Handling of User Ids containing non ascii characters

[Ingo Klöcker]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thomas Gebhardt wrote:
> recently I had some trouble importing a key generated by NAI PGP 7.0
> with an user id containing non ascii characters, namely the German
> umlaut "ö".
>
> These are the observations (The user id as showed by NAI PGP
> is "Peter Knöll ...."):
>
> 1. gpg --list-keys shows "Peter Kn\x6cl ...." as user id,
>    i.e., ö -> \x6c

Their seems to be a problem with the encoding. My user id also includes 
an 'ö'. I generated the key with GnuPG. gpg --list-keys gives me "Ingo 
Klöcker ..." as user id.

You could try gpg --list-keys --with-colons. Then the user id is 
printed in UTF-8 encoding. In UTF-8 the 'ö' becomes 'ö'.

> As far as I can see RFC2440 allows utf8 encoded unicode characters
> in the user id. I cannot judge whether the encoding is correct
> in the original key.

I just tried gpg --with-colons on the key data you included in your 
messages. This gives me "Peter Knöll ...". So the problem is that PGP 
saved the 'ö' as 'ö' instead of saving the UTF-8 representation of it. 
It seems that PGP 7.0 is not RFC2440 compliant. No wonder that GnuPG 
gets confused when it tries to interpret 'öl' as UTF-8 encoded 
character.

I'm really glad I used a RFC2440 compliant program to generate my 
OpenPGP keys.
Thanks for this great program, Werner!

Regards,
Ingo

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.3 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE6fvrHGnR+RTDgudgRAjYQAJ9mwkWmRliKgMmfQB3X3u5tClVhoACgxnGV
SK8BTJWDswtFHkuv5nnEmJk=
=URlK
-----END PGP SIGNATURE-----