On Monday 20 December 2004 10:25, Yedidyah Bar-David wrote:
> On Sun, Dec 19, 2004 at 11:10:33PM +0200, Naomi Schor wrote:
> > > > > 0001120 3234 3220 3030 0a34 2020 e2c6 2ee0 7874
> > > >
> > > > All plain ASCII except those three characters here.
> > > >
> > > > Appears to be cp1255: Gimel, Segol, Alef
>
> [snip]
>
> > > what I see on the Windows screen is : Ayin, Dalet, Vav ("ido")
>
> [snip]
>
> I must be having Deja Vu ...
>
> This is the very same symptom someone raised here a few months ago,
> with unzip.
>
> Try this:
> iconv -f iso8859-1 -t cp850 | iconv -f cp862 -t iso8859-8
>
> Why does this happen? I do not know. Either smbclient Or Windows
> wrongly think this is cp850, and converts to iso8859-1. You can try
> and sniff (or some other such low-level debugging) if you are
> interested, or simply put this in your script.
Samba has a configuration directive "dos charset" which is by default
set to cp850. Apparently on Naomi's system, Samba cannot determine the
charset automaticly and falls back to the dos charset defined.
I tried reproducing said problem on my system and failed - my console is
utf-8, so I have "unix charset" in the samba configuration set to
utf-8, and I have no "dos charset" configured. With Samba 3.0.9 and
Windows XP I get the correct hebrew characters using smbclient on the
console.
I don't have access to a non NT 5.x Windows system, so I cannot test
other configurations but currently I assume that Naomi's Windows
machine does not advertise its character set correctly and doesn't use
unicode. Its likley that setting "dos charset = cp862" in the smb.conf
file will resolve this issue.
--
Oded
::..
Don't let people drive you crazy when you know it's in walking distance.
=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
