bug#47767: 28.0.50; japanese-hankaku unnatural or misconversion results

bug#47767: 28.0.50; japanese-hankaku unnatural or misconversion results

[Kazuhiro Ito]

1. japanese-hankaku[-region] with ASCII-ONLY option converts Japanese
panctuation characters and prolonged sound mark to ascii.

(japanese-hankaku "ケーキ、ドーナツ。" t)
-> "ケ-キ,ド-ナツ."

the result is very unnatural because "ー", "、" and "。" are normally
used among Japanese characters which are not converted to hankaku in
this case.  I hope they are kept as is (the result should be
"ケーキ、ドーナツ。").


2. japanese-hankaku[-region] without ASCII-ONLY option and
japanese-zenkaku[-region] fails to convert Latin punctuations.

(japanese-zenkaku "A, B, C.")
-> "Ａ、　Ｂ、　Ｃ。"

(japanese-hankaku "Ａ，Ｂ，Ｃ．")
-> "A､B､C｡"

They should be "Ａ，　Ｂ，　Ｃ．" and "A,B,C." respectively.


Below patch fixes problems.


diff --git a/lisp/language/japan-util.el b/lisp/language/japan-util.el
index 3f1fb2b749..8b80599d99 100644
--- a/lisp/language/japan-util.el
+++ b/lisp/language/japan-util.el
@@ -96,9 +96,9 @@ japanese-kana-table
 	  (put-char-code-property jisx0201 'jisx0208 katakana)))))
 
 (defconst japanese-symbol-table
-  '((?\　 ?\ ) (?， ?, ?､) (?． ?. ?｡) (?、 ?, ?､) (?。 ?. ?｡) (?・ nil ?･)
+  '((?\　 ?\ ) (?， ?,) (?． ?.) (?、 nil ?､) (?。 nil ?｡) (?・ nil ?･)
     (?： ?:) (?； ?\;) (?？ ??) (?！ ?!) (?゛ nil ?ﾞ) (?゜ nil ?ﾟ)
-    (?´ ?') (?｀ ?`) (?＾ ?^) (?＿ ?_) (?ー ?- ?ｰ) (?— ?-) (?‐ ?-)
+    (?´ ?') (?｀ ?`) (?＾ ?^) (?＿ ?_) (?ー nil ?ｰ) (?— ?-) (?‐ ?-)
     (?／ ?/) (?＼ ?\\) (?〜 ?~)  (?｜ ?|) (?‘ ?`) (?’ ?') (?“ ?\") (?” ?\")
     (?\（ ?\() (?\） ?\)) (?\［ ?\[) (?\］ ?\]) (?\｛ ?{) (?\｝ ?})
     (?〈 ?<) (?〉 ?>) (?\「 nil ?\｢) (?\」 nil ?\｣)


-- 
Kazuhiro Ito

bug#47767: 28.0.50; japanese-hankaku unnatural or misconversion results

[Lars Ingebrigtsen]

Kazuhiro Ito <kzhr@d1.dion.ne.jp> writes:

> 1. japanese-hankaku[-region] with ASCII-ONLY option converts Japanese
> panctuation characters and prolonged sound mark to ascii.
>
> (japanese-hankaku "ケーキ、ドーナツ。" t)
> -> "ケ-キ,ド-ナツ."
>
> the result is very unnatural because "ー", "、" and "。" are normally
> used among Japanese characters which are not converted to hankaku in
> this case.  I hope they are kept as is (the result should be
> "ケーキ、ドーナツ。").
>
> 2. japanese-hankaku[-region] without ASCII-ONLY option and
> japanese-zenkaku[-region] fails to convert Latin punctuations.
>
> (japanese-zenkaku "A, B, C.")
> -> "Ａ、　Ｂ、　Ｃ。"
>
> (japanese-hankaku "Ａ，Ｂ，Ｃ．")
> -> "A､B､C｡"
>
> They should be "Ａ，　Ｂ，　Ｃ．" and "A,B,C." respectively.
>
> Below patch fixes problems.

Thanks; applied to Emacs 28.

I don't know Japanese, but since nobody else has piped up about this
patch in three weeks, I'm applying it.  If other Japanese-speaking
people disagree with this patch, we can revert it.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no