutually exclusive.
This was inconsistent: the actual statusline used [statusline] as
message prefix, while other parts of the terminal OSD used [cplayer]
(and MSGL_STATUS). This commit makes it consistent.
Note that we can't use mp_msg, because it's not async-signal safe (we
might be running other threads while forking, so only functions
specified to be async-signal safe can be called, and this doesn't
include stdio; mp_msg acquires a mutex too).
Also, always print a \n before running the program to flush the status
line. The effect is that a program running successfully as well as the
error message will effectively start on a new line.
Quvi subtitles are considered external subtitles (simply because they're
separate from the audio/video stream), but for the sake of subtitle
auto-selection, they should not be considered external.
Change this so that quvi subtitles are treated like muxed subtitles
(with default flag never set). This means subtitles won't be selected by
default, unless explicitly requested with --sid or --slang.
Because of this commit there were problems displaying the frmase in their right
order.
This reverts commit 96e75d234a.
Conflicts:
video/out/gl_wayland.c
video/out/wayland_common.h
The changes in the vo_wayland_ontop function have no effect on the workaround.
Somehow the problem just disappeared. I guess it is because of the new control
function in gl_wayland.c where the resize happens immediatly after the event
dispatch/flush.
demux_subreader.c contains the old MPlayer subtitle parser, and I have
absolutely no confidence in this (very crappy) code. There might be
one or two security risks associated with running that code on
arbitrary input.
This solves the issue where we would not receive any frame events. The
difference to my earlier tests is that now it looks like eglSwapBuffers uses
it's own event queue or something similiar along the lines. Becaues the
performance is the same as without any redraw callback.
At the moment there are visual glitches when we resize the window. This happens
because in wayland there a special function for resizing EGL windows.
To prevent the glitches move the egl_context to the wayland state in
wayland_common.h and add a new control function to gl_wayland.c to wrap the
vo_wayland_control function to check for resize events.
With the new control wrapper the glitches are gone and the resizing is fluid.
The reason a segmentation happend here was because we couldn't get the
requested minor version. The major version is enough for differentiating
between OpenGL 3 and OpenGL 2. If it fails there is still a fallback to any
version available.
Also add a warning if we use the fallback.
Streams like CDDA have special requirements in what quantities data can
be read: you can only read a sector at once, not more and not less. The
stream_peek() function didn't respect that and set less (used internal
buffer size of 2048 bytes, instead of CD sector size of 2352 bytes), so
no data was read and EOF was accidentally set, making playback with
cdda:// fail.
This is a regression since commit 9a723f, but that commit merely exposed
the issue (the redundant seek would clear the EOF flag).
Add check in old-configure as well. Reformat the check to use a maximum of 80
columns in the wscript.
Signed-off-by: Stefano Pigozzi <[email protected]>
Set the flag CODEC_FLAG_OUTPUT_CORRUPT by default. Note that there is
also CODEC_FLAG2_SHOW_ALL, which is older, but this seems to be ffmpeg
only.
Note that whether you want this enabled depends on the user. Some might
prefer that only good frames are output, while others want the decoder
to try as hard as possible to output _anything_. Since mplayer/mpv is
rather the kind of player that tries hard instead of being "clever", set
the new default to override libavcodec's default.
A nice way to test this is switching video tracks. Since mpv doesn't
wait for the next key frame, it'll start feeding the decoder with a
packet from the middle of the stream.
The Haali Matroska splitter is basically the reference implementation
for this crap, and it knows only:
application/vnd.ms-opentype
application/x-font-ttf
application/x-truetype-font
Two of them were missing in our code. One of them, "application/x-font",
is probably plain incorrect, but I can't really tell.
Also see: http://www.cccp-project.net/beta/test_files/fontsample.mkv
This is probably useful.
Note that this includes a small, stupid hack to prevent loading of the
config file if vf_lavfi is not available. The profile by default uses
vf_lavfi, and the config parser will output errors if vf_lavfi is not
available.
As another caveat, we install the example profile even if encoding is
disabled (though we don't load it, since this would print errors).
stream_read_complete() fails if the file is larger than the requested
maximum size. But input.c didn't check for this case, and no indication
that something went wrong was printed.
This reverts commit 877303aaa9.
The OpenGL 2.1 fallback for vo_opengl didn't work. Two things come
together: 1. trying to create an OpenGL 3.0 context will fail with
a GLXBadFBConfig error, and 2. X errors are fatal by default. Since
the reverted commit removed the X error handler, the mpv process was
killed, instead of continuing for the fallback.
(Note that this commit is not an exact inverse commit, since mp_msg
changed, but it does about the same thing.)
Apparently, Matroska packs TrueHD packets in a way lavc doesn't expect.
This broke decoding of some files [1] completely. A short look at the
libavcodec parser shows that parsing this ourselves would probably be
too much work, so make use of the libavcodec parser API.
[1] http://www.cccp-project.net/beta/test_files/mzero_truehd_sample.mkv
query_format was setting state even if wasn't the correct thing to do. Somehow
it worked by pure luck (until commit e6e6b88b6d).
Fix the initialization by setting state inside of reconfig.
Currently KDE will copy a media file into a temporary folder instead of
trying to stream it if a KIO slave location file is started. This change
will tell KDE to make mpv try to directly play the file. Perhaps the
proper flags should be added according to the individual enabled
features of the build but I suggest that be for the future.
Signed-off-by: wm4 <wm4@nowhere>
Hide --enable variants from [autodetect]'ed options and --enable/--disable
variants for [enable]'d/[disable]'d options. The hidden options are still
usable, just hidden for more readability.
If the utf8 string used to create the NSString for title was invalid utf8,
-stringWithUTF8String returned nil and triggered an assertion in Cocoa's
framework code.
Sanitize the utf8 string and if the sanitation wasn't enough just avoid
crashing by not setting a title.
Fixes#406
This applies the usual logic of resetting stream selections to default
when switching to a file with a different track layout. (This is to
prevent selecting random streams.)
Also, make sure that a track can't be selected twice. While this might
work in some situations, it certainly won't work with subtitles demuxed
from a stream.
Fixes#425.
This is relatively hacky, but it's Christmas, so it's ok. This does two
things: 1. allow selecting two subtitle tracks, and 2. include a hack
that renders the second subtitle always as toptitle. See manpage
additions how to use this.
Normally, there can be only one demuxer stream active for each demuxer
of an external file, but this assumption will be broken for multiple
subtitles support.
For some reason, this checked whether there are external tracks at all
before doing any seeks. Possibly this was to avoid multiple
get_main_demux_pts() calls, but calling this multiple times shouldn't be
too bad.
Normally we shouldn't load these files. But for some reason it was added
in commit b784346e some years ago, and disabling this hack would
probably be an inconvenience. So just print a warning.
The only thing that used mp_load_per_file_config() was inside
configfiles.c too, so remove the declaration from core.h and move the
function before its use.
Code for loading "[vo.vdpau]" profiles and similar. The messages printed
on loading change, but other than that, everything should behave about
the same.
Including the "." in the returned extension was too inconvenient. I
think originally, the semantics were supposed to work like in Python,
but screw this.
Also, return NULL instead of "" on failure (which is what its only user
actually seems to expect).
How embarrassing...
This code is inactive for all VOs other than vo_vdpau. For vo_vdpau,
this caused various issues, such as stuttering after about an hour of
running mpv; see github issue #403.
Note that this will print a difference even with perfect sync, because
the code queues the frames _between_ vsync, probably for error margin
(though I don't understand why it uses the exact values chosen).
I find this annoying. It's the reason common/version.c exists at all.
options.c did this for the user agent, which contains the version
number. Because not including version.h means you can't build the user
agent and use it in mp_default_opts anymore, do something rather awkward
in main.c to initialize the default user agent.
Make it so --msglevel extends previous --msglevel uses, instead of
overwriting them. Do this by literally appending the --msglevel option
value to the previous one.
There's a single mp_msg() in path.c, but all path lookup functions seem
to depend on it, so we get a rat-tail of stuff we have to change. This
is probably a good thing though, because we can have the path lookup
functions also access options, so we could allow overriding the default
config path, or ignore the MPV_HOME environment variable, and such
things.
Also take the chance to consistently add talloc_ctx parameters to the
path lookup functions.
Also, this change causes a big mess on configfiles.c. It's the same
issue: everything suddenly needs a (different) context argument. Make it
less wild by providing a mp_load_auto_profiles() function, which
isolates most of it to configfiles.c.
The TV code pretends to be part of stream/, but it's actually demuxer
code too. The audio_in code is shared between the TV code and
stream_radio.c, so stream_radio.c needs a small hack until stream.c is
converted.
This is pretty nasty, because FFmpeg/Libav is yet another library with a
global message callback. We do something with mutexes trying to get it
done, but of course we can't actually solve this problem. If more than
one library in a process use FFmpeg/Libav, only one of them will get log
messages.
Always pass around mp_log contexts in the option parser code. This of
course affects all users of this API as well.
In stream.c, pass a mp_null_log, because we can't do it properly yet.
This will be fixed later.
Until now, there were two functions to add input sources (stuff like
stdin input, slave mode, lirc, joystick). Unify them to a single
function (mp_input_add_fd()), and make sure the associated callbacks
always have a context parameter.
Change the lirc and joystick code such that they take store their state
in a context struct (probably worthless), and use the new mp_msg
replacements (the point of this refactoring).
Additionally, get rid of the ugly USE_FD0_CMD_SELECT etc. ifdeffery in
the terminal handling code.
The mp_select_decoders() function doesn't have a log context (and I
don't want to give it one), so get rid of the mp_msg error message by
enhancing the semantics such that the syntax error is replaced by a new
feature. Now doing "--ad=something" will enable all decoders in the
"something" module, same as "--ad=something:*". Pretty useless, but gets
rid of the annoyance.
Remove the nonsensical print_lock too.
Things that are called from the option validator are not converted yet,
because the option parser doesn't provide a log context yet.
This ended up a little bit messy. In order to get a mp_log everywhere,
mostly make use of the fact that va_surface already references global
state anyway.
This removes the messages printed on unknown pixel format messages.
Passing a mp_log to them would be too messy. Actually, this is a good
change, because in the past we often had trouble with these messages
printed too often (causing terminal spam etc.), and printing warnings or
error messages on the caller sides is much cleaner.
vd_lavc.c had a change earlier to print an error message if a decoder
outputs an unsupported pixel format.
This requires the caller to provide a mp_log in order to see error
messages. Unfortunately we don't do this in most places, but I guess we
have to live with it.
lcms2 has a global message callback for error reporting. If you don't
set this, these error messages are silently thrown away. I think we
still want the error messages, so we have to do dumb stuff to avoid
clashes. This doesn't handle the case if another library in the same
process sets the message callback, but at least this should exclude
possible memory errors when running multiple instances of mpv.
This has similar problems as the ALSA message callback, though in theory
we could use the Display handle to find the right mpv instance from the
global callback. It still wouldn't work if another library happens to
set the error handler at the same time. There doesn't seem much of an
advantage overriding the error handler (though it used to be required),
so remove it.
This could output additional, potentially useful error messages. But the
callback is global and not library-safe, and would require us to add
additional state. Remove it, because it's obviously too much of a pain.
Also, it seems ALSA prints stuff to stderr anyway.
I don't know under which circumstances this can error (other than a
broken format string). It seems it won't return an error code on I/O
errors, so maybe broken format strings are the only case. Either way,
don't continue if an error is returned.
We have certain race conditions coming from doing multiple fprintf()
calls (setting up colors etc.). I'm not sure whether it would be worth
changing to code such that we do only one fprintf() call (and assume
this synchronizes access), but considering it would be hard to do
(Windows compatibility, ...), and that stdio uses per FILE locks anyway,
this is simpler and probably not less efficient. Also, there's no
problem handling the weird statusline special case this way.
Note that mp_msg_* calls which are silent won't acquire the lock, and
acquiring the lock happens on actual output only (which is slow and
serialized anyway).
Basically, reimplement --msglevel. Instead of making the new msg code
use the legacy code, make the legacy code use the reimplemented
functionality.
The handling of the deprecated --identify switch changes. It temporarily
stops working; this will be fixed in later commits.
The actual sub-options syntax (like --msglevel-vo=...) goes away, but I
bet nobody knew about this or used this anyway.
Instead of making msg.c an ifdef hell for unix vs. windows code, move
the code to separate functions defined in terminal-unix.c/terminal-
win.c.
Drop the code that selects random colors for --msgmodule prefixes.
So you can pass a command as list of strings (each item is an argument),
instead of having to worry about escaping and such.
These functions also take an argument for the default command flags. In
particular, this allows setting saner defaults for commands sent by
program code.
Expose this to Lua as mp.send_commandv command (suggestions for a better
name welcome). The Lua version doesn't allow setting the default command
flags, but it can still use command prefixes. The default flags are
different from input.conf, and disable OSD and property expansion.
This comes with a real change in behavior: now the signal handler is set
only when the terminal input code is active (e.g. not with
--no-consolecontrols), but this should be ok.
Remove these because I'm too lazy to convert them to proper
STREAM_CTRLs. Considering that probably nobody uses radio://, caring
about this is a complete waste of time. I will add these commands back
if someone asks for them, but I don't expect this to happen.
Apparently this has been broken for a year or so. The were three
reasons for the breakage here:
1. The window dragging hack prevented any DOWN event from
passing through since it always returned before we even got
the button.
2. The window style had CS_DBLCLKS in its flags, so we did not
get any DOWN events when the OS had detected a double click
(instead expecting us to handle a DBL event).
3. We never sent any mouse buttons when mouse movement handling
was disabled.
request_channels has been deprecated for years (request_channel_layout
is the replacement), but it appears it's still needed despite the
deprecation at least on older libavcodec versions.
So still set request_channels, but to it with the avoption API, which
hides the deprecation warning. This should also prevent mpv getting
trashed when libavcodec happens to bump its major version.
In my opinion, config.h inclusions should be kept to a minimum. MPlayer
code really liked including config.h everywhere, though, even in often
used header files. Try to reduce this.
Since m_option.h and options.h are extremely often included, a lot of
files have to be changed.
Moving path.c/h to options/ is a bit questionable, but since this is
mainly about access to config files (which are also handled in
options/), it's probably ok.
The tmsg stuff was for the internal gettext() based translation system,
which nobody ever attempted to use and thus was removed. mp_gtext() and
set_osd_tmsg() were also for this.
mp_dbg was once enabled in debug mode only, but since we have log level
for enabling debug messages, it seems utterly useless.
Nothing actually used the returned length. Since the remaining time can
easily become 0 or negative (e.g. incorrectly estimated file duration),
the time_remaining function still needs 2 return values, though.
"-" could skip optional arguments. I think this was a pretty bad idea,
because it introduced a weird special case.
I'll remove the special syntax, but keep compatibility for the "seek"
and "screenshot" commands.
If input is not waiting for select(), writing to the wakeup pipe is
wasteful, and, if there are many wakeups, might even block the wakeup
threads if the pipe gets full.
However, if it's waiting for select(), the wakup pipe must be used to
unblock the waiting thread. Actually there's a small race condition: we
might determine that the main thread is in select(), and write to the
wakeup pipe (whether we do this while unlocked or locked doesn't really
matter). Then, the main thread might leave select() before reading from
the wakup pipe. This should be harmless, because at worst more wakeups
than needed happen, but never fewer.
In order to use bare mp_log contexts, I find myself creating dummy
structs (with a single "log" field) to use the MP_ERR() etc. macros,
which hardcode the idiom that a context struct has a log field. On the
other hand, just using mp_msg_log() is too much typing (and I want to
rename it to mp_msg() when the transition is done), so it seems nice to
have message printing macros that use mp_log directly.
There's no reason why we should do this. For some reason, the existing
code reset terminal handling to default after unloading a file, just to
initialize it again when loading a new file.
Might be related to github issue #412, although I don't think it helps,
since the default SIGTERM handler _should_ kill the mpv process.
(It's still a nice simplification, though.)
The OSD style settings depend on the PlayRes, simply because all style
values are implicitly scaled by the PlayResY in libass. Also, the OSC
changes the PlayResY in certain situations, so something could go wrong.
But not sure if this actually matters in practice.
This simplifies things, although it is slightly less efficient (probably
uses a bit more memory).
This also happens to fix that the OSC dropped the libass cache on every
frame.
This doesn't have much value. It can't be accessed by anything else than
the actual subtitle renderer (sd_ass.c). sd_ass.c could create the
renderer itself, except that we apparently want to save memory (and some
font loading time) when using ordered chapters or multiple subtitle
tracks.
The end of the current segment will be the end of the file if there is
no next segment. Normally, this didn't matter much, since UNIX files
allow seeking past the end of the file. But when opening files from
HTTP, this would print confusing error messages. So explicitly check for
EOF before trying to read a segment.
Add the mp_get_user_path() function, and make it expand special path
prefixes. Use it for some things in mpv which take filenames
(--input-config, --screenshot-template, opengl icc-profile suboption).
This allows accessing files in the mpv config dir without hardcoding the
config path by prefixing the path with ~~/. Details see manpage
additions.
Before this, they were displayed forever. Since some dvd screens seem
not to allow escaping from the still frame using the menu, this could
get you stuck forever.
On dvdnav, caching kind of works but not really. (Not our fault, at
least not fully. It's due to libdvdnav being slightly misdesigned; see
previous commit for some explanations.)
The TV code is implemented in the demuxer, and the stream implementation
is just a wrapper, so caching makes no sense here.
No idea why this was disabled. It was in the original MPlayer code,
which doesn't make much sense to me, because using the MPlayer stream
cache seems 100% broken due to design issues.
Not sure how this should work. Using libdvdnav with a large FIFO doesn't
make sense either: data and control commands use the same stream, so if
you want to send input to libdvdnav, you just have to read from
libdvdnav all the time to get a reaction, which is not compatible with
maintaining a buffer that could remain full for a long time.
I have no idea either whether this improves or worsens anything, though
it might be more correct.
EOF is a special case. Normally, the reader will block until the cache
thread has new data. Obviously we don't want to do this on EOF, because
we'd potentially block forever. On the other hand, EOF will put the
cache thread into a waiting state, so if EOF recovers, this will happen
at a "later" point. This is bad if there is some kind of external event
that ends the EOF condition. In this case, a steram_read() call would
still return EOF. Make it so that the reader waits at least for one
iteration of the cache trying to rad a new block.
Also adjust some debug messages to not print file positions in hex.
There was already something similar in the code that did the actual
seek, but I think seeking to the same position could still trigger an
actual seek due to weid interaction with the buffer.
Use the scaled video size (i.e. as shown on the window) as reference for
zoom. This is the easiest way to fix different width/height scale
factors as they happen when zooming video with a pixel aspect ratio
other than 1:1.
Also fix the unscaled mode, so that it 1. doesn't scale even with
--video-zoom, and 2. doesn't scale by small amounts when the video is
cropped by making the window smaller than the video.
Since dvb input is not seekable anyway, this would probably just flush
the stream buffer, but at least it'll avoid a bogus message about
seeking errors.
This code tried to pass a still monotonic (even if not strictly
monotonic) PTS to the player, but as a result it remained stuck on
the PTS before a reset (since the PTS was lower).
This readds a more or less completely new dvdnav implementation, though
it's based on the code from before commit 41fbcee. Note that this is
rather basic, and might be broken or not quite usable in many cases.
Most importantly, navigation highlights are not correctly implemented.
This would require changes in the FFmpeg dvdsub decoder (to apply a
different internal CLUT), so supporting it is not really possible right
now. And in fact, I don't think I ever want to support it, because it's
a very small gain for a lot of work. Instead, mpv will display fake
highlights, which are an approximate bounding box around the real
highlights.
Some things like mouse input or switching audio/subtitles stream using
the dvdnav VM are not supported.
Might be quite fragile on transitions: if dvdnav initiates a transition,
and doesn't give us enough mpeg data to initialize video playback, the
player will just quit.
This is added only because some users seem to want it. I don't intend to
make mpv a good DVD player, so the very basic minimum will have to do.
How about you just convert your DVD to proper video files?
The code to remove a section from the active section array wasn't
correct (it should have tried to copy the elements in reverse), so just
replace it with a macro that does the intended thing.
Add some debug output to print the section stack.
Until now, the player didn't care to drain frames on video reconfig.
Instead, the VO was reconfigured (i.e. resized) before the queued frames
finished displaying. This can for example be observed by passing
multiple images with different size as mf:// filename. Then the window
would resize one frame before image with the new size is displayed. With
--vo=vdpau, the effect is worse, because this VO queues more than 1
frame internally.
Fix this by explicitly draining buffered frames before video reconfig.
Raise the display time of the last frame. Otherwise, the last frame
would be shown for a very short time only. This usually doesn't matter,
but helps when playing image files. This is a byproduct of frame
draining, because normally, video timing is based on the frames queued
to the VO, and we can't do that with frames of different size or format.
So we pretend that the frame before the change is the last frame in
order to time it. This code is incorrect though: it tries to use the
framerate, which often doesn't make sense. But it's good enough to test
this code with mf://.
This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's
architecture that the VO is just a (terminating) video filter. It didn't
really work for us with respect to video timing (the "end" of the video
chain isn't really made for video timing, and making it do so would be
awkward), and now we're removing it entirely. We will be able to fix
some things, such as properly draining video on reconfiguration.
I don't think this has any reason to exist. It's likely that this used
to be required by the old direct rendering infrastructure. (See
git blame output.)
This should help fixing some issues (like not draining video frames
correctly on reinit), as well as decoupling the decoder, filter chain,
and VO code.
I also wanted to make the hardware video decoding fallback work properly
if software-only video filters are inserted. This currently has the
issue that the fallback is too violent, and throws away a bunch of
demuxer packets needed to restart software decoding properly. But
keeping "backup" packets turned out as too hacky, so I'm not doing this,
at least not yet.
Libav 9 still uses the unprefixed PIX_FMT_... symbols, but they will
probably be removed some time in the future.
There are some other deprecations we have yet to take care of, but
there are no clear replacements yet.
The previous RING_BUFFER_COUNT value, 64, would have ao_wasapi buffer 64
frames of audio in the ring buffer; a delay of 1280ms, which is clearly
overkill for everything. A value of 8 buffers 8 frames for a total of
160ms.
When get_space was converted to returning samples instead of bytes, a
unit type mismatch in get_delay's calculation returned bogus values. Fix
by converting get_space's value back to bytes.
Fixes playback with ao_wasapi when reaching EOF, or seeking past it.
There are 3 users of the image format option type: demux_raw,
vf_format, vf_noformat. Allow the hwaccel formats (like vdpau etc.)
in general, so that the filters can use it. This won't work for
demux_raw, so explicitly reject these formats there.
Remove the inconsistent, duplicated, and insufficient scale filter
insertion code, and do it in one place instead. This also compensates
for the earlier removal of vf_match_csp() (which was in fact duplicated
code).
The algorithm to determine where to insert a filter etc. is probably the
same, though it also comes with some changes that should make debugging
easier when trying to figure out why a chain is failing to configure.
Add an "in" pseudo filter, which makes insertion of conversion filters
easier. Also change the vf->reconfig signature. At a later point, I'll
probably change format negotiation such that the generic filter code
will choose the output format, so having separate in and out params will
be useful.
Reason: I never liked it being recursive. Generally, this seems to
cause more problems than trouble, and is less flexible for access
outside of the chain.
I don't think we need these flags anymore. Simplify the code and get rid
of the vf_format struct.
There still is the vf_format.configured field, but this can be replaced
by checking for a valid image format.
This adds vf_chain, which unlike vf_instance refers to the filter chain
as a whole. This makes the filter API less awkward, and will allow
handling format negotiation better.
This function improves automatic filter insertion, but this really
should be done by the generic filter code. Remove vf_match_csp() and all
code using it as preparation for that.
This commit temporarily makes handling of filter insertion worse for
now, but it will be fixed with the following commits.
If sys/soundcard.h is actually linux/soundcard.h then it supports only OSSv3
API. This may happen when OSSLIBDIR == /usr while forgetting to replace
sys/soundcard.h from glibc.
However, after fa620ff waf prefers native implementation which is inferior
on Linux. To fix try making waf prefer oss-audio-4front. It's quite unusual
to have 4Front OSS installed where native implementation is superior, anyway.
Signed-off-by: bugmen0t <@>
Make the false positives path also undef the 4Front define.
Signed-off-by: Stefano Pigozzi <[email protected]>
Fixes#396
Bug introduced by commit 6fb020f5. It doesn't always happen, since it is
caused by the playloop and cocoa UI code running in separate threads.
Fixes#398.
Uncompressed rar archives can be transparently opened, but the filename
the player doesn't have the direct filename (but something starting
with rar://... instead). This will lead to external subtitles not
being loaded.
This doesn't handle multi-volume rar files, but in that cases just use
the --autosub-match=fuzzy option.
Fixes#397 on github.
If only coreaudio was activativated and not cocoa, the build failed for
missing CoreFoundation.
Signed-off-by: Stefano Pigozzi <[email protected]>
Fixes#395
They didn't do anything.
vf_screenshot.c actually did release the previous image, but that's not
really required. At worst you could take a screenshot and get an old
frame when there's no new frame yet.
The --flip option flipped the image upside-down, by trying to use VO
support, or if not available, by inserting a video filter. I'm not sure
why it existed. Maybe it was important in ancient times when VfW based
decoders output an image this way (but even then, flipping an image is a
free operation by negating the stride).
One nice thing about this is that it provided a possible path for
implementing video orientation, which is a feature we should probably
support eventually. The important part is that it would be for free for
VOs that support it, and would work even with hardware decoding.
But for now get rid of it. It's useless, trivial, stands in the way, and
supporting video orientation would require solving other problems first.
In particular, this disables mpeg4. There are some files out there that
use GMC, a usually rarely used and ineffective feature, which is not
supported by most hardware decoders. In these cases the hw decoder
outputs garbage, while software decoding works perfectly fine. We can't
really fallback to software decoding in these cases, because we don't
know that something is wrong in the first place. I can't see any
advantages of hw decoding of mpeg4, so it's better to disable it.
This can be reproduced with:
mpv short.wav -af 'lavfi="aecho=0.8:0.9:5000|6800:0.3|0.25"'
An audio file that is just 1-2 seconds long should play for 8-9 seconds,
which audible echo towards the end.
The code assumes that when playing with AF_FILTER_FLAG_EOF, the filter
will either produce output, or has all remaining data flushed. I'm not
really sure whether this really works if there are multiple filters with
EOF handling in the chain. To handle it correctly, af_lavfi should retry
filtering if 1. EOF flag is set, 2. there were input samples, and 3. no
output samples were produced. But currently it seems to work well enough
anyway.
The new signature is actually closer to how it actually works, and
someone who is not familiar to the API and how it works might make fewer
fatal mistakes with the new signature than the old one. Pretty weird.
Do this to sneak in a flags parameter, which will later be used to flush
remaining data of at least vf_lavfi.
This will make af_channels output a channel layout that is compatible
with any destination layout. Not sure if that's a good idea though,
since the way the AO choses a layout is perhaps less predictable. On the
other hand, using the old MPlayer standard layouts doesn't make much
sense either. We'll see whether this improves or breaks someone's use
case.
Apparently this stopped working after some planar changes (broken format
negotiation). Radically change option parsing in an incompatible way.
Suggest alternatives to this filter, since it barely has any importance
anymore.
Normally, audio decoder don't have a decoder delay, so the code was
fine. But FFmpeg supports multithreaded decoding for some audio codecs,
which introduces such a delay.
The delay means that we won't get decoded audio for the first few
packets, and that we need to do something to get the trailing audio
still buffered in the decoder when reaching EOF.
Two changes are needed to deal with the delay:
- If EOF is reached, pass a "flush" packet to the decoder to return the
buffered audio. Such a flush packet is automatically setup when
calling mp_set_av_packet() with a NULL packet.
- Use the PTS returned by the decoder, instead of the packet's. This is
important to get correct timestamps for decoded audio. Ignoring this
would result into offsetting the audio playback time by the decoder
delay. Note that we can still use the timestamp of the first packet
to get the timestamp for the start of the audio.
If the timebase is set, it's used for converting the packet timestamps.
Otherwise, the previous method of reinterpret-casting the mpv style
double timestamps to libavcodec style int64_t timestamps is used.
Also replace the kind of awkward mp_get_av_frame_pkt_ts() function by
mp_pts_from_av(), which simply converts timestamps in a way the old
function did. (Plus it takes a timebase parameter, similar to the
addition to mp_set_av_packet().)
Note that this should not change anything yet. The code in ad_lavc.c and
vd_lavc.c passes NULL for the timebase parameters. We could set
AVCodecContext.pkt_timebase and use that if we want to give libavcodec
"proper" timestamps.
This could be important for ad_lavc.c: some codecs (opus, probably mp3
and aac too) have weird requirements about doing decoding preroll on the
container level, and thus require adjusting the audio start timestamps
in some cases. libavcodec doesn't tell us how much was skipped, so we
either get shifted timestamps (by the length of the skipped data), or we
give it proper timestamps. (Note: libavcodec interprets or changes
timestamps only if pkt_timebase is set, which by default it is not.)
This would require selecting a timebase though, so I feel uncomfortable
with the idea. At least this change paves the way, and will allow some
testing.
Don't bother explaining the sample format naming schema. The "ne" bit is
outdated anyway, and anyone who has to use this option will be able to
understand the naming schema just by looking at the names too.
This prevents waf from running test programs after compilation. A better
approach would be to only remove this option if the check actually errors,
but we are using this only for Lua anyway.
The vf_eq context contains a very large lookup table, and the method of
setting default values caused the vf_eq context to be included in the
compiled code.
vf_stereo3d now uses vf_lavfi, if mpv was compiled with libavfilter.
vf_swapuv is hereby undeprecated. It's too trivial to wrap it with
libavfilter, and it's also too useless that even typing this commit
message is not really worth the time to spend on it.
Just in case someone expects these are unchanged just because they're
not mentioned in changes.rst anywhere. Documenting all of these changes
would be too much work and not helpful either.
All filters now either use the generic option parser, or don't have
options. This finally finishes a transition started in 2003 (see git
commit 33b62af947).
Why are MPlayer devs so monumentally lazy? Sorry, but this takes the
cake. You had 10 years.
Mostly backwards compatible, we don't change much because we just want
to get rid of the legacy option string handling.
You can't pass an aspect as first argument anymore.
Apparently you can get this with: stereo3d=ab[2]{l,r}:sbs[2]{l,r}
So it seems the filter is redundant and can be removed.
Also see FFmpeg commit 2f11aa141a01.
Unfortunately, this forces filtering both luma and chroma, because
otherwise we'd have to deal with libavfilter's vf_noise weird handling
of YUV vs. RGB formats. Would we e.g. filter luma only, it would filter
red in RGB mode only, because it goes by component and there's no way to
distinguish YUV and RGB by just using the filter's options.
Also remove the ability to disable deinterlacing at runtime. You can
still disable deinterlacing at runtime by using the ``D`` key and its
automatical filter insertion/removal.
This will allow old filter to run libavfilter instead by calling
vf_lw_set_graph(), which turns the filter into a wrapper, using a given
libavfilter graph.
Later commits use that to automatically "reroute" a bunch of filters to
libavfilter. We want to get rid of the old MPlayer filter code, because
it's bad an unmaintained, but we still don't want to force everyone to
use vf_lavfi, so this solution will do for a while.
Previous API worked under the assumption that download_image is always called
after map_image. In practice this is true, but it's better to have a much
generic API that doesn't depend on the order in which the functions are called.
The hwdec driver can be loaded, even if it's not used (e.g. when playing
a file with no hardware decoding after one with it enabled).
Also, check whether dlimage is NULL. Since this will do call into the
native hwdec API, there's a chance a driver could fail doing this, it's
better to check the return value, even if this case currently can't
happen.
mpv was hardcoded to always consider the right Alt key as Alt Gr, but there
are parituclar combinations of platforms and keyboard layouts where it's more
convenient to treat the right Alt as a keyboard modifier just like the left
one.
Fixes#388
The harder work was done in the previous commits. After that this feature comes
out almost for free.
The only problem is I can't get the textures created with CGLTexImageIOSurface2D
to download properly, thus the code performs download using some CoreVideo APIs.
If someone knows why download of textures created with CGLTexImageIOSurface2D
doesn't work please contact me :)
This adds support for packed YUV formats (YUVY and UYVY) using the extension
GL_APPLE_rgb_422. While supporting this formats on their own is not that
important (considering most video is planar YUV) they are used for
interoperability with IOSurfaces.
Next commit will use this formats to render VDA hardware decoded frames through
IOSurface and OpenGL interoperability.
roaraudio has some sort of sndio emulation, but apparently its header
file is either blatantly broken, or an old version from the past. The
sio_onvol() function has the wrong return type (void instead of int),
and the SIO_DEVANY symbol is missing entirely. This broke the build,
because the configure check was successful anyway.
This allows vo_opengl to use GL_TEXTURE_RECTANGLE textures, either by
enabling it with the 'rectangle-textures' sub-option, or by having a
hwdec backend force it. By default it's off.
The _only_ reason we're adding this is because VDA can export rectangle
textures only.
We got a crash in libavutil when encoding with Y8 (GRAY8). The reason
was that libavutil was copying an Y8 image allocated by us, and expected
a palette. This is because GRAY8 is a PSEUDOPAL format. It's not clear
what PSEUDOPAL means, and it makes literally no sense at all. However,
it does expect a palette allocated for some formats that are not
paletted, and libavutil crashed when trying to access the non-existent
palette.
This is for key bindings that use multiple mouse buttons at once. (Yes,
this is weird, but MPlayer always had this feature, and apparently
there are people using it!)
Before this commit, clicking another mouse button while still holding
the previous mouse button forced the command bound to the previous
mouse button to be emitted. This is usually needed to make sure the
input consumer (the player and the OSC) stays in sync with the actual
mouse button state. If there's no command sent, the OSC in particular
would think the button is still held down. However, sending the command
is undesired behavior if you want to use these multiple-key binds.
Solve this by emitting commands in this situation only if a key down
command was sent earlier. Since mouse button key bindings are normally
executed on key-up only, this happens with special commands like
script_dispatch only (used by the OSD to track mouse buttons, but
also used for other OSC bindings).
See github issue #390.
This is not strictly needed anymore. (On the other hand, it's not really
possible to do hw decoding with vo_null, because the VO is still
responsible for opening the hw decoder API, but that's another story.)
There are some use cases for this. For example, you can use it to set
defaults of automatically inserted filters (like af_lavrresample). It's
also useful if you have a non-trivial VO configuration, and want to use
--vo to quickly change between the drivers without repeating the whole
configuration in the --vo argument.
The "run" input command does fork+exec to spawn child processes. But it
doesn't cleanup the child processes, so they are left as zombies until
mpv terminates. Leaving zombie processes around is not very nice, so
employ a simple trick to let pid 1 take care of this: we fork twice, and
when the first fork exits, the second fork becomes orphaned and becomes
pid 1's child. It becomes pid 1's responsibility to cleanup the process.
The advantage is that we don't need extra logic to cleanup the spawned
process, which could have an arbitrary lifetime.
This is e.g. described here: http://yarchive.net/comp/zombie_process.html
Also use _exit() instead of exit(). It's not really sane to run cleanup
handlers (atexit() etc.) inside a forked process.
This is needed so that new processes (created with fork+exec) don't
inherit open files, which can be important for a number of reasons.
Since O_CLOEXEC is relatively new (POSIX.1-2008, before that Linux
specific), we #define it to 0 in io.h to prevent compilation errors on
older/crappy systems. At least this is the plan.
input.c creates a pipe. For that, add a mp_set_cloexec() function (which
is based on Weston's code in vo_wayland.c, but more correct). We could
use pipe2() instead, but that is Linux specific. Technically, we have a
race condition, but it won't matter.
See the changes in input.rst for explanations.
Technically speaking, this also gets rid of some undefined behavior:
passing NULL as a vararg (execl()) is always a bug.
So e.g.
show_text abc#def
will now print "abc#def" instead of "#def". It's simpler, more
consistent with how ";" and other things are handled, and also
possibly avoids bothering the user with extra escaping.
The OSS checks were a big mess and quite buggy. This reimplementes them using
a declarative approach and clearly distinguishing between the various OSS
implementations. The code should now almost be auto-documenting.
We currently support the following implementations of OSS:
* platform-specific (with `sys/soundcard.h`)
* SunAudio (default on NetBSD and useable on OpenBSD even if we have sndio
support there).
* 4Front (default on FreeBSD)
Since now each OSS check also checks for the appropriate soundcard header,
remove the old soundcard check.
Many thanks to @bugmen0t for in depth info about all the BSDs.
Check #380 and #359 for more info on this commit.
PIX_FMT_* -> AV_PIX_FMT_* (except some pixdesc constants)
enum PixelFormat -> enum AVPixelFormat
Losen some version checks in certain newer pixel formats.
av_pix_fmt_descriptors -> av_pix_fmt_desc_get
This removes support for FFmpeg 1.0.x, which is even older than
Libav 9.x. Support for it probably was already broken, and its
libswresample was rejected by our build system anyway because it's
broken.
Mostly untested; it does compile with Libav 9.9.
This partially reverts commit 7d152965. It turns out that at least some
ALSA drivers (at least snd-hda-intel) report incorrect audio delay with
non-native sample rates, even if the sample rate is only very slightly
different from the native one.
For example, 48000Hz is fine on my hda-intel system, while both 8000Hz
and 47999Hz lead to a delay off by 40ms (according to mpv's A/V
difference display), which suggests that something in ALSA is
calculating the delay using the wrong sample rate.
As an additional problem, with ALSA resampling enabled, using
48001Hz/float/2ch fails, while 49000Hz/float/2ch or 48001Hz/s16/2ch
work. With resampling disabled, all these cases work obviously, because
our own resampler doesn't just refuse any of these formats.
Since some people want to use the ALSA resampler (because it's highly
configurable, supports multiple backends, etc.), we still allow enabling
ALSA resampling with an ao_alsa suboption.
These were confined to the video path, but resetting them even if no
video is available shouldn't really matter. Always resetting them makes
the logic easier to follow, I think.
Sometimes, vf_pullup hanged on seek. This was because it never was
properly reset. Old timestamps messed up the timestamp calculations,
which made the player show frames for a ridiculously long time, which is
perceived as pausing or hanging.
Recently, the check was moved, so it was printed only for source video
PTS (since that's easier, and filters should normally behave sane). But
it turns out it's trivial to print a warning in the filter case too by
reusing the code that normally checks for PTS forward jumps without
needing any additional code, so, fine, restore warning in this case.
The old ffmpeg vdpau support code uses separate vdpau pixel formats for
each decoder (pretty much because mplayer's architecture sucked), which
just gets into the way. Force the old decoder's output to IMGFMT_VDPAU,
and remove IMGFMT_IS_VDPAU() where we can remove it.
This should completely remove the differences betwene the old and new
vdpau decoder outside of the decoder.
PIX_FMT_VDA_VLD and PIX_FMT_VAAPI_VLD were never used anywhere. I'm not
sure why they were even added, and they sound like they are just for
compatibility with XvMC-style decoding, which sucks anyway.
Now that there's only a single vaapi format, remove the
IMGFMT_IS_VAAPI() macro. Also get rid of IMGFMT_IS_VDA(), which was
unused.
pthreads should be available anywhere. Even if not, for environment
without threads a pthread wrapper could be provided that can't actually
start threads, thus disabling features that require threads.
Make pthreads mandatory in order to simplify build dependencies and to
reduce ifdeffery. (Admittedly, there wasn't much complexity, but maybe
we will use pthreads more in the future, and then it'd become a real
bother.)
Introduce a mp_cmd_def struct to define commands, instead of using
mp_cmd for this. This way each command parameter can be a m_option,
instead of m_option plus some more stuff.
Define the ARG_ macros directly in terms of the OPT_ macros. Not sure if
this makes it easier to read (maybe not, even if it looks simpler), but
at least it makes it easier to add other option types.
Another idea was adding a name for each parameter (so you could have
named parameters), but not today.
The hr-seek code assumes that when seeking the demuxer, the first image
decoded after the seek will have a PTS exactly equal to the demuxer seek
target time, or before that target time. Incorrect timestamps,
implicitly dropped initial frames, or broken files/demuxers can all
break this assumption, and lead to hr-seek missing the seek target.
Generally, this is not much a problem (the user won't notice being off
by one frame), but it really shows when using the backstep feature. In
this case, backstepping would simply hang.
Add a simple hack that basically forces a minimal value for the --hr-
seek-demuxer-offset option (which is 0 by default) when doing a
backstep-seek. The chosen minimum value is arbitrary. There's no perfect
value, though in general it should perhaps be slightly longer than the
frametime, which the chosen value is more than enough for typical
framerates.
So, FFmpeg/Libav requires us to figure out video timestamps ourselves
(see last 10 commits or so), but the methods it provides for this aren't
even sufficient. In particular, everything that uses AVI-style DTS (avi,
vfw-muxed mkv, possibly mpeg4-in-ogm) with a codec that has an internal
frame delay is broken. In this case, libavcodec will shift the packet-
to-image correspondence by the codec delay, meaning that with a delay=1,
the first AVFrame.pkt_dts is not 0, but that of the second packet. All
timestamps will appear shifted. The start time (e.g. the time displayed
when doing "mpv file.avi --pause") will not be exactly 0.
(According to Libav developers, this is how it's supposed to work; just
that the first DTS values are normally negative with formats that use
DTS "properly". Who cares if it doesn't work at all with very common
video formats? There's no indication that they'll fix this soon,
either. An elegant workaround is missing too.)
Add a hack to re-enable the old PTS code for AVI and vfw-muxed MKV.
Since these timestamps are not reorderd, we wouldn't need to sort them,
but it's less code this way (and possibly more robust, should a demuxer
unexpectedly output PTS).
The original intention of all the timestamp changes recently was
actually to get rid of demuxer-specific hacks and the old timestamp
sorting code, but it looks like this didn't work out. Yet another case
where trying to replace native MPlayer functionality with FFmpeg/Libav
led to disadvantages and bugs. (Note that the old PTS sorting code
doesn't and can't handle frame dropping correctly, though.)
Bug reports:
https://trac.ffmpeg.org/ticket/3178https://bugzilla.libav.org/show_bug.cgi?id=600
Using --start with files that use DTS only, or which simply have broken
PTS timestamps, would incorrectly drop frames and possibly not execute
the seek correctly.
Add yet another heuristic to detect this. The intent is that --start and
hr-seeks in general should work correctly, but in order to keep things
fast, we still want to allow frame dropping during hr-seek if there are
no problems doing so. Do this by disabling frame dropping by default,
but re-enabling it if there are no problems found for a while. As a
consequence, --start might be somewhat slower, but normal user
interaction should remain as fast as before.
Note that there's something subtle about the added code: the
has_broken_packet_pts field is checked even before the first packet is
fed to dec_video.c, so the field must not be set to 0 right on start.
It's not initially set to 0 anyway, because the heuristic requires
decoding some images before enabling frame drop anyway.
Note 2: it's not clear whether frame dropping during hr-seek really
helps; I didn't benchmark it.
Previous code was using the values of the AudioChannelLabel enum directly to
create the channel bitmap. While this was quite smart it was pretty unreadable
and fragile (what if Apple changes the values of those enums?).
Change it to use a 'dumb' conversion table.
The d_video->pts field was a bit strange. The code overwrote it multiple
times (on decoding, on filtering, then once again...), and it wasn't
really clear what purpose this field had exactly. Replace it with the
mpctx->video_next_pts field, which is relatively unambiguous.
Move the decreasing PTS check to dec_video.c. This means it acts on
decoder output, not on filter output. (Just like in the previous commit,
assume the filter chain is sane.) Drop the jitter vs. reset semantics;
the dec_video.c determined PTS never goes backwards, and demuxer
timestamps don't "jitter".
Also get rid of the PTS check _after_ filters. This means if there's a
video filter which unsets PTS, no warning will be printed. But we assume
that all filters are well-behaved enough by now.
Refactor the PTS handling code to make it cleaner, and to separate the
bits that use PTS sorting.
Add a heuristic to fall back to DTS if the PTS us non-monotonic. This
code is based on what FFmpeg/Libav use for ffplay/avplay and also
best_effort_timestamp (which is only in FFmpeg). Basically, this 1. just
uses the DTS if PTS is unset, and 2. ignores PTS entirely if PTS is non-
monotonic, but DTS is sorted.
The code is pretty much the same as in Libav [1]. I'm not sure if all of
it is really needed, or if it does more than what the paragraph above
mentions. But maybe it's fine to cargo-cult this.
This heuristic fixes playback of mpeg4 in ogm, which returns packets
with PTS==DTS, even though the PTS timestamps should follow codec
reordering. This is probably a libavformat demuxer bug, but good luck
trying to fix it.
The way vd_lavc.c returns the frame PTS and DTS to dec_video.c is a bit
inelegant, but maybe better than trying to mess the PTS back into the
decoder callback again.
[1] https://git.libav.org/?p=libav.git;a=blob;f=cmdutils.c;h=3f1c667075724c5cde69d840ed5ed7d992898334;hb=fa515c2088e1d082d45741bbd5c05e13b0500804#l1431
These used the suffix _resync_stream, which is a bit misleading. Nothing
gets "resynchronized", they really just reset state.
(Some audio decoders actually used to "resync" by reading packets for
resuming playback, but that's not the case anymore.)
Also move the function in dec_video.c to the top of the file.
The code stopped at kAudioChannelLabel_TopBackRight and missed mapping for
5 more channel labels. These are in a completely different order that the mpv
ones so they must be mapped manually.
This includes the case when lavc decodes audio with more than 8
channels, which our audio chain currently does not support.
the changes in ad_lavc.c are just simplifications. The code tried to
avoid overriding global parameters if it found something invalid, but
that is not needed anymore.
When using PIC on x86 (eg with hardened toolchains) the ebx register is
reserverd and cannot be used in assembly code.
For vf_eq we allow the compiler to use memory as input.
For vf_noise we temoporarily borrow the ebp register.
This fixes#361.
Signed-off-by: Natanael Copa <[email protected]>
This was broken by the recent commits. Apparently realvideo timestamps
are severely mangled, and Matroska _of course_ doesn't have the sane,
umangled timestamps, but something unusable. The existing unmangling
code in demux_mkv.c didn't output proper timestamps either. Instead,
it was something weird that triggered sorting. Without sorting (it was
disabled by default recently), you'd get decreasing PTS warnings
In order to fix this, steal some code from libavcodec. Basically copy
the contents of rv34_parser.c (with some changes), which makes
everything magically work. (Maybe it would be better to use the
libavcodec parser API, but I don't want to do that just for this. An
alternative idea would be refusing to read files that have realvideo
tracks, and delegate this to demux_lavf.c, but maybe that's too redical
too.)
I wish I hadn't notice this...
Resampling with non-ancient ALSA setups works fine, so there is no
need to keep this around. Furthermore, as of writing, the default
builtin resampler used by many ALSA setups (taken from libspeex)
actually has higher quality than the default resampling modes of
avresample and swresample.
Apparently just 5 packets is not enough for the initial audio decode
(which is needed to find the format). The old code (before the recent
refactor) appeared to use 5 packets, but there were apparently other
code paths which in the end amounted to more than 5 packets being read.
The sample that failed (see github issue #368) needed 9 packets.
Fixes#368.
These packets have to be explicitly dropped, because usually libavcodec
uses 0-sized packets to flush delayed frames, meaning just passing
through these packets would have bad consequences.
Normally, libavformat doesn't output 0-sized packets anyway. But I don't
want to take any chances, so don't delete it, and just move it out of
the way to demux.c.
Incidentally this seems wrong in the configure as well because only gl-win32
depends on it instead of also making direct3d depend on this (as I did here).
If the value for --cache-on-pause is larger than --cache-min, and the
cache runs below --cache-on-pause, but above --cache-min, the logic
would demand to pause the player and then unpause immediately again.
This doesn't make much sense, and alternating the pause state in each
playloop iteration has negative consequences. Add an explicit check to
avoid this situation.
This means the code that tries to figure out the timestamp from
demuxer and decoder output is now all in dec_video.c. We set the
final timestamp on the returned image (mp_image.pts), as well as
the d_video->pts field.
The way the player uses d_video->pts field is still a bit messy. Maybe
this could be cleaned up later.
It appears PTS sorting was useful only for avi files (and VfW-muxed
mkv). Maybe it was historically also important for decoders with broken
or non-existent PTS reordering (win32 codecs?). But now that we handle
demuxers which outputs DTS only correctly, it just seems dead weight.
Disable it by default. The --pts-association-mode option is now forced
to always use the decoder's PTS value. You can still enable the old
default (auto) or force sorting. But we will probably remove this option
entirely at some point.
Make demux_mkv export timestamps at DTS when it's in VfW mode. This is
needed to get correct timestamps with the new default mode. demux_lavf
already does that.
This was needed to determine PTS from DTS, but the previous commits
make it unnecessary.
The builtin genpts hack was used for DVD, because libavformat's genpts
essentially went amok on DVD timestamp resets. See commit 65d87091 for
details.
Having the DTS directly can be useful for restoring PTS values.
The avi file format doesn't actually store PTS values, just DTS. An
older hack explicitly exported the DTS as PTS (ignoring the [I assume]
genpts generated non-sense PTS), which is not necessary anymore due to
this change.
Now the --no-correct-pts mode is like the normal mode, just with
different timestamp calculations. The semantics should be about the
same as before this commit.
Before this commit, this mode estimated the frame time by subtracting
successive packet PTS values. This is complete non-sense for video
codecs which use reordering. The code compensated frame times for these
non-sense using the FPS value, but confused the rest of the player with
non-sense jumping around timestamps. So, all in all this mode is not
very useful.
Repurpose this mode for fixed frame rate playback. This gives almost the
same behavior as the old mode with forced framerate (--fps option). The
result is simpler and often more robust.
Instead of passing the PTS as separate field, pass it as part of the
usual data structures. Basically, this removes strange artifacts from
the API. (It's not finished, though: the final decoded PTS goes through
strange paths, and filter_video() finally overwrites the decoded
mp_image's pts field with it.)
We also stop using libavcodec's reordered_opaque fields, and use
AVPacket.pts and AVFrame.pkt_pts. This is slightly unorthodox, because
these pts fields are not "really" opaque anymore, yet we treat them as
such. But the end result should be the same, and reordered_opaque is
marked as partially deprecated (it's not clear whether it's really
deprecated).
This normally shouldn't happen. It does happen with VfW-muxed mkv files,
and would normally happen with avi files (except that we force the
correct association mode using an explicit hack for avi).
This is usually prepended by warnings like:
Decreasing video pts: 0.125000 < 0.167000
and after the switch, there should be no warnings anymore.
Background: avi likes to use DTS for timestamps instead of PTS.
Basically, there are no proper timestamps in the file, only frame
numbers. And Matroska is insane and stores the made-up DTS instead
of a proper PTS. This will be handled properly later.
Original commit was implemented differently by @bugmen0t. The problem here was
that the waf API was called directly, instead of using our own check_cc (which
defaults, among other things, to non mandatory checks).
- without Utils.* always returns empty string
- subprocess doesn't need extra quoting for sh -c
- "source" is a bash'ism, not in POSIX sh
- most shell commands embed newline at the end
Apparently Cocoa precise scrolling generates a lot of spurious events with
a delta that is equal to 0.0. Make sure that they are discarded and not added
to the input queue.
Even though this only known to happen with Cocoa, I implemented this at core
level since it makes sense in general.
Fixes: #310
waf wrongly installs resources to BundleRoot/Resources instead of
BundleRoot/Contents/Resources. Will post a patch to them later. Either way,
this must worked around until the next waf patch release.
It seems like a good idea not to generate any additional network traffic
and wait times if we don't have to.
Also print the URL it's downloading from.
Note that if we require a newer waf release, there will be a problem.
Running ./bootstrap.py won't get the newest waf version anymore in case
the old version is in the source dir. Not sure how to handle this.
waf apparently only appends a pkgconfig flag if it doesn't already exist in
the lib storage. Since our configure often checks for multiple libraries in
one call we want to keep the flags as is. This is especially important to
always keep stuff like -lm in the right place.
This is not needed anymore, because we decided that the PAR of the
decoded video matters, and not the PAR of the filtered video that
arrives at the VO.
In Python sets are unordered, so iterating them after converting to a list
always leads to different results. The code iterated on them to collect all
the flags to pass to the compiler, and since the order of the flags changed,
waf would rebuild all of the C files. Seems like in Python 2 this worked as
expected by pure chance.
This commit stores the sets as lists, and converts them to sets when the set
operations are needed.
Fixes#363
When mpv is started with some video filters set (--vf is used), and
hardware decoding is requested, and hardware decoding would be possible,
but is prevented due to video filters that accept software formats only,
the fallback didn't work properly sometimes.
This fallback works rather violently: it tries to initialize the filter
chain, and if it fails it throws away the frame decoded using the
hardware, and retries with software. The case that didn't work was when
decoding the current packet didn't immediately lead to a new frame. Then
the filter chain wouldn't be reinitialized, and the playloop would stop
playback as soon as it encounters the error flag.
Fix this by resetting the filter error flag (back to "uninitialized"),
which is a rather violent, but somewhat working solution.
The fallback in general should perhaps be cleaned up later.
In theory, we can't really do this, because we don't know when a spdif
frame ends. Spdif transports compressed audio through audio setups that
were originally designed for PCM only (which includes the audio filter
chain, the AO API, most audio output APIs, etc.), and to reach this
goal, spdif pretends to be PCM. Compressed data frames are padded with
zeros, until a certain data rate is reached, which corresponds to a
pseudo-PCM format with 2 bytes per sample and 2 channels at 48000 Hz.
Of course an actual spdif frame is significantly larger than a frame
of the PCM format it pretends to be, so cutting audio data on frame
boundaries (as according to the pseudo-PCM format) merely yields an
incomplete and broken frame, not audio that plays for the desired
duration.
However, sending an incomplete frame might still be much better than the
current behavior, which simply ignores --end/--length (but still lets
the video end at the exact end time).
Should this result in trouble with real spdif receivers, this commit
probably has to be reverted.
If the --fps option was given (MPOpts->force_fps), the demuxer FPS value
was overwritten with the forced value. This was fine, since the demuxer
value wasn't needed anymore. But with the recent changes not to write to
the demuxer stream headers, we don't want to do this anymore. So
maintain the (forced/updated) FPS value in dec_video->fps.
The removed code in loadfile.c is probably redundant, and an artifact
from past refactorings.
Note that sub.c will now always use the demuxer FPS value, instead of
the user override value. I think this is fine, because it used the
demuxer's video size values too. (And it's rare that these values are
used at all.)
Now the actual decoder doesn't need to care about this anymore, and it's
handled in generic code instead. This simplifies vd_lavc.c, and in
particular we don't need to detect format changes in the old way
anymore.
The only reason why these structs were dynamically allocated was to
avoid recursive includes in stheader.h, which is (or was) a very central
file included by almost all other files. (If a struct is referenced via
a pointer type only, it can be forward referenced, and the definition of
the struct is not needed.) Now that they're out of stheader.h, this
difference doesn't matter anymore, and the code can be simplified.
Also sneak in some sanity checks.
This used to be needed to access the generic stream header from the
specific headers, which in turn was needed because the decoders had
access only to the specific headers. This is not the case anymore, so
this can finally be removed again.
Also move the "format" field from the specific headers to sh_stream.
Use sh_stream over sh_sub. Use dec_sub (and mpctx->d_sub) instead of the
stream header. This aligns the subtitle code with the recent audio and
video refactoring.
sh_sub still has the decoder context, though. This is because we want to
avoid reinit when switching segments with ordered chapters. (Reinit is
fast, except for creating the ASS_Renderer, which in turn triggers
fontconfig.) Not sure how much this matters, though, because the initial
segment switch will lazily initialize the decoder anyway.
This is similar to the sh_audio commit.
This is mostly cosmetic in nature, except that it also adds automatical
freeing of the decoder driver's state struct (which was in
sh_video->context, now in dec_video->priv).
Also remove all the stheader.h fields that are not needed anymore.
This drops the --pp option, which was probably broken for a while. The
option automatically inserted the "pp" filter. The value passed to it
was ignored (which is probably broken, it always selected maximal
quality).
Inserting this filter can be done simply with --vf=pp, so this is not
needed anymore.
This means most code accessing this struct must now include hwdec.h
instead of dec_video.h. I just put it into dec_video.h at first because
I thought a separate file would be a waste, but it's more proper to do
it this way, as there are too many files which include dec_video.h only
to get the mp_hwdec_info definition.
This was printed before something was decoded, and thus was not really
correct. Also, this code is hilariously broken:
/* Assume FOURCC if all bytes >= 0x20 (' ') */
if (sh_audio->format >= 0x20202020)
mp_msg(MSGT_IDENTIFY, MSGL_INFO,
"ID_AUDIO_FORMAT=%.4s\n", (char *)&sh_audio->format);
Time to kill it.
This information can be accessed through properties instead.
sh_audio is supposed to contain file headers, not whatever was decoded.
Fix this, and write the decoded format to separate fields in the decoder
context, the dec_audio.decoded field. (Note that this field is really
only needed to communicate the audio format from decoder driver to the
generic code, so no other code accesses it.)
Move all state that basically changes during decoding or is needed in
order to manage decoding itself into a new struct (dec_audio).
sh_audio (defined in stheader.h) is supposed to be the audio stream
header. This should reflect the file headers for the stream. Putting the
decoder context there is strange design, to say the least.
Apparently some packaging systems (homebrew does this for example) change
`sys.path` before they run any Python script to ensure that only the correct
Python modules can be loadable.
We need to make sure cwd is in `sys.path` since we need to load
`wscript_build.py` from there.
Simply override the cprogram Task's __str__ method with our own implementation.
This is way easier on the eyes when compiling mpv during development, since
warnings are not pushed outside of your average screenful of content.
vo_opengl always loads the hwdec backend lazily, so hwdec_request_api()
has to be called to possibly load it. This makes vf_vavpp work with
software decoding. (Hardware decoding loads the backend before the
filter is initialized, so this case is different.)
Also, the VFCTRL_GET_HWDEC_INFO call doesn't need to be checked. If it
fails, the info will be left blank.
This initialized only the load_api and load_api_ctx fields, and left the
other fields as they were. This failed with vf_vavpp, which assumed all
fields are initialized.
This commit adds a new build system based on waf. configure and Makefile
are deprecated effective immediately and someday in the future they will be
removed (they are still available by running ./old-configure).
You can find how the choice for waf came to be in `DOCS/waf-buildsystem.rst`.
TL;DR: we couldn't get the same level of abstraction and customization with
other build systems we tried (CMake and autotools).
For guidance on how to build the software now, take a look at README.md
and the cross compilation guide.
CREDITS:
This is a squash of ~250 commits. Some of them are not by me, so here is the
deserved attribution:
- @wm4 contributed some Windows fixes, renamed configure to old-configure
and contributed to the bootstrap script. Also, GNU/Linux testing.
- @lachs0r contributed some Windows fixes and the bootstrap script.
- @Nikoli contributed a lot of testing and discovered many bugs.
- @CrimsonVoid contributed changes to the bootstrap script.
This was forgotten when the parser for mplayer2 EDL files was removed.
Change the header of the mpv EDL format to include a '#', so a naive
parser could skip the header as comment. (Maybe this is questionable;
on the other hand, if it can be simpler, why not.)
Also, strip the header in demux_edl.c before passing on the data, so the
header check doesn't need to be duplicated in tl_mpv_edl.c.
Edit Decision Lists (EDL) allow combining parts from multiple source
files into one virtual file. MPlayer had an EDL format (which sucked),
which mplayer2 tried to improve with its own format (which sucked). As
logic demands, mpv introduces its very own format (which sucks).
The new format should actually be much simpler and easier to use, and
its implementation is simpler and smaller too.
The intention of the existing code was trying to match demuxer-reported
stream IDs, instead of using possibly arbitrary ordering of the frontend
track list. But EDL files can consist of quite different files, for
which trying to match the stream IDs doesn't always make sense.
This didn't have any consequences, other than suddenly reinitializing
video when it works again (such as with EDL timeline mixing video and
audio-only files).
The AF control commands used an elaborate and unnecessary organization
for the command constants. Get rid of all that and convert the
definitions to a simple enum. Also remove the control commands that
were not really needed, because they were not used outside of the
filters that implemented them.
And by "cleanup", I mean "remove". Actually, only remove the parts that
are redundant and doxygen noise. Move useful parts to the comment above
the function's implementation in the C source file.
Some decoders used to read packets and decode data when calling
resync_audio_stream(). This required a special case in mp_seek() for
audio. (A comment mentions liba52, which is long gone; but until
recently ad_mpg123.c actually exposed this behavior.)
No decoder does this anymore, and resync_audio_stream() works similar
as resync_video_stream(). Remove the special case.
The priv struct is now allocated by talloc in stream.c. It doesn't need
to be manually freed, and using free() instead of talloc_free() probably
crashes.
When the decoder detects a format change, it overwrites the values
stored in sh_audio (this affects the members sample_format, samplerate,
channels). In the case when the old audio data still needs to be
played/filtered, the audio format as identified by sh_audio and the
format used for the decoder buffer can mismatch. In particular, they
will mismatch in the very unlikely but possible case the audio chain is
reinitialized while old data is draining during a format change.
Or in other words, sh_audio might contain the new format, while the
audio chain is still configured to use the old format.
Currently, the audio code (player/audio.c and init_audio_filters) access
sh_audio to get the current format. This is in theory incorrect for the
reasons mentioned above. Use the decoder buffer's format instead, which
should be correct at any point.
Commit 22b3f522 not only redid major aspects of audio decoding, but also
attempted to fix audio format change handling. Before that commit, data
that was already decoded but not yet filtered was thrown away on a
format change. After that commit, data was supposed to finish playing
before rebuilding filters and so on.
It was still buggy, though: the decoder buffer was initialized to the
new format too early, triggering an assertion failure. Move the reinit
call below filtering to fix this.
ad_mpg123.c needs to be adjusted so that it doesn't decode new data
before the format change is actually executed.
Add some more assertions to af_play() (audio filtering) to make sure
input data and configured format don't mismatch. This will also catch
filters which don't set the format on their output data correctly.
Regression due to planar_audio branch.
Simulate proper handling of AOPLAY_FINAL_CHUNK. Print when underruns
occur (i.e. running out of data). Add some options that control
simulated buffer and outburst sizes.
All this is useful for debugging and self-documentation. (Note that
ao_null always was supposed to simulate an ideal AO, which is the reason
why it fools people who try to use it for benchmarking video.)
demuxer->filepos contains the byte offset of the last read packet. This
is so that the player can estimate the current playback position, if no
proper timestamps are available. Simplify it to use demux_packet->pos in
the generic demuxer code, instead of bothering every demuxer
implementation about it.
(Note that this is still a bit incorrect: it relfects the position of
the last packet read by the demuxer, not that returned to the user. But
that was already broken, and is not that trivial to fix.)
This was originally added for better seeking where libavformat's seek
function won't work well: files with timestamp resets. In these cases,
the code tried to calculate an average bitrate, and then do byte based
seeks by multiplying the seek target time with the bitrate.
Apparently this was unreliable enough that the code was just commented
(and other parts became inactive). Get rid of it.
Note that the player still does byte based seeks in these cases when
doing percent-seeks.
This should allow it to select better fallback formats, instead of
picking the first encoder sample format if ao->format is not equal to
any of the encoder sample formats.
Not sure what is supposed to happen if the encoder provides no
compatible sample format (or no sample format list at all), but in this
case ao_lavc.c still fails gracefully.
The added function af_format_conversion_score() can be used to select
the best sample format to convert to in order to reduce loss and extra
conversion work.
It calculates a "loss" score when going from one format to another, and
for each conversion that needs to be done a certain score is subtracted.
Thus, if you have to convert from one format to a set of other formats,
you can calculate the score for each conversion, and pick the one with
the highest score.
Conversion between int and float is considered the worst case. One odd
consequence is that when converting from s32 to u8 or float, u8 will be
picked.
Test program used to develop this follows:
#define MAX_FMT 200
struct entry {
const char *name;
int score;
};
static int compentry(const void *px1, const void *px2)
{
const struct entry *x1 = px1;
const struct entry *x2 = px2;
if (x1->score > x2->score)
return 1;
if (x1->score < x2->score)
return -1;
return 0;
}
int main(int argc, char *argv[])
{
for (int n = 0; af_fmtstr_table[n].name; n++) {
struct entry entry[MAX_FMT];
int entries = 0;
for (int i = 0; af_fmtstr_table[i].name; i++) {
assert(i < MAX_FMT);
entry[entries].name = af_fmtstr_table[i].name;
entry[entries].score =
af_format_conversion_score(af_fmtstr_table[i].format,
af_fmtstr_table[n].format);
entries++;
}
qsort(&entry[0], entries, sizeof(entry[0]), compentry);
for (int i = 0; i < entries; i++) {
printf("%s -> %s: %d \n", af_fmtstr_table[n].name,
entry[i].name, entry[i].score);
}
}
}
These must be written even if there was no "final frame", e.g. due to
the player being exited with "q".
Although the issue is mostly of theoretical nature, as most audio codecs
don't need the final encoding calls with NULL data. Maybe will be more
relevant in the future.
This used to be in bytes, now it's in samples. Divide the value by 8
(assuming a typical audio format, float samples with 2 channels).
Fix some editing mistake or non-sense about the extra buffering added
(1<<x instead of x<<5).
Also sneak in a s/MPlayer/mpv/.
This changes option parsing as well as filter defaults slightly. The
default is now to encode to spdif (this is way more useful than writing
raw AC3 - what was this even useful for, other than writing broken ac3
-in-wav files?). The bitrate parameter is now always in kbps.
Apparently this was completely broken after commit 22b3f522. Basically,
this locked up immediately completely while decoding the first packet.
The reason was that the buffer calculations confused bytes and number of
samples. Also, EOF reporting was broken (wrong return code).
The special-casing of ad_mpg123 and ad_spdif (with DECODE_MAX_UNIT) is a
bit annoying, but will eventually be solved in a better way.
This could cause the bundle to recache stuff because of differences with
configuration of other software using fonconfig. The defaults OS X directories
should be added to fontconfig at build time (through configure).
Slightly simplifies memory management. This might make adding a demuxer
cache wrapper easier at a later point, because you can just copy the
complete stream header, without worrying that the wrapper will free the
individual stream header fields.
The existing code tried to remove the "extra" profile flags for h264.
FF_PROFILE_H264_INTRA doesn't matter for us at all, because it's set
only for profiles the vdpau/vaapi APIs don't support.
The FF_PROFILE_H264_CONSTRAINED flag on the other hand is added to
H264_BASELINE, except that it makes the file a real subset of H264_MAIN
and H264_HIGH. Removing that flag would select the BASELINE profile,
which appears to be rarely supported by hardware decoders. This means we
accidentally rejected perfectly hardware decodable files. Use MAIN for
it instead.
(vaapi has explicit support for CONSTRAINED_BASELINE, but it seems to be
a new thing, and is not reported as supported where I tried. So don't
bother to check it, and do the same as on vdpau.)
See github issue #204.
This used to be needed for teletext support. Teletext commit has been
removed (see commit ebaaa41f), and it appears this code is inactive.
It was just forgotten with the removal. Get rid of it completely.
Untested. (Like all changes to the TV code.)
Signed-off-by: wm4 <wm4@nowhere>
Significant modifications over the original patch by not overriding
syscalls with macros ("#define open v4l2open") for fallback, but the
other way around ("#define v4l2open open"). As consequence, the calls
have to be replaced throughout the file.
Untested, although the original patch probably was tested.
Apparently this is not portable to FreeBSD. It turns out that we
(probably) don't use any symbols defined by this header directly, so
the includes are not needed.
These use the _oldargs_ hack, which failed in combination with playback
resume. Make it work.
It would be better to port all filters to new option parsing, but that's
obviously too much work, and most filters will probably be deleted and
replaced by libavfilter in the long run.
ao_null should simulate a "perfect" AO, but framestepping behaved quite
badly with it. Framstepping usually exposes problems with AOs dropping
their buffers on pause, and that's what happened here.
Most libavcodec decoders output non-interleaved audio. Add direct
support for this, and remove the hack that repacked non-interleaved
audio back to packed audio.
Remove the minlen argument from the decoder callback. Instead of
forcing every decoder to have its own decode loop to fill the buffer
until minlen is reached, leave this to the caller. So if a decoder
doesn't return enough data, it's simply called again. (In future, I
even want to change it so that decoders don't read packets directly,
but instead the caller has to pass packets to the decoders. This fits
well with this change, because now the decoder callback typically
decodes at most one packet.)
ad_mpg123.c receives some heavy refactoring. The main problem is that
it wanted to handle format changes when there was no data in the decode
output buffer yet. This sounds reasonable, but actually it would write
data into a buffer prepared for old data, since the caller doesn't know
about the format change yet. (I.e. the best place for a format change
would be _after_ writing the last sample to the output buffer.) It's
possible that this code was not perfectly sane before this commit,
and perhaps lost one frame of data after a format change, but I didn't
confirm this. Trying to fix this, I ended up rewriting the decoding
and also the probing.
libav* is generally freaking horrible, and might do bad things if the
data pointer passed to it are not aligned. One way to be sure that the
alignment is correct is allocating all pointers using av_malloc().
It's possible that this is not needed at all, though. For now it might
be better to keep this, since the mp_audio code is intended to replace
another buffer in dec_audio.c, which is currently av_malloc() allocated.
The original reason why this uses av_malloc() is apparently because
libavcodec used to directly encode into mplayer buffers, which is not
the case anymore, and thus (probably) doesn't make sense anymore.
(The commit subject uses the word "cargo cult", after all.)
Remove the awkward planarization. It had to be done because the AC3
encoder requires planar formats, but now we support them natively.
Try to simplify buffer management with mp_audio_buffer.
Improve checking for buffer overflows and out of bound writes. In
theory, these shouldn't happen due to AC3 fixed frame sizes, but being
paranoid is better.
The format negotiation is the same, except don't confusingly copy the
input format into af->data, just to overwrite it later. af->data should
alwass contain the output format, and the existing code was just a very
misguided use of the af_test_output() helper function.
Just set af->data to the output format immediately, and modify the input
format properly.
Also, if format negotiation fails (and needs another iteration), don't
initialize the libavcodec encoder.
Before this commit, the af_instance->mul/delay values were in bytes.
Using bytes is confusing for non-interleaved audio, so switch mul to
samples, and delay to seconds. For delay, seconds are more intuitive
than bytes or samples, because it's used for the latency calculation.
We also might want to replace the delay mechanism with real PTS
tracking inside the filter chain some time in the future, and PTS
will also require time-adjustments to be done in seconds.
For most filters, we just remove the redundant mul=1 initialization.
(Setting this used to be required, but not anymore.)
Since ao_openal simulates multi-channel audio by placing a bunch of
mono-sources in 3D space, non-interleaved audio is a perfect match for
it. We just have to remove the interleaving code.
ALSA supports non-interleaved audio natively using a separate API
function for writing audio. (Though you have to tell it about this on
initialization.) ALSA doesn't have separate sample formats for this,
so just pretend to negotiate the interleaved format, and assume that
all non-interleaved formats have an interleaved companion format.
Replace the code that used a single buffer with mp_audio_buffer. This
also enables non-interleaved output operation, although it's still
disabled, and no AO supports it yet.
Implementation wise, this could be much improved, such as using a
ringbuffer that doesn't require copying data all the time. This is
why we don't use mp_audio directly instead of mp_audio_buffer.
This comes with two internal AO API changes:
1. ao_driver.play now can take non-interleaved audio. For this purpose,
the data pointer is changed to void **data, where data[0] corresponds to
the pointer in the old API. Also, the len argument as well as the return
value are now in samples, not bytes. "Sample" in this context means the
unit of the smallest possible audio frame, i.e. sample_size * channels.
2. ao_driver.get_space now returns samples instead of bytes. (Similar to
the play function.)
Change all AOs to use the new API.
The AO API as exposed to the rest of the player still uses the old API.
It's emulated in ao.c. This is purely to split the commits changing all
AOs and the commits adding actual support for outputting N-I audio.
Allocate af_instance->data in generic code before filter initialization.
Every filter needs af->data (since it contains the output
configuration), so there's no reason why every filter should allocate
and free it.
Remove RESIZE_LOCAL_BUFFER(), and replace it with mp_audio_realloc_min().
Interestingly, most code becomes simpler, because the new function takes
the size in samples, and not in bytes. There are larger change in
af_scaletempo.c and af_lavcac3enc.c, because these had copied and
modified versions of the RESIZE_LOCAL_BUFFER macro/function.
No AO can handle these, so it would be a problem if they get added
later, and non-interleaved formats get accepted erroneously. Let them
gracefully fall back to other formats.
Most AOs actually would fall back, but to an unrelated formats. This is
covered by this commit too, and if possible they should pick the
interleaved variant if a non-interleaved format is requested.
Based on earlier work by Stefano Pigozzi.
There are 2 changes:
1. Instead of mp_audio.audio, mp_audio.planes[0] must be used.
2. mp_audio.len used to contain the size of the audio in bytes. Now
mp_audio.samples must be used. (Where 1 sample is the smallest unit
of audio that covers all channels.)
Also, some filters need changes to reject non-interleaved formats
properly.
Nothing uses the non-interleaved features yet, but this is needed so
that things don't just break when doing so.
This affects 64 bit floats and big endian integer PCM variants
(basically crap nobody uses). Possibly not all MS-muxed files work, but
I couldn't get or produce any samples.
Remove a bunch of format tags that are not needed anymore. Most of these
were used by demux_mov, which is long gone. Repurpose/abuse 'twos' as
mpv-internal tag for dealing with the PCM variants mentioned above.
Now to shift audio pts when outputting to e.g. avi, you need an explicit
facility to insert/remove initial samples, to avoid initial regions of
the video to be sped up/slowed down.
One such facility is the delay filter in libavfilter.
My main problem with this is that the output format will be incorrect.
(This doesn't matter right, because there are no samples output.)
This assumes all audio filters can deal with len==0 passed in for
filtering (though I wouldn't see why not).
A filter can still signal an error by returning NULL.
af_lavrresample has to be fixed, since resampling 0 samples makes
libavresample fail and return a negative error code. (Even though it's
not documented to return an error code!)
When blending OSD and subtitles onto the video, we write bogus alpha
values. This doesn't normally matter, because these values are normally
unused and discarded. But at least on Wayland, the alpha values are used
by the compositor and leads to transparent windows even with opaque
video on places where the OSD happens to use transparency.
(Also see github issue #338.)
Until now, the alpha basically contained garbage. The source factor
GL_SRC_ALPHA meant that alpha was multiplied with itself. Use GL_ONE
instead (which is why we have to use glBlendFuncSeparate()). This should
give correct results, even with video that has alpha. (Or at least it's
something close to correct, I haven't thought too hard how the
compositor will blend it, and in fact I couldn't manage to test it.)
If glBlendFuncSeparate() is not available, fall back to glBlendFunc(),
which does the same as the code did before this commit. Technically, we
support GL 1.1, but glBlendFuncSeparate is 1.4, and I guess we should
try not to crash if vo_opengl_old runs on a system with GL 1.1 drivers
only.
Apparently we were using FFmpeg-specific APIs. I have no idea whether
this code is correct on both FFmpeg and Libav (no examples, bad
doxygen... why do they even complaint aht people are using their APIs
incorrectly?), but it appears to work on FFmpeg. That was also the case
before commit ebc4ccb though, where it used internal libavformat
symbols.
Untested on Libav, Travis will tell us.
Set the PulseAudio stream title, just like the VO window title is set.
Refactor update_vo_window_title() so that we can use it for AOs too.
The ao_pulse.c bit is stolen from MPlayer.
In theory, af_volume could use separate volume levels for each channel.
But this was never used anywhere.
MPlayer implemented something similar before (svn r36498), but kept the
old path for some reason.
This member was redundant. sh_audio->sample_format indicates the sample
size already.
The TV code is a bit strange: the redundant sample size was part of the
internal TV interface. Assume it's really redundant and not something
else. The PCM decoder ignores the sample size anyway.
Also do some cosmetic changes, like merging definition and
initialization of local variables.
Remove an annoying debug mp_msg() from af_open(). It just printed the
command line parameters; if this is really needed, it could be added
to af.c instead (similar as to what vf.c does).
Helps with readability. Also remove the ctx_opt_set_* helper macros and
use av_opt_set_* directly (I think these macros were used because the
lines ended up too long, but this commit removes two indentation levels,
giving more space).
This should allow to make format negotiation much simpler, since it
takes the responsibility to compare actual input and accepted input
formats from the filters. It's also backwards compatible. Filters which
have expensive initialization still can use the old method.
I have no idea what these do, but apparently they are needed to inform
ALSA about spdif configuration. First, replace the literal constant "6"
for the AES0 parameter with the symbolic constants from the ALSA
headers (the final value is the same). Second, copy paste some funky
looking parameter setup from VLC's alsa output for setting the AES1,
AES2, AES3 parameters. (The code is actually not literally copy-pasted,
but does exactly the same.)
My small but non-zero hope is that this could make DTS-HD work, or at
least work into that direction. I can't test spdif stuff though, and
for DTS-HD not even opening the ALSA device succeeds on my system.
Using spdif with alsa requires adding magic parameters to the device
name, and the existing code tried to deal with the situation when the
user wanted to add parameters too.
Rewrite this code, in particular remove the duplicated parameter string
as preparation for the next commit. The new code is a bit stricter, e.g.
it doesn't skip spaces before and after '{' and '}'. (Just don't add
spaces.)
This accessed tons of private libavformat symbols all over the place.
Don't do this and convert all code to proper public APIs. As a
consequence, the code becomes shorter and cleaner (many things the code
tried are done by libavformat APIs).
It's probably better if all auto-inserted filters are removed when doing
an af_add operation. If they're really needed, they will be
automatically re-added.
Fix the error message. It used to be for an actual internal error, but
now it happens when format negotiation fails, e.g. when trying to use
spdif and real audio filters.
Note that the change in seek_reset is not entirely equivalent: we even
drop the remainder of buffered audio when seeking. This should be more
correct, because the whole point of the reset_ao parameter is to control
whether audio queued for output should be dropped or not.
ao_lavc.c accesses ao->buffer, which I consider internal. The access was
done in ao_lavc.c/uninit(), which tried to get the left-over audio in
order to write the last (possibly partial) audio frame. The play()
function didn't accept partial frames, because the AOPLAY_FINAL_CHUNK
flag was not correctly set, and handling it otherwise would require an
internal FIFO.
Fix this by making sure that with gapless audio (used with encoding),
the AOPLAY_FINAL_CHUNK is set only once, instead when each file ends.
Basically, move the hack in ao_lavc's uninit to uninit_player.
One thing can not be entirely correctly handled: if gapless audio is
active, we don't know really whether the AO is closed because the file
ended playing (i.e. we want to send the buffered remainder of the audio
to the AO), or whether the user is quitting the player. (The stop_play
flag is overwritten, fixing that is perhaps not worth it.) Handle this
by adding additional code to drain the AO and the buffers when playback
is quit (see play_current_file() change).
Test case: mpv avdevice://lavfi:sine=441 avdevice://lavfi:sine=441 -length 0.2267 -gapless-audio
It spams these in verbose mode. It's caused by format negotiation code
in af.c. It's for the mpv format to ffmpeg-format case, and that one is
very uninteresting. (The ffmpeg supported audio formats are practically
never extended.)
This was supposed to handle preemption better. I still think the current
state isn't very nice, since the decoder can "accidentally" call the
previous render function after preemption (instead of calling the
reloaded function), so there might be issues. But all in all, this
dummy_render function is a bit confusing, and still not entirely
correct, so it's not worth it.
Turn the sample format definitions into an enum. (The format bits are
still macros.) The native endian versions of the new definitions don't
have a NE suffix anymore, although there are still compatibility defines
since too much code uses the NE variants.
Rename the format bits for special formats to help to distinguish them
from the actual definitions, e.g. AF_FORMAT_AC3 to AF_FORMAT_S_AC3.
We found that the stretching - although it usually improves the looks of
the fonts - is incorrect.
On DVD, subtitles can cover the full area of the picture, and they have
the same pixel aspect as the movie itself.
Too bad many commercially released DVDs use bitmap fonts made with the
wrong pixel aspect (i.e. assuming 1:1) - --stretch-dvd-subs will make
these more pretty then.
The previous code used the output video's pixel aspect for stretching
purposes, breaking rendering with e.g. -vf scale in the chain. Now
subtitles are stretched using the input video's pixel aspect only,
matching the intentions of the original subtitle author.
I overlooked the fact that the ffmpeg/libav build system only supports `--cc`
and completly ignores $CC. Hopefully this makes the build times a little
faster.
Fixes#332
This removes "--hwdec=crystalhd".
I doubt anyone even tried to use this. But even if someone wants to
use it, the decoders can still be explicitly invoked with e.g.:
--vd=lavc:h264_crystalhd
The only advantage our special code provided was fallback to
software decoding. (But I'm not sure how the ffmpeg crystalhd
pseudo-decoder actually behaves.)
Removing this will allow some simplifications as soon as we don't need
vdpau_old.c anymore.
This drops autorepeated key events for a number of commands. This should
help with slow situations accidentally triggering too many repeats. (I'm
not sure if this actually happened to users - maybe.)
It's not clear whether MP_CMD_COMMAND_LIST (multiple commands on one
binding separated by ';') should be repeated, or whether it should try
to do something clever. For now, disallow autorepeat with it.
This uses vdpau OpenGL interop to convert a vdpau surface to a texture.
Note that this is a bit weak and primitive. Deinterlacing (or any other
form of vdpau postprocessing) is not supported. vo_opengl chroma scaling
and chroma sample position are not supported. Internally, the vdpau
video surfaces are converted to a RGBA surface first, because using the
video surfaces directly is too complicated. (These surfaces are always
split into separate fields, and the vo_opengl core expects progressive
frames or frames with weaved fields.)
Instead of checking for resolution and image format changes, always
fully reinit on any parameter change. Let init_video do all required
initializations, which simplifies things a little bit.
Change the gl_video/hardware decoding interop API slightly, so that
hwdec initialization gets the full image parameters.
Also make some cosmetic changes.
These formats are helpful for distinguishing surfaces with and without
alpha. Unfortunately, Libav and older version of FFmpeg don't support
them, so code will break. Fix this by treating these formats specially
on the mpv side, mapping them to RGBA on Libav, and unseting the alpha
bit in the mp_imgfmt_desc struct.
Somehow the new parser ends up much smaller. Much of it is because we
don't parse some additional information. We just skip it, instead of
parsing it and then throwing it away.
More importantly, we use the physical order of entries, instead of
trying to sort them by entry number. Each "File" entry is followed by a
number that is supposed to be the entry number, and "File1" is first.
(Should it turn out that this is really needed, an additional field
should be added to playlist_entry, and then qsort().)
Make TOOLS/matroska.pl output structs with fields sorted by name in
ebml_types.h to make the order of fields deterministic. Fix warnings in
demux_mkv.c caused by the first struct fields switching between scalar
and struct types due to non-deterministic ebml_types.h field order.
Since it's deterministic now, this shouldn't change anymore.
The warnings produced by the compilers are bogus, but we want to silence
them anyway, since this could make developers overlook legitimate
warnings.
What commits 7b52ba8, 6dd97cc, 4aae1ff were supposed to fix. An earlier
attempt sorted fields in the generated C source file, not the header
file. Hopefully this is the last commit concerning this issue...
Before the bitstream_buffers field was deprecated, you had to free it,
otherwise you would leak memory.
(Although vdpau.c uses a new API, they managed to introduce a new
deprecation this quickly. This is a complaint.)
This introduces a memory leak of 12 bytes per file on every file on some
_older_ libavcodec versions. This is minor enough that I don't care.
Video has up to 4 textures, if you include obscure formats with alpha.
This means alpha formats could always overwrite the first scaler
texture, leading to corrupted video display. This problem was recently
brought to light, when commit 571e697 started to explicitly unbind all 4
video textures, which broke rendering for non-alpha formats as well.
Fix this by reserving the correct number of texture units.
mpvcore/player/playloop.c: In function 'seek':
mpvcore/player/playloop.c:209:54: warning: declaration of 'seek' shadows a global declaration [-Wshadow]
mpvcore/player/playloop.c:209:12: warning: shadowed declaration is here [-Wshadow]
mpvcore/player/playloop.c: In function 'queue_seek':
mpvcore/player/playloop.c:360:25: warning: declaration of 'seek' shadows a global declaration [-Wshadow]
mpvcore/player/playloop.c:209:12: warning: shadowed declaration is here [-Wshadow]
Signed-off-by: wm4 <wm4@nowhere>
VA-API's OpenGL/GLX interop is pretty bad and perhaps slow (renders a
X11 pixmap into a FBO, and has to go over X11, probably involves one or
more copies), and this code serves more as an example, rather than for
serious use. On the other hand, this might be work much better than
vo_vaapi, even if slightly slower.
Most hardware decoding APIs provide some OpenGL interop. This allows
using vo_opengl, without having to read the video data back from GPU.
This requires adding a backend for each hardware decoding API. (Each
backend is an entry in gl_hwdec_vaglx[].) The backends expose video data
as a set of OpenGL textures.
Add infrastructure to support this. The next commit will add support for
VA-API.
The configure followed 5 different convetions of defines because the next guy
always wanted to introduce a new better way to uniform it[1]. For an
hypothetic feature 'hurr' you could have had:
* #define HAVE_HURR 1 / #undef HAVE_DURR
* #define HAVE_HURR / #undef HAVE_DURR
* #define CONFIG_HURR 1 / #undef CONFIG_DURR
* #define HAVE_HURR 1 / #define HAVE_DURR 0
* #define CONFIG_HURR 1 / #define CONFIG_DURR 0
All is now uniform and uses:
* #define HAVE_HURR 1
* #define HAVE_DURR 0
We like definining to 0 as opposed to `undef` bcause it can help spot typos
and is very helpful when doing big reorganizations in the code.
[1]: http://xkcd.com/927/ related
Instead of having each demuxer do it (only demux_mkv actually did...),
let generic code determine whether the file is seekable. This requires
adding exceptions to demuxers where the stream is not seekable, but the
demuxer is.
Sort-of try to improve handling of unseekable files in the player. Exit
early if the file is determined to be unseekable, instead of resetting
all decoders and then performing a pointless seek.
Add an exception to allow seeking if the file is not seekable, but the
stream cache is enabled. Print a warning in this case, because seeking
outside the cache (which we can't prevent since the demuxer is not aware
of this problem) still messes everything up.
Pointless, using stream->start_pos/end_pos instead.
demux_mf was the only place where this was used specially, but we can
rely on timestamps instead for this case.
Never check s->seek (except in init), because it'd have to check
s->flags anyway. Also, for fast skippable streams (like pipes), don't
set the bit that indicates support for seek forward.
Make sure s->end_pos is always 0 for unseekable streams. Lots of code
outside of stream.c uses this to check seeking support.
Not really cargo cult, but an unexplainable, needless difference that
just exists to annoy us.
Fixes that gcc on MinGW treats format specifiers in MSVC mode. Just why?
Why?
There are some Microsoft Windows symbols which are traditionally used by
the mplayer core, because it used to be convenient (avi was the big
format, using binary windows decoders made sense...). So these symbols
have the exact same definition as the Windows one, and if mplayer is
compiled on Windows, the symbols from windows.h are used.
This broke recently just because some files were shuffled around, and
the symbols defined in ms_hdr.h collided with windows.h ones. Since we
don't have windows binary decoders anymore, there's not the slightest
reason our symbols should have the same names. Rename them to reduce the
risk for collision, and to fix the recent regression.
Drop WAVEFORMATEXTENSIBLE, because it's mostly unused. ao_dsound defines
its own version if the windows headers don't define it, and ao_wasapi is
not available on systems where this symbol is missing.
Also reindent ms_hdr.h.
Now that matroska.pl generates struct fields in deterministic order,
this should be the last time I change this.
(gcc and clang shouldn't warn about this line of code, but since they
do, we want to workaround and silence the warning anyway.)
Newer versions of perl randomize the hash used for hashes every time
it's run; this makes the order of the fields be non-deterministic. Tack
a sort there to make it deterministic. Needed to fix (or allow fixing) a
buggy gcc warning.
mpv crashed when linked files were not found. The reason was that the
chapters array contained some uninitialized data.
I have no idea how this code works (after the merge). The old code
actually seems to remove missing chapters, while the new code just
leaves them unintiialized. Work around the crash by initializing the
chapters array (and a bunch of other things) with 0, which means the
missing chapter will be located at 00:00:00 and have no name.
There is a regression since commit af0306d.
We had some code for checking profiles earlier, which was removed in
commits 2508f38 and adfb71b. These commits mentioned that (working) hw
decoding was sometimes prevented due to profile checking, but I can't
find the samples anymore that showed this behavior. Also, I changed my
opinion, and I think checking the profiles is something that should be
done for better fallback to software decoding behavior.
The checks roughly follow VLC's vdpau profile checks, although we do
not check codec levels. (VLC's profile checks aren't necessarily
completely correct, but they're a welcome help anyway.)
Add a --vd-lavc-check-hw-profile option, which skips the profile check.
Unfortunately, we can't avoid this warning 100%, because ebml_info is
written by a Perl script. I think the script writes the struct fields in
random order (thanks Perl), so there's no way to know whether the first
struct field is a scalar or a struct.
At least {0} is always valid here, even if it shows a warning. (The
compilers are wrong, see e.g. [1].)
[1] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53119
This one really did bite me hard (see previous commit), so enable it by
default.
Fix some cases of shadowing throughout the codebase. None of these
change behavior, and all of these were correct code, and just tripped up
the warning.
E.g. "-vf scale=848:480" set the w argument twice, instead of setting w
and then h.
This was caused by accidental shadowing of a local variable.
Regression since probably 4cd143e.
As preparation for resizing the window with input commands in the
following commit.
Since there are already so many functions which somehow resize the
window, add the word "highlevel" to the name of this new function.
DVD subs (rarely) have subtitle events without end timestamp. The
duration is unknown, and they should be displayed until they're
replaced by the next event.
FFmpeg fails hard to make us aware whether duration is unknown or
actually 0, so we can't distinguish between these two cases. It fails
at this twice: AVPacket.duration is set to 0 if duration is unknown,
and AVSubtitle.end_display_time has the same issue.
Add a hack that considers all bitmap subtitles with duration==0 as
events with uknown length. I'd rather accidentally display a hidden
subtitle (if they exist at all), instead of not displaying random
subtitles at all.
See github issue #325.
The code was selecting PA_CHANNEL_POSITION_MONO for MP_SPEAKER_ID_FC,
which is correct only with the "mono" channel layout, but not anything
else. Remove the mono entry, and handle mono separately.
See github issue #326.
We mixed the "old" AVFrame management functions (avcodec_alloc_frame,
avcodec_free_frame) with reference counting. This doesn't work
correctly; you must use av_frame_alloc and av_frame_free. Of course
ffmpeg doesn't warn us about the bad usage, but will just mess up
things silently. (Thanks a lot...)
While the alloc function seems to be 100% compatible, the free function
will do bad things, such as freeing memory that might still be
referenced by another frame. I didn't experience any actual bugs, but
maybe that was pure luck.
This stopped working when the code was changed to create a window even
if --wid is used.
It appears we can't create our own window in this case, because in X11
there is no difference between a window with the root window as parent,
and a window that is managed by the WM. So make this (kind of worthless)
special case use the root window itself.
Just doing this because mp_osd.h and osd.c is not consistent.
There are some other header files (command.h and screenshot.h), but
since I don't feel too good about inflating mp_core.h, I'm not merging
them, at least not yet.
I considered making a header file for each .c file, but decided against
it. Asking around, not making separate headers was deemed acceptable. In
the end, all of these depend on MPContext and store state inside of it,
so separate headers aren't all that useful anyway.
mplayer.c was a bit too big. Split it into multiple files. I hope the
way it's split makes sense. Maybe some things don't make too much sense,
or go against intuition. These will fixed as soon as I notice them.
Some files are a bit questionable (misc.c, osd.c, configfiles.c), and
suggestions how to organize this better are welcome.
Regressions are possible due to reorganized include statements.
Obviously I didn't just copy mplayer.c's orgy of include statements, but
recreated them for each file. It's easily possible that there are
oversights and mistakes, which will show up on other platforms.
There is one actual change: the public avutil.h include is removed from
encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I
consider using libavutil too much as dangerous, because the set of
include files they recursively pull in is rather arbitrary and is
different between FFmpeg and Libav.
All these files access mp_core.h and MPContext, and form the actual
player application. They should be all in one place, and separate
from the other sources that are mere utility helpers.
Preparation for splitting mplayer.c into multiple smaller parts.
On systems that provide legacy OpenGL (up to 2.1), but not GL3 and
later, creating a GL3 context will fail. We then revert to legacy GL.
Apparently the error message printed when the GL3 context creation
fails is confusing. We could just silence it, but there's still a X
error ("X11 error: GLXBadFBConfig"), which would be quite hard to
filter out. For one, it would require messing with the X11 error
handler, which doesn't even carry a context pointer (for application
private data), so we don't even want to touch it. Instead, change
the error message to inform the user what's actually happening: a
fallback to an older version of OpenGL.
When starting mpv with nohup, file descriptor 0 seems to be invalid for
some reason. (I'm not quite sure why it should be... /proc/pid/fd/0
seems to indicate it's just /dev/null, and using /dev/null explicitly
shows that it works just fine.)
select() will always immediately return, and this causes mpv to burn CPU
without reason. Fix this by treating it as EOF when read() returns
EBADF.
Also add EINVAL to this condition, because it seems like a good idea.
This might actually fix an issue with DVB channel switching, because
that uses some sort of hack to reinitialize the demuxer, which causes
the subtitle renderer to initialize twice. As consequence, the assert in
add_subtitle_fonts_from_sources() would trigger.
I didn't actually test the DVB case, because I don't have the hardware.
Disable autorepeat for the "add"/"cycle" commands when changing
properties that are choices (such as fullscreen, pause, and more).
In these cases, autorepeat is not very useful, and can rather harmful
effects (like triggering too many repeated commands if command
execution is slow).
(Actually, the commands are just dropped _after_ being repeated, since
input.c itself does not know enough about the commands to decide whether
or not they should be repeated.)
Trying to toggle the border during fullscreen (with "cycle border")
would leave the window stuck without border, and it couldn't be
restored. This was because vo_x11_decoration() always excepted to be
called when toggling the state, and thus confusing the contents of the
olddecor variable. Add got_motif_hints to hopefully prevent this.
Also, when changing the border, don't take fs in account. May break on
older/broken WMs, but all in all is in fact more robust and simpler,
because you do not need to update the border state manually when
returning from fullscreen.
Defining names like min, max etc. in an often used header is not really
a good idea.
Somewhat similar to MPlayer svn commit 36491, but don't use libavutil,
because that typically causes us sorrow.
Roughly follows MPlayer svn commits 36492 and 36493. We also remove
the volume peak reporting. (There are much better libavfilter filters
for this, I think.)
Not up to my quality standards (need ASS boundingboxes), so disabled by default and undocumented for now.
Can be enabled with seektooltip=yes in plugin_osc.conf
There are certain cases where mpv will automatically set options, such
as per-file configs, per protocol/VO/AO/extension profiles, and
watch_later resume configs. All these were overwriting the user's
options, even when they were specified on command line.
Add something that explicitly preserves command line options. This
means you can now actually use the command line to override any
options that the playback resume functionality backups and restores.
It still happily overrides options set at runtime (e.g. changed via
properties while playing a file; then playing the next file might
override them again), but maybe that's not a problem with typical use.
It's true that ALSA uses alloca() in some of its API functions, but
since this is hidden behind macros in the ALSA headers, we have no
reason to include alloca.h ourselves.
Might help with portability (FreeBSD).
In cases we're just listing options or checking their values (as it
happens with -vo/-vf etc. suboption parsing), we don't need to allocate
abd initialize the actual option struct. All we're interested in is
the list of options.
The default behavior of weston changed some time ago to not fill the surface
black for fullscreen windows.
Now let mpv draw the whole screen in fullscreen mode.
Keep track of the default values directly, instead of creating a new
instance of the option struct just to get the defaults.
Also get rid of the special handling of m_obj_desc.init_options.
Instead, handle it purely by the option parser. Originally, I wanted to
handle --vo=opengl-hq and --vo=direct3d_shaders with this (by making
them aliases to the real VOs with a different preset), but since --vo
=opengl-hq=help prints the wrong values (as consequence of the
simplification), I'm not doing that, and instead use something
different.
Might reduce memory overhead, and is also less annoying.
Since pointers to m_config_options are kept around, this assumes
they're added only during initialization. Otherwise you'd get
dangling pointers.
Minor simplification. String options are now not allowed to use the
same variable/field anymore. (Also affects other "dynamic" options
which require memory allocation.)
This seems to be a problem only in OS X 10.9. I guess they improved the
general speed of the Cocoa startup and suddenly mpv core takes more time
than the Cocoa thread to initialize.
Fixes#285
(hopefully!)
Drop the author and comment fields. They were completely unused - not
even printed in verbose mode, just dead weight.
Also use designated initializers and drop redundant flags.
Set the input/output format in filter init. This doesn't change anything
functionally, but it makes the forced format show up in the filter chain
init verbose output (which sometimes prints the filter chain before all
filters have been configured).
af_format is the old audio conversion filter. It could do all possible
conversions supported by the audio chain. However, ever since the
addition of af_lavrresample, most conversions are done by
libav/swresample, and af_format is used as fallback.
Separate out the fallback cases and remove af_format. af_convert24 does
24 bit <-> 32 bit conversions, while af_convertsignendian does sign and
endian conversions. Maybe the way the conversions are split sounds a bit
odd. But the former changes the size of the audio data, while the latter
is fully in-place, so there's at least different buffer management.
This requires a quite complicated algorithm to make sure all these
"partial" conversion filters can actually get from one format to
another. E.g. s24le->s32be always requires convertsignendian and
convert24, but af.c has no idea what the intermediate format should
be. So I added a graph search (trying every possible format and
filter) to determine required format and filter. When I wrote this,
it seemed this was still better than messing everything into
af_lavrresample, but maybe this is overkill and I'll change my
opinion. For now, it seems nice to get rid of af_format though.
The AC3->IEC61937 conversion isn't supported anymore, but I don't think
this is needed anywhere. Most AOs test all formats explicitly, or use
the AF_FORMAT_IS_IEC61937() macro (which includes AC3).
One positive consequence of this change is that conversions always
include dithering (done by libav/swresample), instead of possibly going
through af_format, which doesn't do anything fancy.
Rename af_force to af_format. It's essentially compatible with command
line uses of af_format. We retain a compatibility alias for af_force.
At least not with ffmpeg.
Honestly, I have no idea how little endian AC3 works at all, since
ao_pcm doesn't do anything special about it, and treats it like s16le.
Maybe it's broken and ffmpeg has special logic to detect it.
This is obviously not needed, and just creates potential for bad
breakages (e.g. what happens if libquvi tries to open a normal filename
as http URL?).
Note that for simplicity, we still pass file:// URIs to quvi, and we
don't exclude other protocol prefixes either. In general, we don't know
what protocols quvi might support, so we don't try to second-guess it.
(Even though in practice, it's probably only "http" and "https".)
The code did not set and unset the current context inside sync sections. I am
not sure if this was an actual problem but this is better since the context is
linked to a single thread. In my brief tests this seems to avoid garbage to
show up in fullscreen.
Instead of removing dragging we now test if it we should drag the window or
not. Because if the OSC shows up we can not drag the window because that would
cause mouse events that makes the OSC disappear.
gcc and clang happen to allow {} to default-initialize a struct, but
strictly speaking, C99 requires at least {0}. In one case, we use {{0}},
but that's only because gcc as well as clang are too damn stupid not
to warn about {0}, which is a perfectly valid construct in this case.
(Sure is funny, don't warn about the non-standard case, but warn about
another standard conform case.)
Leaving these braces away just because the syntax allows them is really
obnoxious. It removes the visual cues which help understanding the code
at the first look.
For the record,
if (cond)
something();
is ok, as long as there's no else branch, and the if body is one
physical line. But everything else should have braces.
This was probably not a real problem. But it's not entirely clear
whether this could actually happen or not, so it's better to be
defensive. The code is now also somewhat easier to understand.
If cache was enabled, using ordered chapters could easily crash. The
reason is that enable_cache() reopens the demuxer and closes the old
one. The code after that (reading m->ordered_chapters etc.) then
accessed freed data.
This commit also avoids enabling cache for files which are not used
(which would make opening much slower).
Implement MP_GROW_ARRAY using MP_TARRAY_GROW. MP_GROW_ARRAY is basically
the earlier version of MP_TARRAY_GROW, and had different semantics. When
I added MP_TARRAY_GROW, I didn't dare to change it, but I think all code
that relied on the exact semantics of MP_GROW_ARRAY is gone now, or the
difference doesn't matter anyway. The semantic change is that now
(n+1)*2 elements are preallocated instead of n*2.
Also, implement MP_TARRAY_GROW using MP_RESIZE_ARRAY, which saves 1 line
of code.
In future, these macros should be part of TA.
Was disabled by default, was never used, internal support was
inconsistent and poor, and there has been virtually no interest in
creating translations.
And I don't even think that a terminal program should be translated.
This is something for (hypothetical) GUIs.
This adds support for ChapterSegmentEditionUID (pull request #258),
and also fixes issue #278 (pull request #292).
In fact, this is a straight merge of pr/292, which also contains pr/258.
When adding or removing frames to avoid 1-frame seeks on chapter
boundaries, the end of the chapter needs to be pushed or pulled by the
same amount to keep the intended end frame the same.
Note that you still need --vd-lavc-o='strict=-2' to enable the decoder.
Also, there's no guarantee that all required features for HEVC demuxing
are actually implemented, nor that the current muxing schema is the
final one.
This could cause the OSC to be displayed without mouse interaction: for
example, starting mpv with --fs, and putting the mouse to where the OSC
area is beforehand, would cause the OSC to appear and stay visible. We
don't want that. The simplest solution is not generating artificial
mouse move events from mouse enter events, because they make the OSC
think the mouse was actually moved.
Also see commit 0c7978c, where handling of mouse enter events was added.
This was supposed to fix certain corner cases, but they're not relevant
anymore due to changes in OSC behavior.
Commit 9777047 fixed this as well (by resetting the mouse state on
MOUSE_LEAVE), but all the behavior reverted with this commit as perhaps
a bad idea. It wasn't very robust, made it hard to distinguish real
events from artificial ones, and finally made the mouse cursor more
often visible than needed. (Now switching between workspaces doesn't
make the cursor visible again when switching to a fullscreened mpv.)
Now you can pretend the config file is quite literally command line
values dumped into a file, e.g.
--option1=value
--option2=value
...
although the underlying mechanisms are quite different.
Until now it used both char[] and bstr variants in the same code, which
was nasty. For example, the next commit would have additionally required
using memmove() to remove the prefix from the char[] string.
This fixes handling of e.g. "--vf=lavfi=[ noise ]" when used with
playback resume functionality. The spaces made it bug out, and there are
more cases where it could potentially break.
We could always escape for simplicity, but for now make old and new mpv
builds somewhat interoperable and use this new escaping only if needed.
This parses "%len%string" escapes, where string can contain any
characters. This method of escaping has also been used in other parts
of mplayer and mpv, so it's not a new idea.
(Also, don't confuse with URL encoding.)
Needed by the following commit.
With the way I've been doing releases in the release/0.1 branch, this
proved completely useless. You need to write the VERSION file anyway,
and since we use github's automatically generated tarballs (via its
release system), the VERSION file needs to be in the git revision of
a release anyway. git master on the other hand displayed the v0.1.0
tag, because the tag is within the master branch.
This essentially reverts commit b27f65a. Now just print the contents
of the VERSION file if it exists, and likewise, we append the git
revision if it exists. (Do that even if VERSION exists, because this
way we can tell if someone is using the release branch at an in-between
point where no new release has been made yet.)
Also change what the FFmpeg version info looks like, and additionally
dump lavfi/lavr/lswr versions. (Don't bother with libavdevice and
libpostproc, they're not important enough.)
Unfortunately, there's no "single" FFmpeg/Libav version due to fatal
braindeath on the FFmpeg/Libav side. We can't map the versions to
releases either (it simply isn't accessible anywhere).
The mouse area that covers the OSC is now only activated when the OSC is actually visible, to make sure the mouse still hides if it happens to be parked in the OSC area without making the OSC show up.
vo_image didn't handle OSD redrawing correctly anymore after OSD
redrawing behavior was changed in commit ed9295c (or maybe it has been a
problem for a longer time, and only showed up now). Basically, flip_page
was called unexpectedly and when no image was stored, which made it
crash trying to access the image. This could happen when for example
provoking OSD redrawing by pausing while using --vo=image, or by using
this command line: mpv --vo=image '-vf=lavfi="select=not(mod(n\,3))"'
Fix by removing the code that pretends vo_image can redraw OSD, and by
removing the framestepping fallback, which could make bad things happen
if the VO didn't support OSD redrawing. By now, there aren't any real
VOs that can't redraw the OSD properly, so this code is not needed and
just complicates things like vo_image.
This change likely will also be useful for vo_lavc (encoding).
Libass is technically an optional dependency, but in practice users
tend to disable libass accidentally or for the hell of it to get
something "minimal", without being aware of the consequences.
The deadzone-size is now by default zero, so movement on the entire window will make the OSC show up. To avoid it showing up by randomly moving mice, the option 'minmousemove' controls how many pixels movement (default: 1) between ticks (frames) are necessary to make the OSC show up.
The deadzone can be reenabeled by setting the option 'deadzonesize' (default: 0 = no deadzone, 1 = entire area between OSC and opposite window border), to restore the old behavior, set it to ~0.92.
The OSC will hide immediately when leaving the window or entering the deadzone (if existing) or after the time specified with 'hidetimeout' (default: 500ms) passed without any new movement. Set to negative value to disabling auto-hide (thus restoring old behavior). The OSC will never hide if hovered by the mouse.
Now that talloc has been removed, the license can be switched back to
GPLv2+. Actually, there never was a GPLv2+ licensed MPlayer (fork or
not) until now, but removal of some GPLv2-only code makes this possible
now. Rewrite the Copyright file to explain the reasons for the licenses
MPlayer and forks use. The old Copyright file didn't contain anything
interesting anymore, and all information it contained is available at
other places in the source tree.
The reason for the license change itself is that it should improve
interoperability with differently licensed code in general.
This essentially reverts commit 1752808.
There are multiple reasons to do this. One big reason is the license:
talloc is LGPLv3+, which forces mpv to be licensed as GPLv3+.
Another one is that our talloc copy contains modifications, which makes
it essentially incompatible with upstream talloc (in particular, our
version aborts on out of memory conditions - well, it wasn't my idea).
Updating from upstream is also a bit involved - the talloc source is
not really organized in a way to allow copying it into projects (and
this isn't an intended use-case).
Finally, talloc is kind of big and bloated. The replacement halves the
amount of code - mainly because we didn't use all talloc features. It's
even more extreme if you compare upstream talloc (~4700 lines) and the
new allocator without talloc compat (~900 lines).
The replacement provides all features we need. It also doesn't clash
with talloc. (The talloc compatibility wrapper uses macros to avoid
introducing linker-level symbols which could clash with libtalloc.)
It also tries to lower the overhead (only 4 words opposed to 10 words
in talloc for leaf nodes in release mode). Debugging features like leak
reporting can be enabled at compile time and add somewhat more overhead.
Though I'm not sure whether the overhead reduction was actually
successful: allocations with children need an "extra" header, which adds
plenty of overhead, and it turns out that almost half of all allocations
have children. Maybe the implementation could be simplified and the
extra header removed - even then, overhead would be lower than talloc's.
Currently, debugging features can be entirely deactivated by defining
NDEBUG - I'm not sure if anything defines this directly yet, though.
Unlike in talloc, the leak reporting stuff is thread-safe. (That's also
why it's far less elegant, and requires extra list pointers.)
Comes with a compatibility layer, so no changes to mpv source code
are needed. The idea is that we will pretend to be using talloc for
a while, so that we can revert to our old talloc implementation at
any time for debugging purposes.
Some inspiration was taken from Mesa's ralloc:
http://cgit.freedesktop.org/mesa/mesa/tree/src/glsl/ralloc.h
This is another talloc replacement, but lacks some features we need
(getting size of an allocation, debugging features, being able to
access children in the dtor).
There's some information in ta/README what will happen next and how the
transition is expected to progress.
Change talloc destructor so that they can never signal failure, and
don't return a status code. This makes our talloc copy even more
incompatible to upstream talloc, but on the other hand this is
preparation for getting rid of talloc entirely.
(The talloc replacement in the next commit won't allow the talloc_free
equivalent to fail, and the destructor return value would be useless.
But I don't want to change any mpv code either; the idea is that the
talloc replacement commit can be reverted for some time in order to
test whether the talloc replacement introduced a regression.)
This only shows any differences when mpv isn't frontmost and is in fullscreen.
Cmd+Tab overlay is still at a higher level as to avoid complete usability fail.
Before this, they were shown on terminal only. Now they use the OSD
mechanism, which shows them on the video window or the terminal,
depending on settings and what's available.
Changing volume when audio is disabled was a feature request (github
issue #215), and was introduced with commit 327a779.
But trying to fix github issue #280 (volume is not correct in no-audio
mode, and if audio is re-enabled, the volume set in no-audio mode isn't
set), I concluded that it's not worth the trouble and the current
implementation is questionable all around. (For example, you can't
change the real volume in no-audio mode, even if the AO is open - this
could happen with gapless audio.) It's hard to get right, and the
current mixer code is already hilariously overcomplicated. (Virtually
all of mixer.c is an amalgamation of various obscure corner cases.)
So just remove this feature again.
Note that "options/volume" and "options/mute" still can be used in
idle mode to adjust the volume used next time, though these properties
can't be used during playback and thus not in audio-only mode.
Querying the volume still "works" in audio-only mode, though it can
return bogus values.
The argument or this change is that --loop should set how often the
file is played, not the number of additional repeats.
Based on pull request 277, with additions to the manpage and removal
of "--loop=0".
Signed-off-by: wm4 <wm4@nowhere>
The primary effect of this commit is that it fixes playback resume if
libquvi 0.9 and German locale are used, and libkdecore is on the system.
(See github issue #279.)
libquvi uses libproxy to determine system-wide proxy settings. libproxy
in turn loads libkdecore (if present) to determine KDE proxy settings.
libkdecore has a global constructor, which calls setlocale(LC_ALL, ""),
switching the current locale from "C" to what the user's settings. This
breaks some basic C string processing functions. Note that the locale
won't be set on program start, but only when libproxy calls dlopen() on
its config_kde module, which actually causes libkdecore to be loaded and
initialized.
In particular, with German locale, snprintf() would use "," instead of
"." when formatting float values, which in combination with playback
resume, would lead to parse errors the next time mpv was started, which
is how this issue was found.
I'd consider this a bug with libkdecore or at least libproxy. No library
should ever even touch locale: it might break basic expectations on C
string handling functions, might override program settings, and it's not
thread-safe. But this is so nasty and severe, that a quick hack to fix
it hurts less.
See github issue #279 and KDE bug #325902.
Before this patch, those will cause a crash:
mpv -playlist /dev/null
mpv -playlist <(bla) # if the result of bla is empty
Signed-off-by: Mohammad Alsaleh <[email protected]>
Signed-off-by: Stefano Pigozzi <[email protected]>
previously, mp_ring_read_cb() would allow reading past the end of the
ringbuffer when buffered > available. mp_ring_read() had logic for
splitting the read into two, which I duplicated into mp_ring_read_cb().
Since ordered editions can reference external files, when an ordered
chapter references another ordered edition, we have to go and figure out
what is is referencing as well.
When playing a Matroska file with ordered chapters, it may reference
another file by edition uid. When this edition is also ordered, it may
reference other files. When this occurs, the new segment/edition pair is
added to the list of sources to search for.
To support edition references in matroska chapters, editions need to be
remembered for each chapter and source. To facilitate easier management
of these now-paired uids, a single structure is used.
First, don't try to seek if the result is 0 (i.e. nothing found, or
subtitle event happens to be exactly on spot).
Second, since we never can make sure that we actually seek to the exact
subtitle PTS (seeking "snaps" to video PTS), offset the seek by 10ms.
Since most subtitle events are longer than 10ms, this should work fine.
The OSC will now display cache fill status between the timecodes, but only if it's below 48% to not clutter the interface with erratically changing values.
By default, the displayed value is multiplied by 2 to not confuse users who are unfamillar with the inner workings of the caching system. This can be disabled using the iAmAProgrammer=true setting.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@36461 b3059339-0415-0410-9bf9-f77b7e298cf2
Fixes playback of http://mpg123.org/test/44and22.mp3
Cherry-picked from MPlayer SVN rev. #36461, a patch by
Thomas Orgis, committed by by Reimar Döffinger.
You couldn't jump to the first entry in the playlist, if that entry was
marked as "too short". But this is usually fine, because it doesn't get
into the way of the user. (This feature is meant to allow being able to
jump backwards through the playlist even if a file immediately stops
playback right after initialization, not to prevent playing these files
at all.)
Also, apply the skip logic when wrapping around the playlist when going
backwards. This happens with --loop, because looping pretends the
playlist is infinitely repeated forwards and backwards (as long as the
playlist_prev/next commands are concerned).
Also add some comments.
When for example switching off the video stream, and --force-window is
used, forcefully reconfigure the VO. This will reset the size, and will
clear the window with black.
Needed some effort, because you don't always want to clear the window
on "discontinuity" points like going to a next file: this would
introduce some flicker.
glXGetVisualFromFBConfig() specifies specifies that it can return NULL
if there is no associated X visual. Instead of crashing, let
initialization fail. I'm not sure if this is actually supposed to work
with a fallback visual (passing a NULL visual to vo_x11_config_vo_window
would just do this), but let's play safe for now.
Apparently this can happen when trying to use vo_opengl over a remote
X display.
Output silence to the output buffer during underruns. This removes small
occasional glitches that happen before the AUHAL is actually paused from the
`audio_pause` call.
Fixes#269
Reverts a small change made in commit ed9295c. This is needed, because
otherwise mplayer.c/update_video_attached_pic() thinks it never has to
update the picture after initialization. (Maybe there would be more
elegant ways to handle this, but not without adding extra state.)
Didn't handle VO events, didn't handle OSD message management.
There is probably still some strangeness left. --idle mode was never
made for direct interaction.
We can render subtitles if a VO is open. Whether we're decoding video
(i.e. if mpctx->sh_video is set) doesn't really matter.
Subtitle display with --force-window still doesn't quite work yet,
because there's nothing to actually force redrawing of subtitles in this
mode.
This commit adds the --force-window option, which will cause mpv always
to create a window when started. This can be useful when pretending that
mpv is a GUI application (which it isn't, but users pretend anyway), and
playing audio files would run mpv in the background without giving a
window to control it.
This doesn't actually create the window immediately: it only does so
only after initializing playback and when it is clear that there won't
be any actual video. This could be a problem when starting slow or
completely stuck network streams (mpv would remain frozen in the
background), or if video initialization somehow is stuck forever in
an in-between state (like when the decoder doesn't output a video
frame, but doesn't return an error either). Well, we can pretend only
so much that mpv is a GUI application.
Seeking normally resets the VO and throws away the currently displayed
frame, so if you seek outside the video with --keep-open enabled, the
window would normally be "stuck" and not redraw properly, because there
is no source video frame that could be redrawn. To deal with this, a
precise seek to the position of the last displayed video frame was
issued.
This usually worked, but it can fail for formats where seeking is broken
or not possible (consider reading from a pipe).
Fix this by changing the semantics for vo_seek_reset(): now the video
frame is remembered even after seeking.
Note that this changes behavior a little when trying to seek outside of
a file with --keep-open enabled. Since no actual seek is done anymore,
the video will remain "frozen" on the previous position, and you can't
unpause or framestep to see the video between current position and
actual end of the video. If users complain, I might revert this commit.
vo_vdpau is the only VO which implements VOCTRL_RESET. Redrawing the
last output frame is hard, because the output could consist of several
source video frames with certain types of post-processing
(deinterlacing). Implement redrawing as special case by keeping the
previous video frames aside until at least one new frame is decoded.
This improves the previous commit, but is separate, because it's rather
complicated.
Before, a VO could easily refuse to respond to VOCTRL_REDRAW_FRAME,
which means the VO wouldn't redraw OSD and window contents, and the
player would appear frozen to the user. This was a bit stupid, and makes
dealing with some corner cases much harder (think of --keep-open, which
was hard to implement, because the VO gets into this state if there are
no new video frames after a seek reset).
Change this, and require VOs to always react to VOCTRL_REDRAW_FRAME.
There are two aspects of this: First, behavior after a (successful)
vo_reconfig() call, but before any video frame has been displayed.
Second, behavior after a vo_seek_reset().
For the first issue, we define that sending VOCTRL_REDRAW_FRAME after
vo_reconfig() should clear the window with black. This requires minor
changes to some VOs. In particular vaapi makes this horribly
complicated, because OSD rendering is bound to a video surface. We
create a black dummy surface for this purpose.
The second issue is much simpler and works already with most VOs: they
simply redraw whatever has been uploaded previously. The exception is
vdpau, which has a complicated mechanism to track and filter video
frames. The state associated with this mechanism is completely cleared
with vo_seek_reset(), so implementing this to work as expected is not
trivial. For now, we just clear the window with black.
Trying to connect multiple mpv clients to JACK with the
JackUseExactName option would fail unless the user manually
specifies a unique client name. This changes the behavior
to automatically generate a unique name if the requested
one is already in use.
vo_x11 had a clever trick to implement a video equalizer: it requested a
DirectColor visual. This is a X11 mechanism which allows you to specify
a lookup table for each color channel. Effectively, this is a safe
override for the graphic card's gamma ramp. If X thinks the window
deserves priority over other windows in the system, X would temporarily
switch the gamma ramp so that DirectColor visuals can be displayed as
the application intends. (I'm not sure what the exact policy is, but in
practice, this meant the equalizer worked when the mouse button was
inside the window.)
But all in all, this is just lots of useless code for a feature that is
rarely ever useful. Remove it and use the libswscale equalizer instead.
(This comes without a cost, since vo_x11 already uses libswscale.)
One worry was that using DirectColor could have made it work better in
8-bit paletted mode. But this is not the case: there's no difference,
and in both cases, the video looks equally bad.
Considering that we use version control, maintaining an AUTHORS file
doesn't make much sense, and indeed it was barely touched since the time
mplayer2 forked from MPlayer. None of the contributors to mplayer2 or
mpv are listed in this file (even uau's work is not appropriately
reflected, despite having an entry in this file).
I'm removing this file because it's obviously pointless to keep around,
and moreover it's misleading.
In theory removing this file might be questionable: the git log reaches
back into CVS and SVN times, which don't have separate committer and
author fields for each comment. The commit messages could be incomplete
and not mention the original author, while the AUTHORS file somehow
could. I doubt this is a real concern, since you can always reconstruct
the original AUTHORS file from git.
Another concern is that this could be interpreted as disrespectful to
the people mentioned in this file; but it isn't meant to be
disrespectful.
Also point out a slight irregularity in the Copyright file.
Commit 884c179 attempted to make it possible to skip backwards through
the playlist, even for files which fail to intitialize, or play for a
very short time. This was also used to prevent mpv from looping forever
and doing nothing if --loop=inf is used, and no file in the playlist is
playable.
This broke looping of very short files, because mpv was assuming that
this case happened. But there are legitimate use cases.
Fix this by making the looping case special. Instead of checking whether
playback was "very short", check whether something could be decoded and
displayed/played. If yes, allow looping.
After rebasing my dev branch it turned out that the code deadlocked on
recursive calls of `vo_control`. Make the locking code a little bit smarter
by making always skip locking/unlocking if we are executing a chunck of
code that is already synchronized with `dispatch_sync`.
Split the code to several files. The GUI elements now each have they own files
and private state. The original code was a mess to respect the retarded mplayer
convention of having everything in a single file.
This commit also seems to fix the long running bug of artifacts showing
randomly when going fullscreen using nVidia GPUs.
Don't allocate a VAImage and a mp_image every time. VAImage are cached
in the surfaces themselves, and for mp_image an explicit pool is
created. The retry loop runs only once for each surface now.
This also makes use of vaDeriveImage() if possible.
Until now, mouse positions were just passed to the core as-is, even if
the mouse coordinates didn't map to any useful coordinate space, like
OSD coordinates. Lua scripting (used by the OSC, the only current user
of mouse input) had to translate mouse coordinates manually to OSD space
using mp_get_osd_mouse_pos(). This actually didn't work correctly in
cases mouse coordinates didn't map to OSD (like vo_xv): the mouse
coordinates the OSC got were correct, but input.c was still expecting
"real" mosue coordinates for mouse areas.
Fix this by converting to OSD coordinates before passing the mouse
position to the core.
Do this so that MOUSE_LEAVE can't be combined with other keys. (E.g.
keep 'w' pressed, then move the mouse outside of the mpv window; it will
print a warning what w-MOUSE_LEAVE is not mapped.)
It appears the last run missed all mp_tmsg().
Since the command parser also prints messages, it needs a new parameter
to pass a log context. We go the easy way and just require the input_ctx
to be passed, and use its mp_log.
There is uninitialized memory access if the actual size isn't passed
along. In the worst case, this can cause a source to be loaded against
the uninitialized memory, causing a false count of found versus required
sources, preventing the "Failed to find ordered chapter part" message.
This is pretty much a hack for the OSC. It will allow it to rely on a
somewhat predictable style, instead of having to overwrite all user
OSD settings manually with override tags.
Nothing really accesses it. Subtitle initialization actually does in a
somewhat meaningful way, but there container size is probably fine, as
subtitles were always initialized before the first video frame was
decoded.
Now writing -1 to the 'aspect' property resets the video to the auto
aspect ratio. Returning the aspect from the property becomes a bit more
complicated, because we still try to return the container aspect ratio
if no frame has been decoded yet.
This function would probably be useful in other places too.
I'm not sure why vd.c doesn't apply the aspect if it changes size by
less than 4 pixels. Maybe it's supposed to avoid ugly results with bad
scalers if the difference is too small to be noticed normally.
This is preliminary. There are still tons of issues, and any aspect
of scripting may change in the future. I decided to merge this
(preliminary) work now because it makes it easier to develop it, not
because it's done. lua.rst is clear enough about it (plus some
sarcasm).
This requires linking to Lua. Lua has no official pkg-config file, but
there are distribution specific .pc files, all with different names.
Adding a non-pkg-config based configure test was considered, but we'd
rather not.
One major complication is that libquvi links against Lua too, and if
the Lua version is different from mpv's, you will get a crash as soon
as libquvi uses Lua. (libquvi by design always runs when a file is
opened.) I would consider this the problem of distros and whoever
builds mpv, but to make things easier for users, we add a terrible
runtime test to the configure script, which probes whether libquvi
will crash. This is disabled when cross-compiling, but in that case
we hope the user knows what he is doing.
This time it broke because I didn't actually test compiling vo_vaapi.c,
and it was using a macro from mp_image.h, which implicitly assumed
FFALIGN was available. Screw that too, and copy the definition of
ffmpeg's FFALIGN to MP_ALIGN_UP, and move these macros to mp_comnon.h.
The code using FFSWAP was moved from vo_vaapi.c to vaapi.c, which didn't
include libavutil/common.h anymore, just libavutil/avutil.h. The header
avutil.h doesn't include common.h recursively in Libav, so it broke
there.
Add FFSWAP as MPSWAP in mp_common.h (copy pasted from ffmpeg) to make
sure this doesn't happen again. (This kind of stuff happens all too
often, so screw libavutil.)
This code is actually quite inefficient: it reuses the (slow, simple)
screenshot code. It uses an inefficient method to read the image
(vaGetImage() instead of vaDeriveImage()), allocates new memory for
each frame that is read, and it tries all image formats again each
time.
Also, in my tests it always picked NV12 as image format, which is not
ideal if you actually want to filter the video, and vo_xv can't handle
this format without conversion either.
However, a user confirmed that it worked for him, so everything is fine.
This will allow GPU read-back with process_image.
We have to restructure how init_vo() works. Instead of initializing the
VO before process_image is called, rename init_vo() to
update_image_params(), and let it update the params only. Then we really
initialize the VO after process_image.
As a consequence of these changes, already decoded hw frames are
correctly unreferenced if creation of the filter chain fails. This
could trigger assertions on VO uninitialization, because it's not
allowed to reference hw frames past VO lifetime.
Merged from pull request #246 by xylosper. Minor cosmetic changes, some
adjustments (compatibility with older libva versions), and manpage
additions by wm4.
Signed-off-by: wm4 <wm4@nowhere>
This code was made inactive some months ago. At this time it wasn't
entirely clear whether this code was still needed, but now I'm pretty
sure it isn't. Even if it is, it didn't work anymore.
This wasn't enough and could lead to a cut off message shown on OSD.
Just make it dynamic, since we already use dynamic memory allocation
at this point anyway.
Moving the window was convenient but generates a MOUSE_LEAVE event
which it shouldn't. Now we remove it, because it is still possible
to move the window in weston with MOD+BTN0.
The problem with DVD/BD and playback resume is that most often, the
filename is just "dvd://", while the actual path to the DVD disk image
is given with --dvd-device. But playback resume works on the filename
only.
Add a pretty bad hack that includes the path to the disk image if the
filename starts with dvd://, and the same for BD respectively. (It's a
bad hack, but I want to go to bed, so here we go. I might revert or
improve it later, depending on user feedback.)
We have to cleanup the global variable mess around the dvd_device.
Ideally, this should go into MPOpts, but it isn't yet. Make the code
paths in mplayer.c take MPOpts anyway.
By default, libavformat uses UDP for rtsp playback. This doesn't work
very well. Apparently the reason is that the buffer sizes libavformat
chooses for UDP are way too small, and switching to TCP gets rid of this
issue entirely (thanks go to Reimar Döffinger for figuring this out).
In theory, you can set buffer sizes as libavformat options, but that
doesn't seem to help.
Add an option to select the rtsp transport, and make TCP the default.
Also remove an outdated comment from stream.c.
Mainly for debugging. Usually, we just set options for all possible
protocols, and we can't really know whether a certain protocol is used
beforehand. That's also the reason why avio_open2() takes a dictionary,
instead of letting the user set options directly with av_opt_set(). Or
in other words, we don't know whether an option that could be set is an
error or not, thus we print the messages only at verbose level.
This is for properties that normally show a bar, and thus do not show an
OSD message (as per classic mplayer behavior). Setting an extra_msg
allows showing an OSD message anyway, except if OSD messages are
explicitly suppressed.
This refactors the whole show_property_osd() function a bit, and
replaces the weird sep field with a more general method.
Allows for example: --status-msg='${?pause==yes:(Paused) } ...' to
emulate the normal terminal status line. It's useful in other situations
too.
I'm a bit worried about extending this mini-DSL, and sure hope nobody
will implement a generic formula evaluator at some point in the future.
But for now we're probably safe.
In most cases, it's better if deinterlacing happens before any other
filtering, so prepend the filter to the user's filter list, instead
of appending it.
Instead of hardcoding a single filter. This might be helpful for
modeling the vaapi deinterlacer as a video filter. The idea is that a
software deinterlacer would be tried first, and if that fails (because
vaapi hardware decoding uses HW surfaces, which a software deinterlacer
does not accept), the vaapi filter would be tried.
Normally, we need this for Xutf8LookupString(). But we can just fall
back to XLookupString(). In fact, the code for this was already there,
the code was just never tested and was actually crashing when active
(see commit 2115c4a).
Calling them separately doesn't really make sense, and all existing
calls to them usually combined them. One subtitle difference was that
af_init() didn't wipe the filter chain if initialization of the chain
itself failed, but that didn't really make sense anyway.
Also remove af_init() from the code for setting balance in mixer.c. The
mixer should be in the initialized state only if audio is fully
initialized, so the af_init() call made no sense.
Note that the filter "editing" code in command.c doesn't really do a
nice job of handling errors in case recreating an _old_ (known to work)
filter chain unexpectedly fails, and this obscure/rare case might be
differently handled after this change.
The volume is set as soon as the audio chain is created again. This
works only in softvol mode. For system wide volume or otherwise
externally user controllable volume, this code is intentionally
disabled. It would be extremely weird if changing volume (while audio is
not initialized) would do nothing, and then suddenly change it when the
audio chain is created.
There's another odd corner case: the user-set volume will be thrown away
if it's set before the _first_ audio chain initialization. This is
because the volume restore logic recognizes a change from nothing to
softvol or an AO, and circumventing that would require additional code.
Also, we don't even know the start volume before that point.
Forcing the volume with --volume will can override the volume set during
no-audio mode, depending on the situation.
Note that this is intentionally never done if the AO or softvolume is
different, or if the current volume control method is thought to control
system wide volume (such as ALSA) or otherwise user controllable (such
as PulseAudio). The intention is to keep things robust and to avoid
messing with the user's audio settings as far as possible, while still
providing the ability to resume volume if it makes sense.
Commit broke text subtitles without embedded fonts. Will look for a better
solution later. Revert it for now, since I'm starting to get bug reports.
This reverts commit 4a9f618d9f.
XOpenIM can fail to find a valid input method, in which case it
returns NULL. Passing a NULL pointer to XCreateIC would cause a
crash, so fail VO init before that happens.
Before this commit there was just an error message, but the file descriptor was
still open. Now we close the file descriptor and prevent it from calling
endlessly. Also a CLOSE_WIN event is sent which closes the window eventually if
the action of CLOSE_WIN is set to quit or quit_watch_later.
Improves display of images and video with alpha channel, especially if
the transparent regions contain (supposed to be invisible) garbage
color values.
Refactor how mixer.c does volume/mute restoration and initialization.
Move to handling of --volume and --mute to mixer.c. Simplify the
implementation of these and hopefully fix bugs/strange behavior related
to using them as file-local options (this uses a somewhat dirty trick:
the option values are reverted to "auto" after initialization). Put most
code related to initialization and volume restoring in probe_softvol()
and restore_volume(). Having this code all in one place is less
confusing.
Instead of trying to detect whether to use softvol at runtime, detect it
at initialization time using AOCONTROL_GET_VOLUME (same with mute,
AOCONTROL_GET_MUTE). This implies we expect SET_VOLUME/SET_MUTE to work
if the GET variants work. Hopefully this is always the case.
This is also preparation for being able to change volume/mute settings
if audio is disabled, and for allowing restoring value with playback
resume.
Softvol always used a linear multiplier for volume control. This was
converted to dB, and then back to linear in af_volume. Remove this non-
sense. We still try to keep the command line argument to af_volume in
dB, though.
This is to avoid the 30s hang while mpv caches fonts. In practice all the
fonts an average user is going to use are embedded in mkv files so there is
no reason to build fontconfig's cache on all of OS X system directories.
I might add something similar for terminal usage, but I am highly undecided.
It's quite unlikely, but functions like mp_find_user_config_file() can
return NULL, e.g. if $HOME is unset.
Fix all the code that didn't check for this correctly yet.
Remove the ifdef hell from mp_find_user_config_file(). Move the win32
specific code (for MinGW and Cygwin) to path-win.c. The behavior should
be about the same, but I can't be sure due to lack of testing and
because the old path.c code was hard to follow. (I expect those who care
about windows will fix things, should issues pop up - sorry.)
One difference is that the new code will always force MPV_HOME. It looks
like the old code preferred the mpv config dir in the exe dir if it
exists.
Also, make sure MP_PATH_MAX has enough space, even if the equivalent
wchar_t string is not 0-terminated with PATH_MAX (because apparently the
winapi doesn't require this). (Actually, maybe we should just kill all
uses of PATH_MAX/MP_PATH_MAX.)
This is a bit "hard", because getenv() returns a static string, and we
can't just return an allocated string. We also want getenv() to be
thread-safe if possible. (If the mpv core is going to be more threaded,
we sure do want the lower layers to be thread-safe as well.)
The homepath variable was static, and its value was set to a stack
buffer. This means a second invocation of the function would trigger
undefined behavior. Moreover the stack buffer always went out of scope
before homepath was used.
This is mainly to avoid spurious cursor states due to the mouse moving inside
or outside the window as a result of the window resize (with cmd-0/1/2).
This avoids complex logic and triggers a mouse move so that the player
recomputes the correct cursor state based on the autohide configuration of
the user.
This keeps the state in sync with the current state in cocoa_common. Infact the
cocoa code in mpv can decide wether it really wants to hide the cursor based on
the result of the `canHideCursor` method (this is so that the cursor is only
hidden when hovering on the video window).
Turns out that these checks were for versions of OS X that mpv doesn't even
support anymore. So just remove the checks since they cause a deprecation
warning.
GetCurrentProcess() is deprecated on 10.9. Make a universal solution by
checking OS version number.
get_system_version() function is the recommended Apple way of getting the
OS version, since Gestalt is also deprecated (and does pretty much the same
thing anyway)
Updating HIDRemote.m to use a similar function would allow to get rid of the
2 other warnings.
This is supposed to reduce the amount of useless error messages shown
during initialization of vo_opengl. If multiple backends are compiled,
usually only one of them will work. For example, on Linux both X and
Wayland backends can be compiled, but usually either Wayland or X is
running. Then, if Wayland is not running, but X is, trying to initialize
the Wayland backend should not spam the terminal with error messages.
Signed-off-by: Andreas Sinz <[email protected]>
Call update_subtitles() on every iteration of the playloop, so that
subtitle packets are read as soon as possible, instead of every time a
video frame is displayed. This helps in case the packet queue is swamped
with subtitle packets, which can happen with certain insane mkv files.
The change will simply cause the subtitle queue to be emptied on each
playloop iteration.
The timestamps update_subtitles() uses for display are the same before
and after this commit. (Important for files which have subtitle packets
with timestamps or duration not set.)
In insane files with a very huge number of subtitle events, and if the
--demuxer-mkv-subtitle-preroll option is given, seeking can still
overflow the packet queue. Normally, the subtitle_preroll variable
specifies the maximum number of packets that can be added. But once this
number is reached, the normal seeking behavior is enabled, which will
add all subtitle packets with the right timestamps to the packet queue.
At this point the next video keyframe can still be quite far away, with
enough subtitle packets on the way to overflow the packet queue.
Fix this by always setting an upper limit of subtitle packets read
during seeking. This should provide additional robustness even if the
preroll option is not used.
This means that even with normal seeking, at most 500 subtitle packets
are demuxed. Packets after that are discarded.
One slightly questionable aspect of this commit is that subtitle_preroll
is never reset in audio-only mode, but that is probably ok.
The mpv.desktop file is taken from the Arch package [1]. It appears to
be based on the mplayer2 git mplayer.desktop file (e.g. very similar
MimeType field), with minor modifications applied by Arch package
maintainers.
Note that for now, this doesn't show a terminal (Terminal=false), which
might not always be ideal. For example, if the file is audio only, or
if VO initialization fails for some reason, mpv will run in the
background and play audio without showing a window. But users prefer
running it without terminal, and don't want to play audio files with
it.
Maybe a --force-window option will be added in the future, which would
always create a VO window, and compensate for these issues.
[1] https://aur.archlinux.org/packages/mpv-git/
These would have to be updated manually all the time. Replacing them
automatically would be possible, but additional work, and would force
regeneration of the manpage way too often.
We decided that we don't need these fields.
Instead of containing a format string within %w{...}, simply allow %w
to specify one item of a time format string. This is simpler, more like
other format specifiers (%t), and probably easier to use too.
This is for situations when repeated attempts at playing a playlist
entry failed, and playlist navigation becomes impossible due to that.
For example, it wasn't possible to skip backwards past an unplayable
playlist entry:
mpv file1.mkv doesntexist.mkv file3.mkv
You couldn't skip back to file1.mkv from file3.mkv. When running a
single "playlist_prev" command, doesntexist.mkv would be played, which
would fail to load. As reaction to the failure to load it, the next file
would be played, which is file3.mkv.
To make this even worse, the file could successfully load, but run only
for a split second. So just loading successfully isn't good enough.
Attempt to solve this by marking problematic playlist entries as failed,
and by having playlist_prev skip past such playlist entries. We define
failure as not being able to play more than 3 seconds (or failing to
initialize to begin with). (The 3 seconds are in real time, not file
duration.)
"playlist_prev force" still exhibits the old behavior.
Additionally, use the same mechanism to prevent pointless infinite
reloading if none of the files on the playlist exist. (See github issue
All in all, this is a heuristic, and later adjustments might be
necessary.
Note: forward skips (playlist_next) are not affected at all. (Except for
the interaction with --loop.)
This is commonly used to disable the screensaver with broken/non-
standard X screensavers. During pause, the screensaver should not be
disabled, so not calling this command while paused seems sensible.
See github issue #236.
They are rarely useful in my opinion.
This commit was mainly motivated by this message:
Video uses a non-standard and wasteful way to store B-frames ('packed B-frames'). Consider using a tool like VirtualDub or avidemux to fix it.
It's what's left over from the "Invalid and inefficient vfw-avi..."
warning that used to be printed when playing avi/divx files. Although
the new message is much better, it's still rather useless and poses
more questions than it answers. Besides, nobody wants to remux a file
when playing it, especially not if playback appears to be completely
fine. (There are some claims that these files raise CPU usage, but even
my old crappy CPU can decode low res avi/divx files at real time at
about x35 playback speed.)
I have a sample where some final chapters are missing. This was causing a
segmentation fault when trying to fetch chapter times for them.
This makes the code ignore those chapters.
The quicktime html scripting guide suggests to wrap urls not
necesarly associated with quicktime in a .mov file.
(so that when <embed>ing videos quicktime would be forced.)
These mov files may contain several "Text Hacks".
One of these is RTSPtext.
The suggested/allowed format (as regex) is like:
RTSPtext[ \r]RTSP://url
See also p.51 of:
https://developer.apple.com/library/mac/documentation/QuickTime/Conceptual/QTScripting_HTML/QTScripting_HTML.pdf
In reality there are also files like (e.g. zdfmediathek.de):
RTSPtext\nrtsp://url\n\n
Lets handle these files as a playlist with one element.
In init_vo(), if sh->aspect is 0 or last_sample_aspect_ratio is set,
sh->aspect is overwritten. With software decoding fallback behaviour,
this makes the aspect ratio from container ignored since
last_sample_aspect_ratio is already set in first try with hardware
decoding.
The --deinterlace option does on playback start what the "deinterlace"
property normally does at runtime. You could do this before by using the
--vf option or by messing with the vo_vdpau default options, but this
new option is supposed to be a "foolproof" way.
The main motivation for adding this is so that the deinterlace property
can be restored when using the video resume functionality
(quit_watch_later command).
Implementation-wise, this is a bit messy. The video chain is rebuilt in
mpcodecs_reconfig_vo(), where we don't have access to MPContext, so the
usual mechanism for enabling deinterlacing can't be used. Further,
mpcodecs_reconfig_vo() is called by the video decoder, which doesn't
have access to MPContext either. Moving this call to mplayer.c isn't
currently possible either (see below). So we just do this before frames
are filtered, which potentially means setting the deinterlacing every
frame. Fortunately, setting deinterlacing is stable and idempotent, so
this is hopefully not a problem. We also add a counter that is
incremented on each reconfig to reduce the amount of additional work per
frame to nearly zero.
The reason we can't move mpcodecs_reconfig_vo() to mplayer.c is because
of hardware decoding: we need to check whether the video chain works
before we decide that we can use hardware decoding. Changing it so that
this can be decided in advance without building a filter chain sounds
like a good idea and should be done, but we aren't there yet.
I did commit 86c05655d by thinking `mpv` already removed the `mpv` from
argc/argv. It actually is still there, so the argc must be 1 to check for no
arguments.
Thanks to @Nyx0uf for pointing out the bug and for testing on 10.9!
File opening through Finder, apparently drops `--psn` arguments on Mavericks
and just uses no args. Modify the code to account for that case.
This wasn't tested on 10.9 itself (I don't have a paid dev account), but it
*should* work if I understood the problem correctly.
Problem: I own the buffer and I destroyed while still being displayed.
Solution: Add a temporary buffer and destroy it when the next buffer is
attached.
Previously, mpv incorrectly used the %HOME% environment variable on
MinGW to determine the current user’s home directory. This is wrong;
the correct variable to use would be %HOMEPATH%, which would however
still be wrong since application data goes into the application data
directory, not the user’s home. This patch makes it use the local
AppData path instead of reading an environment variable.
This however exposed another problem (which also affected users who
actually had the %HOME% variable set):
b2c2fe7a37 (discussed in issue #95) introduced some changes that
make mpv load user config files from the executable path on Windows.
The problem with this change is that config_dir was still declared
static, so once a config file had been found in the executable path,
it would set config_dir to an empty string, so mpv would dump e.g.
watch_later data straight into the user’s home. This commit also
fixes that.
One side effect of this is that mpv no longer considers the “mpv”
subdirectory in the executable path (that behavior resulted from
the homedir variable always being empty), unless it is somehow
unable to determine the local AppData path.
This is mostly related to the fullscreen behaviour. cecbd8864 introduces an
option to make mpv behave like a OSX user would expect. This commit changes
the Cocoa parts of the code to be consistent with the behaviour on X11. Old
behaviour is still available through the option mentioned in cecbd8864.
There is still custom logic in the cocoa backend and it can probably be moved
to core:
* Don't perform autohide if the mouse is down
* Don't perform autohide outside of the video window
Fixes#218 (by accident)
libquvi 0.4 doesn't allow us listing the formats supported by a
streaming site without doing additional network accesses, so switching
formats was not supported with it. (It's different with libquvi 0.9.)
But the most important case is switching between SD and HD. Usually,
--quvi-format=default will get SD, while --quvi-format=best gives HD.
Use this, and pretend that an URL supported by libquvi 0.4 supports both
of these. "cycle quvi-format" will switch between these. If the user
specifies something else via --quvi-format, this is included in the list
of switchable formats additionally to "default" and "best".
It's annoying for users if you can't get a list of options with --help,
but on the other hand, printing all options would be overkill. So just
mentioned --list-options.
Retrieve per-chapter metadata, but don't do much with it. We just make
the metadata of the _current_ chapter available as chapter-metadata
property. Returning the full chapter list with metadata would be no
problem, except that the property interface isn't really good with
structured data, so it's not available for now.
Not sure if it's worth it, but it was requested via github issue #201.
Consider the cluster used for prerolling contains an insane amount of
subtitle packets. Then the demuxer packet queue would be full of
subtitle packets, and demux.c would refuse to read any further packets -
including video and audio packets, resulting in EOF. Since everything
involving Matroska and subtitles is 100% insane, this can actually
happen.
Fix this by putting a limit on the number of subtitle packets read by
preroll, and throw away any further packets if the limit is exceeded. If
this happens, the preroll mechanism will stop working, but the player's
operation is unaffected otherwise.
Cherry picked from various commits in lua_experiment by ChrisK2.
The metrics of the OSD symbols change slightly, possibly due to the
font editor that was used, and the metrics were not correct to begin
with. (But the real reason seems unknown.) Remove the rescaling of
the OSD font in ASS_USE_OSD_FONT, because the height more or less fits
now. (This change wasn't in the lua_experiment branch.)
run_playloop() is already stuffed enough. This function is still quite
big, but all the other code shares various variables, so it's not as
easy to split.
This option makes the cursor always visible in windowed mode.
Apparently, this is what (some?) Windows and OSX users expect. It's
disabled by default for now.
Restructure the cursor hide logic a bit for this purpose.
Let all key events go through mp_input_feed_key() internally, and also
do double click and MP_INPUT_RELEASE_ALL handling there. Move
check_autorepeat() to where it's actually used.
This caused the OSC to be always visible at startup on X11:
- EnterNotify event send a mouse event to input.c
- OSC has not completely initialized yet, and no mouse area is set
- mouse event is dispatched to "showhide" OSC section
- OSC becomes visible, regardless of mouse position
Fix this by treating the mouse area as empty if it's not set, instead of
infinite as it was before this commit. This means an input section must
set a mouse area to receive mouse events at all. We also have to change
the default section to receive mouse events with the new behavior.
Also, if MOUSE_MOVE is unmapped (or mapped to something that doesn't
parse), and produces no command, the mouse position wouldn't be updated
(because the mouse position is bound to input commands), so we have to
generate a dummy command in this case.
(This matters only for the OSC, On Screen Controller, which isn't merged
yet, so these changes shouldn't have much effect right now.)
Until now, any command was dropped as soon as the input queue was full,
and the command was not an abort command (i.e. a command that exits the
player or goes to the next file).
This could cause issues with key down events (especially mouse buttons)
not being released.
Change it so that key up events can never be dropped. This is a bit
involved, because we know whether a key maps to an abort command only
after interpreting it, and interpreting it changes global state, which
in turn requires undoing the event if the input is dropped. Refactor
the code a bit to move more functionality into interpret_key() to make
this easier.
Remove options which are too obscure and most likely not very useful in
general, or update them to something more modern. Add some comments
about how configuration files work in general.
Even if a subtitle was explicitly loaded with -sub, it was still auto-
loaded (if auto-loading applied to that file). Fix this by explicitly
checking whether a file is already loaded.
The check is maximal naive and just compares the filenames as strings.
The change in find_subfiles.c is so that "-sub something.ass" happens to
work (auto-loading prepended a "./" to it, so the naive filename
comparison check didn't work).
External vobsubs usually come as .idx/.sub pairs. Loading the .idx file
implicitly loads the .sub file, whereas loading the .sub file will kind
of work, but miss important information such as subtitle resolution. Or
in other words, if the .idx file exists, adding the .sub file as track
is useless and confusing.
Explicitly remove .sub file from the auto-load suntitle list in these
cases. Standalone .sub files are still loaded.
We also drop that weird logic that excluded .utf8 files from being
loaded if -subcp was in use. I hope the associated use case didn't make
much sense to begin with. If not, we could still implement it properly,
instead of this weird hack.
Use mp_path_exists() to check for existence of a file (which in turn
uses stat()), instead of opening and closing it. The difference is that
if we don't have sufficient permissions to read the subtitle files, we
will loudly complain. Personally, I prefer this behavior.
Until now, options could be accessed as properties via "options/name",
but the access was read-only. Change it so that write access is possible
in --idle mode. (All options have to support setting options at that
time, simply because of the way MPlayer designed per-file options.
During playback, normal properties take care of changing things,
including things backed by options.)
This is work in progress. There are some issues: at least setting the
"vf" and "af" options won't work for strange reasons.
Pretty useful for people writing userscripts for web browsers. Links starting
with 'mpv://' are forwarded to the mpv OSX bundle. The leading 'mpv://' is
stripped from the recived url and the rest of the string is inserted as is in
the playlist.
The --volume option accepted values up to 10000, but internally, the
value is always clipped to 0-100 range. What makes this even worse is
that --softvol-max suggests that it extends the range of --volume, which
is not the case. (And passing a volume larger than 100 to --volume
didn't even print a warning.)
The really funny thing about this commit is that this code is added on
top of another work around. Basically, subtitle seeking in libavformat
is completely broken. To make it useful, we have to add yet another
workaround.
The basic problem is that libavformat's subtitle seeking code always
uses the stream time base, instead of AV_TIME_BASE if stream index -1 is
passed to the avformat_seek_file() function.
Fixes github issue #216. Hopefully this will be fixed in ffmpeg too at
some point.
This affects MOUSE_MOVE and MOUSE_LEAVE. Both are needed internally
(such as for the OSC), but not really useful for input.conf. Since the
warning has the purpose of notifying the user that a key is unmapped and
what key name to use for setting up a binding in input.conf, the warning
is rather useless in this case. It's also annoying in combination with
the
--no-input-default-bindings option, since that removes the default
bindings to "ignore" for these keys.
When enabling --save-position-on-quit, playback position stored not only
on quit, but in any case playback of a file was stopped. This includes
going to the next file with playlist navigation commands.
After some discussion on IRC, it turned out that nobody thought this was
good behavior. Disable it, and really make it save only on quit.
Maybe the option is useless now, as the user could remap the CLOSE_WIN
key binding. On the other hand, CLOSE_WIN sounds and _is_ a bit obscure.
Before this commit, the player didn't write resume info if the playback
position was within the first or last percent of the file.
This was sometimes annoying, and with playlist resume can lead to
unintuitive behavior. (It wouldn't resume the playlist if the currently
played file was at 0-1% or 99-100%, even if you were in the middle of a
playlist.)
Moreover, the "percent > 99" check is a bit bogus anyway, because 100
(in integer) is rarely reached.
Drop the check, and make sure using --save-position-on-quit won't write
resume info when reaching EOF. The latter check is needed to make sure
playback of the file starts at beginning when playing it again after
EOF.
The previous check just searched for a "://" substring. This was quite
bad, because "://" can be a valid part of a path. Later, I added
special handling for filenames starting with "." and "/", so that you
could reliably pass arbitrary filenames to mpv, no matter how messed
up they were.
Even though it doesn't really matter in practise anymore, this is still
crap, so add a more reliable check instead.
This includes the case of passing multiple files to command line
(internally this is the same as loading a playlist).
Resuming works by finding the first playlist entry that can be resumed.
Alternative implementations would be possible, such as hashing the
playlist contents. But this implementation is simpler, and doesn't have
the disadvantage that changes to the playlist (like appending entries)
will throw away the resume point.
This makes loading large playlists a bit slower, because it has to look
into ~/.mpv/watch_later/ for every entry. Loading a 15000 entries
playlist now increases from 150ms to 400ms. Considering you rarely load
playlists this big with mpv (because it's impractical considering the
terminal and non-GUI nature of the player), this is probably ok.
Undo URL percent encoding if the filename appears to be an URL. This
will fix display of the actual filename in some cases.
We don't put any effort into checking whether the URL is really percent
encoded, because we don't really know how the protocol handler is going
to interpret the URL. For stream_lavf, we probably can't know. Still,
from the perspective of this commit, it seems to make sense to assume
they are percent encoded.
icon_size is the number of array items of type long, not bytes. Change
the type of icon_size to int, because size_t makes you think of byte
quantities too quickly.
As an unrelated change, change the (char *) cast to (unsigned char *),
because it matches the common XChangeProperty idiom better.
MPlayer handles this correctly, because MPlayer still has the FourCC
codec dispatch (codecs.conf). We need to handle this case specially,
because the libavformat rawvideo decoder will of course not eat mjpeg.
mjpeg is the only supported format, though. (Even MPlayer needs to
convert between V4L2 formats and MPlayer FourCCs, and mjpeg is the only
non-raw format.)
The option list contains an empty string member with this option value,
so ignore that. I'm not sure whether the option list should maybe be
empty in this case, but it could be the wrong thing in case of other
options.
This happens by default with pausing: if a file was paused when doing
quit_watch_later, then resume and unpause, then the file played after
that would start in paused mode. This is because the pause option is
backed up at thr wrong place, so it backs up the state from resuming,
instead of whatever it was set to before that.
This prevents keys to become stuck due to changing keyboard modifiers during
the key down. Not the prettiest approach but event `x11_common` does it like
this.
Fixes#210
The png file added to etc/ are taken from the link mentioned in commit
303096b, except that they have been converted to 16 bit, sRGB (with
color profile info dropped, if there was one), and transparent pixels
reset for better compression.
The file x11_icon.bin is generated by gen-x11-icon.sh. I'm adding it to
the git repo directly, because the script requires ImageMagick, and we
don't want to make building even more complicated.
The way how this is done is basically a compromise between effort
required in x11_common.c and in gen-x11-icon.sh. Ideally, x11_icon.bin
would be directly in the format as required by _NET_WM_ICON, but trying
to write the binary width/height values from shell would probably be a
nightmare, so here we go.
The zlib code in x11_common.c is lifted from demux_mkv.c, with some
modifications (like accepting a gzip header, because I don't know how to
make gzip write raw compressed data).
This allows other threads to use mp_input_put_key without blocking if the
playloop is doing the 500ms select call (i.e.: during pause).
Makes Cocoa GUI responsive again (regression since 2d363c3).
If the input section is enabled with MP_INPUT_ALLOW_VO_DRAGGING, then
the VO will be allowed to drag the window with mouse button down +
mouse move even if the mouse is inside the section's mouse area.
If the mpv window is unfocus, clicking on the OSC should focus the
window (done by the window manager) and allow interaction with the OSC.
But somehow X sends a spurious LeaveNotify event, immediately followed
by an EnterNotify event. This happens at least with IceWM. The result is
that the OSC will disappear (due to receiving MOUSE_LEAVE). The OSC will
stay invisible, because EnterNotify isn't handled, and there's nothing
that could make the OSC appear again.
Solve this by handling EnterNotify. We cause a redundant MOUSE_MOVE
event to be sent, which triggers the code to make the OSC visible. We
have to remove the code from input.c, which ignores redundant mouse move
events.
Since the code ignoring redundant mouse move events is still needed on
Windows, move that code to w32_common.c. The need for this is documented
in the code, also see commit 03fd2fe. (The original idea was to save
some code by having this code in the core, but now it turns out that
this didn't quite work out.)
Normally, moving the mouse outside of the mouse area of an input section
will send mouse events somewhere else (because input section mouse areas
are similar to windows/widgets in real GUI toolkits). This was done even
if a mouse button was held down. This is quite different from how GUI
toolkits behave.
Change the code so that if a mouse button is down, the mouse area of the
current input section can't be left. Releasing the mouse button (while
the mouse pointer is outside of the mouse area) will actually leave the
mouse area.
As a side-effect, this commit also tests more often whether the current
mouse input section is valid. This is needed to make releasing a mouse
button trigger the mouse input section change.
These keys can be found on various "multimedia" and "internet" keyboard.
X defines many keycodes, so I'm not adding all, just what I found on my
own keyboard.
Other key codes can be added on request.
This is basically a libavcodec API oddity: it can happen that
avcodec_decode_audio4() returns 0 (meaning 0 bytes were consumed). It
requires you to feed the complete packet again to decode the full
packet, and to successfully decode the following packets.
We ignored this case with the argument that there's the danger of an
endless decode loop (because nothing of that packet is apparently
decoded, so it would retry forever), but change it in order to decode
mpc8 files correctly.
Also add some comments to explain the mess.
I suspect most users will just copy etc/input.conf when they want to
remap some default bindings. But usually this means the user even copies
bindings he doesn't care about, and it's better if the user maps only
the bindings in his input.conf the user intends to remap.
Comment all bindings in etc/input.conf. Since this file also defines the
builtin defaults and is baked into the mpv binary, we have to do
something to get them anyway, even though they are commented. Do this by
having input.c "uncomment" the bindings in the baked in input.conf. (Of
course this is done only for the builtin config, not configs loaded from
disk.)
The previous code was locking only the input queue. That was too weak since it
didn't protect the input_ctx data structure. So remove the locking on the queue
and lock all the public functions that interact with the input_ctx.
The private functions and public functions that do not act on the input_ctx
(there are quite some functions doing mp_cmd manipulations for instance) are
not locked.
Some changes by wm4. Use a recursive mutex, and restructure some code to be
less annoying with locks, such as converting code to single return, or
making use of the recursive mutex.
Generate a mouse down event on the first click so that one can interact
with the OSC directly as opposed to wasting the first click in order to focus
the window.
I would like to thank Chris Ward (@tenzerothree, http://tenzerothree.com/) for
working on the art for these icons and bringing some eye candy to the project.
The PSDs made by Chris are available on our Dropbox [1], along with the exports
I made to create OSX and Windows icons. The PSDs are almost completly vector
and all the resolutions look really similar, except the 16px favicon which was
handcrafted to look better and more recognizeable on the smaller pixel budget.
For Mac OS X the icons were created using iconutils on the PNGs iconsets
exported from the PSDs. These even support retina resolutions (except 512@2x).
For Windows the .ico file was created with imagemagick. The included images
are 16px, 24px, 32px, 48px 64px, 256px. These are the resolutions listed on
MSDN for supporting Windows XP [2] and Windows versions based on Aero [3].
Only 32bit PNGs were used since it is 2013.
For Linux nothing changed yet, even though @wm4 talked about using the PNGs
directly there. This will probably be dealt with in a later commit.
[1]: https://www.dropbox.com/sh/yelfoj9tbft7o06/A8vOT6JKaG
[2]: http://msdn.microsoft.com/en-us/library/ms997636.aspx
[3]: http://msdn.microsoft.com/en-us/library/aa511280.aspx
This probably has been broken since bbc865a: a test was added that uses
a FBO, but it's always run, even if FBOs were not detected. On the other
hand, fbotex_init() just runs into an assert. Fix the test that
triggered this condition, and make fbotex_init() "nicer" by just failing
if FBOs are not available.
Port it from playlist_parser.c to demux_playlist.c. Also, change the m3u
parser to drop whitespace from the trailing part of the line (will make
it work properly with windows line endings).
(I hoped that this would make MMS URIs with http instead of mmsh
prefixes work, but it doesn't. Instead, it leads to a playlist loop. So
solving this issue would require a change in ffmpeg, probably.)
Until now, stream_peek() read only the bare minimum it had to read from
the stream. But this could cause problems, such as being very
inefficient when peeking a lot, or conflicting with ability to seek
back. (The latter issue can be caused by peeking a few bytes, and then
doing a stream_read() with a size that is 1 byte longer: this would read
the peeked data, then call stream_fill_buffer(), which throws away the
previously peeked bytes - so you can't seek back anymore. This is
mitigated by a hack in demux_open(): it peeks a full buffer, to avoid
that peeking/reading during demuxer probing [or before that, in a stream
filter] can cause the buffer to be dropped.)
The original condition was too weak, requiring only the header. The header is
installed is FFmpeg regardless of the presence of VDA on the system, so just
perform a check on the `ff_vda_create_decoder` function.
Make use of opaque regions on non-alpha formats. This allows the compositor to
improve the drawing of the surface, because he can discard everything behind
the window when drawing.
Odd video sizes if pixel formats with chroma subsampling and PBOs were
used, garbage was rendered. This was because the PBO path created
buffers with an unpadded size, and then tried to upload a padded
image to it. Fix it by explicitly setting the padded size. (As with
the non-PBO path, we rely that image allocations are somehow padded,
which is normally the case.)
Apparently, it is popular to store large files in uncompressed rar
archives. Extracting files is not practical, and some media players
suport playing directly from uncompressed rar (at least VLC and some
DirectShow components).
Storing or accessing files this way is completely idiotic, but it is
a common practice, and the ones subjected to this practice can't do
much to change this (at least that's what I assume/hope). Also, it's
a feature request, so we say yes.
This code is mostly taken from VLC (commit f6e7240 from their git tree).
We also copy the way this is done: opening a rar file by itself yields
a playlist, which contains URLs to the actual entries in the rar file.
Compressed entries are simply skipped.
Add a stream filter concept, in which streams can be opened on top of
an underlying "source" stream. Change the open code to make this
easier, and also to account for some mechanisms that will be needed
for this.
The following commit will add stream_rar, which contains such a stream
filter.
This is really not needed. While we really can't take a loaded buffer
over to the cache, there's no reason why the cache couldn't read this
buffer normally.
On the other hand, this code could cause trouble when probing from a
stream before the cache has been enabled.
The way the url_options field was handled was not entirely sane: it's
actually a flexible array member, so it points to garbage for streams
which do not initialize this member (it just points to the data right
after the struct, which is garbage in theory and practice). This was
not actually a problem, since the field is only used if priv_size is
set (due to how this stuff is used). But it doesn't allow setting
priv_size only, which might be useful in some cases.
Also, make the protocols array not a fixed size array. Most stream
implementations have only 1 protocol prefix, but stream_lavf.c has
over 10 (whitelists ffmpeg protocols). The high size of the fixed
size protocol array wastes space, and it is _still_ annoying to
add new prefixes to stream_lavf (have to bump the maximum length),
so make it arbitrary length.
The two changes (plus some more cosmetic changes) arte conflated into
one, because it was annoying going over all the stream implementations.
Modeled after the old playlist_parser.c, but actually new code, and it
works a bit differently.
Demuxers (and sometimes streams) are the component that should be used
to open files and to determine the file format. This was already done
for subtitles, but playlists still use a separate code path.
af_str2fmt_short(), which is used by the command line option parser,
allowed passing a hex number. The user could set arbitrary integers as
internal audio formats, even formats which don't exist or make no sense.
This is not very useful, so get rid of it.
Having to use -1 for that is generally quite annoying.
Audio formats are created from bitmasks, and it can't be excluded that
0 is not a valid format. Fix this by adjusting AF_FORMAT_I so that it
is never 0. Along with AF_FORMAT_F and the special formats, all valid
formats are covered and guaranteed to be non-0.
It's possible that this commit will cause some regressions, as the
check for invalid audio formats changes a bit.
A wayland output based on shared memory. This video output is useful for x11
free systems, because the current libGL in mesa provides GLX symbols. It is also
useful for embedded systems where the wayland backend for EGL is not
implemented like the raspberry pi.
At the moment only rgb formats are supported, because there is still no
compositor which supports planar formats like yuv420p. The most used compositor
at the moment, weston, supports only BGR0, BGRA and BGR16 (565).
The BGR16 format is the fastest to convert and render without any noticeable
differences to the BGR32 formats. For this reason the current (very basic)
auto-detection code will prefer the BGR16 format. Also the weston source code
indicates that the preferred format is BGR16 (RGB565).
There are 2 options:
* default-format (yes|no) Which uses the BGR32 format
* alpha (yes|no) For outputting images and videos with transparencies
The obtained information from the shm listener isn't used by anything and is
also wrong now in wayland git master branch because of new shm formats which
need a different way of saving the supported formats.
Moves a good chunk of the resizing code to wayland_common.c. This makes it
possible to share it with future video drivers.
It doesn't resizit it immediatly, it calcutlates the new position and size and
then shedules a resizing event. This removes the ugly callback and void pointer
from the wayland data structure.
Regression since 18b6c01d92. That commit changed the colorspace handling to
always reinit the video output. Since the CVPixelBuffers are lazily created,
VOCTRL_SET_YUV_COLORSPACE was always called when the CVPixelBufferRef was NULL.
Since CoreVideo functions do not complain when called on NULL, no one noticed
that CVBufferSetAttachment, which stored the color matrix meta data was called
on NULL.
Until now, video output levels (obscure feature, like using TV screens
that require RGB output in limited range, similar to YUY) still required
handling of VOCTRL_SET_YUV_COLORSPACE. Simplify this, and use the new
mp_image_params code. This gets rid of some code. VOCTRL_SET_YUV_COLORSPACE
is not needed at all anymore in VOs that use the reconfig callback. The
result of VOCTRL_GET_YUV_COLORSPACE is now used only used for the
colormatrix related properties (basically, for display on OSD). For
other VOs, VOCTRL_SET_YUV_COLORSPACE will be sent only once after config
instead of twice.
This affects VOs which just reuse the mp_image from draw_image() to
return screenshots. The aspect of these images is never different
from the aspect the screenshots should be, so there's no reason to
adjust the aspect in these cases.
Other VOs still need it in order to restore the original image
attributes.
This requires some changes to the video filter code to make sure that
the aspect in the passed mp_images is consistent.
The changes in mplayer.c and vd_lavc.c are (probably) not strictly
needed for this commit, but contribute to consistency.
mp_image_set_params() doesn't check whether the colorspace parameters
are consistent (e.g. setting YUV colorspaces with RGB formats), and
shouldn't need to.
The way this was added to FFmpeg is less than ideal, because it requires
text parsing in the Matroska demuxer. But in order to use the FFmpeg
webvtt-to-ass converter, we still have to mimic this in some way. We do
this by putting the parsing into sd_lavc_conv.c, before the subtitle
packet is passed to libavcodec. At least this keeps the ugliness out of
unrelated code.
There is some change that FFmpeg will fix their design eventually.
Instead of rewriting the parsing code, we simply borrow it from FFmpeg's
Matroska demuxer.
cocoa_common contains some locking calls to support video outputs that support
live resizing (at this moment only vo=opengl).
These should not be used unless the VO declares it is multithreaded by
registering the resize_redraw callback used for live resizing.
Fixes#200
Use the new MP_ macros for some AOs instead of mp_msg.
Not all AOs are converted, and some only partially. In some cases, some
additional cosmetic changes are made.
Otherwise, this would just try to demux a good chunk of the file, even
though the operation can't succeed anyway.
This caused some pretty strange issues, where perfectly valid use cases
would print a "Too many packets in the demuxer packet queue..." message.
The rawaudio demuxer read one frame per packet, basically a few bytes,
which caused insane overhead. (I found this when I couldn't play raw
audio without dropouts when using -v, which printed a line per packet
read.)
Fix this and read 1 second of audio per packet. This is a regression
since cfa5712 (merging of demux_rawaudio and demux_rawvideo).
Instead of always skipping in STREAM_BUFFER_SIZE blocks, allow an
arbitrary size. This allows - in theory - faster forward seeking in
pipes.
(Maybe not a very significant change, but it reduces the number of
things that depend on STREAM_BUFFER_SIZE for no good reason. Though
we still use that value as minimum read size.)
stream_file.c contains some code meant for forward seeking with pipes.
This simply reads data until the seek position is reached. Move this
code to stream.c. This stops stream_file from doing strange things
(messing with stream internals), and removes the code duplication too.
We also make stream_seek_long() use the new skip code. This is shorter
and much easier to follow than the old code, which basically did strange
things.
Decoding H264 using Video Decode Acceleration used the custom 'vda_h264_dec'
decoder in FFmpeg.
The Good: This new implementation has some advantages over the previous one:
- It works with Libav: vda_h264_dec never got into Libav since they prefer
client applications to use the hwaccel API.
- It is way more efficient: in my tests this implementation yields a
reduction of CPU usage of roughly ~50% compared to using `vda_h264_dec` and
~65-75% compared to h264 software decoding. This is mainly because
`vo_corevideo` was adapted to perform direct rendering of the
`CVPixelBufferRefs` created by the Video Decode Acceleration API Framework.
The Bad:
- `vo_corevideo` is required to use VDA decoding acceleration.
- only works with versions of ffmpeg/libav new enough (needs reference
refcounting). That is FFmpeg 2.0+ and Libav's git master currently.
The Ugly: VDA was hardcoded to use UYVY (2vuy) for the uploaded video texture.
One one end this makes the code simple since Apple's OpenGL implementation
actually supports this out of the box. It would be nice to support other
output image formats and choose the best format depending on the input, or at
least making it configurable. My tests indicate that CPU usage actually
increases with a 420p IMGFMT output which is not what I would have expected.
NOTE: There is a small memory leak with old versions of FFmpeg and with Libav
since the CVPixelBufferRef is not automatically released when the AVFrame is
deallocated. This can cause leaks inside libavcodec for decoded frames that
are discarded before mpv wraps them inside a refcounted mp_image (this only
happens on seeks).
For frames that enter mpv's refcounting facilities, this is not a problem
since we rewrap the CVPixelBufferRef in our mp_image that properly forwards
CVPixelBufferRetain/CvPixelBufferRelease calls to the underying
CVPixelBufferRef.
So, for FFmpeg use something more recent than `b3d63995` for Libav the patch
was posted to the dev ML in July and in review since, apparently, the proposed
fix is rather hacky.
Regression since ff3b98d11c. The window positioning code relied on the
visibleFrame's height without taking into account the dock's presence.
Also moved the constraining code to the proper method that overrides the
original NSWindow behaviour. This avoids having to check for border since the
constraining is performed by Cocoa only for titled windows.
Fixes#190
Too many people thought "D:" really meant number of dropped frames. But
it's actually the number of frames where the playloop thought it'd be
a good idea to drop them. Of course this does nothing if frame dropping
is disabled, but even with normal frame dropping, this doesn't indicate
whether a frame was _really_ dropped. (Looks like libavcodec doesn't
even give us this information reliably? The decode function can return
no frame in case of codec delay due to threading and such.)
Originally, the objective of this commit was changing --edition to be
1-based, but this was cancelled. I'm still leaving the change to
demux_mkv.c though, which is now only of cosmetic nature.
We don't need to store the offsets of the options corresponding to the
properties, because the option-property bridge knows about this already.
The check against 1000 was for the case if e.g. the --brightness option
is not used. Always overwrite the option value instead, both when
querying and setting the property. (This is needed to make the settings
persistent even if vf_eq is used and the video chain is reinitialized.)
This commit assumes that VFCTRL_SET_EQUALIZER is always paired with
VFCTRL_GET_EQUALIZER (likewise for VOCTRL), which is the case.
Using -vf eq and changing brightness, contrast, etc. using key bindings
with e.g. "add brightness 1" didn't work well: with step width 1, the
property gets easily "stuck". This is a rounding problem: e.g. setting
gamma to 3 would actually make it report that gamma is set to 2, so
the "add" command will obviously never reach 3 with a step width of 1.
Fix this by storing the parameters as integers.
This was broken in cac7702. This commit effectively changed these
properties to use the value as reported by vf_eq, instead of the
previously set value for the "add" command. This was more robust,
but not very correct either, so we keep the new behavior and make
vf_eq report its parameters more accurately.
The display, window, keyboard and cursor structures are now cleanly and
logically separated. Also could prevent a future bug where no shm format is set
when the cursor image is loaded (Never happened until now).
Add --video-align-x/y, --video-pan-x/y, --video-scale options and
properties. See the additions to the manpage for description and
semantics.
These transformations are intentionally done on top of panscan. Unlike
the (now removed) --panscanrange option, this doesn't affect the default
panscan behavior. (Although panscan itself becomes kind of useless if
the new options are used.)
This option allowed you to extend the range of the panscan controls, so
that you could essentially use it to scale the video. This will be
replaced by a separate option to set the zoom factor directly.
See github issue #194.
Unfortunately, this breaks the property that going back in the playlist
always works as expected. This changes, because the playlist_prev
command will work on the reshuffled playlist, instead of loading the
previously played files in order. If this ever becomes an issue, I
might revert this commit.
Now the code does the same as the original MPlayer VAAPI patch, instead
of trying to map the profiles exactly.
See previous commit for justification and discussion.
Instead, do what MPlayer did all these years. It worked for them, so
there's probably no reason to change this.
Apparently fixes playback with some files, where the VDPAU decoder does
not formally support a profile, but decoding works with a more powerful
profile anyway.
Though note that MPlayer did this because it couldn't do it in a better
way (no decoder reported profiles available when creating the VDPAU
decoder), so it's not entirely clear whether this is a good idea. An
alterbative implementation might try to map the profiles exactly, and
do some fall backs if the exact profile is not available. But this
hack-solution works too.
Apparently this was dropped some years ago, but judging from MPlayer's
handling of this, the original code wasn't so great anyway. The new
code handling clearing of panscan borders correctly, and integrates
better with the YUV path. (Although the VDPAU API sure makes this
annoying with its separate surface types for RGB.)
Note that we create 5 surfaces for some reason - I don't think this
makes too much sense (because we can't use the deinterlacer with RGB
surfaces), but at least it reduces the amount of differences with
the YUV code path.
Clearing the borders is done by drawing a single black pixel over the
window. This sounds pretty dumb, but it appears to work well, and
there is no other API for that. (One could try to use the video mixer
for this purpose, since it has all kinds of features, including
compositing multiple RGBA surfaces and clearing the window background.
But it would require an invisible dummy video surface to make the
video mixer happy, and that's getting too messy.)
When panscan was used, i.e. the video is cropped to make the video fill
the screen if video and screen aspects don't match, screenshots
contained only the visible region of the source video, stretched to
original video size.
Consider:
mpv file1.mkv file2.mkv
and file1.mkv is restored from an earlier session when quit_watch_later
was used. Then all restored options were reset when file2.mkv is played,
even if the user changed them during playback. This affects for example
the fullscreen setting.
Make it so that after finishing a resumed file, the previously restored
settings are not reset again. (Which means only resuming will forcefully
overwrite the settings.)
If close to chapter start, skipping back goes to previous chapter (no change).
If more than <threshold> seconds in, skipping back will now go to the beginning
of the current chapter instead.
The threshold is set by the new option --chapter-seek-threshold and defaults to
5 seconds. A negative value disables the new functionality.
The VDPAU default colorkey, although it seems to be driver specific, is
usually green. This is a pretty annoying color, and you usually see it
briefly (as flashes) if the VDPAU window resizes.
Change it to some shade of black. The new default color is close to what
MPlayer picks as colorkey (and apparently it worked well for them):
VdpColor vdp_bg = {0.01, 0.02, 0.03, 0};
Since our OPT_COLOR can set 8 bit colors only, we use '#020507' instead,
which should be the same assuming 8 bit colors.
Obviously, you can't use black, because black is a way too common color,
and would make it too easy to observe the colorkey effect when e.g.
moving a terminal with black background over the video window.
Formally, this sets the "background color" of the presentation queue.
But in practice, this color is also used as colorkey.
This commit doesn't change the VDPAU default yet.
Not actually useful. This would break whenever a new text subtitle
format would be added, which requires a binary->text transformation.
(mov_text is one such format; disable it.) In general, we would have
to know which packet formats are binary, which we don't, so the only
reasonable way to handle this is a white list.
One example in ao.rst used the old syntax with mangled device names.
Fix it. Mention some shell related caveats. Explicitly mention the
change of device name syntax in changes.rst, because it seems to be
a common issue.
Normally, --subcp always forces conversion. This really always forces
conversion, even if the UTF-8 check on the input succeeds.
Extend the --subcp to allow codepages as fallback if UTF-8 doesn't
work. So, for example --subcp=utf8:cp1250 will use UTF-8 if the input
looks like UTF-8, and will fall back to use cp1250 if the UTF-8 check
fails.
I think this should actually be the default, but on the other hand,
this changes the semantics of the option, and a user would actually
expect --subcp to force conversion, rather than silently using UTF-8
if that happens to work.
Don't accept overlong sequences. Don't accept codepoints past the
maximum unicode codepoint. Don't accept the UTF-16 surrogate codepoints.
I'm not sure if there are more codepoints that are defined to be
invalid, but we just want to make libavcodec happy, so this is enough.
(libavcodec's subtitle converter checks for valid UTF-8 and throws up
and dies if it's not - now we want to use bstr_sanitize_utf8_latin1() to
force valid UTF-8, so the strictness of our UTF-8 parser has to match at
least that of the libavcodec's check.)
I'm not sure whether the min test is actually 100% correct.
Note that libavcodec also treats BOM codepoints as invalid. This is
definitely a bug: the BOM is really just "zero-width non-breaking space"
redefined by Microsoft, but it is perfectly valid to appear in the
middle of a string. Official Unicode has merely deprecated the old
usage of the BOM codepoint, and didn't make it illegal. Besides, the
string could be from the start of a file, so even this check doesn't
make sense even with libavcodec's insane logic. We don't copy this bug.
Broken UTF-8 in this context means we treat it as UTF-8, but we also
interpret broken UTF-8 sequences as Latin1.
Also, run our own UTF-8 check function before the charset detectors.
This prevents from ENCA's UTF-8 check possibly messing up (like
detecting 7-bit clean UTF-8 as ASCII, or other things). It also takes
care of UTF-8 detection if no charset detector (ENCA, libguess) is
compiled in, and it lets us deal better with cut-off UTF-8 sequences.
Libav's <libavcodec/vdpau.h> header uses some libavocdec symbols without
forward-declaring them and without including the headers declaring them.
FFmpeg's header for this is fine.
About this issue, it would be better if the surfaces could be allocated
with the real size, and the vdpau video mixer could be created with that
size as well. That would be a bit hard, because the real surface size
had to be communicated to vdpau. So I'm going with this solution. vaapi
seems to be fine with either surface size, so there's hopefully no
problem.
Instead of passing AVFrame. This also moves the mysterious logic about
the size of the allocated image to common code, instead of duplicating
it everywhere.
The fix_overlaps_and_gaps() function in dec_sub.c fixes small gaps or
overlaps between subtitle events. However, sometimes it could happen
that the corrected subtitle events could overlap by 1ms due to bad
rounding, making libass shift subtitles to reduce collisions. (The
second subtitle will be shown above the previous one, even if both
subtitles are visible only for 1ms.)
sd_ass.c rounds the timestamps when converting to integers for unknown
reasons. I think it would work fine without that rounding, but since
I have no clue why it rounds, and since it could be needed to ensure
correct timestamps with ASS subtitles demuxed from Matroska, I'd rather
not touch it. So the solution is to use already rounded timestamps to
calculate the new subtitle duration in fix_overlaps_and_gaps().
See github issue #182.
This adds precise scrolling support. I ran some tests and it seems a little
bit smoother and well.. precise. The defaults are rebindable using: AXIS_UP,
AXIS_DOWN, AXIS_LEFT and AXIS_RIGHT.
If pthreads are enabled the input queue accesses are regulated by acquiring
a mutex. This is useful for platforms like OS X, where the events are created
in the cocoa thread and added to the queue to then be dequeued in the playloop
thread.
This reverts commit 689a25003f, with some
adjustments to code that was added after that commit.
I just messed up big time. We don't need this, and in fact the commit
confused straight and premultiplied alpha at one point (just a simple
inverted condition due to an oversight), which is why it looked like
it was working.
In commit 2827295 I wrote:
Also, libva can't decide whether it accepts straight or premultiplied
alpha for OSD sub-pictures [...]
That was just me messing up and being severely confused by my own bugs.
VA API uses premultiplied alpha, which by the way is nice and
thoughtful of the VA API devs.
Well, this was stupid. But in the end, I'm glad that I could actually
reduce codesize by a good amount again.
This is based on the MPlayer VA API patches. To be exact it's based on
a very stripped down version of commit f1ad459a263f8537f6c from
git://gitorious.org/vaapi/mplayer.git.
This doesn't contain useless things like benchmarking hacks and the
demo code for GLX interop. Also, unlike in the original patch, decoding
and video output are split into separate source files (the separation
between decoding and display also makes pixel format hacks unnecessary).
On the other hand, some features not present in the original patch were
added, like screenshot support.
VA API is rather bad for actual video output. Dealing with older libva
versions or the completely broken vdpau backend doesn't help. OSD is
low quality and should be rather slow. In some cases, only either OSD
or subtitles can be shown at the same time (because OSD is drawn first,
OSD is prefered).
Also, libva can't decide whether it accepts straight or premultiplied
alpha for OSD sub-pictures: the vdpau backend seems to assume
premultiplied, while a native vaapi driver uses straight. So I picked
straight alpha. It doesn't matter much, because the blending code for
straight alpha I added to img_convert.c is probably buggy, and ASS
subtitles might be blended incorrectly.
Really good video output with VA API would probably use OpenGL and the
GL interop features, but at this point you might just use vo_opengl.
(Patches for making HW decoding with vo_opengl have a chance of being
accepted.)
Despite these issues, decoding seems to work ok. I still got tearing
on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also
tested with the vdpau vaapi wrapper on a nvidia system; however this
was rather broken. (Fortunately, there is no reason to use mpv's VAAPI
support over native VDPAU.)
This is for VAAPI support. VAAPI does not support premultiplied alpha
for OSD. (Normally, we prefer premultiplied, because it has better
behavior on scaling.)
I'm not sure whether blending in the ASS->RGBA part is correct and I
didn't test it extensively.
Change how the HW decoding stuff is organized, the way it's initialized
in particular. Instead of duplicating the list of supported codecs for
hwaccel decoders, add a probe function which allows each decoder to
report whether it supports a given codec.
Add an "auto" choice to the --hwdec option, which automatically enables
hardware decoding if libavcodec and/or the VO supports it.
What mpv prints on the terminal changes a bit. Now it will just print
a single line whether hw decoding is used or not (and nothing at all if
no hw decoding at all was requested). The pretty violent fallback from
hw decoding to software decoding is still quite verbose and evil-looking
though.
2013-08-11 23:59:18 +02:00
496 changed files with 44150 additions and 32678 deletions
wm4