Re: Charsets on Debian

[Juliusz Chroboczek <jch@pps.jussieu.fr>]

Er... that's not strictly X related, what's a better mailing list?

SB> I recognize that on my US Debian system, characters 0-127 are
SB> simply ASCII, and that characters 160-255 are ISO-8859-1.
 
SB> What standard covers characters 128-159?

ISO 2022, or perhaps ISO 6429.  Or something else.

(The nice thing about standards etc.)

Codes 128 through 159 are control characters, not printable
characters.  A number of character sets, however, do use these
positions for printable characters; they are known as ``unstructured''
character sets in ISO 2022 terminology (most notable among those are
the Microsoft codepages and Big 5).

SB> Also, how do I change my charset?

Change your locale to one with a different charset, and pray.

SB> Wouldn't it be better if the Debian X packages used ISO-8859-15 by
SB> default?

There's no such thing as a ``default'' charset; there's the locale's
charset.  The default Western-European locales (fr_FR, de_DE, etc.)
use ISO 8859-1, but there are variant locales that use ISO 8859-15
(fr_FR@euro, etc.) and other variants that use UTF-8 (fr_FR.UTF8 etc.)

(Actually I'm lying: there is a default locale, called ``C'' or POSIX,
and hence a default charset.  Under current releases of libc, the
charset of the C locale is ASCII -- only codes 0 through 127.)

Making fr_FR etc. use ISO 8859-15 requires a simple manipulation as
root using the localedef utility.  However, doing so would break a
certain amount of software (notably Xlib).  So don't.

                                        Juliusz