A reference is kept in the queue and freed after use.

These filters are designed for storing and transmitting video sequences
with alpha using higher-efficiency codecs such as x264 which don't
natively support an alpha channel. 'alphaextract' takes an input stream
with an alpha channel and returns a video containing just the alpha
component as a grayscale value; 'alphamerge' takes an RGB or YUV stream
and adds an alpha channel recovered from a second grayscale stream.

Signed-off-by: Steven Robertson <steven@strobe.cc>
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

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>

This reduces code duplication and prevents stale pointers from remaining
on the link.

This reduces code duplication and prevents stale pointers from remaining
on the link.

The buffer is passed on to the next filter, so we shouldn't keep any
pointers to it.

Avoid pointless filter ref/unref. Just use the reference passed to
start_frame().

Directly forwarding the input buffer to the next filter means that
drawtext no longer owns any references to it and thus shouldn't refer to
it in any way.

The buffer is passed on to the next filter, so we shouldn't keep any
pointers to it.

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

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

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

This one is tricky. Consider a graph with two sink links, A and B.
request_oldest() requests on A, and A returns EOF. This EOF causes
a filter to flush pending frames to B; they are unrequested.
Then request_oldest() moves on to B, and B returns EOF.

The doxy says, and the heap implementations assumes,
it is in AV_TIME_BASE units.

Reduce log clutter, consistent with 1a49a169eb.

Signed-off-by: Steven Robertson <steven@strobe.cc>
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

movie currently forwards EOF (or begins looping) immediately upon
reaching the last packet in the input stream, which can leave frames in
the decoder. This change first tries to read any remaining packets from
the decoder before forwarding EOF.

Signed-off-by: Steven Robertson <steven@strobe.cc>
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

Use the init_opaque callback for the purpose.

Fix regression introduced in a5e8c41c28. In particular, fix lavfi
device in case a conversion to the supported (packed) formats is needed.

This will be used by filters which require an opaque field to be passed
in input. Should be required only for filters which imply a programmatic
use.

This is possibly a temporary solution, to be removed when we'll have a
clearer and better agreememnt of how/if to pass binary data for
initializing a filter.

See thread:
Subject: [FFmpeg-devel][PATCH] lavfi: add init2 callback
Date: Fri,  6 Jul 2012 01:22:17 +0200

Make sure that an output frame has really been produced before returning
0.

Make sure that an output frame has really been produced before returning
0.

Also make this the default value.

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

Fix signature mismatch warnings.

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

Input on/off state can change in request_samples(), which can result in
a state where only the first input is active. get_available_samples()
will then return 0, and request_frame() will fail with EAGAIN even
though there is data on the single active input.

Take this into account and check the number of active inputs again after
calling request_samples().

0 should only be returned when there was at least one output frame.

The file contains the code for the audio buffersink as well.

This tool uses lavfi internal symbols not accessible in shared
libraries.  TESTPROGS are linked statically to allow them use of
library internals not normally exported.

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