L'octidi 18 brumaire, an CCXXVI, Hendrik Leppkes a écrit : > In the meantime, it does no real harm to add one more sidedata type to > the numerous list of types we already have, and lets people carry on > with their work.
It does harm, look at the other patch: 5 lines of useless clutter out of 10 lines for the feature, using side data for that features has a 100% overhead. And that is not only for the code here, it has the same consequences for the applications that will use that API as well: they have to do the same errors checks, they have to handle a type-pruned pointer, check the size of the data, etc. This is a terrible API, there is no doubt that people would prefer a single field. By all means let us discuss all this. But if Paul maintains his veto, I maintain mine. Please persuade Paul to behave like an adult of make good on his trice-repeated promise of forking. Regards, -- Nicolas George
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
