>>>> "C" == Christian Kirsch
<ck(a)held.mind.de> writes:
C> On Tue, Jun 23, 1998 at 07:16:17PM +0200, vroonhof(a)math.ethz.ch wrote:
> I just found that under an old XEmacs 20.4 (no Mule, Motif) I
have
> lying aroung I can reproduce this, but not under 21.1-Mule-XIMMotif.
> I set some breakpoints (unfortunately this not a debug compile) and I
> think I found it.
>
C> Yup. It works ok with xim=motif/mule, and doesn't work with
C> xim=xlib/no mule -- I could reproduce this with Linux here.
xim currently requires mule.
C> This is the only change I made, but IIRC it inflates the binary. And I
C> don't really see the point: Why should one use a Motif input method
C> and Mule, if X provides all the necessary functionality by itself?
It's historical. I added all of the input method handling stuff
specifically to support Japanese on Solaris with Motif. For them
there is no choice but to use mule. You are correct that non-mule
XEmacs should do a better job of supporting Latin-1 properly. That
will require a major code reorg that won't happen till 21.1 at the
earliest. In the meantime, you could always put Multi_key on a key of
its own using xmodmap. Then x-compose.el should work properly.
Or simply,
(define-key function-key-map [your key sequence here] 'compose-map)
should also work.
Martin