Fixes Ticket1474

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fix Ticket #1617, revealing a regression I introduced in 8f3eebd.

We need to make sure no stream is added in between Ogg context save and
restore operations (because it would likely lead to a mismatch between
ogg->nstreams and AVFormatContext->nb_streams after the restore op).
This is the reason the ogg->state check is added in ogg_new_stream().
Before this patch, checking for ogg->headers was preventing this:
ogg->headers is always set before any ogg save/restore (though, it was
also preventing from creating the stream when necessary).

Reviewed-by: Tomas Härdin <tomas.hardin@codemill.se>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes bug 372.

Fixes out of array read, should fix ffprove fate failure

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Prior to this it could theoretically become negative and read
out of array.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

fixes regression from the xing toc support

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This occurs with fuzzed mpeg-ts files. set_codec_from_probe_data() is
called with a zeroed AVProbeData since no packet made through for
specific stream.

Approved-by: Tjoppen
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reviewed-by: Tomas Härdin <tomas.hardin@codemill.se>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Frames described by this grouping are the starter of a closed or
an open GOP.
This is useful for open GOP of H.264 stream which is not described
by sync sample atom.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Support Matroska native formatting.

On demuxing prepend a Frame container atom (32bit big endian encoded
frame size and 'icpf' string).
On muxing remove it.

Stricter validation, explicitly exit on misparsing and some error
forwarding from the ebml parsing functions used.

Make MATROSKA_TRACK_ENCODING_COMP_HEADERSTRIP behave like
the other encodings and spare few lines of boilerplate code.

Support Matroska native formatting.

On demuxing reconstruct the 36-bytes QuickTime atom that the ALAC
decoder expects by prepending the "atom size", "tag" and
"tag version" fields missing from the Matroska's CodecPrivate
element.

On muxing remove the initial 12 bytes

Sample files are available:
http://www.bunkus.org/videotools/mkvtoolnix/samples/alac/alac-in-matroska.mka
and the CoreAudio file it was created from with today's mkvmerge:
http://www.bunkus.org/videotools/mkvtoolnix/samples/alac/alac-in-matroska-source.caf

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Make all the compression encodings behave the same way.

Fixes Ticket1742

Based on patch by: Nikola Vranic <nikola.vranic@rt-rk.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This fixes stream copy from a format that already has incompatible
codec tags set. The chained ismv muxer exports this same codec tag
list, so set it on this one as well, to allow the caller (and
lavf common code) to set them correctly.

Signed-off-by: Martin Storsjö <martin@martin.st>

A quite widespread pattern in the demuxer is read a 32bit unsigned
integer and then loop till this value is reached.

Checking for EOF prevents pathological situations.

The compiler fails to figure out that enc->codec_type can only
have 3 different values.
Thus when an if/else is encountered it triggers on the possibility
of the else case has not initialized the flags variable.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

they keep showing up if i grep for non av assert()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This also fixes a potential crash since s->pb can be NULL.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

It would have been done anyway in the av_dict_set() call.

This simplifies the code and avoid a warning because of assigning a
const string from ff_id3v1_genre_str to a non-const variable.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

This also avoids a memleak.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Also re-use the allocated buffer instead of re-allocating a new one.

Fix near infinite loop in stsd parsing.
Bug found by: Diana Elena Muscalu

The size is unsigned according the specification.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

nstreams is assumed to be 1 at that point, so the loop is pointless.