[tlug] xemacs and japanese

["Stephen J. Turnbull" <stephen@example.com>]

Niels Kobschaetzki writes:

 > I'm running xemacs 21.4.21 (yp, I'm trying to learn it ;) ) and use in
 > Xfce scim-anthy as IME. According to google I shall use:
 > m-x set-input-method RET japanese-anthy

I don't think XEmacs groks SCIM, except maybe as an XIM input method.
If you have the XMODIFIERS variable set correctly, XEmacs/Mule will
use XIM automatically.

However, I do not recommend this, as XIM interferes with XEmacs's key
bindings.  I haven't looked at Emacs to see if this is true, but I
think Handa cheated RMS there as well.  AFAIK Europeans use Xkbd to
input 8-bit encodings directly, and Chinese et al use Quail.  So I
suspect our XIM implementation is as good as theirs.

 > But using m-x there is no match for set-input-method.

That means you don't have a Mule build; that XEmacs cannot process
Japanese at all.  (Well, it could, but you could write a doctoral
dissertation on making it do so.)  You need to get an XEmacs that has
Mule built in.  Also, I strongly recommend upgrading to 21.5.28 or
current Mercurial tip, because 21.4 has very poor support for Unicode,
and it won't get better.

On Linux, I use japanese-canna (built-in, not XIM) because that is
what I have always used on Linux.  On Mac OS X, I use japanese-anthy,
setting up in .xemacs/init.el as follows

(setq anthy-lisp-dir "/usr/local/anthy/lib/xemacs/site-lisp/anthy/"
      anthy-exec-dir "/usr/local/anthy/bin/")
(push anthy-lisp-dir load-path)
(push anthy-exec-dir exec-path)
(let ((load-path (list anthy-lisp-dir)))
  (load-library "leim-list"))
(load-library "anthy")

(depending on you have Anthy installed you may not need all of this).