Re: EXIF support in PNG

Very recently I saw a bug report about ImageMagick; someone writing a PNG
with GIMP mixed up resolution units and put 11000 pixels per cm, resulting
in IM running out of memory trying to expand the image.  I'm not suggesting
that that's a bug that we need to fix, but it is s security issue to be
aware of.

Sorry, I can't find that bug report now...

Glenn

Glenn

On Thu, Jan 12, 2017 at 10:51 PM, John Bowler <
[email protected]> wrote:

> This github project:
>
> https://github.com/jbowler/jpegtopng
>
> Now contains a Makefile and source to build three programs:
>
> jpegtopng: requires an installed -ljpeg and -lz
> jpegdump: stand-alone hack program to dump a JPEG
> pngdumpmeta: requires -lpng and -lexiv2
> The first file takes any JPEG with an APP1 EXIF marker and converts it
> to an exIf chunk matching Willems previously supplied example image.
>
> The second is just a way of looking inside a JPEG, at least for the
> markers it understands (it also produces lots of warnings with -Wall).
>
> The third is the real example program; it decodes the exIf chunk
> (that's just a few lines of code, mostly recognizing that it is an
> exIf from the header) then it prints the contents using the Exiv2
> library; that's about 20 lines of code as well.
>
> It also prints an ISO XMP chunk.  Quite a lot of the code is to verify
> UTF-8 data (that's in the directory hutf8/), XMP is just raw XML so it
> just prints so long as it is valid.
>
> It will also print text chunks; either a list of the keywords or the
> text one of them contains.  It prints the text in UTF-8; so it
> converts the Latin1 text in tEXt or zTXt chunks into UTF-8 on-the-fly.
>
> It only does one thing at once.  The first and last programs read from
> stdin and write to stdout (no file management).
>
> Next I'm going to look at adding support for the text chunk used to
> output Exif from Image/GraphicsMagick and DigiKam.  The sample code in
> the pngexif directory already does that but I haven't tested it.  It's
> 132 lines of code as opposed to 59 when the data is in an unknown
> chunk; decoding the format about doubles the amount of code required
> (the rest of it is comments and looking though the text or unknown
> chunk lists.)
>
> I haven't verified compilation against libpng12 yet, but I don't
> anticipate any problems.
>
> John Bowler
>
> ------------------------------------------------------------
> ------------------
> Developer Access Program for Intel Xeon Phi Processors
> Access to Intel Xeon Phi processor-based developer platforms.
> With one year of Intel Parallel Studio XE.
> Training and support from Colfax.
> Order your platform today. http://sdm.link/xeonphi
> _______________________________________________
> png-mng-misc mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/png-mng-misc
>

------------------------------------------------------------------------------
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi

_______________________________________________
png-mng-misc mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/png-mng-misc