samba codepage/char set problem
 

Hi all,

I set up a linux box (rh9) with samba (2.2.7 - rebuilt with ldap support, does not work as domain controller), Samba should give several windows users access to their directories.

Everything works fine but region specific characters. Names of any file or directory that contain region specific characters appear "crippled" on each other system (win created file -> liunx doesn't see correct name, even cannot access the file , linux created file -> win doesn't see the correct filename).

I tried following smb.conf settings:
codepage = 850
character set = iso8859-1
valid chars = ä:Ä ö:Ö ü:Ü

plus some variatons like
codepage = 437
character set = iso8859-15
...

Nothing did effect my issue.

Does anyone can giv me a hint, what else I can try?

Thanks in advance,
Clemens vo Musil

P.S: The concrete region, in which this happened is germany, the "problem childs" are äöüÄÖÜß. If you see them "crippled" , too: aouAOU with two dots above and a special s.

In [global] section I've put:
character set = ISO8859-2
client code page = 852
And it works for Polish codings.

It is UTF-8!

I found out, that the conversion to UTF-8 on my linux box must fail. Unfotunately, Samba seems not to support UTF-8 at "character set".

Is it possible, to add it?

Clemente, if your system support UTF-8 (it does) the files might be coded in CP850 (example) format. You have to use a program that will convert the files in UTF-8 format (or another format that won't mess with your system).

convmv will convert a file from a format to another. Convert every samba file you have to UTF-8, or CP850. Check it on a single file first, but you have a dry-run (test mode) integrated anyway. This prog is a must-have and saved my life twice now.

google convmv