UNIX for Advanced & Expert Users

All,

I'm trying to grasp how to use Unicode with/in Unix. I've made progress on some fronts, for example, when uploading files to my server I can use the intermediary language to convert the file to UTF-8.

I'm having trouble getting Samba to do this (I'm using "unix charset" in smb.conf); however, I thought I'd ask a questions that's even more bare:

How would I create (touch) a file that is named in UTF-8? For example, the terminal would display: groß

Thanks much--I've been banging my head against this.

P.S. I'm using RedHat Linux and locale tells me I'm en_US.UTF-8.

The attached screenshot shows how I made it. I think it depends on the shell being used. I used bash and touched the file without problems. As bash is default on Linux, it is likely to work for you too.

Ask again if you have further queries.

N.B. This snapshot is taken on Fedora Core 5.

Excellent! Thank you. What do you call the $' ' syntax?

Perhaps you could help me with another question.... I have a file on my Windows 98 desktop called "frere_jacques.mp3". I opened charmap, copied the "é", and renamed the file to "frére_jacques.mp3". When I copy the file onto my Unix box via Samba, an ls shows "fr?re_jacques.mp3" and od -cx shows:

Code:

0000000   f   r 202   r   e   _   j   a   c   q   u   e   s   .   m   p
        7266 7282 5f65 616a 7163 6575 2e73 706d
0000020   3  \n
        0a33
0000022

I don't understand this result.

Samba (3.0.22) is configured with:

unix charset = UTF8
display charset = UTF8
dos charset = (I've tried ASCII and UTF8)

By the way, the groß file that I created shows as "gro_(black bar)" through Samba. Something is awry...

Thanks again.

Quite obviously is not creating the file in Unicode. I am not familiar with Samba Unicode handling but from your quoted results it appears that it has converted the filename into ISO-8859-1 instead.

From your "od", é has the (Extended) ASCII charcode of 82 (in hex), which matches the output of "od" that you quoted. Therefore, your "od" has shown an ASCII representation. How did you generate the 'od'? Piping ls output to od? Also in UTF-8, that character should be represented by 2 bytes, not 1.

I also tried on my system (both machines Linux and samba-mountable), and I could touch the groß file and have the name properly shown on both systems. The smb.conf has no special Unicode-related configuration. Is it possible that your Windows side has problems with configuration or support?

If all else fails, you may need to visit Samba-related newsgroups or other user communities for help.

Bernard,

As usual, it was something silly on my part. I had installed a newer version of Samba, but certain links and rc.d scripts were still pointing to the old. This is working great

Thank you!