Daniel Cantarín (12021-12-12): > This quickly went from exciting to depressing. :(
I find this whole situation depressing too. > I guess some recent-times archeology will also be needed in order to > discuss this, as the patchset is 23 versions now and has 21 parts. I'll give > it some time starting from tomorrow. If you Mr. George happen to > remember where to look at, please kindly point me in the right direction, > in order to avoid reading the full patchset history of messages. Of > course, I'll read the whole thing otherwise, no problem. > > Also, from your last comment, I feel you're kinda pissed off with this > patchset (or with softworkz, obviously). Yet, I've seen several answers > from you years ago (like, about a decade now) where you were actually > involved in the whole "subtitles in lavfi" thing. I mean: I've searched for > this several times from years by now, and most of the time you were > involved in the debates. I would say then this is something you care > and have some knowledge about. > > If I were doing all the work softworkz is doing, I'll be pissed off with the > backslash, no doubt about it. And if I were dealing with this use case for > the last 10 years and people keep not doing what I'm convinced should > be done, I'll be very pissed off too. So, this is kinda understandable. > But I find it to be a pity that this use cases get blocked by people just > not getting along. There has to be some way around it. > > Isn't there anything that can be done to combine both your knowledge > and softworkz's will to tackle the issue? Perhaps laying down some base > design that should be honored, while other people like softworkz or I > could just progressively implement it? I have told Soft Works they were going at it the wrong way when they posted the very first patch. At the time, I was working on refactoring that needs to happen before libavfilter can be extended with a third media type. I gave Soft Works advice on how to help for tasks that are necessary to implement subtitles the right way. Instead of listening to my expertise, they persisted. If they wasted six months of effort and counting, they can blame only themselves. I still intend to work on this, I have clear plans on how to go forward, but right now I find the whole mess too depressing to even consider investing time in it. > I bet that if softworkz code actually do the job, then it can't be THAT far > away from an acceptable design, and thus that distance should be > workable. Unfortunately, it is that far. They designed and tested their patches with file inputs and simple grpahs. We need libavfilter to work with complex graphs where timestamps are changed, frames arrive in the wrong order on filters, etc. More importantly, they completely bypassed the format negotiation system, letting the responsibility fall onto users and making it impossible to have the many necessary utility filters without unacceptable code duplication. Regards, -- Nicolas George
signature.asc
Description: PGP signature
_______________________________________________ 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".
