Found by John Villamil <johnv@matasano.com>

MinGW doesn't have sleep, only _sleep (which is deprecated),
Sleep (which is defined in winbase.h and not in the standard
C headers) and usleep.
Signed-off-by: Martin Storsjö <martin@martin.st>

It can also optionally split the file into individual fragments,
which allows it to be served from any web server without any
server side support.
Signed-off-by: Martin Storsjö <martin@martin.st>

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

For ismv/isma, the time scale might not be the same as the sample
rate.
Signed-off-by: Martin Storsjö <martin@martin.st>

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

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

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

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

This fixes calculation of trackDuration if the MOVIentry array
is cleared. This is required by the fragmentation support in the
next patch.
Signed-off-by: Martin Storsjö <martin@martin.st>

Current code only writes 8 pixels of vertical edge for YUV422, which
causes MC artifacts when subsequent frames use data from that edge.

When the length is the escape value (15), the new length is calculated by
15 + get_bits(8) - 1.
Signed-off-by: Alex Converse <alex.converse@gmail.com>

There is only one stream, so it will always be index 0.
This also allows for removal of the private context.

also allows for removing some unused context fields

also allows for removing some unused context fields

Time stamps are not read from the bitstream, so no wrapping is needed.

Fixes decoding of the handful of "death scream" samples from
"Command and Conquer" that use SND1. e.g.
http://samples.libav.org/game-formats/westwood-aud/nuyell1.aud
http://samples.libav.org/game-formats/westwood-aud/nuyell7.aud

Stereo ADPCM IMA WS is planar for VQA version 3 and 2-sample interleaved for
VQA version 2.

The stereo layout and extradata is significantly different from that in
Westwood IMA ADPCM, so a separate codec_id is warranted.

Fix setting of bit_rate and bits_per_coded_sample.
Ask for sample for unsupported VQA versions.

According to Multimedia Wiki documentation, the frame rate is not always 15,
and the actual value can be found in the header.

Also remove unused context field, audio_frame_counter.

The format has no coded timestamps and must use packet durations to generate
them.

Although it has been deprecated for a long time, its intended
replacement (request_channel_layout) is not actually used anywhere, so
request_channels is currently the only way to access that functionality.

It has been deprecated only recently.

It has been deprecated only recently.

Error recognition flags can and should be set directly with AVOptions.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Found by John Villamil <johnv@matasano.com> in fuzzed rv20 in mkv files.

Signed-off-by: Alex Converse <alex.converse@gmail.com>

Signed-off-by: Alex Converse <alex.converse@gmail.com>

Signed-off-by: Alex Converse <alex.converse@gmail.com>