1997-02-20 07:02:49 +00:00
|
|
|
|
;;; characters.el --- set syntax and category for multibyte characters
|
|
|
|
|
|
|
2007-01-21 03:53:13 +00:00
|
|
|
|
;; Copyright (C) 1997, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007
|
2005-05-13 06:03:46 +00:00
|
|
|
|
;; Free Software Foundation, Inc.
|
2006-12-13 01:13:58 +00:00
|
|
|
|
;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
|
2007-01-21 03:53:13 +00:00
|
|
|
|
;; 2005, 2006, 2007
|
2005-05-13 06:03:46 +00:00
|
|
|
|
;; National Institute of Advanced Industrial Science and Technology (AIST)
|
|
|
|
|
|
;; Registration Number H14PRO021
|
2003-09-08 12:53:41 +00:00
|
|
|
|
;; Copyright (C) 2003
|
2002-03-01 02:06:19 +00:00
|
|
|
|
;; National Institute of Advanced Industrial Science and Technology (AIST)
|
|
|
|
|
|
;; Registration Number H13PRO009
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; Keywords: multibyte character, character set, syntax, category
|
|
|
|
|
|
|
|
|
|
|
|
;; This file is part of GNU Emacs.
|
|
|
|
|
|
|
|
|
|
|
|
;; GNU Emacs is free software; you can redistribute it and/or modify
|
|
|
|
|
|
;; it under the terms of the GNU General Public License as published by
|
|
|
|
|
|
;; the Free Software Foundation; either version 2, or (at your option)
|
|
|
|
|
|
;; any later version.
|
|
|
|
|
|
|
|
|
|
|
|
;; GNU Emacs is distributed in the hope that it will be useful,
|
|
|
|
|
|
;; but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
|
;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
|
;; GNU General Public License for more details.
|
|
|
|
|
|
|
|
|
|
|
|
;; You should have received a copy of the GNU General Public License
|
1997-02-23 09:20:52 +00:00
|
|
|
|
;; along with GNU Emacs; see the file COPYING. If not, write to the
|
2005-07-04 17:55:18 +00:00
|
|
|
|
;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
|
|
|
|
|
|
;; Boston, MA 02110-1301, USA.
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;;; Commentary:
|
|
|
|
|
|
|
2001-07-15 19:53:53 +00:00
|
|
|
|
;;; Code:
|
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;;; Predefined categories.
|
|
|
|
|
|
|
|
|
|
|
|
;; For each character set.
|
|
|
|
|
|
|
2007-02-13 02:47:32 +00:00
|
|
|
|
(define-category ?a "ASCII graphic characters 32-126 (ISO646 IRV:1983[4/0])")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(define-category ?l "Latin")
|
|
|
|
|
|
(define-category ?t "Thai")
|
|
|
|
|
|
(define-category ?g "Greek")
|
|
|
|
|
|
(define-category ?b "Arabic")
|
|
|
|
|
|
(define-category ?w "Hebrew")
|
|
|
|
|
|
(define-category ?y "Cyrillic")
|
|
|
|
|
|
(define-category ?k "Japanese katakana")
|
|
|
|
|
|
(define-category ?r "Japanese roman")
|
|
|
|
|
|
(define-category ?c "Chinese")
|
|
|
|
|
|
(define-category ?j "Japanese")
|
|
|
|
|
|
(define-category ?h "Korean")
|
|
|
|
|
|
(define-category ?e "Ethiopic (Ge'ez)")
|
|
|
|
|
|
(define-category ?v "Vietnamese")
|
|
|
|
|
|
(define-category ?i "Indian")
|
1997-02-26 12:03:35 +00:00
|
|
|
|
(define-category ?o "Lao")
|
1997-05-12 05:30:27 +00:00
|
|
|
|
(define-category ?q "Tibetan")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; For each group (row) of 2-byte character sets.
|
|
|
|
|
|
|
1998-06-20 22:06:30 +00:00
|
|
|
|
(define-category ?A "Alpha-numeric characters of 2-byte character sets")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(define-category ?C "Chinese (Han) characters of 2-byte character sets")
|
1998-06-20 22:06:30 +00:00
|
|
|
|
(define-category ?G "Greek characters of 2-byte character sets")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(define-category ?H "Japanese Hiragana characters of 2-byte character sets")
|
|
|
|
|
|
(define-category ?K "Japanese Katakana characters of 2-byte character sets")
|
|
|
|
|
|
(define-category ?N "Korean Hangul characters of 2-byte character sets")
|
1998-06-20 22:06:30 +00:00
|
|
|
|
(define-category ?Y "Cyrillic characters of 2-byte character sets")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(define-category ?I "Indian Glyphs")
|
|
|
|
|
|
|
|
|
|
|
|
;; For phonetic classifications.
|
|
|
|
|
|
|
|
|
|
|
|
(define-category ?0 "consonant")
|
1997-07-02 12:59:43 +00:00
|
|
|
|
(define-category ?1 "base (independent) vowel")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(define-category ?2 "upper diacritical mark (including upper vowel)")
|
|
|
|
|
|
(define-category ?3 "lower diacritical mark (including lower vowel)")
|
|
|
|
|
|
(define-category ?4 "tone mark")
|
1997-07-02 12:59:43 +00:00
|
|
|
|
(define-category ?5 "symbol")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(define-category ?6 "digit")
|
|
|
|
|
|
(define-category ?7 "vowel-modifying diacritical mark")
|
1997-02-26 12:03:35 +00:00
|
|
|
|
(define-category ?8 "vowel-signs")
|
|
|
|
|
|
(define-category ?9 "semivowel lower")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; For filling.
|
|
|
|
|
|
(define-category ?| "While filling, we can break a line at this character.")
|
|
|
|
|
|
|
1997-11-11 03:24:18 +00:00
|
|
|
|
;; For indentation calculation.
|
2003-02-13 15:56:09 +00:00
|
|
|
|
(define-category ?\s
|
1998-01-22 01:42:20 +00:00
|
|
|
|
"This character counts as a space for indentation purposes.")
|
1997-11-11 03:24:18 +00:00
|
|
|
|
|
1998-06-20 22:06:30 +00:00
|
|
|
|
;; Keep the following for `kinsoku' processing. See comments in
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;; kinsoku.el.
|
|
|
|
|
|
(define-category ?> "A character which can't be placed at beginning of line.")
|
|
|
|
|
|
(define-category ?< "A character which can't be placed at end of line.")
|
|
|
|
|
|
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; Combining
|
|
|
|
|
|
(define-category ?^ "Combining diacritic or mark")
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;;; Setting syntax and category.
|
|
|
|
|
|
|
|
|
|
|
|
;; ASCII
|
|
|
|
|
|
|
2002-03-05 00:06:48 +00:00
|
|
|
|
;; All ASCII characters have the category `a' (ASCII) and `l' (Latin).
|
|
|
|
|
|
(modify-category-entry '(32 . 127) ?a)
|
|
|
|
|
|
(modify-category-entry '(32 . 127) ?l)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-12-05 14:28:21 +00:00
|
|
|
|
;; Deal with the CJK charsets first. Since the syntax of blocks is
|
|
|
|
|
|
;; defined per charset, and the charsets may contain e.g. Latin
|
|
|
|
|
|
;; characters, we end up with the wrong syntax definitions if we're
|
|
|
|
|
|
;; not careful.
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-06-12 00:16:54 +00:00
|
|
|
|
;; Chinese characters (Unicode)
|
2004-04-20 08:32:39 +00:00
|
|
|
|
(modify-category-entry '(#x2E80 . #x312F) ?|)
|
|
|
|
|
|
(modify-category-entry '(#x3190 . #x33FF) ?|)
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(modify-category-entry '(#x3400 . #x9FAF) ?C)
|
|
|
|
|
|
(modify-category-entry '(#x3400 . #x9FAF) ?c)
|
|
|
|
|
|
(modify-category-entry '(#x3400 . #x9FAF) ?|)
|
|
|
|
|
|
(modify-category-entry '(#xF900 . #xFAFF) ?C)
|
|
|
|
|
|
(modify-category-entry '(#xF900 . #xFAFF) ?c)
|
|
|
|
|
|
(modify-category-entry '(#xF900 . #xFAFF) ?|)
|
2003-12-29 07:50:51 +00:00
|
|
|
|
(modify-category-entry '(#x20000 . #x2AFFF) ?|)
|
|
|
|
|
|
(modify-category-entry '(#x2F800 . #x2FFFF) ?|)
|
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; Chinese character set (GB2312)
|
|
|
|
|
|
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'chinese-gb2312 "_" #x2121 #x217E)
|
|
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'chinese-gb2312 "_" #x2221 #x227E)
|
|
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'chinese-gb2312 "_" #x2921 #x297E)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-06-06 11:34:24 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?c)
|
2002-08-19 12:13:45 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?A #x2330 #x2339)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?A #x2341 #x235A)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?A #x2361 #x237A)
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?H #x2421 #x247E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?K #x2521 #x257E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?G #x2621 #x267E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?Y #x2721 #x277E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'chinese-gb2312 ?C #x3021 #x7E7E)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; Chinese character set (BIG5)
|
|
|
|
|
|
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'big5 ?c)
|
2002-08-19 12:13:45 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'big5 ?C #xA259 #xA25F)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'big5 ?C #xA440 #xC67E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'big5 ?C #xC940 #xF9DF)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; Chinese character set (CNS11643)
|
|
|
|
|
|
|
2002-06-06 11:34:24 +00:00
|
|
|
|
(dolist (c '(chinese-cns11643-1 chinese-cns11643-2 chinese-cns11643-3
|
|
|
|
|
|
chinese-cns11643-4 chinese-cns11643-5 chinese-cns11643-6
|
|
|
|
|
|
chinese-cns11643-7))
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry c ?c)
|
2002-08-19 12:13:45 +00:00
|
|
|
|
(if (eq c 'chinese-cns11643-1)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry c ?C #x4421 #x7E7E)
|
2003-12-29 07:50:51 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry c ?C)))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2003-09-08 12:53:41 +00:00
|
|
|
|
;; Japanese character set (JISX0201, JISX0208, JISX0212, JISX0213)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'katakana-jisx0201 ?k)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'latin-jisx0201 ?r)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(dolist (l '(katakana-jisx0201 japanese-jisx0208 japanese-jisx0212
|
|
|
|
|
|
japanese-jisx0213-1 japanese-jisx0213-2))
|
2003-12-29 07:50:51 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry l ?j))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; Unicode equivalents of JISX0201-kana
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(let ((range '(#xff61 . #xff9f)))
|
|
|
|
|
|
(modify-category-entry range ?k)
|
|
|
|
|
|
(modify-category-entry range ?j)
|
|
|
|
|
|
(modify-category-entry range ?\|))
|
2001-12-07 14:47:37 +00:00
|
|
|
|
|
|
|
|
|
|
;; Katakana block
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(let ((range '(#x30a0 . #x30ff)))
|
|
|
|
|
|
;; ?K is double width, ?k isn't specified
|
|
|
|
|
|
(modify-category-entry range ?K)
|
|
|
|
|
|
(modify-category-entry range ?\|))
|
2001-12-07 14:47:37 +00:00
|
|
|
|
|
|
|
|
|
|
;; Hiragana block
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(let ((range '(#x3040 . #x309d)))
|
2002-06-12 00:16:54 +00:00
|
|
|
|
;; ?H is actually defined to be double width
|
|
|
|
|
|
;;(modify-category-entry range ?H)
|
2003-12-29 07:50:51 +00:00
|
|
|
|
(modify-category-entry range ?\|)
|
2002-06-12 00:16:54 +00:00
|
|
|
|
)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;; JISX0208
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'japanese-jisx0208 "_" #x2121 #x227E)
|
|
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'japanese-jisx0208 "_" #x2821 #x287E)
|
|
|
|
|
|
(let ((chars '(?ー ?゛ ?゜ ?ヽ ?ヾ ?ゝ ?ゞ ?〃 ?仝 ?々 ?〆 ?〇)))
|
2002-05-14 23:15:58 +00:00
|
|
|
|
(dolist (elt chars)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-syntax-entry (car chars) "w")))
|
2002-06-12 00:16:54 +00:00
|
|
|
|
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'japanese-jisx0208 ?A #x2321 #x237E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'japanese-jisx0208 ?H #x2421 #x247E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'japanese-jisx0208 ?K #x2521 #x257E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'japanese-jisx0208 ?G #x2621 #x267E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'japanese-jisx0208 ?Y #x2721 #x277E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'japanese-jisx0208 ?C #x3021 #x7E7E)
|
|
|
|
|
|
(modify-category-entry ?ー ?K)
|
|
|
|
|
|
(let ((chars '(?゛ ?゜)))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(while chars
|
|
|
|
|
|
(modify-category-entry (car chars) ?K)
|
|
|
|
|
|
(modify-category-entry (car chars) ?H)
|
|
|
|
|
|
(setq chars (cdr chars))))
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(let ((chars '(?ヽ ?ヾ ?ゝ ?ゞ ?〃 ?仝 ?々 ?〆 ?〇)))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(while chars
|
|
|
|
|
|
(modify-category-entry (car chars) ?C)
|
|
|
|
|
|
(setq chars (cdr chars))))
|
|
|
|
|
|
|
|
|
|
|
|
;; JISX0212
|
|
|
|
|
|
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'japanese-jisx0212 "_" #x2121 #x237E)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
;; JISX0201-Kana
|
2002-06-06 11:34:24 +00:00
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(let ((chars '(?｡ ?､ ?･)))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(while chars
|
|
|
|
|
|
(modify-syntax-entry (car chars) ".")
|
|
|
|
|
|
(setq chars (cdr chars))))
|
|
|
|
|
|
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(modify-syntax-entry ?\｢ "(｣")
|
|
|
|
|
|
(modify-syntax-entry ?\｣ "(｢")
|
2001-03-09 05:59:56 +00:00
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;; Korean character set (KSC5601)
|
|
|
|
|
|
|
2002-06-06 11:34:24 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?h)
|
2002-06-12 00:16:54 +00:00
|
|
|
|
|
|
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'korean-ksc5601 "_" #x2121 #x227E)
|
2002-09-23 14:03:06 +00:00
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'korean-ksc5601 "_" #x2621 #x277E)
|
|
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'korean-ksc5601 "_" #x2830 #x287E)
|
|
|
|
|
|
(map-charset-chars #'modify-syntax-entry 'korean-ksc5601 "_" #x2930 #x297E)
|
2002-08-19 12:13:45 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?A #x2330 #x2339)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?A #x2341 #x235A)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?A #x2361 #x237A)
|
2002-06-12 00:16:54 +00:00
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?G #x2521 #x257E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?H #x2A21 #x2A7E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?K #x2B21 #x2B7E)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'korean-ksc5601 ?Y #x2C21 #x2C7E)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2002-12-05 14:28:21 +00:00
|
|
|
|
;; These are in more than one charset.
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(let ((parens (concat "〈〉《》「」『』【】〔〕〖〗〘〙〚〛"
|
|
|
|
|
|
"︵︶︷︸︹︺︻︼︽︾︿﹀﹁﹂﹃﹄"
|
|
|
|
|
|
"（）［］｛｝"))
|
|
|
|
|
|
open close)
|
|
|
|
|
|
(dotimes (i (/ (length parens) 2))
|
|
|
|
|
|
(setq open (aref parens (* i 2))
|
|
|
|
|
|
close (aref parens (1+ (* i 2))))
|
|
|
|
|
|
(modify-syntax-entry open (format "(%c" close))
|
|
|
|
|
|
(modify-syntax-entry close (format ")%c" open))))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2002-12-05 14:28:21 +00:00
|
|
|
|
;; Arabic character set
|
1997-02-26 12:03:35 +00:00
|
|
|
|
|
2002-12-05 14:28:21 +00:00
|
|
|
|
(let ((charsets '(arabic-iso8859-6
|
|
|
|
|
|
arabic-digit
|
|
|
|
|
|
arabic-1-column
|
|
|
|
|
|
arabic-2-column)))
|
|
|
|
|
|
(while charsets
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry (car charsets) ?b)
|
|
|
|
|
|
(setq charsets (cdr charsets))))
|
|
|
|
|
|
(modify-category-entry '(#x600 . #x6ff) ?b)
|
|
|
|
|
|
(modify-category-entry '(#xfb50 . #xfdff) ?b)
|
|
|
|
|
|
(modify-category-entry '(#xfe70 . #xfefe) ?b)
|
1997-02-26 12:03:35 +00:00
|
|
|
|
|
2002-12-05 14:28:21 +00:00
|
|
|
|
;; Cyrillic character set (ISO-8859-5)
|
|
|
|
|
|
|
|
|
|
|
|
(modify-syntax-entry ?№ ".")
|
|
|
|
|
|
|
|
|
|
|
|
;; Ethiopic character set
|
|
|
|
|
|
|
2006-07-20 13:01:24 +00:00
|
|
|
|
(modify-category-entry '(#x1200 . #x1399) ?e)
|
|
|
|
|
|
(modify-category-entry '(#x2d80 . #x2dde) ?e)
|
2002-12-05 14:28:21 +00:00
|
|
|
|
(let ((chars '(?፡ ?። ?፣ ?፤ ?፥ ?፦ ?፧ ?፨ ?<3F><EFBFBD><EFBFBD><EFBFBD> ?<3F><EFBFBD><EFBFBD><EFBFBD> ?<3F><EFBFBD><EFBFBD><EFBFBD> ?<3F><EFBFBD><EFBFBD><EFBFBD> ?<3F><EFBFBD><EFBFBD><EFBFBD> ?<3F><EFBFBD><EFBFBD><EFBFBD>)))
|
|
|
|
|
|
(while chars
|
|
|
|
|
|
(modify-syntax-entry (car chars) ".")
|
|
|
|
|
|
(setq chars (cdr chars))))
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'ethiopic ?e)
|
|
|
|
|
|
|
|
|
|
|
|
;; Hebrew character set (ISO-8859-8)
|
|
|
|
|
|
|
|
|
|
|
|
(modify-syntax-entry #x5be ".") ; MAQAF
|
|
|
|
|
|
(modify-syntax-entry #x5c0 ".") ; PASEQ
|
|
|
|
|
|
(modify-syntax-entry #x5c3 ".") ; SOF PASUQ
|
|
|
|
|
|
(modify-syntax-entry #x5f3 ".") ; GERESH
|
|
|
|
|
|
(modify-syntax-entry #x5f4 ".") ; GERSHAYIM
|
|
|
|
|
|
|
|
|
|
|
|
;; Indian character set (IS 13194 and other Emacs original Indian charsets)
|
|
|
|
|
|
|
|
|
|
|
|
(modify-category-entry '(#x901 . #x970) ?i)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'indian-is13194 ?i)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'indian-2-column ?i)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
1997-02-26 12:03:35 +00:00
|
|
|
|
;; Lao character set
|
|
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#xe80 . #xeff) ?o)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'lao ?o)
|
1997-02-26 12:03:35 +00:00
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(let ((deflist '(("ກ-ຮ" "w" ?0) ; consonant
|
2002-05-09 06:44:06 +00:00
|
|
|
|
("ະາຳຽເ-ໄ" "w" ?1) ; vowel base
|
|
|
|
|
|
("ັິ-ືົໍ" "w" ?2) ; vowel upper
|
|
|
|
|
|
("ຸູ" "w" ?3) ; vowel lower
|
2003-09-08 12:53:41 +00:00
|
|
|
|
("່-໋" "w" ?4) ; tone mark
|
2002-05-09 06:44:06 +00:00
|
|
|
|
("ຼຽ" "w" ?9) ; semivowel lower
|
|
|
|
|
|
("໐-໙" "w" ?6) ; digit
|
|
|
|
|
|
("ຯໆ" "_" ?5) ; symbol
|
1997-02-26 12:03:35 +00:00
|
|
|
|
))
|
|
|
|
|
|
elm chars len syntax category to ch i)
|
|
|
|
|
|
(while deflist
|
|
|
|
|
|
(setq elm (car deflist))
|
|
|
|
|
|
(setq chars (car elm)
|
|
|
|
|
|
len (length chars)
|
|
|
|
|
|
syntax (nth 1 elm)
|
|
|
|
|
|
category (nth 2 elm)
|
|
|
|
|
|
i 0)
|
|
|
|
|
|
(while (< i len)
|
|
|
|
|
|
(if (= (aref chars i) ?-)
|
|
|
|
|
|
(setq i (1+ i)
|
1998-02-04 11:25:47 +00:00
|
|
|
|
to (aref chars i))
|
|
|
|
|
|
(setq ch (aref chars i)
|
1997-02-26 12:03:35 +00:00
|
|
|
|
to ch))
|
|
|
|
|
|
(while (<= ch to)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(unless (string-equal syntax "w")
|
|
|
|
|
|
(modify-syntax-entry ch syntax))
|
1997-02-26 12:03:35 +00:00
|
|
|
|
(modify-category-entry ch category)
|
|
|
|
|
|
(setq ch (1+ ch)))
|
1998-02-04 11:25:47 +00:00
|
|
|
|
(setq i (1+ i)))
|
1997-02-26 12:03:35 +00:00
|
|
|
|
(setq deflist (cdr deflist))))
|
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
;; Thai character set (TIS620)
|
|
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#xe00 . #xe7f) ?t)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'thai-tis620 ?t)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
|
|
|
|
|
(let ((deflist '(;; chars syntax category
|
2002-05-09 06:44:06 +00:00
|
|
|
|
("ก-รลว-ฮ" "w" ?0) ; consonant
|
|
|
|
|
|
("ฤฦะาำเ-ๅ" "w" ?1) ; vowel base
|
|
|
|
|
|
("ัิ-ื็๎" "w" ?2) ; vowel upper
|
|
|
|
|
|
("ุ-ฺ" "w" ?3) ; vowel lower
|
2003-09-08 12:53:41 +00:00
|
|
|
|
("่-ํ" "w" ?4) ; tone mark
|
2002-05-09 06:44:06 +00:00
|
|
|
|
("๐-๙" "w" ?6) ; digit
|
|
|
|
|
|
("ฯๆ฿๏๚๛" "_" ?5) ; symbol
|
1997-02-20 07:02:49 +00:00
|
|
|
|
))
|
|
|
|
|
|
elm chars len syntax category to ch i)
|
1997-05-12 05:30:27 +00:00
|
|
|
|
(while deflist
|
|
|
|
|
|
(setq elm (car deflist))
|
|
|
|
|
|
(setq chars (car elm)
|
|
|
|
|
|
len (length chars)
|
|
|
|
|
|
syntax (nth 1 elm)
|
|
|
|
|
|
category (nth 2 elm)
|
|
|
|
|
|
i 0)
|
|
|
|
|
|
(while (< i len)
|
|
|
|
|
|
(if (= (aref chars i) ?-)
|
|
|
|
|
|
(setq i (1+ i)
|
1998-02-04 11:25:47 +00:00
|
|
|
|
to (aref chars i))
|
|
|
|
|
|
(setq ch (aref chars i)
|
1997-05-12 05:30:27 +00:00
|
|
|
|
to ch))
|
|
|
|
|
|
(while (<= ch to)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(unless (string-equal syntax "w")
|
|
|
|
|
|
(modify-syntax-entry ch syntax))
|
1997-05-12 05:30:27 +00:00
|
|
|
|
(modify-category-entry ch category)
|
|
|
|
|
|
(setq ch (1+ ch)))
|
1998-02-04 11:25:47 +00:00
|
|
|
|
(setq i (1+ i)))
|
1997-05-12 05:30:27 +00:00
|
|
|
|
(setq deflist (cdr deflist))))
|
|
|
|
|
|
|
|
|
|
|
|
;; Tibetan character set
|
|
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#xf00 . #xfff) ?q)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'tibetan ?q)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'tibetan-1-column ?q)
|
1997-05-12 05:30:27 +00:00
|
|
|
|
|
|
|
|
|
|
(let ((deflist '(;; chars syntax category
|
2002-06-02 20:32:26 +00:00
|
|
|
|
("ཀ-ཀྵཪ" "w" ?0) ; consonant
|
|
|
|
|
|
("ྐ-ྐྵྺྻྼ<E0BEBB><E0BEBC><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>" "w" ?0) ;
|
2002-05-09 06:44:06 +00:00
|
|
|
|
("<EFBFBD><EFBFBD><EFBFBD><EFBFBD>-<2D><><EFBFBD><EFBFBD>" "w" ?0) ;
|
|
|
|
|
|
("<EFBFBD><EFBFBD><EFBFBD><EFBFBD>-<2D><><EFBFBD><EFBFBD>" "w" ?0) ;
|
2002-06-02 20:32:26 +00:00
|
|
|
|
("ིེཻོཽྀ" "w" ?2) ; upper vowel
|
|
|
|
|
|
("ཾྂྃ྆྇ྈྉྊྋ" "w" ?2) ; upper modifier
|
|
|
|
|
|
("༙<EFBFBD><EFBFBD><EFBFBD><EFBFBD>྄ཱུ༵༷" "w" ?3) ; lowel vowel/modifier
|
2003-09-08 12:53:41 +00:00
|
|
|
|
("" "w" ?3) ; invisible vowel a
|
2002-06-02 20:32:26 +00:00
|
|
|
|
("༠-༩༪-༳" "w" ?6) ; digit
|
|
|
|
|
|
("་།-༒༔ཿ" "." ?|) ; line-break char
|
|
|
|
|
|
("་།༏༐༑༔ཿ" "." ?|) ;
|
|
|
|
|
|
("༈་།-༒༔ཿ༽༴" "." ?>) ; prohibition
|
|
|
|
|
|
("་།༏༐༑༔ཿ" "." ?>) ;
|
|
|
|
|
|
("ༀ-༊༼࿁࿂྅" "." ?<) ; prohibition
|
|
|
|
|
|
("༓༕-༘༚-༟༶༸-༻༾༿྾྿-࿏" "." ?q) ; others
|
1997-05-12 05:30:27 +00:00
|
|
|
|
))
|
|
|
|
|
|
elm chars len syntax category to ch i)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(while deflist
|
|
|
|
|
|
(setq elm (car deflist))
|
|
|
|
|
|
(setq chars (car elm)
|
|
|
|
|
|
len (length chars)
|
|
|
|
|
|
syntax (nth 1 elm)
|
|
|
|
|
|
category (nth 2 elm)
|
|
|
|
|
|
i 0)
|
|
|
|
|
|
(while (< i len)
|
|
|
|
|
|
(if (= (aref chars i) ?-)
|
|
|
|
|
|
(setq i (1+ i)
|
1998-02-04 11:25:47 +00:00
|
|
|
|
to (aref chars i))
|
|
|
|
|
|
(setq ch (aref chars i)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
to ch))
|
|
|
|
|
|
(while (<= ch to)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(unless (string-equal syntax "w")
|
|
|
|
|
|
(modify-syntax-entry ch syntax))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(modify-category-entry ch category)
|
|
|
|
|
|
(setq ch (1+ ch)))
|
1998-02-04 11:25:47 +00:00
|
|
|
|
(setq i (1+ i)))
|
1997-02-20 07:02:49 +00:00
|
|
|
|
(setq deflist (cdr deflist))))
|
|
|
|
|
|
|
|
|
|
|
|
;; Vietnamese character set
|
|
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
;; To make a word with Latin characters
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'vietnamese-viscii-lower ?l)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'vietnamese-viscii-lower ?v)
|
|
|
|
|
|
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'vietnamese-viscii-upper ?l)
|
|
|
|
|
|
(map-charset-chars #'modify-category-entry 'vietnamese-viscii-upper ?v)
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
2000-02-13 23:50:32 +00:00
|
|
|
|
(let ((tbl (standard-case-table))
|
|
|
|
|
|
(i 32))
|
|
|
|
|
|
(while (< i 128)
|
2002-06-02 20:32:26 +00:00
|
|
|
|
(let* ((char (decode-char 'vietnamese-viscii-upper i))
|
|
|
|
|
|
(charl (decode-char 'vietnamese-viscii-lower i))
|
|
|
|
|
|
(uc (encode-char char 'ucs))
|
|
|
|
|
|
(lc (encode-char charl 'ucs)))
|
|
|
|
|
|
(set-case-syntax-pair char (decode-char 'vietnamese-viscii-lower i)
|
|
|
|
|
|
tbl)
|
|
|
|
|
|
(if uc (modify-category-entry uc ?v))
|
|
|
|
|
|
(if lc (modify-category-entry lc ?v)))
|
2000-02-13 23:50:32 +00:00
|
|
|
|
(setq i (1+ i))))
|
|
|
|
|
|
|
2002-12-05 14:28:21 +00:00
|
|
|
|
|
|
|
|
|
|
;; Latin
|
|
|
|
|
|
|
|
|
|
|
|
(modify-category-entry '(#x80 . #x024F) ?l)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2000-12-13 02:01:05 +00:00
|
|
|
|
(let ((tbl (standard-case-table)) c)
|
|
|
|
|
|
|
2002-10-05 18:51:37 +00:00
|
|
|
|
;; Latin-1
|
|
|
|
|
|
|
|
|
|
|
|
;; Fixme: Some of the non-word syntaxes here perhaps should be
|
|
|
|
|
|
;; reviewed. (Note that the following all implicitly have word
|
|
|
|
|
|
;; syntax: ¢£¤¥¨ª¯²³´¶¸¹º.) There should be a well-defined way of
|
|
|
|
|
|
;; relating Unicode categories to Emacs syntax codes.
|
2005-04-21 05:57:27 +00:00
|
|
|
|
|
|
|
|
|
|
;; NBSP isn't semantically interchangeable with other whitespace chars,
|
|
|
|
|
|
;; so it's more like punctation.
|
|
|
|
|
|
(set-case-syntax ? "." tbl)
|
2002-10-05 18:51:37 +00:00
|
|
|
|
(set-case-syntax ?¡ "." tbl)
|
|
|
|
|
|
(set-case-syntax ?¦ "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?§ "." tbl)
|
|
|
|
|
|
(set-case-syntax ?© "_" tbl)
|
|
|
|
|
|
(set-case-syntax-delims 171 187 tbl) ; « »
|
|
|
|
|
|
(set-case-syntax ?¬ "_" tbl)
|
|
|
|
|
|
(set-case-syntax ? "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?® "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?° "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?± "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?µ "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?· "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?¼ "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?½ "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?¾ "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?¿ "." tbl)
|
|
|
|
|
|
(let ((c 192))
|
|
|
|
|
|
(while (<= c 222)
|
|
|
|
|
|
(set-case-syntax-pair c (+ c 32) tbl)
|
|
|
|
|
|
(setq c (1+ c))))
|
|
|
|
|
|
(set-case-syntax ?× "_" tbl)
|
|
|
|
|
|
(set-case-syntax ?ß "w" tbl)
|
|
|
|
|
|
(set-case-syntax ?÷ "_" tbl)
|
|
|
|
|
|
;; See below for ÿ.
|
2000-12-13 02:01:05 +00:00
|
|
|
|
|
|
|
|
|
|
;; Latin Extended-A, Latin Extended-B
|
|
|
|
|
|
(setq c #x0100)
|
|
|
|
|
|
(while (<= c #x0233)
|
|
|
|
|
|
(and (or (<= c #x012e)
|
|
|
|
|
|
(and (>= c #x014a) (<= c #x0177)))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(zerop (% c 2))
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (1+ c) tbl))
|
2000-12-13 02:01:05 +00:00
|
|
|
|
(and (>= c #x013a)
|
|
|
|
|
|
(<= c #x0148)
|
|
|
|
|
|
(zerop (% c 2))
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair (1- c) c tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(setq c (1+ c)))
|
2006-11-22 01:05:20 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
;; In some languages, such as Turkish, U+0049 LATIN CAPITAL LETTER I
|
|
|
|
|
|
;; and U+0131 LATIN SMALL LETTER DOTLESS I make a case pair, and so
|
|
|
|
|
|
;; do U+0130 LATIN CAPITAL LETTER I WITH DOT ABOVE and U+0069 LATIN
|
|
|
|
|
|
;; SMALL LETTER I.
|
|
|
|
|
|
|
|
|
|
|
|
;; We used to set up half of those correspondence unconditionally,
|
|
|
|
|
|
;; but that makes searches slow. So now we don't set up either half
|
|
|
|
|
|
;; of these correspondences by default.
|
|
|
|
|
|
|
|
|
|
|
|
;; (set-downcase-syntax ?İ ?i tbl)
|
|
|
|
|
|
;; (set-upcase-syntax ?I ?ı tbl)
|
|
|
|
|
|
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?IJ ?ij tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ĵ ?ĵ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ķ ?ķ tbl)
|
2002-09-05 22:57:13 +00:00
|
|
|
|
(set-case-syntax-pair ?Ÿ ?ÿ tbl)
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?Ź ?ź tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ż ?ż tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ž ?ž tbl)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; Latin Extended-B
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?Ɓ ?ɓ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƃ ?ƃ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƅ ?ƅ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɔ ?ɔ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƈ ?ƈ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɖ ?ɖ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɗ ?ɗ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƌ ?ƌ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǝ ?ǝ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ə ?ə tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɛ ?ɛ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƒ ?ƒ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɠ ?ɠ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɣ ?ɣ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɩ ?ɩ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɨ ?ɨ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƙ ?ƙ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɯ ?ɯ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɲ ?ɲ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ɵ ?ɵ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ơ ?ơ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƣ ?ƣ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƥ ?ƥ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ʀ ?ʀ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƨ ?ƨ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ʃ ?ʃ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƭ ?ƭ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ʈ ?ʈ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ư ?ư tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ʊ ?ʊ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ʋ ?ʋ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƴ ?ƴ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƶ ?ƶ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ʒ ?ʒ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƹ ?ƹ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƽ ?ƽ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?DŽ ?dž tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Dž ?dž tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?LJ ?lj tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Lj ?lj tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?NJ ?nj tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Nj ?nj tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǎ ?ǎ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǐ ?ǐ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǒ ?ǒ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǔ ?ǔ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǖ ?ǖ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǘ ?ǘ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǚ ?ǚ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǜ ?ǜ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǟ ?ǟ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǡ ?ǡ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǣ ?ǣ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǥ ?ǥ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǧ ?ǧ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǩ ?ǩ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǫ ?ǫ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǭ ?ǭ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǯ ?ǯ tbl)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; 01F0; F; 006A 030C; # LATIN SMALL LETTER J WITH CARON
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?DZ ?dz tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Dz ?dz tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǵ ?ǵ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƕ ?ƕ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ƿ ?ƿ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǹ ?ǹ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǻ ?ǻ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǽ ?ǽ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ǿ ?ǿ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȁ ?ȁ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȃ ?ȃ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȅ ?ȅ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȇ ?ȇ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȉ ?ȉ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȋ ?ȋ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȍ ?ȍ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȏ ?ȏ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȑ ?ȑ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȓ ?ȓ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȕ ?ȕ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȗ ?ȗ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ș ?ș tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ț ?ț tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȝ ?ȝ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȟ ?ȟ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȣ ?ȣ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȥ ?ȥ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȧ ?ȧ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȩ ?ȩ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȫ ?ȫ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȭ ?ȭ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȯ ?ȯ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȱ ?ȱ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ȳ ?ȳ tbl)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
|
2000-12-13 02:01:05 +00:00
|
|
|
|
;; Latin Extended Additional
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#x1e00 . #x1ef9) ?l)
|
2000-12-13 02:01:05 +00:00
|
|
|
|
(setq c #x1e00)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(while (<= c #x1ef9)
|
|
|
|
|
|
(and (zerop (% c 2))
|
|
|
|
|
|
(or (<= c #x1e94) (>= c #x1ea0))
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (1+ c) tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
|
2000-12-13 02:01:05 +00:00
|
|
|
|
;; Greek
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#x0370 . #x03ff) ?g)
|
2000-12-13 02:01:05 +00:00
|
|
|
|
(setq c #x0370)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(while (<= c #x03ff)
|
|
|
|
|
|
(if (or (and (>= c #x0391) (<= c #x03a1))
|
|
|
|
|
|
(and (>= c #x03a3) (<= c #x03ab)))
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c 32) tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(and (>= c #x03da)
|
|
|
|
|
|
(<= c #x03ee)
|
|
|
|
|
|
(zerop (% c 2))
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (1+ c) tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(setq c (1+ c)))
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?Ά ?ά tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Έ ?έ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ή ?ή tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ί ?ί tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ό ?ό tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ύ ?ύ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ώ ?ώ tbl)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; Armenian
|
|
|
|
|
|
(setq c #x531)
|
|
|
|
|
|
(while (<= c #x556)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c #x30) tbl)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
|
2000-12-13 02:01:05 +00:00
|
|
|
|
;; Greek Extended
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#x1f00 . #x1fff) ?g)
|
2000-12-13 02:01:05 +00:00
|
|
|
|
(setq c #x1f00)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(while (<= c #x1fff)
|
|
|
|
|
|
(and (<= (logand c #x000f) 7)
|
|
|
|
|
|
(<= c #x1fa7)
|
|
|
|
|
|
(not (memq c '(#x1f50 #x1f52 #x1f54 #x1f56)))
|
|
|
|
|
|
(/= (logand c #x00f0) 7)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair (+ c 8) c tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(setq c (1+ c)))
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?Ᾰ ?ᾰ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ᾱ ?ᾱ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὰ ?ὰ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ά ?ά tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?ᾼ ?ᾳ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὲ ?ὲ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Έ ?έ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὴ ?ὴ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ή ?ή tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?ῌ ?ῃ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ῐ ?ῐ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ῑ ?ῑ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὶ ?ὶ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ί ?ί tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ῠ ?ῠ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ῡ ?ῡ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὺ ?ὺ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ύ ?ύ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ῥ ?ῥ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὸ ?ὸ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ό ?ό tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ὼ ?ὼ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ώ ?ώ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?ῼ ?ῳ tbl)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2000-12-13 02:01:05 +00:00
|
|
|
|
;; cyrillic
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#x0400 . #x04FF) ?y)
|
2000-12-13 02:01:05 +00:00
|
|
|
|
(setq c #x0400)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(while (<= c #x04ff)
|
|
|
|
|
|
(and (>= c #x0400)
|
|
|
|
|
|
(<= c #x040f)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c 80) tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(and (>= c #x0410)
|
|
|
|
|
|
(<= c #x042f)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c 32) tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(and (zerop (% c 2))
|
|
|
|
|
|
(or (and (>= c #x0460) (<= c #x0480))
|
|
|
|
|
|
(and (>= c #x048c) (<= c #x04be))
|
|
|
|
|
|
(and (>= c #x04d0) (<= c #x04f4)))
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(set-case-syntax-pair c (1+ c) tbl))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(setq c (1+ c)))
|
2002-05-09 06:44:06 +00:00
|
|
|
|
(set-case-syntax-pair ?Ӂ ?ӂ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ӄ ?ӄ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ӈ ?ӈ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ӌ ?ӌ tbl)
|
|
|
|
|
|
(set-case-syntax-pair ?Ӹ ?ӹ tbl)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2000-12-13 02:01:05 +00:00
|
|
|
|
;; general punctuation
|
|
|
|
|
|
(setq c #x2000)
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(while (<= c #x200b)
|
|
|
|
|
|
(set-case-syntax c " " tbl)
|
|
|
|
|
|
(setq c (1+ c)))
|
2002-09-23 22:21:46 +00:00
|
|
|
|
(while (<= c #x200F)
|
|
|
|
|
|
(set-case-syntax c "." tbl)
|
|
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
;; Fixme: These aren't all right:
|
2004-04-16 12:51:06 +00:00
|
|
|
|
(setq c #x2010)
|
|
|
|
|
|
(while (<= c #x2016)
|
|
|
|
|
|
(set-case-syntax c "_" tbl)
|
|
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
;; Punctuation syntax for quotation marks (like `)
|
|
|
|
|
|
(while (<= c #x201f)
|
|
|
|
|
|
(set-case-syntax c "." tbl)
|
|
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
;; Fixme: These aren't all right:
|
2000-11-27 12:22:29 +00:00
|
|
|
|
(while (<= c #x2027)
|
|
|
|
|
|
(set-case-syntax c "_" tbl)
|
|
|
|
|
|
(setq c (1+ c)))
|
2002-09-23 22:21:46 +00:00
|
|
|
|
(while (<= c #x206F)
|
|
|
|
|
|
(set-case-syntax c "." tbl)
|
|
|
|
|
|
(setq c (1+ c)))
|
2000-11-27 12:22:29 +00:00
|
|
|
|
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; Roman numerals
|
|
|
|
|
|
(setq c #x2160)
|
|
|
|
|
|
(while (<= c #x216f)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c #x10) tbl)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
|
2002-10-05 18:51:37 +00:00
|
|
|
|
;; Fixme: The following blocks might be better as symbol rather than
|
|
|
|
|
|
;; punctuation.
|
2002-09-23 22:21:46 +00:00
|
|
|
|
;; Arrows
|
|
|
|
|
|
(setq c #x2190)
|
2002-09-25 00:00:33 +00:00
|
|
|
|
(while (<= c #x21FF)
|
|
|
|
|
|
(set-case-syntax c "." tbl)
|
2002-09-23 22:21:46 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
;; Mathematical Operators
|
|
|
|
|
|
(while (<= c #x22FF)
|
2002-09-25 00:00:33 +00:00
|
|
|
|
(set-case-syntax c "." tbl)
|
2002-09-23 22:21:46 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
;; Miscellaneous Technical
|
|
|
|
|
|
(while (<= c #x23FF)
|
2002-09-25 00:00:33 +00:00
|
|
|
|
(set-case-syntax c "." tbl)
|
2002-09-23 22:21:46 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
;; Control Pictures
|
|
|
|
|
|
(while (<= c #x243F)
|
2002-09-25 00:00:33 +00:00
|
|
|
|
(set-case-syntax c "_" tbl)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
|
|
|
|
|
|
;; Circled Latin
|
|
|
|
|
|
(setq c #x24b6)
|
|
|
|
|
|
(while (<= c #x24cf)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c 26) tbl)
|
|
|
|
|
|
(modify-category-entry c ?l)
|
|
|
|
|
|
(modify-category-entry (+ c 26) ?l)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
|
|
|
|
|
|
;; Fullwidth Latin
|
|
|
|
|
|
(setq c #xff21)
|
|
|
|
|
|
(while (<= c #xff3a)
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(set-case-syntax-pair c (+ c #x20) tbl)
|
|
|
|
|
|
(modify-category-entry c ?l)
|
|
|
|
|
|
(modify-category-entry (+ c #x20) ?l)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
(setq c (1+ c)))
|
|
|
|
|
|
|
|
|
|
|
|
;; Combining diacritics
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#x300 . #x362) ?^)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
;; Combining marks
|
2002-05-23 18:16:52 +00:00
|
|
|
|
(modify-category-entry '(#x20d0 . #x20e3) ?^)
|
2001-12-07 14:47:37 +00:00
|
|
|
|
|
|
|
|
|
|
;; Fixme: syntax for symbols &c
|
|
|
|
|
|
)
|
2004-04-16 12:51:06 +00:00
|
|
|
|
|
|
|
|
|
|
(let ((pairs
|
2004-04-19 07:58:26 +00:00
|
|
|
|
'("⁅⁆" ; U+2045 U+2046
|
|
|
|
|
|
"⁽⁾" ; U+207D U+207E
|
|
|
|
|
|
"₍₎" ; U+208D U+208E
|
|
|
|
|
|
"〈〉" ; U+2329 U+232A
|
|
|
|
|
|
"⎴⎵" ; U+23B4 U+23B5
|
|
|
|
|
|
"❨❩" ; U+2768 U+2769
|
|
|
|
|
|
"❪❫" ; U+276A U+276B
|
|
|
|
|
|
"❬❭" ; U+276C U+276D
|
|
|
|
|
|
"❰❱" ; U+2770 U+2771
|
|
|
|
|
|
"❲❳" ; U+2772 U+2773
|
|
|
|
|
|
"❴❵" ; U+2774 U+2775
|
|
|
|
|
|
"⟦⟧" ; U+27E6 U+27E7
|
|
|
|
|
|
"⟨⟩" ; U+27E8 U+27E9
|
|
|
|
|
|
"⟪⟫" ; U+27EA U+27EB
|
|
|
|
|
|
"⦃⦄" ; U+2983 U+2984
|
|
|
|
|
|
"⦅⦆" ; U+2985 U+2986
|
|
|
|
|
|
"⦇⦈" ; U+2987 U+2988
|
|
|
|
|
|
"⦉⦊" ; U+2989 U+298A
|
|
|
|
|
|
"⦋⦌" ; U+298B U+298C
|
|
|
|
|
|
"⦍⦎" ; U+298D U+298E
|
|
|
|
|
|
"⦏⦐" ; U+298F U+2990
|
|
|
|
|
|
"⦑⦒" ; U+2991 U+2992
|
|
|
|
|
|
"⦓⦔" ; U+2993 U+2994
|
|
|
|
|
|
"⦕⦖" ; U+2995 U+2996
|
|
|
|
|
|
"⦗⦘" ; U+2997 U+2998
|
|
|
|
|
|
"⧼⧽" ; U+29FC U+29FD
|
|
|
|
|
|
"〈〉" ; U+3008 U+3009
|
|
|
|
|
|
"《》" ; U+300A U+300B
|
|
|
|
|
|
"「」" ; U+300C U+300D
|
|
|
|
|
|
"『』" ; U+300E U+300F
|
|
|
|
|
|
"【】" ; U+3010 U+3011
|
|
|
|
|
|
"〔〕" ; U+3014 U+3015
|
|
|
|
|
|
"〖〗" ; U+3016 U+3017
|
|
|
|
|
|
"〘〙" ; U+3018 U+3019
|
|
|
|
|
|
"〚〛" ; U+301A U+301B
|
|
|
|
|
|
"﴾﴿" ; U+FD3E U+FD3F
|
|
|
|
|
|
"︵︶" ; U+FE35 U+FE36
|
|
|
|
|
|
"︷︸" ; U+FE37 U+FE38
|
|
|
|
|
|
"︹︺" ; U+FE39 U+FE3A
|
|
|
|
|
|
"︻︼" ; U+FE3B U+FE3C
|
|
|
|
|
|
"︽︾" ; U+FE3D U+FE3E
|
|
|
|
|
|
"︿﹀" ; U+FE3F U+FE40
|
|
|
|
|
|
"﹁﹂" ; U+FE41 U+FE42
|
|
|
|
|
|
"﹃﹄" ; U+FE43 U+FE44
|
|
|
|
|
|
"﹙﹚" ; U+FE59 U+FE5A
|
|
|
|
|
|
"﹛﹜" ; U+FE5B U+FE5C
|
|
|
|
|
|
"﹝﹞" ; U+FE5D U+FE5E
|
|
|
|
|
|
"（）" ; U+FF08 U+FF09
|
|
|
|
|
|
"［］" ; U+FF3B U+FF3D
|
|
|
|
|
|
"｛｝" ; U+FF5B U+FF5D
|
|
|
|
|
|
"｟｠" ; U+FF5F U+FF60
|
|
|
|
|
|
"｢｣" ; U+FF62 U+FF63
|
2004-04-16 12:51:06 +00:00
|
|
|
|
)))
|
|
|
|
|
|
(dolist (elt pairs)
|
|
|
|
|
|
(modify-syntax-entry (aref elt 0) (string ?\( (aref elt 1)))
|
|
|
|
|
|
(modify-syntax-entry (aref elt 1) (string ?\) (aref elt 0)))))
|
|
|
|
|
|
|
1997-02-20 07:02:49 +00:00
|
|
|
|
|
1998-01-22 01:42:20 +00:00
|
|
|
|
;; For each character set, put the information of the most proper
|
1999-05-26 12:29:18 +00:00
|
|
|
|
;; coding system to encode it by `preferred-coding-system' property.
|
1998-01-22 01:42:20 +00:00
|
|
|
|
|
2002-05-23 18:16:52 +00:00
|
|
|
|
;; Fixme: should this be junked?
|
1998-01-22 01:42:20 +00:00
|
|
|
|
(let ((l '((latin-iso8859-1 . iso-latin-1)
|
|
|
|
|
|
(latin-iso8859-2 . iso-latin-2)
|
|
|
|
|
|
(latin-iso8859-3 . iso-latin-3)
|
|
|
|
|
|
(latin-iso8859-4 . iso-latin-4)
|
|
|
|
|
|
(thai-tis620 . thai-tis620)
|
|
|
|
|
|
(greek-iso8859-7 . greek-iso-8bit)
|
|
|
|
|
|
(arabic-iso8859-6 . iso-2022-7bit)
|
|
|
|
|
|
(hebrew-iso8859-8 . hebrew-iso-8bit)
|
|
|
|
|
|
(katakana-jisx0201 . japanese-shift-jis)
|
|
|
|
|
|
(latin-jisx0201 . japanese-shift-jis)
|
|
|
|
|
|
(cyrillic-iso8859-5 . cyrillic-iso-8bit)
|
|
|
|
|
|
(latin-iso8859-9 . iso-latin-5)
|
|
|
|
|
|
(japanese-jisx0208-1978 . iso-2022-jp)
|
2006-11-27 04:40:25 +00:00
|
|
|
|
(chinese-gb2312 . chinese-iso-8bit)
|
|
|
|
|
|
(chinese-gbk . chinese-gbk)
|
|
|
|
|
|
(gb18030-2-byte . chinese-gb18030)
|
|
|
|
|
|
(gb18030-4-byte-bmp . chinese-gb18030)
|
|
|
|
|
|
(gb18030-4-byte-smp . chinese-gb18030)
|
|
|
|
|
|
(gb18030-4-byte-ext-1 . chinese-gb18030)
|
|
|
|
|
|
(gb18030-4-byte-ext-2 . chinese-gb18030)
|
1998-01-22 01:42:20 +00:00
|
|
|
|
(japanese-jisx0208 . iso-2022-jp)
|
|
|
|
|
|
(korean-ksc5601 . iso-2022-kr)
|
|
|
|
|
|
(japanese-jisx0212 . iso-2022-jp)
|
|
|
|
|
|
(chinese-big5-1 . chinese-big5)
|
|
|
|
|
|
(chinese-big5-2 . chinese-big5)
|
|
|
|
|
|
(chinese-sisheng . iso-2022-7bit)
|
|
|
|
|
|
(ipa . iso-2022-7bit)
|
|
|
|
|
|
(vietnamese-viscii-lower . vietnamese-viscii)
|
|
|
|
|
|
(vietnamese-viscii-upper . vietnamese-viscii)
|
|
|
|
|
|
(arabic-digit . iso-2022-7bit)
|
|
|
|
|
|
(arabic-1-column . iso-2022-7bit)
|
|
|
|
|
|
(lao . lao)
|
|
|
|
|
|
(arabic-2-column . iso-2022-7bit)
|
|
|
|
|
|
(indian-is13194 . devanagari)
|
2001-11-25 02:34:41 +00:00
|
|
|
|
(indian-glyph . devanagari)
|
1998-01-22 01:42:20 +00:00
|
|
|
|
(tibetan-1-column . tibetan)
|
1999-02-03 06:26:56 +00:00
|
|
|
|
(ethiopic . iso-2022-7bit)
|
2006-11-27 04:40:25 +00:00
|
|
|
|
(chinese-cns11643-1 . iso-2022-cn)
|
|
|
|
|
|
(chinese-cns11643-2 . iso-2022-cn)
|
1998-01-22 01:42:20 +00:00
|
|
|
|
(chinese-cns11643-3 . iso-2022-cn)
|
|
|
|
|
|
(chinese-cns11643-4 . iso-2022-cn)
|
|
|
|
|
|
(chinese-cns11643-5 . iso-2022-cn)
|
|
|
|
|
|
(chinese-cns11643-6 . iso-2022-cn)
|
|
|
|
|
|
(chinese-cns11643-7 . iso-2022-cn)
|
|
|
|
|
|
(indian-2-column . devanagari)
|
1999-10-14 09:49:12 +00:00
|
|
|
|
(tibetan . tibetan)
|
|
|
|
|
|
(latin-iso8859-14 . iso-latin-8)
|
|
|
|
|
|
(latin-iso8859-15 . iso-latin-9))))
|
1998-01-22 01:42:20 +00:00
|
|
|
|
(while l
|
1999-05-26 12:29:18 +00:00
|
|
|
|
(put-charset-property (car (car l)) 'preferred-coding-system (cdr (car l)))
|
1998-01-22 01:42:20 +00:00
|
|
|
|
(setq l (cdr l))))
|
1998-10-19 00:40:10 +00:00
|
|
|
|
|
|
|
|
|
|
|
1998-10-21 11:50:56 +00:00
|
|
|
|
;; Setup auto-fill-chars for charsets that should invoke auto-filling.
|
2006-10-26 02:31:05 +00:00
|
|
|
|
;; SPACE and NEWLINE are already set.
|
2005-12-28 07:22:45 +00:00
|
|
|
|
|
|
|
|
|
|
(set-char-table-range auto-fill-chars '(#x3041 . #x30FF) t)
|
|
|
|
|
|
(set-char-table-range auto-fill-chars '(#x3400 . #x4DB5) t)
|
|
|
|
|
|
(set-char-table-range auto-fill-chars '(#x4e00 . #x9fbb) t)
|
|
|
|
|
|
(set-char-table-range auto-fill-chars '(#xF900 . #xFAFF) t)
|
|
|
|
|
|
(set-char-table-range auto-fill-chars '(#xFF00 . #xFF9F) t)
|
|
|
|
|
|
(set-char-table-range auto-fill-chars '(#x20000 . #x2FFFF) t)
|
|
|
|
|
|
|
2002-03-01 02:06:19 +00:00
|
|
|
|
|
2006-10-26 02:31:05 +00:00
|
|
|
|
;;; Setting char-width-table. The default is 1.
|
|
|
|
|
|
|
|
|
|
|
|
;; 0: non-spacing, enclosing combining, formatting, Hangul Jamo medial
|
|
|
|
|
|
;; and final characters.
|
|
|
|
|
|
(let ((l '((#x00AD . #x00AD)
|
|
|
|
|
|
(#x0300 . #x036F)
|
|
|
|
|
|
(#x0483 . #x0489)
|
|
|
|
|
|
(#x0591 . #x05BD)
|
|
|
|
|
|
(#x05BF . #x05BF)
|
|
|
|
|
|
(#x05C1 . #x05C2)
|
|
|
|
|
|
(#x05C4 . #x05C5)
|
|
|
|
|
|
(#x05C7 . #x05C7)
|
|
|
|
|
|
(#x0600 . #x0603)
|
|
|
|
|
|
(#x0610 . #x0615)
|
|
|
|
|
|
(#x064B . #x065E)
|
|
|
|
|
|
(#x0670 . #x0670)
|
|
|
|
|
|
(#x06D6 . #x06E4)
|
|
|
|
|
|
(#x06E7 . #x06E8)
|
|
|
|
|
|
(#x06EA . #x06ED)
|
|
|
|
|
|
(#x070F . #x070F)
|
|
|
|
|
|
(#x0711 . #x0711)
|
|
|
|
|
|
(#x0730 . #x074A)
|
|
|
|
|
|
(#x07A6 . #x07B0)
|
|
|
|
|
|
(#x07EB . #x07F3)
|
|
|
|
|
|
(#x0901 . #x0902)
|
|
|
|
|
|
(#x093C . #x093C)
|
|
|
|
|
|
(#x0941 . #x0948)
|
|
|
|
|
|
(#x094D . #x094D)
|
|
|
|
|
|
(#x0951 . #x0954)
|
|
|
|
|
|
(#x0962 . #x0963)
|
|
|
|
|
|
(#x0981 . #x0981)
|
|
|
|
|
|
(#x09BC . #x09BC)
|
|
|
|
|
|
(#x09C1 . #x09C4)
|
|
|
|
|
|
(#x09CD . #x09CD)
|
|
|
|
|
|
(#x09E2 . #x09E3)
|
|
|
|
|
|
(#x0A01 . #x0A02)
|
|
|
|
|
|
(#x0A3C . #x0A3C)
|
|
|
|
|
|
(#x0A41 . #x0A4D)
|
|
|
|
|
|
(#x0A70 . #x0A71)
|
|
|
|
|
|
(#x0A81 . #x0A82)
|
|
|
|
|
|
(#x0ABC . #x0ABC)
|
|
|
|
|
|
(#x0AC1 . #x0AC8)
|
|
|
|
|
|
(#x0ACD . #x0ACD)
|
|
|
|
|
|
(#x0AE2 . #x0AE3)
|
|
|
|
|
|
(#x0B01 . #x0B01)
|
|
|
|
|
|
(#x0B3C . #x0B3C)
|
|
|
|
|
|
(#x0B3F . #x0B3F)
|
|
|
|
|
|
(#x0B41 . #x0B43)
|
|
|
|
|
|
(#x0B4D . #x0B56)
|
|
|
|
|
|
(#x0B82 . #x0B82)
|
|
|
|
|
|
(#x0BC0 . #x0BC0)
|
|
|
|
|
|
(#x0BCD . #x0BCD)
|
|
|
|
|
|
(#x0C3E . #x0C40)
|
|
|
|
|
|
(#x0C46 . #x0C56)
|
|
|
|
|
|
(#x0CBC . #x0CBC)
|
|
|
|
|
|
(#x0CBF . #x0CBF)
|
|
|
|
|
|
(#x0CC6 . #x0CC6)
|
|
|
|
|
|
(#x0CCC . #x0CCD)
|
|
|
|
|
|
(#x0CE2 . #x0CE3)
|
|
|
|
|
|
(#x0D41 . #x0D43)
|
|
|
|
|
|
(#x0D4D . #x0D4D)
|
|
|
|
|
|
(#x0DCA . #x0DCA)
|
|
|
|
|
|
(#x0DD2 . #x0DD6)
|
|
|
|
|
|
(#x0E31 . #x0E31)
|
|
|
|
|
|
(#x0E34 . #x0E3A)
|
|
|
|
|
|
(#x0E47 . #x0E4E)
|
|
|
|
|
|
(#x0EB1 . #x0EB1)
|
|
|
|
|
|
(#x0EB4 . #x0EBC)
|
|
|
|
|
|
(#x0EC8 . #x0ECD)
|
|
|
|
|
|
(#x0F18 . #x0F19)
|
|
|
|
|
|
(#x0F35 . #x0F35)
|
|
|
|
|
|
(#x0F37 . #x0F37)
|
|
|
|
|
|
(#x0F39 . #x0F39)
|
|
|
|
|
|
(#x0F71 . #x0F7E)
|
|
|
|
|
|
(#x0F80 . #x0F84)
|
|
|
|
|
|
(#x0F86 . #x0F87)
|
|
|
|
|
|
(#x0F90 . #x0FBC)
|
|
|
|
|
|
(#x0FC6 . #x0FC6)
|
|
|
|
|
|
(#x102D . #x1030)
|
|
|
|
|
|
(#x1032 . #x1037)
|
|
|
|
|
|
(#x1039 . #x1039)
|
|
|
|
|
|
(#x1058 . #x1059)
|
|
|
|
|
|
(#x1160 . #x11FF)
|
|
|
|
|
|
(#x135F . #x135F)
|
|
|
|
|
|
(#x1712 . #x1714)
|
|
|
|
|
|
(#x1732 . #x1734)
|
|
|
|
|
|
(#x1752 . #x1753)
|
|
|
|
|
|
(#x1772 . #x1773)
|
|
|
|
|
|
(#x17B4 . #x17B5)
|
|
|
|
|
|
(#x17B7 . #x17BD)
|
|
|
|
|
|
(#x17C6 . #x17C6)
|
|
|
|
|
|
(#x17C9 . #x17D3)
|
|
|
|
|
|
(#x17DD . #x17DD)
|
|
|
|
|
|
(#x180B . #x180D)
|
|
|
|
|
|
(#x18A9 . #x18A9)
|
|
|
|
|
|
(#x1920 . #x1922)
|
|
|
|
|
|
(#x1927 . #x1928)
|
|
|
|
|
|
(#x1932 . #x1932)
|
|
|
|
|
|
(#x1939 . #x193B)
|
|
|
|
|
|
(#x1A17 . #x1A18)
|
|
|
|
|
|
(#x1B00 . #x1B03)
|
|
|
|
|
|
(#x1B34 . #x1B34)
|
|
|
|
|
|
(#x1B36 . #x1B3A)
|
|
|
|
|
|
(#x1B3C . #x1B3C)
|
|
|
|
|
|
(#x1B42 . #x1B42)
|
|
|
|
|
|
(#x1B6B . #x1B73)
|
|
|
|
|
|
(#x1DC0 . #x1DFF)
|
|
|
|
|
|
(#x200B . #x200F)
|
|
|
|
|
|
(#x202A . #x202E)
|
|
|
|
|
|
(#x2060 . #x206F)
|
|
|
|
|
|
(#x20D0 . #x20EF)
|
|
|
|
|
|
(#x302A . #x302F)
|
|
|
|
|
|
(#x3099 . #x309A)
|
|
|
|
|
|
(#xA806 . #xA806)
|
|
|
|
|
|
(#xA80B . #xA80B)
|
|
|
|
|
|
(#xA825 . #xA826)
|
|
|
|
|
|
(#xFB1E . #xFB1E)
|
|
|
|
|
|
(#xFE00 . #xFE0F)
|
|
|
|
|
|
(#xFE20 . #xFE23)
|
|
|
|
|
|
(#xFEFF . #xFEFF)
|
|
|
|
|
|
(#xFFF9 . #xFFFB)
|
|
|
|
|
|
(#x10A01 . #x10A0F)
|
|
|
|
|
|
(#x10A38 . #x10A3F)
|
|
|
|
|
|
(#x1D167 . #x1D169)
|
|
|
|
|
|
(#x1D173 . #x1D182)
|
|
|
|
|
|
(#x1D185 . #x1D18B)
|
|
|
|
|
|
(#x1D1AA . #x1D1AD)
|
|
|
|
|
|
(#x1D242 . #x1D244)
|
|
|
|
|
|
(#xE0001 . #xE01EF))))
|
|
|
|
|
|
(dolist (elt l)
|
|
|
|
|
|
(set-char-table-range char-width-table elt 0)))
|
|
|
|
|
|
|
|
|
|
|
|
;; 2: East Asian Wide and Full-width characters.
|
|
|
|
|
|
(let ((l '((#x1100 . #x115F)
|
|
|
|
|
|
(#x2329 . #x232A)
|
|
|
|
|
|
(#x2E80 . #x303E)
|
|
|
|
|
|
(#x3040 . #xA4CF)
|
|
|
|
|
|
(#xAC00 . #xD7A3)
|
2002-05-14 07:48:11 +00:00
|
|
|
|
(#xF900 . #xFAFF)
|
2006-10-26 02:31:05 +00:00
|
|
|
|
(#xFE30 . #xFE6F)
|
2006-11-08 12:15:57 +00:00
|
|
|
|
(#xFF01 . #xFF60)
|
2006-10-26 02:31:05 +00:00
|
|
|
|
(#xFFE0 . #xFFE6)
|
|
|
|
|
|
(#x20000 . #x2FFFF)
|
|
|
|
|
|
(#x30000 . #x3FFFF))))
|
2002-05-14 07:48:11 +00:00
|
|
|
|
(dolist (elt l)
|
2006-10-26 02:31:05 +00:00
|
|
|
|
(set-char-table-range char-width-table elt 2)))
|
2002-05-25 23:30:45 +00:00
|
|
|
|
|
|
|
|
|
|
;; Other double width
|
2006-10-26 02:31:05 +00:00
|
|
|
|
;;(map-charset-chars
|
|
|
|
|
|
;; (lambda (range ignore) (set-char-table-range char-width-table range 2))
|
|
|
|
|
|
;; 'ethiopic)
|
|
|
|
|
|
;; (map-charset-chars
|
|
|
|
|
|
;; (lambda (range ignore) (set-char-table-range char-width-table range 2))
|
|
|
|
|
|
;; 'tibetan)
|
2002-05-25 23:30:45 +00:00
|
|
|
|
(map-charset-chars
|
|
|
|
|
|
(lambda (range ignore) (set-char-table-range char-width-table range 2))
|
|
|
|
|
|
'indian-2-column)
|
|
|
|
|
|
(map-charset-chars
|
|
|
|
|
|
(lambda (range ignore) (set-char-table-range char-width-table range 2))
|
|
|
|
|
|
'arabic-2-column)
|
1998-01-22 01:42:20 +00:00
|
|
|
|
|
2002-06-06 11:34:24 +00:00
|
|
|
|
(optimize-char-table (standard-case-table))
|
|
|
|
|
|
(optimize-char-table (standard-category-table))
|
|
|
|
|
|
(optimize-char-table (standard-syntax-table))
|
|
|
|
|
|
|
2002-09-23 22:21:46 +00:00
|
|
|
|
;; The Unicode blocks actually extend past some of these ranges with
|
|
|
|
|
|
;; undefined codepoints.
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(let ((script-list nil))
|
|
|
|
|
|
(dolist
|
|
|
|
|
|
(elt
|
2002-08-19 10:58:11 +00:00
|
|
|
|
'((#x0000 #x007F latin)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#x00A0 #x036F latin)
|
|
|
|
|
|
(#x0370 #x03E1 greek)
|
|
|
|
|
|
(#x03E2 #x03EF coptic)
|
|
|
|
|
|
(#x03F0 #x03F3 greek)
|
|
|
|
|
|
(#x0400 #x04FF cyrillic)
|
|
|
|
|
|
(#x0530 #x058F armenian)
|
|
|
|
|
|
(#x0590 #x05FF hebrew)
|
|
|
|
|
|
(#x0600 #x06FF arabic)
|
|
|
|
|
|
(#x0700 #x074F syriac)
|
2006-08-28 02:32:19 +00:00
|
|
|
|
(#x07C0 #x07FA nko)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#x0780 #x07BF thaana)
|
|
|
|
|
|
(#x0900 #x097F devanagari)
|
|
|
|
|
|
(#x0980 #x09FF bengali)
|
|
|
|
|
|
(#x0A00 #x0A7F gurmukhi)
|
|
|
|
|
|
(#x0A80 #x0AFF gujarati)
|
|
|
|
|
|
(#x0B00 #x0B7F oriya)
|
|
|
|
|
|
(#x0B80 #x0BFF tamil)
|
|
|
|
|
|
(#x0C00 #x0C7F telugu)
|
|
|
|
|
|
(#x0C80 #x0CFF kannada)
|
|
|
|
|
|
(#x0D00 #x0D7F malayalam)
|
|
|
|
|
|
(#x0D80 #x0DFF sinhala)
|
|
|
|
|
|
(#x0E00 #x0E5F thai)
|
|
|
|
|
|
(#x0E80 #x0EDF lao)
|
|
|
|
|
|
(#x0F00 #x0FFF tibetan)
|
|
|
|
|
|
(#x1000 #x105F myanmar)
|
|
|
|
|
|
(#x10A0 #x10FF georgian)
|
|
|
|
|
|
(#x1100 #x11FF hangul)
|
2006-07-20 13:01:24 +00:00
|
|
|
|
(#x1200 #x139F ethiopic)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#x13A0 #x13FF cherokee)
|
|
|
|
|
|
(#x1400 #x167F canadian-aboriginal)
|
|
|
|
|
|
(#x1680 #x169F ogham)
|
|
|
|
|
|
(#x16A0 #x16FF runic)
|
|
|
|
|
|
(#x1780 #x17FF khmer)
|
|
|
|
|
|
(#x1800 #x18AF mongolian)
|
|
|
|
|
|
(#x1E00 #x1EFF latin)
|
|
|
|
|
|
(#x1F00 #x1FFF greek)
|
2003-11-21 23:55:01 +00:00
|
|
|
|
(#x2000 #x27FF symbol)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#x2800 #x28FF braille)
|
2006-07-20 13:01:24 +00:00
|
|
|
|
(#x2D80 #x2DDF ethiopic)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#x2E80 #x2FDF han)
|
|
|
|
|
|
(#x2FF0 #x2FFF ideographic-description)
|
|
|
|
|
|
(#x3000 #x303F cjk-misc)
|
|
|
|
|
|
(#x3040 #x30FF kana)
|
|
|
|
|
|
(#x3100 #x312F bopomofo)
|
|
|
|
|
|
(#x3130 #x318F hangul)
|
|
|
|
|
|
(#x3190 #x319F kanbun)
|
|
|
|
|
|
(#x31A0 #x31BF bopomofo)
|
|
|
|
|
|
(#x3400 #x9FAF han)
|
|
|
|
|
|
(#xA000 #xA4CF yi)
|
|
|
|
|
|
(#xAC00 #xD7AF hangul)
|
2004-02-05 04:31:35 +00:00
|
|
|
|
(#xF900 #xFAFF han)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#xFB1D #xFB4F hebrew)
|
|
|
|
|
|
(#xFB50 #xFDFF arabic)
|
|
|
|
|
|
(#xFE70 #xFEFC arabic)
|
|
|
|
|
|
(#xFF00 #xFF5F cjk-misc)
|
|
|
|
|
|
(#xFF61 #xFF9F kana)
|
|
|
|
|
|
(#xFFE0 #xFFE6 cjk-misc)
|
2006-08-28 02:32:19 +00:00
|
|
|
|
(#x1D000 #x1D0FF byzantine-musical-symbol)
|
|
|
|
|
|
(#x1D100 #x1D1FF musical-symbol)
|
|
|
|
|
|
(#x1D400 #x1D7FF mathematical)
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(#x20000 #x2AFFF han)
|
2002-07-26 04:02:41 +00:00
|
|
|
|
(#x2F800 #x2FFFF han)))
|
|
|
|
|
|
(set-char-table-range char-script-table
|
|
|
|
|
|
(cons (car elt) (nth 1 elt)) (nth 2 elt))
|
|
|
|
|
|
(or (memq (nth 2 elt) script-list)
|
|
|
|
|
|
(setq script-list (cons (nth 2 elt) script-list))))
|
|
|
|
|
|
(set-char-table-extra-slot char-script-table 0 (nreverse script-list)))
|
|
|
|
|
|
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(map-charset-chars
|
2002-11-07 06:27:55 +00:00
|
|
|
|
#'(lambda (range ignore)
|
|
|
|
|
|
(set-char-table-range char-script-table range 'tibetan))
|
|
|
|
|
|
'tibetan)
|
|
|
|
|
|
|
2002-08-01 12:36:17 +00:00
|
|
|
|
|
|
|
|
|
|
;;; Setting word boundary.
|
|
|
|
|
|
|
|
|
|
|
|
(defun next-word-boundary-han (pos limit)
|
|
|
|
|
|
(if (<= pos limit)
|
|
|
|
|
|
(save-excursion
|
|
|
|
|
|
(goto-char pos)
|
|
|
|
|
|
(looking-at "\\cC+")
|
|
|
|
|
|
(goto-char (match-end 0))
|
2002-08-08 13:00:10 +00:00
|
|
|
|
(if (looking-at "\\cH+")
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(goto-char (match-end 0)))
|
|
|
|
|
|
(point))
|
|
|
|
|
|
(while (and (> pos limit)
|
|
|
|
|
|
(eq (aref char-script-table (char-after (1- pos))) 'han))
|
|
|
|
|
|
(setq pos (1- pos)))
|
|
|
|
|
|
pos))
|
|
|
|
|
|
|
|
|
|
|
|
(defun next-word-boundary-kana (pos limit)
|
|
|
|
|
|
(if (<= pos limit)
|
|
|
|
|
|
(save-excursion
|
|
|
|
|
|
(goto-char pos)
|
|
|
|
|
|
(if (looking-at "\\cK+")
|
|
|
|
|
|
(goto-char (match-end 0)))
|
|
|
|
|
|
(if (looking-at "\\cH+")
|
|
|
|
|
|
(goto-char (match-end 0)))
|
2004-04-13 06:20:28 +00:00
|
|
|
|
(if (looking-at "\\ck+")
|
|
|
|
|
|
(goto-char (match-end 0)))
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(point))
|
|
|
|
|
|
(let ((category-set (char-category-set (char-after pos)))
|
|
|
|
|
|
category)
|
2004-04-13 06:20:28 +00:00
|
|
|
|
(if (or (aref category-set ?K) (aref category-set ?k))
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(while (and (> pos limit)
|
2004-04-13 06:20:28 +00:00
|
|
|
|
(setq category-set
|
|
|
|
|
|
(char-category-set (char-after (1- pos))))
|
|
|
|
|
|
(or (aref category-set ?K) (aref category-set ?k)))
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(setq pos (1- pos)))
|
|
|
|
|
|
(while (and (> pos limit)
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(aref (setq category-set
|
2002-08-01 12:36:17 +00:00
|
|
|
|
(char-category-set (char-after (1- pos)))) ?H))
|
|
|
|
|
|
(setq pos (1- pos)))
|
|
|
|
|
|
(setq category (cond ((aref category-set ?C) ?C)
|
|
|
|
|
|
((aref category-set ?K) ?K)
|
|
|
|
|
|
((aref category-set ?A) ?A)))
|
|
|
|
|
|
(when category
|
|
|
|
|
|
(setq pos (1- pos))
|
|
|
|
|
|
(while (and (> pos limit)
|
|
|
|
|
|
(aref (char-category-set (char-after (1- pos)))
|
|
|
|
|
|
category))
|
|
|
|
|
|
(setq pos (1- pos)))))
|
|
|
|
|
|
pos)))
|
|
|
|
|
|
|
|
|
|
|
|
(map-char-table
|
|
|
|
|
|
#'(lambda (char script)
|
|
|
|
|
|
(cond ((eq script 'han)
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(set-char-table-range find-word-boundary-function-table
|
2002-08-01 12:36:17 +00:00
|
|
|
|
char #'next-word-boundary-han))
|
|
|
|
|
|
((eq script 'kana)
|
2003-09-08 12:53:41 +00:00
|
|
|
|
(set-char-table-range find-word-boundary-function-table
|
2002-08-01 12:36:17 +00:00
|
|
|
|
char #'next-word-boundary-kana))))
|
|
|
|
|
|
char-script-table)
|
|
|
|
|
|
|
|
|
|
|
|
(setq word-combining-categories
|
2006-06-20 07:35:06 +00:00
|
|
|
|
'((?l . ?l)
|
|
|
|
|
|
(?C . ?C)
|
|
|
|
|
|
(?C . ?H)
|
|
|
|
|
|
(?C . ?K)))
|
2002-08-01 12:36:17 +00:00
|
|
|
|
|
|
|
|
|
|
(setq word-separating-categories ; (2-byte character sets)
|
|
|
|
|
|
'((?A . ?K) ; Alpha numeric - Katakana
|
|
|
|
|
|
(?A . ?C) ; Alpha numeric - Chinese
|
|
|
|
|
|
(?H . ?A) ; Hiragana - Alpha numeric
|
|
|
|
|
|
(?H . ?K) ; Hiragana - Katakana
|
|
|
|
|
|
(?H . ?C) ; Hiragana - Chinese
|
|
|
|
|
|
(?K . ?A) ; Katakana - Alpha numeric
|
|
|
|
|
|
(?K . ?C) ; Katakana - Chinese
|
|
|
|
|
|
(?C . ?A) ; Chinese - Alpha numeric
|
|
|
|
|
|
(?C . ?K) ; Chinese - Katakana
|
|
|
|
|
|
))
|
|
|
|
|
|
|
1998-01-22 01:42:20 +00:00
|
|
|
|
;;; Local Variables:
|
2002-05-09 06:44:06 +00:00
|
|
|
|
;;; coding: utf-8-emacs
|
1998-01-22 01:42:20 +00:00
|
|
|
|
;;; End:
|
|
|
|
|
|
|
2004-04-16 12:51:06 +00:00
|
|
|
|
;;; arch-tag: 85889c35-9f4d-4912-9bf5-82de31b0d42d
|
2001-07-15 19:53:53 +00:00
|
|
|
|
;;; characters.el ends here
|
