Subject: Re: fs transcoding, was Re: Unicode support in iso9660.
To: Jason Thorpe <email@example.com>
From: Pavel Cahyna <firstname.lastname@example.org>
Date: 11/22/2004 17:28:20
> On Nov 22, 2004, at 8:06 AM, Pavel Cahyna wrote:
> >On Mon, 22 Nov 2004 16:01:37 +0000, Jason Thorpe wrote:
> >>Honestly, I think we need to reserve a chunk of space in the FFS
> >>superblock to specify the encoding that the file system is using.
> >What would be this good for?
> Making it possible for the file system to self-describe. The "what
I still don't see what would be good for, because all applications would
have to use an API to determine the encoding and application authors
probably won't care (until a compatible API exists on Linux).
Also, if the applications already must be modified to use such API, it is
simpler to just modify them to use UTF-8 always, just as Glib/GTK+ does.
> we've got" part of the equation. Legacy file systems that don't
> specify this can continue their free-for-all wrt. encoding that makes
> this problem difficult to solve. New file systems will have a
> reasonable default encoding (UTF-8), and it can be set at newfs time to
> an alternate encoding or "none" (i.e. the legacy case).
And how the encoding information would be used? It would be just passed to
userland with an API (fcntl?) or would it enforce that all newly created
filenames are in the proper encoding?