Andreas Rheinhardt (12021-08-17):
> Well, this has the problem that it adds code in many places, whereas
> static flags for the filter just need code in two places (with zero
> marginal cost for making another filter use this feature, whereas a flag
> to ff_append_... has a nonzero cost even to users not making use of this
> feature). I also like that it uses a static flag for what is essentially
> a static property.

I do not understand your reasoning. If you have one case that is the
majority by a wide margin, then it makes sense to make it the default
and require the other case to be explicit. But that is not what is
happening here.

A quick count indicate that among filters adding pads dynamically, there
are two using allocated names for one using static names. Two-to-one is
not a huge majority, it is one of the cases where it makes more sense to
write it explicitly on both cases.

Furthermore, it is two-to-one for dynamically allocated, which means the
default should be to free(), and the flags should be
FF_FILTER_FLAG_DO_NOT_FREE_(IN|OUT)PADS, which would make the logic
quite ugly.

And honestly, we are talking adding ", 1" or ", 0" to each
ff_insert_(in|out)pad() call. Just the name of the flag requires eight
times as much characters in the source code.

Regards,

-- 
  Nicolas George

Attachment: signature.asc
Description: PGP signature

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

Reply via email to