Font setup on ubuntu?

[Stephen J. Turnbull]

On Tue, May 1, 2012 at 2:18 AM, Raymond Toy <rtoy at google.com> wrote:

> Heh. Where did I get kr from?

ko is the language code, kr is the region (country) code.  You were
very definitely on the right track!  Just remember that even though in
practice the correlation between country and language is very high, we
still have to deal with the many edge cases.  If you need to add a
language or country to your set of usable locales, look for "ISO 636"
and "ISO 3166" tables.  On some distros, there's an "iso-codes"
package which provides this database, I believe.

>>  > (As an aside, I tried this on my mac using other fonts for Japanese,
>>  > Chinese, and Arabic, just for fun.  The display for Japanese and Chinese
>>  > look right, but I can't get arabic to display. I would like, someday, to
>>  > display all of the HELLO page.)

This will pretty much automatically happen when we have Unicode
inside.  Until then, it will be an annoying amount of work to get
right.

>> You and me both! Jeff Sparkes is our great hope.

Heh.  XEmacs-2000 (aka CHISE) does this fine since 1998. Ben didn't
like Tomo Morioka's work (for good reason such as innumerable
segfaults, unfortunately, so I couldn't change his mind about
including it).

> What is the blocker for these other examples?

The internal Mule encoding is based on ISO-2022, which uses a
different charset for every script.  So it needs to know how to map
charsets to fonts.  The database is incomplete.  I don't know why
Aidan's "JIT charset" doesn't work perfectly in this respect, but
unfortunately it doesn't.

The database could be brought up to date, but doing a complete job
isn't really worth the effort IMO.