On Sat, Apr 27, 2013 at 05:12:06PM +0200, Hendrik Leppkes wrote: > On Sat, Apr 27, 2013 at 5:03 PM, Diego Biurrun <[email protected]> wrote: > > > On Sat, Apr 27, 2013 at 04:01:09PM +0200, Luca Barbato wrote: > > > On 04/27/2013 03:10 PM, Diego Biurrun wrote: > > > > --- a/libavcodec/h264_refs.c > > > > +++ b/libavcodec/h264_refs.c > > > > @@ -782,9 +782,9 @@ int ff_h264_decode_ref_pic_marking(H264Context *h, > > GetBitContext *gb, > > > > > > > > if (first_slice && mmco_index != -1) { > > > > h->mmco_index = mmco_index; > > > > - } else if (!first_slice && mmco_index >= 0 && > > > > - (mmco_index != h->mmco_index || > > > > - (i = check_opcodes(h->mmco, mmco_temp, mmco_index)))) > > { > > > > + } else if ((i = check_opcodes(h->mmco, mmco_temp, mmco_index)) || > > > > + !first_slice && mmco_index >= 0 && > > > > + mmco_index != h->mmco_index) { > > > > av_log(h->avctx, AV_LOG_ERROR, > > > > "Inconsistent MMCO state between slices [%d, %d, > > %d]\n", > > > > mmco_index, h->mmco_index, i); > > > > > > what happens to check_opcodes when mmc_index is negative? > > > > I don't know, but that seems to be immaterial: before a negative mmco_index > > was one of the conditions that triggered the invocation of check_opcodes(). > > > I can clearly see a check there for mmco_index >= 0 which was before the > invocation...
Yes, that was a brainfart... Diego _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
