Re: [Gnu-arch-users] How does arch/tla handle encodings?

[Christian von Kietzell]

On Fri, 27 Aug 2004 18:51:03 +0300, Marcus Sundman <address@hidden> wrote:
> > > 1) Does arch/tla keep track of the type/encoding of each file?
> >
> > Isn't this a little out of scope for arch?
> 
> No. If arch doesn't keep track of the encoding of files then you have to
> communicate this piece of metadata information through some other channel
> (e.g. per email). A text file *always* has an encoding associated with it.
> There is no such thing as "plain text".
> 
> > I, for one, use UTF-8 on purpose sometimes, as Pango expects UTF-8
> > strings. It would be very bad for these to get automatically transcoded to
> > some different encoding, even when I'm using ISO-8859-1 for example.
> 
> That is a good example of where the encoding of a file is part of that
> file's semantics. Of course no transcoding would be done for such files.
> (Some verification _could_ be done, though. Since all byte sequences aren't
> valid UTF-8 string it is particularly important for such files to be
> validated.)

The way I see it, encoding of text files is part of the policy of a
group of developers. Arch/tla is just a mechanism of keeping track of
said developers' work.

  Chris