From: (wrong string) ürgen Hestermann <juergen.hestermann_at_gmx.de>
Date: Sat, 28 Aug 2004 12:11:32 +0200

> > The Windows 95 and later kernel store all file names in (16 bit)
> > unicode, [...]
> Thanks for the clarification.
> Apparently, older versions of Linux implicitly translated the unicode
> to Latin1. The newer versions deal with unicode.
> When I mount the Archos wtih option utf=yes, I get unicode.
> Unfortunately, not all tools can handle this yet, so for the time
> being mount option iocharset=iso8859-1 will get me the old behaviour.

I don't know much about Linux but as far as I know under Windows
there are three possible ways for programs to interact with the
OS regarding file names:

1.) Console (text) mode programs normaly use a (system wide) 8 bit
codepage. The codepage can be changed but affects all console
programs.

2.) GUI programs usually use a different 8 bit code page. This leads
to problems if GUI and console programs store and read file names
with characters which are unique to one of the two codepages.

3.) All programs *can* use a unicode (16 bit) approach. This avoids
problems with exotic characters, but if they want to display a file
name on the screen a mapping to a 8 bit codepage is still needed.

So if characters from a file name on a FAT32 file system are displayed
as an (illegal) question mark then I suspect that the program which
displays this file name uses a codepage which misses this special character.

Jürgen Hestermann.



_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox
Received on 2004-08-28