[horde] locales/charsets and backends

[Tony Earnshaw]

søn, 2003-01-12 kl. 14:02 skrev Jan Schneider:

> That was pretty easy, but it becomes a lot more complicated on some other
> backends. LDAP is easy too, but what about other authentication backends for
> example? Let's say usernames, passwords on imap/ftp/pop3 servers? The sork
> modules are affected at most by this but potentially also every Auth driver
> or even IMP where we might manipulate user names.

I made a Norwegian LDAP user, Torbjørn Styggen, with cn=torbjørn,etc,
well enough (that "bjørn" is with the Norwegian (o.k., Danish too)
letter 'oe', o-slash). I put him into Horde's SQL database with that
name and because my locale is nn_NO, I could see his Norwegian name in
MySQL xterm and console login. Strangely enough, I can make a home
directory /u/home/torbjørn, too. But there's no way Posix will accept
8-bit names. So Imap (pam_ldap) can't verify him, since he doesn't have
a Posix-style home directory ) 'chown torbjørn:localusers
/u/home/torbjørn' - "chown: torbjørn:localusers: invalid user"

So we make his LDAP-based uid 7-bit, torbjorn. Then it works.

O.k. As long as all the Posix stuff is 7-bit, it works. He can log into
Horde main as torbjørn and the LDAP cn can be torbjørn (utf8), the
locale and MySQL, but everything Posix - uid, home directory, group,
must be 7-bit. User torbjørn has to log in 7-bit (as torbjorn) to Imp
using IMAP for authentication. Same with user gøran, I'd already tried
this a month ago..

If you were to implement a translation between utf8 and Posix, how would
you translate "ø" to "o"? Worse, if you were to translate "æ",
Norwegian/Danish ae ligature, how would you know to translate with "a"
or "e"? 

Versions:

UW Imapd/c-client 2002a
PHP 4.3.0, Apache/SSL 1.3.27
Openldap 2.1.0
MySQL 3.23.52

(All above self-compiled).

Horde v 3.0-CVs
Imp 4.0-cvs
locale GNU libc 2.2.4

Best,

Tony

--

> Any input how to handle this is welcome. Are there any other backends than
> sql and ldap (not even sure about this one) who support different (even
> multibyte) charsets other than ascii and iso-8859-1? Should we always add a
> charset parameter to a configuration when we deal with external data? Are
> there backends where we can assume a fixed charset?
> 
> I send this to the horde list because I hope for feedback especially from
> folks who deal on a daily basis with "non-standard" charsets, even multibyte
> charsets. All developers live in iso-8859-1 encoded parts of the world :-)
> so we depend on other people's experiences.
> 
> Jan.
> 
> --
> http://www.horde.org - The Horde Project
> http://www.ammma.de - discover your knowledge
> http://www.tip4all.de - Deine private Tippgemeinschaft



-- 

Tony Earnshaw

When all's said and done ...
there's nothing left to say or do.

e-post:		tonni at billy.demon.nl
www:		http://www.billy.demon.nl