[Gnu-arch-users] Re: tla-update-ids lacks support for filenames with spaces

[Jan Hudec]

On Mon, May 24, 2004 at 18:24:29 +0900, Miles Bader wrote:
> Jan Hudec <address@hidden> writes:
> > Well, the language does not need to know it's using unicode. But you
> > need to have charset conversion function that can do unicode available.
> 
> All the ways I'm actually going to _use_ the filename involve giving it
> to some other program in the form it appeared in the first place on disk
> -- if the user's using unicode, then probably as a UTF-8 string, but
> possibly as some other weird encoding that has nothing to do with
> unicode.  It's really annoying to be required to have fancy conversions
> functions just to get back what was there in the first place (possibly
> incorrectly, if the original wasn't actually UTF-8).
> 
> Really all I want is a sort of bare-minimum escaping, that escapes
> newline and space, and I guess that's it (I suppose \0 is a lost cause
> though).

There is nothing in utf-8 multibyte sequences that should cause problems
anywhere, so they don't need to be escaped...

> How about --slightly-escaped?

I believe when the unicode support comes, we will actualy need two
versions of escaping. The escaped unicode name, that goes in the archive
and the minimaly escaped filename as it lives in the filesystem.

-------------------------------------------------------------------------------
                                                 Jan 'Bulb' Hudec 
<address@hidden>

signature.asc
Description: Digital signature