On 8 January 2016 5:27:37 pm IST, "Clément Bœsch" <u...@pkh.me> wrote: >On Fri, Jan 08, 2016 at 05:24:15PM +0530, Anshul wrote: >> >> >> On 6 January 2016 1:55:27 am IST, "Clément Bœsch" <u...@pkh.me> wrote: >> >On Sun, Jan 03, 2016 at 01:07:15PM +0100, Clément Bœsch wrote: >> >[...] >> >> This indeed LGTM, but I'm not the maintainer. >> >> >> > >> >OK I finally understood why it's done that way: >validate_cc_data_pair() >> >alters the pkt data, but the decoder isn't supposed to do that. >> > >> >So this patch is actually incorrect in this state. >> >> >> I did not knew, what could be the disadvantage of changing pkt data, >I >> changed it to save some CPU cycles. > >The packet is considered read-only for the decoder. It appears not to >be >marked as const for whatever bad reason. > >But what CPU cycles would you be saving anyway here? I don't see what >operation you are saving. > >[...]
Not much, just bitwise operations to discard parity bit, every time for conversion and comparison. -Anshul -- Sent from my Android device with K-9 Mail. Please excuse my brevity. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel