[FFmpeg-devel] [PATCH] avcodec: Implement DM PAR Muxer/Demuxer

[Hendrik Leppkes]

On Fri, Jul 5, 2019 at 10:53 PM Tom Needham <06needhamt at gmail.com> wrote:
>
> Samples are Available from
>
> https://transfernow.net/131xk9g4u0jt
>
> Signed-off-by: Thomas Needham <06needhamt at gmail.com>
> ---
>  Changelog                     |    1 +
>  configure                     |    7 +
>  libavformat/Makefile          |   10 +
>  libavformat/adaudio.c         |  137 ++++
>  libavformat/adbinary.c        |  609 ++++++++++++++++
>  libavformat/adcommon.c        | 1106 +++++++++++++++++++++++++++++
>  libavformat/adffmpeg_errors.h |   94 +++
>  libavformat/adjfif.c          |  551 ++++++++++++++
>  libavformat/adjfif.h          |   41 ++
>  libavformat/admime.c          |  822 +++++++++++++++++++++
>  libavformat/adpic.h           |  116 +++
>  libavformat/adraw.c           |  131 ++++
>  libavformat/allformats.c      |    7 +
>  libavformat/ds.c              | 1262 +++++++++++++++++++++++++++++++++
>  libavformat/ds.h              |  135 ++++
>  libavformat/ds_exports.h      |  173 +++++
>  libavformat/dsenc.c           |  488 +++++++++++++
>  libavformat/dsenc.h           |   35 +
>  libavformat/dspic.c           |  317 +++++++++
>  libavformat/libpar.c          | 1030 +++++++++++++++++++++++++++
>  libavformat/libpar.h          |   40 ++
>  libavformat/netvu.c           |  214 ++++++
>  libavformat/netvu.h           |   21 +
>  libavformat/version.h         |    4 +-
>  24 files changed, 7349 insertions(+), 2 deletions(-)

Before anyone spends a huge amount of time reviewing 24(!) files and
over 7000(!) lines of code..

This is an implementation of a proprietary format based on a non-free
library (and for some reason despite using a library still needs over
7000 lines of code?)
We don't really accept non-free proprietary library wrappers like this
into ffmpeg, the only possible exceptions not applying here
(open-source with incompatible license, or hardware libraries, and the
second one is even being debated on still).

Outside of that, the code looks like it needs a serious overhaul, and
the submitter does not appear to be the author of most of it, judging
from the comments in the headers, which makes authorship problematic
at best.

- Hendrik