On Mon, May 1, 2017 at 11:45 PM, Ronald S. Bultje <rsbul...@gmail.com> wrote: > Hi, > > On Mon, May 1, 2017 at 12:44 PM, Ronald S. Bultje <rsbul...@gmail.com> > wrote: > >> Hi, >> >> On Sun, Apr 30, 2017 at 1:16 PM, wm4 <nfx...@googlemail.com> wrote: >> >>> I think it's probably ok to push your current patch, since the original >>> author of the patch who broke this probably won't fix it. >> >> >> I don't like the patch because it essentially duplicates the C >> implementation of the DSP function. If we do that here, are we going to do >> that everywhere? Will we get 2 DSP functions, aligned and unaligned? I >> foresee a mess. >> >> Michael already said that too, I believe. >> > > Let me expand a little bit further: > - if we agree that the API does not require alignment, then we should fix > the DSP function (and others in other filters) to allow non-aligned or > differently-aligned input data. > - if not, then the data alignment should be fixed.
My opinion: - for short term, the data alignment should be fixed. - for long term, the framework should support non-aligned data. No needs for the DSP function to support non-aligned data. Instead ensure the framework to send aligned data to filters/codecs which require it. As Michael suggested, this can be achieved by using flags/required_alignment similar to need_writable. Thank's _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
