[EMAIL PROTECTED] dixit:

>add mirports/sysutils/chkuterm

How about the following idea?

I integrate this into busybox (this will cost us only 1668 bytes in .text,
maybe even less if I can squeeze it, which I probably can), and add the
following to /etc/profile:
| eval $(chkuterm && echo set -U)

This will auto-enable mksh's UTF-8 mode if logging in from an Unicode ter-
minal (via ssh) and set $LC_CTYPE correctly for both Unicode and non-Unicode
cases and DOES NOT depend on locale support.

Benefit: Programmes that know that uClibc doesn't really do setlocale can
still know what kind of umlauts to expect and/or display, for instance,
the package/joe editor can. It won't mangle them then (one of our interns
has had issues with nano and vi when confronted with utf-8, them not knowing
that a character != a byte != a screen column).

As I already said (I'm better repeating), this does not depend on the
uClibc locale support, which weighs in several 100 KiB as I'm told by
wbx@, which I'll add separately and configurably.

The MirBSD boot floppy already uses chkuterm to start a latin1<->utf8
converter automatically on boot, because our userland applications
always expect utf-8 but our text mode terminal isn't quite there yet.


Comments?

bye,
//mirabile
-- 
I believe no one can invent an algorithm. One just happens to hit upon it
when God enlightens him. Or only God invents algorithms, we merely copy them.
If you don't believe in God, just consider God as Nature if you won't deny
existence.              -- Coywolf Qi Hunt
_______________________________________________
freewrt-developers mailing list
[email protected]
https://www.freewrt.org/lists/listinfo/freewrt-developers

Reply via email to