On Thu, Mar 27, 2014 at 02:42:22AM +0100, Vittorio Giovara wrote: > On Wed, Mar 26, 2014 at 7:26 PM, Diego Biurrun <[email protected]> wrote: > > On Wed, Mar 26, 2014 at 05:48:53PM +0100, Vittorio Giovara wrote: > >> --- a/libavcodec/vp8dsp.c > >> +++ b/libavcodec/vp8dsp.c > >> @@ -518,27 +650,35 @@ VP8_BILINEAR(4) > >> > >> -av_cold void ff_vp8dsp_init(VP8DSPContext *dsp) > >> +av_cold void ff_vp8dsp_init(VP8DSPContext *dsp, int vp7) > > > > Why don't you split the dsp stuff in two and call just one or both of > > ff_vp8dsp_init() and ff_vp8dsp_init()? > > I'm not sure I understand. You'd prefer a single separate > ff_vp7dsp_init() and ff_vp8dsp_init() or a single ff_vp78dsp_init() > that calls the relevant parts of the dsp?
ff_vp8dsp_init() and ff_vp78dsp_init() - IIUC there are no VP7-only optimizations. This way you only have to call the right init functions once instead of propagating the is_vp7 information all around. Diego _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
