David,

> I need help because i'm kind of lost ;). I'm trying to configure my
> text console to be able to input text in utf-8. And is 
> working..., except
> for composed characters. After setting the console in Unicode 
> mode, with
> the script 'unicode_start', i'm able to input non us-ascii 
> characters in
> utf-8, but if i try to input a compose character, for example 
> vowels with
> acute tildes that i need to type spanish, i don't get the 
> composed characters
> on the screen. I'm able to read them, though, because cat'ting a file
> with Latin1 characters in utf-8 enconding, shows all the characters
> properly if the console is in unicode mode.

You are not lost at all.  The compose table stored in the kernel uses
8-bit values, and it's probably not going to be fixed.  (How would you
fix it?  Should we keep input characters at 8-bit and change the output
to a UTF-8 string?  UTF-8 input characters/strings too?  More
generalized input methods to support other scripts?)  Anything short of
a full generalized input method would probably be considered a "hack"
and would not get into the kernel.  :-(  But you're welcome to try to
prove me wrong!

You can try my patch p1_conv_8bit_to_uni.patch, found at
http://chris.heathens.co.nz/linux/utf8.html .  This patch keeps the
compose tables 8-bit, but uses a conversion table to convert it to
Unicode when the keyboard is in UNICODE mode.  Obviously this still only
supports characters from one 8-bit character set, but at least the
output is valid UTF-8!

Chris




--
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/linux-utf8/

Reply via email to