[FFmpeg-devel] [PATCH] avutil/opt: Support max > INT64_MAX in write_number() with AV_OPT_TYPE_INT64

[Ronald S. Bultje]

Hi,

On Sun, Oct 25, 2015 at 3:12 PM, wm4 <nfxjfg at googlemail.com> wrote:

> On Sun, 25 Oct 2015 19:46:42 +0100
> Michael Niedermayer <michaelni at gmx.at> wrote:
>
> > From: Michael Niedermayer <michael at niedermayer.cc>
> >
> > This allows for example to set max to UINT64_MAX and set values in that
> > range
> >
> > Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
> > ---
> >  libavutil/opt.c |   12 +++++++++++-
> >  1 file changed, 11 insertions(+), 1 deletion(-)
> >
> > diff --git a/libavutil/opt.c b/libavutil/opt.c
> > index 36eeeb0..157179c 100644
> > --- a/libavutil/opt.c
> > +++ b/libavutil/opt.c
> > @@ -102,7 +102,17 @@ static int write_number(void *obj, const AVOption
> *o, void *dst, double num, int
> >      case AV_OPT_TYPE_INT:   *(int       *)dst= llrint(num/den)*intnum;
> break;
> >      case AV_OPT_TYPE_DURATION:
> >      case AV_OPT_TYPE_CHANNEL_LAYOUT:
> > -    case AV_OPT_TYPE_INT64: *(int64_t   *)dst= llrint(num/den)*intnum;
> break;
> > +    case AV_OPT_TYPE_INT64:
> > +        // We must special case uint64_t here as llrint() does not
> support values
> > +        // outside the int64_t range and there is no portable function
> which does
> > +        // "INT64_MAX + 1ULL" is used as it is representable exactly as
> IEEE double
> > +        // while INT64_MAX is not
> > +        if (o->max > INT64_MAX + 1ULL && num/den > INT64_MAX + 1ULL) {
> > +            *(uint64_t *)dst = (llrint(num/den - (INT64_MAX + 1ULL)) +
> (INT64_MAX + 1ULL))*intnum;
> > +        } else {
> > +            *(int64_t  *)dst = llrint(num/den)*intnum;
> > +        }
> > +        break;
> >      case AV_OPT_TYPE_FLOAT: *(float     *)dst= num*intnum/den;
>  break;
> >      case AV_OPT_TYPE_DOUBLE:*(double    *)dst= num*intnum/den;
>  break;
> >      case AV_OPT_TYPE_RATIONAL:
>
> Introducing AV_OPT_TYPE_UINT64 would be cleaner...


+1.

Ronald