Please decode EXIF comments according to charset

[niol]

Hi,

Here is a preliminary patch (taken from my pet project[1]) which decodes EXIF comment according to the advertised charset. Howerver, it still needs to known the byOrder in order to decode unicode, but I need to fiddle with binding and do ont know yet how to do that. Maybe I'll give it a look in the next days.

(I had sent you an email about it about 2 years ago, but seems you did not receive it)

The patch is against de latest version 0.2.0.

Cheers,

Alex

[niol (alexandre-rossi)]

(My pet project is lazygal : http://sousmonlit.dyndns.org/~niol/playa/oss/projects/lazygal/ )

[Olivier Tilloy (osomon)]

Thanks for the patch!
I'll review it in depth, but from a very quick glance it looks good to me, the only missing bit would be exposing ExifData::byteOrder(), which shouldn't be too hard to do.

[Olivier Tilloy (osomon)]

Alexandre, would you have a set of sample images with comments in various possible encodings (valid and invalid as well if possible)? That would help testing and implementing relevant unit tests for the functionality.

[niol (alexandre-rossi)]

Hi,

> Alexandre, would you have a set of sample images with comments in
> various possible encodings (valid and invalid as well if possible)? That
> would help testing and implementing relevant unit tests for the
> functionality.

Here are the ones that the bug reporter on my program initially supplied :
http://castet.matthieu.free.fr/tmp/test_exif/

Alex

[Olivier Tilloy (osomon)]

Thanks, that’s a good basis to start experimenting with the functionality!

[Olivier Tilloy (osomon)]

Important note: starting with version 0.20, libexiv2 already does the decoding of the comment itself if it’s encoded in unicode.

[Olivier Tilloy (osomon)]

Implemented with revision 342 of the trunk.

[Olivier Tilloy (osomon)]

Alexandre, many thanks for the original patch on which I built to implement this functionality.