[gentoo-user] Umlaut trouble in filenames

[Alex Schuster <wonko@wonkology.org>]

Hi there!

I have a problem with umlauts in file names. They are working fine when I 
create a file, but I have lots of files I moved from my former system that 
show umlauts just as a question mark. I think I transferred them via NFS.

I have LANG=de_DE.utf8@euro set now, before it was unset. Those umlaut files 
refuse to play in applications like amarok2. I can rename them with mv on 
the command line, making use of tab completion, but it's annoying.

So, does anyone have a trick to automate this?

I tried using tr. Using od -t x1 I see that what has to be 'ü' comes out as 
0374 (octal). But "echo ü | od -t o1" gives 0303 0274 which tells me that in 
unicode the umlauts are coded as two bytes. But I think I cannot use tr to 
replace one character by a sequence of two characters.
And with sed, I don't know how to express the source string, it does not 
seem to be capable of octal notation.

The only way I see is to use tr to replace the original ü with some special 
character like €, and then use sed to replace € by ü. And repeat this for 
[äöüÄÖÜß] and all the accented characters which I do not know how to type 
here with nodeadkeys option set in xorg.conf (I'd be interested in how to do 
this, too).

Do you have simpler ideas?

	Wonko