Here are two different issues.
First: UTF8. The image comment (JFIF comment) is saved as UTF8, and UTF8 saves non-ASCII characters in two bytes.

Second: Reading comments to the database. This is only done after importing the image for the first time. After that, the comment is read from the database.
Is that what you mean?

See http://mail.kde.org/pipermail/digikam-users/2006-October/002116.html for information on how digikam handles the various possibilities to store the comment in an image.

First: OK

Second: You mean that, if I edit exiv tags with another images manager, those images needs to be re-imported in digikam?

Thanks for the link, I'll take a look ASAP.

Valerio,

Please give me fresh feedback about this report, epecially using current digiKam implementation from svn. Thanks in advance

Gilles

Hi Gilles, I've made these tests with SVN (trunk) digikam:

Save comments in digikam --> read with exiv from command line --> No problem

Save comments from command line (exiv2) --> read with digikam --> No change made. (Digikam doesn't update its database)

Is it the expected behaviour?

Valerio,

Yes it's normal because something is missing in digiKam core about this problem. There is another file in B.K.O about it...

But the problem of this file is charset. It look fine. Right ?

Gilles

Fine :)

Valerio

ok. thanks closed.

Gilles