The correct invariant is (cur != NULL -> next != NULL). Not the other way around.

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

Based on current version of the asplit filter in FFmpeg written by
Stefano Sabatini and others.

This is in preparation for adding an audio split filter.

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

This code is not only much more powerfull its also simpler

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

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

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

The previous version checked the filter context name,
instead of checking the filter name.
The new version just uses the same type as the input.

Signed-off-by: Nicolas George <nicolas.george@normalesup.org>

The list goes back to being simple integers,
and avfilter_make_format64_list is fixed to work
with the correct structure directly.

The file now contains both an audio and a video split filter.

The move allows to share the init code already used by split.

Most of the code is moved to buffersrc.c
to help Git see the similarities.

src_buffer.c still contains the obsolete compatibility functions.

The format change detection code was removed as it has been
implemented directly in ffmpeg.
It can easily be brought back from the history.

It restores ABI compatibility with libav.

The fields that move were very recent or morally private.

This allow specifying hexadecimal channel layout, and also handle the
error printing internally.

all_channel_layouts.h is not a real header file and breaks make checkheaders.

The difference from this should be pretty small.

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

This fixes a regression that apparently was missed when switching to the
in af resampler

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

Partially based on a patch by Robert Nagy <ronag89@gmail.com>

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

Allow to explicitly specify the channel layout to select.

Inspired-by code from af_resample.c written by Anton Khirnov

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

Issue found by comparing af_resample.c and af_aresample.c

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>

Fixes a null ptr deref

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

Some of the FATE changes are due to off-by-one different rounding being used
(lrintf vs av_rescale_q).
Some fate changes are due to 1 audio frame less being encoded (the new variant seems
matching what qatar does and according to ffprobe its closer to the requested duration)
the mapchan feature sadly is lost in this commit because it depends on resampling
being done in ffmpeg.c which is now moved completely into the av filter layer
-async is broken after this commit, this will be fixed in subsequent commits
the new filter reconfiguration system is flawed and will drop a frame on each
parameter change which is why the nelly moser checksums need updating.

Conflicts:

	ffmpeg.c
	tests/ref/fate/smjpeg

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>

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

The source supports *planar* double, fix regression introduced by me in
4d4098da.

Fix crash.