I had delayed reformatting mplayer.c as I wanted to split it, but
since I didn't come up with a good way to do that I'll clean up the
messy formatting now.
Initialize mpctx->last_chapter_seek to -2 instead of -1. This changes
get_current_chapter() return value to -2 for files which have no
chapters. -2 is used by some commands related to chapters to recognize
files without chapters and return failure without any effect in that
case.
Seeking while paused could result in the current audio pts being
reported incorrectly due to relevant variables not being reinitialized
after the seek until more audio was played. When playing audio-only
files, this meant that current overall playback position could be
reported incorrectly which in turn could break further seeks. Improve
things on two levels: First, store the seek target position and use
that as the current playback position for audio-only files until
things can be reinitialized. Second, try to reinitialize audio
decoding enough to know its current pts even while paused. Also avoid
printing the actual huge negative value of MP_NOPTS_VALUE on the
status line when pts could not be determined.
Change written_audio_pts() and playing_audio_pts() to return
MP_NOPTS_VALUE if no reasonable pts estimate is available. Before they
returned some incorrect value typically around zero (but not
necessarily exactly that).
Allow writing commandline options with two leading dashes. In this
mode a parameter for the option, if any, follows after a '=';
following separate commandline arguments are never consumed as a
parameter to a previous double-dash option.
Flag options may omit parameter and behave like old single-dash
syntax. "--fs=yes", "--fs=no" and "--fs" are all valid; the first two
behave like configuration file "fs=yes" and "fs=no", and last is the
same as old "-fs" (same effect as "--fs=yes").
Make per-file loop option start from --ss position, not always 0.
Do looping in more cases; before looping was only done when
encountering real end of file, now it also happens for example at
--endpos or --frames limits. Also move the --ss option to the option
struct.
The --ass-force-style option was only applied when the main libass
library handle was created. Thus any per-file option changes later had
no effect. Do the ass_set_style_overrides() call in per-file
initialization instead so that possible changes will be applied. Also
move the option variable to the option struct.
Current libass will crash (usually) if you set style overrides to a
nonempty value, then an empty one. It'll be easier to trigger this bug
after this commit, but the problem is not on mplayer2 side. The fix is
trivial so hopefully there will be a fixed libass soon.
Do the global initialization of libavcodec and libavformat
(avcodec_register_all(), av_register_all()) immediately on program
startup and remove the initialization calls from various individual
modules that use libavcodec/libavformat functionality.
Setting O_NONBLOCK on a file descriptor also affects all other fds
that share the same underlying open file description, and in case of
stdin such sharing is likely. Making stdin nonblocking can also make
stdout nonblocking (they may be the same connection to a terminal),
and it can also affect other processes (in "program1 | program2", the
shell may give the same terminal connection to program1 as stdin and
to program2 as stdout, thus program1 making its stdin nonblocking also
turns program2's stdout nonblocking).
To avoid these problems stop making fd 0 nonblocking. After the
previous commit this should no longer cause problems as long as
select() does not spuriously report the fd as readable.
Rework much of the logic related to reading from event sources and
queuing commands. The two biggest architecture changes are:
- The code buffering keycodes in mp_fifo.c is gone. Instead key input
is now immediately fed to input.c and interpreted as commands, and
then the commands are buffered instead.
- mp_input_get_cmd() now always tries to read every available event
from every event source and convert them to (buffered) commands.
Before it would only process new events until one new command became
available.
Some relevant behavior changes:
- Before commands could be lost when stream code called
mp_input_check_interrupt() which read commands (to see if they were
of types that triggered aborts during slow IO tasks) and then threw
them away. This was especially an issue if cache was enabled and slow
to read. Fixed - now it's possible to check whether there are queued
commands which will abort playback of the current file without
throwing other commands away.
- mp_input_check_interrupt() now prints a message if it returns
true. This is especially useful because the failures caused by
aborted stream reads can trigger error messages from other code that
was doing the read; the new message makes it more obvious what the
cause of the subsequent error messages is.
- It's now possible to again avoid making stdin non-blocking (which
caused some issues) without reintroducing extra latency. The change
will be done in a subsequent commit.
- Event sources that do not support select() should now have somewhat
lower latency in certain situations as they will be checked both
before and after select()/sleep in input reading; before the sleep
always happened first even if such sources already had queued
input. Before the key fifo was also handled in this manner (first
key triggered select, but if multiple were read then rest could be
delayed; however in most cases this didn't add latency in practice
as after central code started doing command handling it queried for
further commands with a max sleep time of 0).
- Key fifo limiting is more accurate now: it now counts actual
commands intead of keycodes, and all queued keys are read
immediately from input devices so they can be counted correctly.
- Since keypresses are now interpreted immediately, commands which
change keybindings will no longer affect following keypresses that
have already been read before the command is executed. This should
not be an issue in practice with current keybinding behavior.
After commit 39e373aa8d ("options: allocate dynamic options with
talloc") dynamically allocated options must be allocated with talloc.
Code implementing -use-filename-title still set opts->vo_wintitle to a
value from strdup(), triggering an abort when the option was freed.
Fix.
Make mp_dvdnav_save_smpi more robust and ensure consistency of nav
buffer.
It seems that in_size could be negative sometimes, this would cause
crashes if the malloc somehow succeeded.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@33599 b3059339-0415-0410-9bf9-f77b7e298cf2
This function was left over from older logic to manipulate the
"codec_path" global variable. Now that variable is fully handled by
the general option system, so that the only effect of the function
was to introduce memory leaks in some circumstances. Delete the
useless function.
Move the buffer storing audio data ready to be fed to the audio output
driver from the audio decoder object to the AO object. This will help
encoding code deal with end of input, and may also be useful to
improve other general gapless audio behavior (as AOs which do not
accept chunks smaller than a certain size may keep them in the buffer
while the decoder changes).
Less data may be dropped now when changing audio filters or switching
timeline parts.
Move the call to m_config_free() to be the last thing done before
exiting, otherwise mp_msg() might stop working if options it uses are
freed/reset.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@33380 b3059339-0415-0410-9bf9-f77b7e298cf2
Add some asserts to check that decoders/filters produce complete
samples (byte amounts must be multiples of channels*datatype_size) and
that audio output drivers also accept input in complete units. Fix
ad_pcm which was known to violate this if its last input packet didn't
stop at a sample boundary.
Change ao_pcm to use the new audio output driver API and clean up some
of the code. Rewrite the logic controlling how playback timing works
when using -ao pcm. Deprecate the "fast" suboption; its only effect
now is to print a warning, but it's still accepted so that specifying
it is not an error.
Before, timing with -ao pcm and video enabled had two possible
modes. In the default mode playback speed was rather arbitrary - not
realtime, but not particularly fast. -ao pcm:fast tried to play back
at maximum video playback speed - mostly succeeding, but not quite
guaranteed to work in all cases. Now the default is to play at
realtime speed. The -benchmark option can now be used to get faster
playback (same as the video-only case). In the audio-only case
playback is always maximum speed.
Neither fd 0 slave input (-slave) nor additional opened fds (-input
file=X) were set to nonblocking mode as they should have been. Fix.
Also rename the horribly generic USE_SELECT #define used for a
specific slave input detail.
Commit cbeed30ae8 ("core: wake up a bit less often for audio-only
files") increased the sleep time between audio buffer fills. This
turned out to cause problems on some machines where available audio
buffer sizes are extremely limited (example cases included 85 ms for
stereo and less for multichannel audio). Change the code to check
the amount of buffered audio and shorten sleep times accordingly if
needed.
Such short buffers violate some assumptions made by video timing code,
so they may still cause visible problems in some cases. At least on
some machines using ALSA the problem seems to be caused by bad
configuration defaults (small buffer memory limit which can be
increased).
Analogously to the previous commit, move path handling logic for
loading external vobsub files from mplayer.c to find_subfiles.c.
Based on a commit from Clément Bœsch but fixed and simplified.
Move path handling for loading external subtitle files from mplayer.c
to find_subfiles.c. Now the remaining code in mplayer.c only gets a
list of potential filenames and tries opening those.
Move sub_filenames() and related code from subreader.c to new file
find_subfiles.c. This function is used to find subtitle files that
should be loaded for the current video; this functionality is not
specific to the particular kind of text subtitle handling implemented
in subreader.c.
Also reindent and prettify the moved code a bit.
There is no reason to use manual language list splitting when an
automatic split function is already available.
Some types change from "unsigned char" to "char", but this shouldn't
cause issues since [as]lang settings are unlikely to have characters
above 127.
Add option -osd-fractions which enables display of fractional seconds
when showing the current playback time on OSD.
Based on a patch from Christian <herr.mitterlehner@gsmpaaiml.com> but
with several modifications.
Make the outside interface of audio output handling similar to the
video output one. An AO object is first created, and then methods
called with ao_[methodname](ao, args...). However internally libao2/
still holds all data in globals, and trying to create multiple
simultaneous AO instances won't work.
* edl:
core: support timeline with audio-only files
core: wake up a bit less often for audio-only files
core: audio: cut audio writes at end of timeline part
EDL: add support for new EDL file format
stream.[ch], ass_mp: new stream function for whole-file reads
tl_matroska.c: move the find_files() function here
bstr.[ch], path.[ch]: add string and path handling functions
core: ordered chapters: move timeline creation to timeline/
options: drop support for numeric -demuxer values
cleanup: demuxer.[ch]: remove unused code, make functions static
cleanup: reindent demuxer.h, use struct names for types
Cut audio data written to AO at the point where current timeline part
ends (before, AO buffers were always completely filled, but playback
of the "extra" audio was then cut short by resetting the AO when
switching timeline parts). This doesn't make much difference for
current playback behavior, but will be used by timeline support for
audio-only files and is necessary for future encoding support where
"playback" of written audio cannot be aborted later.
The timeline code previously added to support Matroska ordered
chapters allows constructing a playback timeline from segments picked
from multiple source files. Add support for a new EDL format to make
this machinery available for use with file formats other than Matroska
and in a manner easier to use than creating files with ordered
chapters.
Unlike the old -edl option which specifies an additional file with
edits to apply to the video file given as the main argument, the new
EDL format is used by giving only the EDL file as the file to play;
that file then contains the filename(s) to use as source files where
actual video segments come from. Filename paths in the EDL file are
ignored. Currently the source files are only searched for in the
directory of the EDL file; support for a search path option will
likely be added in the future.
Format of the EDL files
The first line in the file must be "mplayer EDL file, version 2".
The rest of the lines belong to one of these classes:
1) lines specifying source files
2) empty lines
3) lines specifying timeline segments.
Lines beginning with '<' specify source files. These lines first
contain an identifier used to refer to the source file later, then the
filename separated by whitespace. The identifier must start with a
letter. Filenames that start or end with whitespace or contain
newlines are not supported.
On other lines '#' characters delimit comments. Lines that contain
only whitespace after comments have been removed are ignored.
Timeline segments must appear in the file in chronological order. Each
segment has the following information associated with it:
- duration
- output start time
- output end time (= output start time + duration)
- source id (specifies the file the content of the segment comes from)
- source start time (timestamp in the source file)
- source end time (= source start time + duration)
The output timestamps must form a continuous timeline from 0 to the
end of the last segment, such that each new segment starts from the
time the previous one ends at. Source files and times may change
arbitrarily between segments.
The general format for lines specifying timeline segments is
[output time info] source_id [source time info]
source_id must be an identifier defined on a '<' line. Both the time
info parts consists of zero or more of the following elements:
1) timestamp
2) -timestamp
3) +duration
4) *
5) -*
, where "timestamp" and "duration" are decimal numbers (computations
are done with nanosecond precision). Whitespace around "+" and "-" is
optional. 1) and 2) specify start and end time of the segment on
output or source side. 3) specifies duration; the semantics are the
same whether this appears on output or source side. 4) and 5) are
ignored on the output side (they're always implicitly assumed). On the
source side 4) specifies that the segment starts where the previous
segment _using this source_ ended; if there was no previous segment
time 0 is used. 5) specifies that the segment ends where the next
segment using this source starts.
Redundant information may be omitted. It will be filled in using the
following rules:
- output start for first segment is 0
- two of [output start, output end, duration] imply third
- two of [source start, source end, duration] imply third
- output start = output end of previous segment
- output end = output start of next segment
- if "*", source start = source end of earlier segment
- if "-*", source end = source start of a later segment
As a special rule, a last zero-duration segment without a source
specification may appear. This will produce no corresponding segment
in the resulting timeline, but can be used as syntax to specify the
end time of the timeline (with effect equal to adding -time on the
previous line).
Examples:
----- begin -----
mplayer EDL file, version 2
< id1 filename
0 id1 123
100 id1 456
200 id1 789
300
----- end -----
All segments come from the source file "filename". First segment
(output time 0-100) comes from time 123-223, second 456-556, third
789-889.
----- begin -----
mplayer EDL file, version 2
< f filename
f 60-120
f 600-660
f 30- 90
----- end -----
Play first seconds 60-120 from the file, then 600-660, then 30-90.
----- begin -----
mplayer EDL file, version 2
< id1 filename1
< id2 filename2
+10 id1 *
+10 id2 *
+10 id1 *
+10 id2 *
+10 id1 *
+10 id2 *
----- end -----
This plays time 0-10 from filename1, then 0-10 from filename1, then
10-20 from filename1, then 10-20 from filename2, then 20-30 from
filename1, then 20-30 from filename2.
----- begin -----
mplayer EDL file, version 2
< t1 filename1
< t2 filename2
t1 * +2 # segment 1
+2 t2 100 # segment 2
t1 * # segment 3
t2 *-* # segment 4
t1 3 -* # segment 5
+0.111111 t2 102.5 # segment 6
7.37 t1 5 +1 # segment 7
----- end -----
This rather pathological example illustrates the rules for filling in
implied data. All the values can be determined by recursively applying
the rules given above, and the full end result is this:
+2 0-2 t1 0-2 # segment 1
+2 2-4 t2 100-102 # segment 2
+0.758889 4-4.758889 t1 2-2.758889 # segment 3
+0.5 4.4758889-5.258889 t2 102-102.5 # segment 4
+2 5.258889-7.258889 t1 3-5 # segment 5
+0.111111 7.258889-7.37 t2 102.5-102.611111 # segment 6
+1 7.37-8.37 t1 5-6 # segment 7
The select_audio() call was done on the main demuxer, not -audiofile
one (the "if (mpctx->num_sources)" test in the previous code was
always true). Call it on the -audiofile demuxer instead. The
-audiofile stuff still needs a proper cleanup later though.
Windows pthreads requires certain functions to be called to initialize
itself. It can do that through DllMain but no such luck when linked
statically; mplayer needs to call the initialization explicitly.
When doing a precise seek video_out->frame_loaded was left to true
while frames were being skipped. However vo_get_buffered_frame()
always returns success if a frame is already loaded; due to this the
EOF detection in update_video() never triggered, and a hr-seek past
EOF could cause a soft hang (commands were still processed and it was
possible to seek again to exit the loop). This could also happen with
Matroska files using ordered chapters if an underlying file was
actually shorter than the chapter that was supposed to come from it.
Then seeking to a timestamp after the end of the file but before the
end of the chapter would trigger the bug.
Fix the problem by setting frame_loaded to false when we decide to
skip the frame in question.
Add new file timeline/tl_matroska.c. Move the code that parses
ordered chapter information from Matroska files and creates the
timeline structure based on that to the new file.
Initialize the format parameter given to open_stream() in the moved
code. The previous uninitialized value shouldn't have caused any
visible effects.
Remove some unused lines from demuxer.h. Make some demuxer.c functions
static. Move new_ds_stream() declaration from demuxer.h to stream.h
(the function is defined in stream.c). Clean up some code in mplayer.c
that had commented-out free_demuxer_stream() calls.
