On Thu, Jun 23, 2016 at 09:46:38PM +0200, Anton Khirnov wrote: > Quoting Diego Biurrun (2016-06-23 17:05:23) > > On Thu, Jun 23, 2016 at 10:52:46AM -0400, Vittorio Giovara wrote: > > > On Thu, Jun 23, 2016 at 8:46 AM, Diego Biurrun <di...@biurrun.de> wrote: > > > > --- > > > > > > > > This does not seem worth the ifdef soup to me. > > > > > > > > libavcodec/msmpeg4dec.c | 25 ------------------------- > > > > 1 file changed, 25 deletions(-) > > > > > > > > --- a/libavcodec/msmpeg4dec.c > > > > +++ b/libavcodec/msmpeg4dec.c > > > > @@ -778,23 +765,11 @@ int ff_msmpeg4_decode_block(MpegEncContext * s, > > > > int16_t * block, > > > > level = level + rl->max_level[run>>7][(run-1)&63] * > > > > qmul;//FIXME opt indexing > > > > level = (level ^ SHOW_SBITS(re, &s->gb, 1)) - > > > > SHOW_SBITS(re, &s->gb, 1); > > > > LAST_SKIP_BITS(re, &s->gb, 1); > > > > -#ifdef ERROR_DETAILS > > > > - if(run==66) > > > > - av_log(s->avctx, AV_LOG_ERROR, "illegal vlc code > > > > in ESC1 level=%d\n", level); > > > > - else if((i>62 && i<192) || i>192+63) > > > > - av_log(s->avctx, AV_LOG_ERROR, "run overflow in > > > > ESC1 i=%d run=%d level=%d\n", i, run, level); > > > > -#endif > > > > } > > > > } else { > > > > i+= run; > > > > level = (level ^ SHOW_SBITS(re, &s->gb, 1)) - > > > > SHOW_SBITS(re, &s->gb, 1); > > > > LAST_SKIP_BITS(re, &s->gb, 1); > > > > -#ifdef ERROR_DETAILS > > > > - if(run==66) > > > > - av_log(s->avctx, AV_LOG_ERROR, "illegal vlc code > > > > level=%d\n", level); > > > > - else if((i>62 && i<192) || i>192+63) > > > > - av_log(s->avctx, AV_LOG_ERROR, "run overflow i=%d > > > > run=%d level=%d\n", i, run, level); > > > > -#endif > > > > } > > > > if (i > 62){ > > > > i-= 192; > > > > > > how about replacing the #ifdefs with avctx->debug checks > > > > I don't think I can be bothered to be honest. > > More importantly, that could mean a significant performance impact. > Which is probably why this was ifdefed out in the first place.
Can I kill it with fire then? Diego _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel