On 5/5/2017 2:23 PM, Nicolas George wrote: > Le sextidi 16 floréal, an CCXXV, James Almer a écrit : >> stuff you broke > > Please stop spreading wm4's lies. The bug was already there. > > Regards, >
Take a look at https://trac.ffmpeg.org/ticket/6346. That ticket reported a regression in concatdec, where some file started to segfault where it didn't before. A bisect showed the commit that started this was a commit made by myself, even if mostly authored by someone else. That commit was not the cause of the problem, but the one that exposed the underlying bug, as i stated in that ticket. You know what the underlying bug was? The autobsf feature in concatdec was broken by *you* in commits 0cb19c30c6 and b8fa374fb6 (see https://ffmpeg.org/pipermail/ffmpeg-devel/2017-April/210639.html). Now, what was my reaction? I worked around the issue to restore the original behavior. I didn't try to coerce you into fixing what you broke a year ago. I worked around what you broke and i unknowingly exposed, for immediate backporting purposes. Eventually i did go and fixed the underlying bug, as mentioned in the last comment from the ticket, but that's not important here. So, my point? Even if you didn't cause the underlying alignment issue, you exposed it and created a regression. Either work around it yourself, or help others trying to do it for you. Have the decency of not blocking every attempt others give at it. You're right that the real fix needs to go deeper, but that requires a design discussion, and the crashes in release/3.3 can't wait. They need an immediate, and able to backport solution. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel