Seeking in MPEG files with pts resets could fail completely, as it was
always done by timestamps and those of course don't unambiguously
specify a file position in such files. Add basic functionality for
byte-based seeking and playback position reporting, and decide whether
to use that functionality based on a simple heuristic (could be
improved).
When -alang / -slang was specified the numerically first matching
track (if any) was always chosen. This meant that specifying "-alang
eng" could change the track choice even if all tracks were in English,
because now the default flag of tracks was ignored. Change the logic
to take the default flag into account as a secondary sorting key.
The code also accepted prefix matches, so that "-slang g" would match
track language "ger". I think that was not intentional. Change it to
require exact matches.
The Cue entries in typical Matroska files have information for the
video track only. This caused seeks to fail when playing with
-novideo, as demux_mkv tried to use audio track index entries then.
Add a fallback case that uses any index entries without caring what
track they're for if there are no entries specific to the track we're
interested in.
Relative seeks didn't add the current position as they should. Fix.
Note that this had no effect in normal playback case even if the file
had no index, because the "accurate_seek" logic at higher level would
convert all commands to absolute seeks before calling demuxer level.
Various code referred to "mpctx->demuxer" where it should really have
referred to the one used for audio/subtitles in case those differ. Fix
by using "mpctx->d_audio->demuxer" etc instead. Disable the copying of
streams in demux_demuxers; that was a partial workaround for things
referring to the main demuxer (and it wasn't enough anyway).
This fixes, among other things, switching audio tracks within the file
specified by -audiofile.
demux_demuxers doesn't run the normal demuxer.c initialization for new
demuxers. Initialize stream_pts separately (it won't ever be changed
with the current implementation). This at least avoids other code
assuming it was set properly.
Move functions to query current playback position, percentage position
and total video length from from the demuxer layer to top level. The
functions need access to playback state that doesn't belong on the
demuxing level. Make the new functions more capable and simplify some
code that can now rely on them. This fixes some errors in displayed in
OSD and slave mode information when using timeline (ordered chapters).
Use IMGFMT to compare instead of PIX_FMT to avoid issues with the
"JPEG" formats like PIX_FMT_YUVJ422P.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32594 b3059339-0415-0410-9bf9-f77b7e298cf2
For FFmpeg codecs YV12 should always be in the supported format
lists if I420 is.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32593 b3059339-0415-0410-9bf9-f77b7e298cf2
Fix dvd:// fill_buffer function to actually write into the specified
buffer and check that the buffer is sufficiently large.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32591 b3059339-0415-0410-9bf9-f77b7e298cf2
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32592 b3059339-0415-0410-9bf9-f77b7e298cf2
Ensure we queue all subtitle packets before demuxing the next video
packet.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32587 b3059339-0415-0410-9bf9-f77b7e298cf2
fflatm seems to be working, whereas external faad does not work at all
with latm (internal does).
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32585 b3059339-0415-0410-9bf9-f77b7e298cf2
Make scale filter accept all non-hardware-acceleration input formats
we know a PIX_FMT_* for.
It is possible we will not have a conversion for some of these formats,
but since it will just fail differently this approach is better since
it avoids having to expand the explicit list continuously.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32576 b3059339-0415-0410-9bf9-f77b7e298cf2
Also move libfaad codecs.conf entry into the same position as in svn
(no functionality change since ffaac is marked buggy so only faad is
used by default regardless or order).
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32571 b3059339-0415-0410-9bf9-f77b7e298cf2
kvaSetup() resets all the attributes, so there is need to restore them.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32570 b3059339-0415-0410-9bf9-f77b7e298cf2
Add a stream_read_internal() function that reads directly into a given
buffer instead of the stream's internal one. Use this to read directly
into cache memory, avoiding a memcpy(). This requires also adding a
stream_seek_internal() as the normal seek function reads into the
stream's buffer.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@32559 b3059339-0415-0410-9bf9-f77b7e298cf2
When playing avi files take timestamps from the .dts field of the
packet instead of .pts. For avi libavformat returns the original avi
timestamps in .dts; without GENPTS set .pts is sometimes unset, with
GENPTS set it'll contain some made up values which are not correct
pts. Current libavformat also has a bug take makes it loop over the
whole avi file if you use GENPTS with video streams disabled.
The timing code can cope with the avi timestamps even though they're
not really pts.
Some internet radio streams are ogg/vorbis with a new stream for each
track (and the previous stream stops receiving data). Add a hack to
ignore the stream number in packets when input seems to match this
case; all packets will be effectively regarded as parts of the same
audio stream. This workaround can't work perfectly with current lavf
because it will keep adding streams and then hit its 20 stream
maximum; at that point playback will likely fail. Print a warning
about this problem and mention "-demuxer ogg" as a possible
workaround.
Currently there's no check that the vorbis extradata is actually
compatible between the streams.
AVInputFormat name fields consist of comma-separated components, more
of which may be added in backwards-compatible versions. demux_lavf did
a strcmp against the whole field at once; this would break if a
component was added. Change the code to search for individual
component matches within the field instead.
msgfmt creates the output file even if there was a fatal error. Thus
make would stop because of the error return, but the output file would
be left on disk and running make again would continue building other
files. Enable the .DELETE_ON_ERROR special target to automatically
delete target files changed by failing commands. This affects all
rules; I think it is the correct behavior for all existing rules.
