samba - Apr 2001 - client code page setting for win2k

I am having trouble with Samba 2.2.0 on FreeBSD 4.3. I did a simple
install from sources, not the ports collection. Everything seems to work,
except for non-ASCII characters in filenames.

When I use my Windows 2000 system (with English(US) locale) to access an
SMB share on the BSD box, apparently no translation is done on the
filenames. For example, LATIN CAPITAL LETTER U WITH DIAERESIS on the BSD
box (byte 0xDC in iso-8859-1) becomes a LOWER HALF BLOCK on the Win2K box
(byte 0xDC in cp437), whereas it should be translated to 0x9A.

I wrote, regarding the lack of translation of filename
bytes:
> From what I have seen in the documentation, this is the expected default 
> behavior, and I should be able to override it by using the [global] config 
> option
>   client code page = 437
> 
> However, this seems to have no effect whatsoever. I still get the
> graphical box characters, even after stopping and restarting the server.
I finally figured this one out.

If you use "client code page" to indicate the encoding the client is
using, you must also use "character set" to indicate the Unix
system's
filename encoding.

This is in spite of the fact that the smb.conf(5) man page specifically 
says that if you use "character set = ISO8859-1" then you must use
"client  code page = 850". This is crap.

So the answer, for the benefit of the archives, is this:

 [global]
    client code page = 437
    character set = ISO8859-1

Works like a charm.

   - Mike
_____________________________________________________________________________
mike j. brown, software engineer at  |  xml/xslt: http://skew.org/xml/
webb.net in denver, colorado, USA    |  personal: http://hyperreal.org/~mike/