[FFmpeg-devel] [PATCH] lavr: deprecate the entire library

[Rostislav Pehlivanov]

On 25 December 2017 at 06:50, wm4 <nfxjfg at googlemail.com> wrote:

> On Mon, 25 Dec 2017 02:12:38 +0000
> Rostislav Pehlivanov <atomnuker at gmail.com> wrote:
>
> > Signed-off-by: Rostislav Pehlivanov <atomnuker at gmail.com>
> > ---
> >  doc/APIchanges             |  5 +++++
> >  libavresample/avresample.h | 30 +++++++++++++++++++++++++-----
> >  2 files changed, 30 insertions(+), 5 deletions(-)
> >
> > diff --git a/doc/APIchanges b/doc/APIchanges
> > index df79758e86..96fad416d5 100644
> > --- a/doc/APIchanges
> > +++ b/doc/APIchanges
> > @@ -15,6 +15,11 @@ libavutil:     2017-10-21
> >
> >  API changes, most recent first:
> >
> > +2017-xx-xx - xxxxxxx - lavr 4.0.0 - avresample.h
> > +  Deprecate the entire library. It was unmaintained and redundant
> > +  as libswresample did everything it did better, faster, with more
> > +  control and with a better, slightly higher level API.
> > +
>
> I don't strongly disagree with the wording and also with the
> "deprecation", and I think we should build avr by default instead.
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>

Suggest better wording then.
I strongly disagree with your suggestion to make avr the default. That
would undermine all the work that's being and was done on making a better,
faster resampling library just because someone wants to keep compatibility
with a dead project.
For someone who's complaining about the codebase being full of junk you
sure seem lenient towards having 2 resampling libraries and 2 filters to do
resampling, Not to mention its a big security attach surface that's just
waiting to be found. No one wants to work on it, no one has even worked on
it in any capacity in the past 2 or 3 years. Even the original author's
disappeared. And as we've discovered today its also being bitrotten by
lavu. As for the "but many programs do use it!" argument, that's also junk.
You ought to know how much the API's changed over the past years and how
its going to completely break those users in 2 years when the old lavc api
is gone. This is no different, its just another API deprecation. I'd agree
to call it "deprecate the lavr API" if you think its more diplomatic.
Point is, there's no reason to keep it, not political, nor technical. And
many people agree, as the thread last month has shown.