Hi Ronald! > So, since you're stating this as fact, I would like to point out that this > was never formally agreed by any group of FFmpeg developers, and this is > certainly not the official position.
Your message surprises me: The mpegts demuxer (before this discussed patch) accepts all kind of broken streams. Most of the non-trivial (and some of the more trivial) decoders we have accept clearly invalid input provided by users. I remember that in the past, this was one of the main selling arguments in favour of FFmpeg. I was under the impression that the only reason that demuxers and decoders are written like that is that we all want FFmpeg to support invalid streams as long as this doesn't affect valid streams. It also seems difficult to imagine that users typically wouldn't want FFmpeg to behave the way it does. Are you now arguing writing decoders and demuxers like this was a mistake? That libavformat and libavcodec should only accept valid streams that conform to a specification? Am I correct that your argument also applies to the adpcm_ms stream from the Baby Monitor? That user requests for decoding broken streams from whatever devices should be ignored from now on? I just don't understand the whole point of this argumentation, what is it? Carl Eugen _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
