On 2018/1/22 22:15, wm4 wrote: > On Mon, 22 Jan 2018 09:58:56 +0100 > Hendrik Leppkes <h.lepp...@gmail.com> wrote: > >> On Mon, Jan 22, 2018 at 8:55 AM, Jun Zhao <mypopy...@gmail.com> wrote: >>> Hi, all: >>> >>> When I read the code about av_init_packet(), I found we use >>> av_init_packet() in most cases like this: >>> >>> av_init_packet(&enc_pkt); >>> enc_pkt.data = NULL; >>> enc_pkt.size = 0; >>> >>> my question is: is it have any special reason to put >>> AVPakcet.date/size initialization outside av_init_packet()? >>> >>> If don't have special reason, I will submit a patch to put >>> the data/size initialization in av_init_packet() and remove the >>> explicit initialization for data/size field. >>> >> This is in fact intentional behavior, and documented as part of the >> API, so changing that would be an API break. Code can currently rely >> on size/data to not be modified - and I'm sure there are also places >> where it does. > I think nobody likes or actually wants this behavior though, so this is > just not changed for strict API compatibility. Maybe we should replace > it with a new function and deprecate the old one? (Last time we had > this discussion it went nowhere, though.) I agree this, the code depends on old size/date even after call av_init_packet(), which behavior is very strange,I guess most people don't want to use or perceive this type action, but for API compatibility, we can give a new API and deprecate this API now. Other suggestion or comment will be welcome.
Thanks. > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel