[Uim] Key combinations still break input

[TOKUNAGA Hiroyuki]

Hi,

On Sun, 21 Mar 2004 14:23:28 +0100
David Oftedal <david at start.no> wrote:

> I'm using the newest version of uim now, and I'm still strugging with 
> some types of input causing previous input to be reversed.
> 
> The letters q, w, f, z, x and v aren't part any legal sequences in my 
> keymap, so that's why they were causing trouble. Hoever, the letters c
> 
> and y are legal input, and I had to add two empty entries to solve the
> 
> problem:
> 
> ((("y" ))(""))
> ((("c" ))(""))
> 
> Combinations like "yeobuseyo" or "yeobusecho" are working, but if you 
> try any illegal combinations like "yeobusezo" or "yeobuseqo", you'll
> get something like "yeobueso".
> 
> There's one problem I haven't been able to solve though: The letter j 
> causes the previous input to be reversed if preceded by a starting 
> consonant, an a, and an n.
> 
> So hanja = nahja
> ganja = nagja
> banja = nabja
> 
> Etc.
> 
> I haven't been able to find a solution for this yet. Any ideas?
> Perhaps I could make special entries for all consonants + a + n.

I'm not understanging all of your problems yet, but I confirmed that
following patch make possible to input "yeobuseyo". But I'm not
confident in this change doesn't cause another bugs. So please test this
patch. If it's OK, I'll commit.

Index: scm/rk.scm
===================================================================
--- scm/rk.scm  (revision 626)
+++ scm/rk.scm  (working copy)
@@ -123,7 +123,7 @@
           #t)
        (if (not tail-partial)
           (begin
-            (set! res (cadr (rk-lib-find-seq longest-head rule)))
+            (set! res (cadr (rk-lib-find-seq (reverse longest-head) rule)))
             (rk-context-set-seq!
              rkc
              (reverse


Regards,

-- 
TOKUNAGA Hiroyuki
http://kodou.net/