Noriyuki Soda wrote:
> > Why not pretend, for the purposes of creating new files, that the
> > "codepage-dependent-codeset" is UCS-2 too?  I.e. require userland to
> > always supply new filenames in UCS-2?
> 
> Since it makes the filesystem unusable from Windows.
> Note that only one codepage-dependent-codeset can be used per
> system on Windows 9x. (not per file-system, but all filesystems on
> that system.)

The point IMHO is that the kernel would take care about recoding
the filename from UCS-2 or UTF-8 into the codepage-dependant-codeset
used natively by the filesystem.

Jaromir
-- 
Jaromir Dolecek <jdolecek@NetBSD.org>      http://www.ics.muni.cz/~dolecek/
"The only way how to get rid temptation is to yield to it." -- Oscar Wilde