Hi, 2011/9/23 Måns Rullgård <[email protected]>: > Alex Converse <[email protected]> writes: > >> On Thu, Sep 22, 2011 at 5:51 PM, Justin Ruggles >> <[email protected]> wrote: >>> On 09/22/2011 08:33 PM, Måns Rullgård wrote: >>> >>>> Benjamin Larsson <[email protected]> writes: >>>> >>>>> On 09/23/2011 12:10 AM, Diego Biurrun wrote: >>>>>> --- >>>>>> libavformat/utils.c | 1 + >>>>>> 1 files changed, 1 insertions(+), 0 deletions(-) >>>>>> >>>>>> diff --git a/libavformat/utils.c b/libavformat/utils.c >>>>>> index 05d4fda..17b342e 100644 >>>>>> --- a/libavformat/utils.c >>>>>> +++ b/libavformat/utils.c >>>>>> @@ -852,6 +852,7 @@ static int is_intra_only(AVCodecContext *enc){ >>>>>> case CODEC_ID_MJPEG: >>>>>> case CODEC_ID_MJPEGB: >>>>>> case CODEC_ID_LJPEG: >>>>>> + case CODEC_ID_PRORES: >>>>>> case CODEC_ID_RAWVIDEO: >>>>>> case CODEC_ID_DVVIDEO: >>>>>> case CODEC_ID_HUFFYUV: >>>>> >>>>> How about a codec capability instead ? >>>> >>>> WTF is that used for at all? >>> >>> >>> looks to me like a shortcut to not require all the demuxers set the key >>> frame flag like they're supposed to. >> >> For mov files the keyframe box is often missing for intra only >> streams. I'm sure some other formats are similar. > > Then codec_cap flag makes more sense than keeping lists like this > floating around.
I believe this list exists so that we have the freedom to add new codecIDs without implementing the actual decoder (in which case the struct that specifies the flag is missing). In that case, transmuxing will produce the expected results. But yes the list is brittle also... Maybe we need both: a list for those where codecs are still missing and a flag for codecs that exist. Ronald _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
