Re: [FFmpeg-devel] [PATCH v7 1/2] avcodec/pngenc: support writing iCCP chunks

[Niklas Haas <ffmpeg@haasn.xyz>]

On Fri, 01 Apr 2022 15:31:16 +0200 Andreas Rheinhardt <andreas.rheinhardt@outlook.com> wrote:
> Passing the PNGEncContext would be more natural.

Changed.

> Since db57a5370bd37105d389a45b04bf4970802407ec the callers'
> max_packet_size are not size_t any more, but always 64bit (so that
> there's no truncation in case size_t is 32bit).

Changed.

> ULONG_MAX is the maximum of unsigned long, yet deflateBound uses uLong.
> The latter is a currently typedef for unsigned long, but do we want to
> rely on that? The ordinary way to check for whethe a value can be
> represented in a type is by "if (sd->size != (uLong)sd->size)"

Changed. I didn't know about this trick, thanks.

> AV_LOG_WARNING makes no sense given that you error out afterwards.
> (And anyway: Is a log-message really needed for something that will
> never happen in reality?)

I think you're right. For JPEG it makes sense to keep around an error
printout, because the JPEG profile limit is quite small in comparison
(16 MB - the largest ICC profiles I have encountered in the wild are on
the order of megabytes as well).

But PNG supports up to 4 GB ICC profiles, which strains all conceivable
credibility. Somebody would have to be deliberately trying to hit this
limitation. Removed both error printouts (including the one about zlib
misbehaving, equally unlikely).
_______________________________________________
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".