Re: [gexiv2] Python 2/3 and unicode

[Jim Nelson <jim yorba org>]

There are two tickets that are related to this question:

https://bugzilla.gnome.org/show_bug.cgi?id=730136

However, I'm still unsure what encoding this string is in (if it is encoded and not just garbage).

https://bugzilla.gnome.org/show_bug.cgi?id=712484

But this is about detecting the character encoding for IPTC, not EXIF.  (As you mentioned, EXIF only supports ASCII.)

Unfortunately, I'm not a Python hacker and can't provide help any further that direction.  If there's a way to disable (or provide a different codec) in Python, you should take that up with a Python forum.

-- Jim

On Fri, Aug 8, 2014 at 2:45 AM, Jim Easterbrook <jim jim-easterbrook me uk> wrote:

I'm using the Python Gexiv2 bindings to read metadata from some .jpg files. Some of the files have EXIF data in the Latin-1 character set (yes, I know the EXIF spec says ASCII only). In Python 2 I can cope with these files as 'get_tag_string' returns a Python 2 string, which I can then convert to unicode with the latin-1 codec. In Python 3 'get_tag_string' returns a unicode string (as all Python 3 strings are unicode) after converting with the utf-8 codec. This raises the UnicodeDecodeError exception when it encounters some latin-1 characters. Is there any way to change the codec used by 'get_tag_string' or to change its return type from unicode to bytes?

--

Jim Easterbrook <http://www.jim-easterbrook.me.uk/> _______________________________________________ gexiv2-list mailing list gexiv2-list gnome org https://mail.gnome.org/mailman/listinfo/gexiv2-list