Re: problem about XIM

[Jungshik Shin <jshin mailaps org>]

On Sat, 5 Jul 2003, miaocheng wrote:

> I use red hat linux 8.0, the version of XFree86 is 4.0.3 and  the
> desktop manager is Gnome.
>   I have two problems about XIM . One is compatiblity between two
> architectural models of inputing method . one model, as you know , is
> Client/Server input model that is widely used to handle east Asian
> languages inputing, the other is inputing , by means of  keyboard map ,
> letters of other languages which have alphabet . In my pc , I have

 I have successfully used two input models under UTF-8 locales
(under RedHat 8) in tandem (acutally three if gtk2 input module is
counted) To enter Korean alphabetc letters in U+1100 block, I defined
a new xkb for Korean script and use it along with Ami(Korean XIM).
I can mix up not just  my new Xkb for Korean but also any other Xkbs
(French, Russian, Arabic ....) with Ami.

The cause of your problem is likely to be that you're under
zh_CN.gb2312 or zh_CN.Big5 locale (instead of UTF-8 locale) when
Chinput is launched. It's unfortunate that Red Hat 8 didn't come
with CJK UTF-8 locales by default, but you can make up your own
easily, which is what I have used (see
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=75829 )

>      Another question is How to introduce new symbolic names and their
> corresponding value into /usr/X11R6/include/X11/keysymdef.h file .
> Because I want to input Uyghur which letter is adopted to Unicode
> standard . but its letters' symbolic names is not listed in keysymdef.h

   There's absolutely no need to add new symbolic names to the X11
keysym table. You can  simply use 0x0100abcd for U+ABCD.  (that is,
add 0x01000000 to the unicode code point).  See
<http://www.cl.cam.ac.uk/~mgk25/unicode.html>.  Moreover, it takes
new symbolic names a while to be propagated to everyone's desktop
(not everyone is following the CVS of XF86 to say it mildly) while
any recent version of XFree86 supports the convention descibed here.


  Jungshik