[opus] Antw: [PATCH 1/2] op_pcm_seek: fix int64 overflow

[Ulrich Windl]

Hi!

I never inspected the Opus sources, but di you guys think that adding blanks (e.g. around binary operators) is a waste of time and storage space? 

Regards,
Ulrich

>>> James Zern <jzern at google.com> schrieb am 20.11.2017 um 22:07 in Nachricht
<20171120210721.186099-2-jzern at google.com>:
> check for overflow with a negative diff
> ---
>  src/opusfile.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/src/opusfile.c b/src/opusfile.c
> index 72f1272..df326af 100644
> --- a/src/opusfile.c
> +++ b/src/opusfile.c
> @@ -2605,7 +2605,11 @@ int op_pcm_seek(OggOpusFile *_of,ogg_int64_t 
> _pcm_offset){
>           would be better just to do a full seek.*/
>        if(OP_LIKELY(!op_granpos_diff(&diff,gp,pcm_start))){
>          ogg_int64_t discard_count;
> -        discard_count=_pcm_offset-diff;
> +        /*Check for overflow.*/
> +        if(diff<0&&OP_UNLIKELY(OP_INT64_MAX+diff<_pcm_offset)){
> +          discard_count=0;
> +        }
> +        else discard_count=_pcm_offset-diff;
>          /*We use a threshold of 90 ms instead of 80, since 80 ms is the
>             _minimum_ we would have discarded after a full seek.
>            Assuming 20 ms frames (the default), we'd discard 90 ms on 
> average.*/
> -- 
> 2.15.0.448.gf294e3d99a-goog
> 
> _______________________________________________
> opus mailing list
> opus at xiph.org 
> http://lists.xiph.org/mailman/listinfo/opus