On Tue, Jul 30, 2013 at 12:10 PM, Hendrik Leppkes <[email protected]>wrote:
> On Tue, Jul 30, 2013 at 12:00 PM, Vittorio Giovara > <[email protected]> wrote: > > --- > > Resend taking into account Luca and Diego's comments. > > > > libavcodec/h264.c | 198 > ++++++++++++++++++++++++++++++----------------------- > > 1 file changed, 113 insertions(+), 85 deletions(-) > > > > diff --git a/libavcodec/h264.c b/libavcodec/h264.c > > index 2dccb87..effd6d9 100644 > > --- a/libavcodec/h264.c > > +++ b/libavcodec/h264.c > > @@ -415,7 +415,7 @@ int ff_h264_check_intra4x4_pred_mode(H264Context *h) > > av_log(h->avctx, AV_LOG_ERROR, > > "top block unavailable for requested intra4x4 > mode %d at %d %d\n", > > status, h->mb_x, h->mb_y); > > - return -1; > > + return status; > > Status is not an AVERROR code, it'll just be -1, so > AVERROR_INVALIDDATA seems more appropriate? > Ok I'll also take care of the case below. > > > } else if (status) { > > h->intra4x4_pred_mode_cache[scan8[0] + i] = status; > > } > > @@ -431,7 +431,7 @@ int ff_h264_check_intra4x4_pred_mode(H264Context *h) > > av_log(h->avctx, AV_LOG_ERROR, > > "left block unavailable for requested > intra4x4 mode %d at %d %d\n", > > status, h->mb_x, h->mb_y); > > - return -1; > > + return status; > > @@ -2786,7 +2798,8 @@ static int field_end(H264Context *h, int in_setup) > > } > > > > if (avctx->hwaccel) { > > - if (avctx->hwaccel->end_frame(avctx) < 0) > > + err = avctx->hwaccel->end_frame(avctx); > > + if (err < 0) > > av_log(avctx, AV_LOG_ERROR, > > "hardware accelerator failed to decode picture\n"); > > This change seems unnecessary, as the return value is not used for > anything. > In addition, you can skip the additional "ret" variable and keep using > "err" when err is not being clobbered here. > > Ok I'll revert this change. > > > @@ -4062,7 +4082,8 @@ static void loop_filter(H264Context *h, int > start_x, int end_x) > > } > > backup_mb_border(h, dest_y, dest_cb, dest_cr, linesize, > > uvlinesize, 0); > > - if (fill_filter_caches(h, mb_type)) > > + ret = fill_filter_caches(h, mb_type); > > + if (ret) > > continue; > > Since the value is also not used for anything, is this really the better > way? > > Yeah, I'll revert this one too Vittorio _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
