On 4/17/2020 10:54 AM, Derek Buitenhuis wrote: > On 17/04/2020 14:38, Nicolas George wrote: >> James Almer (12020-04-17): >>> The movenc test does not produce any files. It creates bogus packets >>> that are dumped into the AVOutputFormat output using a custom >>> AVIOContext on each stream that just calculates an md5 hash of the data. > > It has an option for this. You could make it default.
Yes, my bad. I didn't look deep enough. > >>> I think the one line change i posted is enough to reproduce the changes, >>> and we don't need the patch author to essentially rewrite the movenc >>> test just to dump everything into a file we can pass to ffprobe that may >>> or may not be parseable to begin with. > See above. There is no need to rewrite. > >> >> Yes, for that I agree. >> >> But urgh, is there a very good reason for the movenc test to be that >> kind of abomination? > > Can you elaborate on why it's an abomination? It's so far been the only thing > that has kept people from breaking movenc API use several times, specifically > in relation to how it behaves for outputting fragments - and ffmpeg.c does not > necessarily exercise these paths. API unit tests are good thing in my books, > as an API-only user who has seen unintentional API breaks because ffmpeg.c > doesn't use something specifically. > > - Derek > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".