Re: [FFmpeg-devel] [PATCH] lavfi: protection against premultiplied alpha (was: The patch series about premultiplied alpha)

From: Nicolas George <george@nsup.org>
To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Subject: Re: [FFmpeg-devel] [PATCH] lavfi: protection against premultiplied alpha (was: The patch series about premultiplied alpha)
Date: Tue, 5 Aug 2025 11:16:34 +0200
Message-ID: <aJHL8nVRBc_AoMa6@phare.normalesup.org> (raw)
In-Reply-To: <aJHJchHFpWBZVJWU@phare.normalesup.org>

Nicolas George (HE12025-08-05):
> > I will repeat myself briefly here that such a flag should be an opt-out, not
> > an opt-in. Supporting premultiplied alpha should be the default assumption
> > unless provable otherwise.
> 
> Absolutely not.
> 
> Do you dereference a pointer when you can prove it cannot be NULL or
> when you cannot prove it can be NULL? It is the same: if you cannot
> prove a filter produces a valid result, then we protect by default. It
> is already too much that it is not blocking when it is unspecified. On
> this I will stand firm.

Just an extra point:

Imagine we had done it opt-out as you suggest, and you were looking at
the list of >300 video filter source files.

Would you have thought to set the opt-out flag to drawbox?

I know you have the integrity to do so, so please take the time to think
about it and answer honestly.

Regards,

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

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