[FFmpeg-devel] [PATCH] libvmaf: exit gracefully if the library fails.

[Ronald S. Bultje]

Hi,

On Mon, Dec 18, 2017 at 11:55 PM, James Almer <jamrial at gmail.com> wrote:

> On 12/8/2017 10:35 PM, Ronald S. Bultje wrote:
> > Fixes trac issue #6884 and Netflix/vmaf issue #124.
> > ---
> >  libavfilter/vf_libvmaf.c | 28 ++++++++++++++++++++++------
> >  1 file changed, 22 insertions(+), 6 deletions(-)
> >
> > diff --git a/libavfilter/vf_libvmaf.c b/libavfilter/vf_libvmaf.c
> > index 7670c51..d628b95 100644
> > --- a/libavfilter/vf_libvmaf.c
> > +++ b/libavfilter/vf_libvmaf.c
> > @@ -61,6 +61,7 @@ typedef struct LIBVMAFContext {
> >      int ssim;
> >      int ms_ssim;
> >      char *pool;
> > +    int error;
> >  } LIBVMAFContext;
> >
> >  #define OFFSET(x) offsetof(LIBVMAFContext, x)
> > @@ -158,17 +159,24 @@ static void compute_vmaf_score(LIBVMAFContext *s)
> >
> >      format = (char *) s->desc->name;
> >
> > -    s->vmaf_score = compute_vmaf(format, s->width, s->height,
> read_frame, s,
> > -                                 s->model_path, s->log_path,
> s->log_fmt, 0, 0,
> > -                                 s->enable_transform, s->phone_model,
> s->psnr,
> > -                                 s->ssim, s->ms_ssim, s->pool);
> > +    s->error = compute_vmaf(&s->vmaf_score, format, s->width, s->height,
>
> This is an awful API break. It broke compilation for everyone using the
> latest stable release of libvmaf. See ticket #6921.
> How hard was for them to add a new function instead?
> compute_vmaf_score() or whatever. You can't just go and change a public
> function signature like this...
>
> This can still implement this properly before a new release is tagged,
> for that matter. Could you suggest that to them?


Both requests made in https://github.com/Netflix/vmaf/issues/124.

Ronald