Hi,
at the beginning I must say that I'm completely lost about character encoding in systems...

I've created x86 appliance based on uClibc and busybox. I also compiled ntfs-3g for ntfs write support. There is almost nothing more but those basic programs.

I can mount ntfs filesystem with default locale (what is default?ANSI?) and all files without national characters (ANSII) are ok.

So when I mount a filesystem with national (Polish, German etc.)
characters I cannot see any files with those characters.

The question is what locale I should use to mount a NTFS? Can I use one universal locale to have my cp/mv/grep tools see every character typed on Earth? The question is more valid because I won't know what language of the system my embedded will mount.

I can say that I don't even need to (and suppose cannot) see those every character, I need them only copied right (I guess I would have to have all those fonts).

BTW creating locales in uClibc is really a pain.

Janusz

I must add that here http://www.ntfs-3g.org/support.html#locale I already found:

"If some of your softwares are not UTF8 ready then they will not show correctly the UTF8 filenames." - so may I expect that busybox tools will copy hashes instead of national characters which effectively change names of files when moved?

Janusz

In theory any UTF8 locale setting is ok but this is not confirmed.

OK, I did it.

I compiled all locales as well as some fonts to display characters which I really want (I don't care of displaying really everyone).

Ideed, I only needed track carefully what rc, keymaps, consolefont scripts in Gentoo does. I was really only confused about displaying in fb console, the rest was already fine.

Thank you.

More clues as to what you did? I need to do the exact same thing. The embedded systems just don't come with locale files. The FAQ entry certainly could use a little more information..