Not only is this easier to read, this also makes the code shorter.

It currently emulates the long-removed
avcodec_decode_audio4/avcodec_decode_video2 APIs, which obfuscates the
actual decoding flow. Restructure the decoding calls so that they
naturally follow the new avcodec_send_packet()/avcodec_receive_frame()
design.

This is not only significantly easier to read, but also shorter.

It is currently handled in the same loop as audio and video, but this
obscures the actual flow, because only one iteration is ever performed
for subtitles.

Also, avoid a pointless packet reference.

process_input_packet() contains two non-interacting pieces of nontrivial
size and complexity - decoding and streamcopy. Separating them makes the
code easier to read.

New placement requires fewer explicit conditions and is easier to
understand.
The logic should be exactly equivalent, since this is the only place
where eof_reached is set for decoding.

Replace the decode_error_stat global with a per-input-stream variable.
Also, print an error message when the error rate is exceeded.

Passing ist=NULL is currently used to identify stream types that do not
decode into AVFrames, i.e. subtitles. That is highly non-obvious -
always pass a non-NULL InputStream and just check the type explicitly.

This cannot be reached, because initialization will fail if decoding is
requested for a stream but no decoder can be found.

It tracks whether the decoder for this stream ever produced any frames
and its only use is for checking whether a filter input ever received a
frame - those that did not are prioritized by the scheduler.

This is awkward and unnecessarily complicated - checking whether the
filtergraph input format is valid works just as well and does not
require maintaining an extra variable.

This special handling for decoder flushing has not been needed since
af1761f7, as the filtergraph actually is drained after that commit.

It has a small fixed size, so it is better to embed it rather than deal
with dynamic allocation.

Avoid extracting it from various remote sources.

It is not accessed outside of ffmpeg_filter.

It is not accessed outside of ffmpeg_filter.

In case no decoder is available, dec_open() called from ist_use() will
fail with 'Decoding requested, but no decoder found', so this check is
redundant.

They are not used outside of ffmpeg_filter.

That is a more appropriate place for that function.

When a filtergraph input receives EOF but never saw any input frames, we
use the fallback parameters. Currently an attempt to actually configure
the filtergraph will happen elsewhere, but there is no reason to
postpone this.

With complex filtergraphs it can happen that the filtergraph is
unconfigured because some other filter than the one we just got EOF on
is missing parameters.

Make sure that the fallback parametes for a given input are only used
when that input is unconfigured.

This is less ad-hoc than checking explicitly in every place that binds
an input stream to a filter or output.

Will be useful in future commits.

Avoids filtering code from digging in demuxer internals.

Do not use a separate counter for CFR timestamps forced with -r used as
an input option. Set durations properly and let estimation code do the
rest.

Otherwise the two values might get desynchronized and lavfi can prefer
the wrong one.

Decoder timebase does not always have to match filter timebase.

They are the same for now, but this may change in the future.

Store decoded frame timestamp directly in AVFrame.pts, there is no
advantage to using a separate local variable for it.

There is only one frame used in decode_video() -- the one output by the
decoder. So there is no point in explicitly calling it the _decoded_
frame.

No point to consume 1st input frames any more after crossfade is over.

No reason to use this prefix here.

Upstream peak detection lost one option and gained one option. Update
code and documentation as required.

This algorithm has once again been refactored, this time leading to a
dropping of the old `tone_mapping_mode` field, to be replaced by a
single tunable hybrid mode with configurable strength.

We can approximately map the old modes onto the new API for backwards
compatibility. Replace deprecated enums by their integer equivalents to
safely preserve this API until the next bump.

Upstream deprecated the old ad-hoc, enum/intent-based gamut mapping API
and added a new API based on colorimetrically accurate gamut mapping
functions.

The relevant change for us is the addition of several new modes, as well
as deprecation of the old options. Update the documentation accordingly.

Explicitly forbid using a non-RGB colorspace with RGB pixel format or
vice versa. This mirrors identical logic from vf_scale.