CIFS character conversions. Please read.

[Thor Thorvaldsen]

Hello,
With Tru64 we could read filenames written by windows with unix applications(Oracle) with correct national characters. A national letter Ø (o with slash) shows up as two bytes in HPUX.

I would prefer not to convert and use the correct alphanumeric presentation.

How come?
And can you help?

Regards,
Thor

[Danny Petterson - DK]

Hm...maybe check and set

locale

LC_LANG and that kind of thing

...I thing we used to use that to on Tru64/DU/OSF1.

[Steven Schweda]

> [...] shows up as two bytes [...]

Shows up where? "ls" output? Other places?

[Peter Nikitka · Accepted solution]

Hi,

I assume you are using a Samba base version 3.x. One of the main changes was (from samba.org):
<<
2) Unicode support. Samba will now negotiate UNICODE on the wire
and internally there is now a much better infrastructure for
multi-byte and UNICODE character sets.
>>

All stuff of the following, you could use for versions 2.2.x in smb.conf to get a mapping of e.g. utf8 to iso8859-1, has been gone:
* character set
* client codepage
* code page directory
* coding system

From the documentation, I see no way to get this mapping back, i.e. seeing a file a WIN-user created on a Samba share with utf-8 A-Umlaut (Ã ) consisting of the bytes '\303\244' as a iso8859-1 A-Umlaut under Unix which is a single byte '\344'.

mfG Peter

[Thor Thorvaldsen]

Heelo,

I´ve changed "unix charset" in smb.conf. Then the filenames created with win client was identical when i logged in the HPUX host.

mvh,
T