Hi, On 23.10.2016 04:10, Ronald S. Bultje wrote: > This is hideous, you're sprinkling emms_c in various places to make some > stupid test pass. The test is morbidly stupid
Can you elaborate why you think the test is bad? > and there is no general > consensus on patterns to be followed as for where to place emms_c. Someone > who doesn't know any better will litter each new decoder with 10-20 calls > to emms_c just because he found that other decoders do it in undocumented, > unexplained and unclear locations also. > > If you want this to be a "thing", you need to design and document carefully > where emms_c is necessary. Then come up with some system that makes this > work by itself. I've said from the beginning that I highly dislike > littering the code with emms_c in individual decoders, and that's exactly > what you're doing here. This is insane. I also don't like adding emms_c in various places, but I don't see a realistic alternative other than not fixing the problem. And I think that would be worse than this patch set. Best regards, Andreas _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
