This allows avoiding negative timestamps.

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

Possibly fixes Ticket1724

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

http://a/b + //c/d = http://c/d
http://a/b?c + ?d = http://a/b?d
http://a/b?c/d + /e = http://a/e

Updated with feedback from Clément Bœsch

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

Fixes crash
Fixes ticket1730

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

The loop was introduced in 64d340c62ad5954c1a834df2d26057135e771774, and
was likely breaking non blocking reads as it busy loops.

This allows using avcodec_free_frame() to free it properly.

In particular, fix crash when the input file contains no packets (e.g. an
ffmeta input).

The channel layout at the demuxer layer is commonly wrong and would
otherwise be used.

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

It could open the wrong decoder.

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.

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

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

The condition should not be &&.

This is consistent with stdio and is what we want to do in all cases.

Fixes a bug in the voc muxer which didn't flush in write_trailer()
previously. This is the cause of the change in the test results.

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

Do not crash if the input format is not allocated yet.

avio_close checks by itself for NULL condition.

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>

Replace "unspecified sample size" with "unspecified frame size". +10l.

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

Make internal small_strptime() function public, and use it in place of
strptime().
This allows to avoid a dependency on strptime() on systems which do not
support it.

In particular, fix trac ticket #992.

read_seek was undeprecated in e5c708151.

This improves dts validity checks and consequently fps detection of files with invalid dts
Fixes Ticket1681

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

Based on code by: Andrey Utkin <andrey.krieger.utkin@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This way it gets executed earlier and updated timestamps become available earlier.

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

Fix "assignment discards ‘const’ qualifier from pointer target type" warnings.

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

This may fix Ticket730

Reviewed-by: Nicolas George <nicolas.george@normalesup.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Mans Rullgard <mans@mansr.com>

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

With minor changes by michael
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

The resolution is in the packets, so decoding must happen.
Since most other formats do not set the dimension, make it
a special case for PGS. If other codecs were to have the
same requirement, using a CODEC_CAP would be preferred.

This also changes behavior as the descriptor table is more complete than
the switch/case it replaces. As well as considering all non video as
intra only

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

If the dts difference is more than 1000 times the average dts
difference, restart the analysis.

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

These are normally initialized to AV_NOPTS_VALUE at the start
of avformat_find_stream_info, but if a new stream is found while
this function is running (e.g. like in mpegts), the newly added
AVStreams didn't have these values properly initalized, leading
to avformat_find_stream_info terminating too soon (when the
first timestamps are far from 0).

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

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

This adds a function to retrieve the number of entries in a
dictionary and updates the places directly accessing what should
be an opaque struct to use this new function instead.

Signed-off-by: Mans Rullgard <mans@mansr.com>

This is limited to the chars that arent filtered by av_log() already
we might filter more aggressively if theres some case where this becomes
needed.

Fixes Ticket1181

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