UTF-8 and filenames

[Toshio Kuratomi]

On Wed, 2007-03-14 at 12:16 -0400, Frank Ch. Eigler wrote:
> jkeating wrote:
> 
> > From the koji dev:
> >
> > [why utf8?] because otherwise you can't really do anything sane with
> > the data.  If it's a mix of utf-8, iso-8859-1, WIN-1259, etc. you
> > have no way of knowing how to display it.
> 
> At worst, a tool can do what "ls -b" does: quoting.  At best, it can
> allow a user to specify the character set/encoding for
> directories/file names.
> 
Quoting is a possibility.  Displaying my ignorance, will there ever be a
time when an encoding maps a character to a value that is valid UTF-8?
If so, then we'll end up with garbage filenames rather than quoting for
those encodings which seems suboptimal.

For specifying the character set, I'd argue that the spec file is the
correct level for that to occur.  If so, why not have the packager
change the filename to UTF-8 instead?

-Toshio
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://listman.redhat.com/archives/fedora-maintainers/attachments/20070314/61b9a062/attachment.sig>