[FFmpeg-devel] [RFC]] swscale modernization proposal

Niklas Haas ffmpeg at haasn.xyz
Tue Sep 17 15:09:51 EEST 2024


On Mon, 08 Jul 2024 09:25:08 -0400 "Ronald S. Bultje" <rsbultje at gmail.com> wrote:
> Hi,
>
> On Mon, Jul 8, 2024 at 8:34 AM Andrew Sayers <ffmpeg-devel at pileofstuff.org>
> wrote:
>
> > We often say e.g. "this struct currently has such-and-such members, but the
> > size is not part of the public API".  So it's not much of a stretch to say
> > "this preset enables such-and-such features, but the value is not part of
> > the
> > public API".
> >
>
> I like that, +1. No other comments for now.

So, the compromise I settled on is to introduce AV_SCALE_PRESET_NONE = 0 and
have that be the default value, with the explicint semantics of preserving
backwards compatibility (including with legacy swscale). That way, test
suites can transparently switch to the new API without breaking behavior,
and we can mark the rest of the named presents as implying semantics which are
not part of the public ABI.


More information about the ffmpeg-devel mailing list