On Tue, 30 Jan 2018 20:16:50 +0700 Muhammad Faiz <mfc...@gmail.com> wrote:
> On Tue, Jan 30, 2018 at 2:58 PM, wm4 <nfx...@googlemail.com> wrote: > > On Tue, 30 Jan 2018 14:24:12 +0700 > > Muhammad Faiz <mfc...@gmail.com> wrote: > > > >> Move REGISTER_FILTER to FILTER_TABLE in configure. > >> Auto generate filter extern and filter table. > >> Sort filter table, use bsearch on avfilter_get_by_name. > >> Define next pointer at filter extern, no need to initialize > >> next pointer at run time, so AVFilter can be set to const. > >> Make avfilter_register always return error. > >> Target checkasm now depends on EXTRALIBS-avformat. > >> > >> Signed-off-by: Muhammad Faiz <mfc...@gmail.com> > >> --- > > > > Pretty nice. Should we still add a new filter listing API that uses the > > same idiom as the BSF API? (Although you generate the .next links at > > configure time it might still be nice to eventually get rid of that.) > > IMHO, it is uglier (i.e. by requiring opaque pointer). I don't know > other people's opinion. Anyway, it can be added later. Yes, because it uses an opaque pointer, it could use the linked list, without changing any of the existing code. I think a simple filter array would be preferable in the end, which would require an iteration function which can pass an index to the user. (Although I agree that this opaque thing is uglier than the current API or an API which just takes an index.) _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel