command line applications don't handle accented characters well

[James Wilkinson]

Colin Brace wrote:
> Paul, I have seen a raft of such problems, even without using ssh. This 
> afternoon, K3b refused to create a temp file for writing to a CD because 
> the title had an accent. This week, I copied a lot of MP3s from my old 
> OS/2 drive via CDs, and I have had a huge amount of work cleaning up the 
> accents in the filenames. In the terminal directory lists, such 
> characters showed up as small boxes; when copying on the command line, 
> they displayed as \###. In Nautilus, they are question marks with 
> "(Invalid Unicode character)" appended to the filename. I have been 
> trying to write some simple shell scripts to clean up filenames with 
> accents, but I am hampered and frustrated by the inconsistency in the 
> way (and/or) my ignorance of how they are handled by the system.
> 
> I have only been using Fedora a couple of months, and hence surprised to 
> read that accented characters were handled better in an early incarnation.

To be fair, what both Colin and Paul have been seeing is the fall-out
from using different character sets.

You're right that it's a pain in the neck. It's going to continue to be
a pain in the neck until everyone standardises around one character set.
And that's going to be some variant of Unicode, simply because there's
nothing else that's internationally acceptable to everyone.

At a quick look, iconv helps to convert file contents; see
http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=100934 for help on
transliterating filenames.

James.

-- 
E-mail address: james |    John's Inverse Law of Physics:
@westexe.demon.co.uk  | You do Physics -- you get inverted.