code.delx.au - gnu-emacs/blob - lisp/international/characters.el

   1 ;;; characters.el --- set syntax and category for multibyte characters
   2
   3 ;; Copyright (C) 1995 Free Software Foundation, Inc.
   4 ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN.
   5
   6 ;; Keywords: multibyte character, character set, syntax, category
   7
   8 ;; This file is part of GNU Emacs.
   9
  10 ;; GNU Emacs is free software; you can redistribute it and/or modify
  11 ;; it under the terms of the GNU General Public License as published by
  12 ;; the Free Software Foundation; either version 2, or (at your option)
  13 ;; any later version.
  14
  15 ;; GNU Emacs is distributed in the hope that it will be useful,
  16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 ;; GNU General Public License for more details.
  19
  20 ;; You should have received a copy of the GNU General Public License
  21 ;; along with GNU Emacs; see the file COPYING.  If not, write to
  22 ;; the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
  23
  24 ;;; Commentary:
  25
  26 ;; This file contains multibyte characters.  Save this file always in
  27 ;; `coding-system-iso-2022-7'.
  28
  29 ;;; Predefined categories.
  30
  31 ;; For each character set.
  32
  33 (define-category ?a "ASCII")
  34 (define-category ?l "Latin")
  35 (define-category ?t "Thai")
  36 (define-category ?g "Greek")
  37 (define-category ?b "Arabic")
  38 (define-category ?w "Hebrew")
  39 (define-category ?y "Cyrillic")
  40 (define-category ?k "Japanese katakana")
  41 (define-category ?r "Japanese roman")
  42 (define-category ?c "Chinese")
  43 (define-category ?j "Japanese")
  44 (define-category ?h "Korean")
  45 (define-category ?e "Ethiopic (Ge'ez)")
  46 (define-category ?v "Vietnamese")
  47 (define-category ?i "Indian")
  48
  49 ;; For each group (row) of 2-byte character sets.
  50
  51 (define-category ?A "Alpha numeric characters of 2-byte character sets")
  52 (define-category ?C "Chinese (Han) characters of 2-byte character sets")
  53 (define-category ?G "Greek characters of 2-byte characters sets")
  54 (define-category ?H "Japanese Hiragana characters of 2-byte character sets")
  55 (define-category ?K "Japanese Katakana characters of 2-byte character sets")
  56 (define-category ?N "Korean Hangul characters of 2-byte character sets")
  57 (define-category ?Y "Cyrillic character of 2-byte character sets")
  58 (define-category ?I "Indian Glyphs")
  59
  60 ;; For phonetic classifications.
  61
  62 (define-category ?0 "consonant")
  63 (define-category ?1 "base vowel")
  64 (define-category ?2 "upper diacritical mark (including upper vowel)")
  65 (define-category ?3 "lower diacritical mark (including lower vowel)")
  66 (define-category ?4 "tone mark")
  67 (define-category ?5 "vowel")
  68 (define-category ?6 "digit")
  69 (define-category ?7 "vowel-modifying diacritical mark")
  70 (define-category ?8 "vowel-signs.")
  71
  72 ;; For filling.
  73 (define-category ?| "While filling, we can break a line at this character.")
  74
  75 ;; Keep the followings for `kinsoku' processing.  See comments in
  76 ;; kinsoku.el.
  77 (define-category ?> "A character which can't be placed at beginning of line.")
  78 (define-category ?< "A character which can't be placed at end of line.")
  79
  80 \f
  81 ;;; Setting syntax and category.
  82
  83 ;; ASCII
  84
  85 (let ((ch 32))
  86   (while (< ch 127)                     ; All ASCII characters have
  87     (modify-category-entry ch ?a)       ; the category `a' (ASCII)
  88     (modify-category-entry ch ?l)       ; and `l' (Latin).
  89     (setq ch (1+ ch))))
  90
  91 ;; Arabic character set
  92
  93 (let ((charsets '(arabic-iso8859-6
  94                   arabic-digit
  95                   arabic-1-column
  96                   arabic-2-column)))
  97   (while charsets
  98     (modify-syntax-entry (make-char (car charsets)) "w")
  99     (modify-category-entry (make-char (car charsets)) ?b)
 100     (setq charsets (cdr charsets))))
 101
 102 ;; Chinese character set (GB2312)
 103
 104 (modify-syntax-entry (make-char 'chinese-gb2312) "w")
 105 (modify-syntax-entry (make-char 'chinese-gb2312 33) "_")
 106 (modify-syntax-entry (make-char 'chinese-gb2312 34) "_")
 107 (modify-syntax-entry (make-char 'chinese-gb2312 41) "_")
 108 (modify-syntax-entry ?\\e$A!2\e(B "(\e$A!3\e(B")
 109 (modify-syntax-entry ?\\e$A!4\e(B "(\e$A!5\e(B")
 110 (modify-syntax-entry ?\\e$A!6\e(B "(\e$A!7\e(B")
 111 (modify-syntax-entry ?\\e$A!8\e(B "(\e$A!9\e(B")
 112 (modify-syntax-entry ?\\e$A!:\e(B "(\e$A!;\e(B")
 113 (modify-syntax-entry ?\\e$A!<\e(B "(\e$A!=\e(B")
 114 (modify-syntax-entry ?\\e$A!>\e(B "(\e$A!?\e(B")
 115 (modify-syntax-entry ?\\e$A!3\e(B ")\e$A!2\e(B")
 116 (modify-syntax-entry ?\\e$A!5\e(B ")\e$A!4\e(B")
 117 (modify-syntax-entry ?\\e$A!7\e(B ")\e$A!6\e(B")
 118 (modify-syntax-entry ?\\e$A!9\e(B ")\e$A!8\e(B")
 119 (modify-syntax-entry ?\\e$A!;\e(B ")\e$A!:\e(B")
 120 (modify-syntax-entry ?\\e$A!=\e(B ")\e$A!<\e(B")
 121 (modify-syntax-entry ?\\e$A!?\e(B ")\e$A!>\e(B")
 122
 123 (modify-category-entry (make-char 'chinese-gb2312) ?c)
 124 (modify-category-entry (make-char 'chinese-gb2312) ?\|)
 125 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 126 (modify-category-entry (make-char 'chinese-gb2312 36) ?H)
 127 (modify-category-entry (make-char 'chinese-gb2312 37) ?K)
 128 (modify-category-entry (make-char 'chinese-gb2312 38) ?G)
 129 (modify-category-entry (make-char 'chinese-gb2312 39) ?Y)
 130 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 131 (let ((row 48))
 132   (while (< row 127)
 133     (modify-category-entry (make-char 'chinese-gb2312 row) ?C)
 134     (setq row (1+ row))))
 135
 136 ;; Chinese character set (BIG5)
 137
 138 (let ((generic-big5-1-char (make-char 'chinese-big5-1))
 139       (generic-big5-2-char (make-char 'chinese-big5-2)))
 140   (modify-syntax-entry generic-big5-1-char "w")
 141   (modify-syntax-entry generic-big5-2-char "w")
 142
 143   (modify-category-entry generic-big5-1-char ?c)
 144   (modify-category-entry generic-big5-2-char ?c)
 145
 146   (modify-category-entry generic-big5-1-char ?C)
 147   (modify-category-entry generic-big5-2-char ?C)
 148
 149   (modify-category-entry generic-big5-1-char ?\|)
 150   (modify-category-entry generic-big5-2-char ?\|))
 151
 152
 153 ;; Chinese character set (CNS11643)
 154
 155 (let ((cns-list '(chinese-cns11643-1
 156                   chinese-cns11643-2
 157                   chinese-cns11643-3
 158                   chinese-cns11643-4
 159                   chinese-cns11643-5
 160                   chinese-cns11643-6
 161                   chinese-cns11643-7))
 162       generic-char)
 163   (while cns-list
 164     (setq generic-char (make-char (car cns-list)))
 165     (modify-syntax-entry generic-char "w")
 166     (modify-category-entry generic-char ?c)
 167     (modify-category-entry generic-char ?C)
 168     (modify-category-entry generic-char ?|)
 169     (setq cns-list (cdr cns-list))))
 170
 171 ;; Cyrillic character set (ISO-8859-5)
 172
 173 (modify-category-entry (make-char 'cyrillic-iso8859-5) ?y)
 174
 175 (let ((c 160))
 176   (while (< c 256)
 177     (modify-syntax-entry (make-char 'cyrillic-iso8859-5 c) "w")
 178     (setq c (1+ c))))
 179 (modify-syntax-entry ?\e,L-\e(B ".")
 180 (modify-syntax-entry ?\e,Lp\e(B ".")
 181 (modify-syntax-entry ?\e,L}\e(B ".")
 182
 183 ;; Ethiopic character set
 184
 185 (modify-category-entry (make-char 'ethiopic) ?e)
 186
 187 ;; European character set (Latin-1,2,3,4,5)
 188
 189 (modify-category-entry (make-char 'latin-iso8859-1) ?l)
 190 (modify-category-entry (make-char 'latin-iso8859-2) ?l)
 191 (modify-category-entry (make-char 'latin-iso8859-3) ?l)
 192 (modify-category-entry (make-char 'latin-iso8859-4) ?l)
 193 (modify-category-entry (make-char 'latin-iso8859-9) ?l)
 194
 195 ;; ISO-8859-1 (Latin-1)
 196 (let ((c 64))
 197   (while (< c 128)                      ; from '\e,A@\e(B' to '\e,A\7f\e(B'
 198     (modify-syntax-entry (make-char 'latin-iso8859-1 c) "w")
 199     (setq c (1+ c)))
 200   (modify-syntax-entry (make-char 'latin-iso8859-1 32) "w") ; NBSP
 201   (modify-syntax-entry ?\e,AW\e(B "_")
 202   (modify-syntax-entry ?\e,Aw\e(B "_")
 203   )
 204
 205 ;; ISO-8859-2 (Latin-2)
 206 (let ((c 190))
 207   (while (< c 255)
 208     (modify-syntax-entry (make-char 'latin-iso8859-2 c) "w")
 209     (setq c (1+ c))))
 210 (let ((chars '(?\e,B!\e(B ?\e,B#\e(B ?\e,B%\e(B ?\e,B&\e(B ?\e,B)\e(B ?\e,B*\e(B ?\e,B+\e(B ?\e,B,\e(B ?\e,B.\e(B ?\e,B/\e(B ?\e,B1\e(B ?\e,B3\e(B ?\e,B5\e(B ?\e,B6\e(B ?\e,B9\e(B ?\e,B:\e(B ?\e,B;\e(B ?\e,B<\e(B)))
 211   (while chars
 212     (modify-syntax-entry (car chars) "w")
 213     (setq chars (cdr chars))))
 214 (modify-syntax-entry (make-char 'latin-iso8859-2 160) "w") ; NBSP
 215 (modify-syntax-entry ?\e,BW\e(B ".")
 216 (modify-syntax-entry ?\e,Bw\e(B ".")
 217
 218 ;; Greek character set (ISO-8859-7)
 219
 220 (modify-category-entry (make-char 'greek-iso8859-7) ?g)
 221
 222 (let ((c 182))
 223   (while (< c 255)
 224     (modify-syntax-entry (make-char 'greek-iso8859-7 c) "w")
 225     (setq c (1+ c))))
 226 (modify-syntax-entry (make-char 'greek-iso8859-7 160) "w") ; NBSP
 227 (modify-syntax-entry ?\e,F7\e(B ".")
 228 (modify-syntax-entry ?\e,F;\e(B ".")
 229 (modify-syntax-entry ?\e,F=\e(B ".")
 230
 231 ;; Hebrew character set (ISO-8859-8)
 232
 233 (modify-category-entry (make-char 'hebrew-iso8859-8) ?w)
 234
 235 (let ((c 224))
 236   (while (< c 251)
 237     (modify-syntax-entry (make-char 'hebrew-iso8859-8 c) "w")
 238     (setq c (1+ c))))
 239 (modify-syntax-entry (make-char 'hebrew-iso8859-8 160) "w") ; NBSP
 240
 241 ;; Indian character set (IS 13194 and other Emacs original Indian charsets)
 242
 243 (modify-category-entry (make-char 'indian-is13194) ?i)
 244 (modify-category-entry (make-char 'indian-2-column) ?I)
 245 (modify-category-entry (make-char 'indian-1-column) ?I)
 246
 247 ;; Japanese character set (JISX0201-kana, JISX0201-roman, JISX0208, JISX0212)
 248
 249 (modify-category-entry (make-char 'katakana-jisx0201) ?k)
 250 (modify-category-entry (make-char 'latin-jisx0201) ?r)
 251 (modify-category-entry (make-char 'japanese-jisx0208) ?j)
 252 (modify-category-entry (make-char 'japanese-jisx0212) ?j)
 253 (modify-category-entry (make-char 'japanese-jisx0208) ?\|)
 254
 255 ;; JISX0208
 256 (modify-syntax-entry (make-char 'japanese-jisx0208) "w")
 257 (modify-syntax-entry (make-char 'japanese-jisx0208 33) "_")
 258 (modify-syntax-entry (make-char 'japanese-jisx0208 34) "_")
 259 (modify-syntax-entry (make-char 'japanese-jisx0208 40) "_")
 260 (let ((chars '(?\e$B!<\e(B ?\e$B!+\e(B ?\e$B!,\e(B ?\e$B!3\e(B ?\e$B!4\e(B ?\e$B!5\e(B ?\e$B!6\e(B ?\e$B!7\e(B ?\e$B!8\e(B ?\e$B!9\e(B ?\e$B!:\e(B ?\e$B!;\e(B)))
 261   (while chars
 262     (modify-syntax-entry (car chars) "w")
 263     (setq chars (cdr chars))))
 264 (modify-syntax-entry ?\\e$B!J\e(B "(\e$B!K\e(B")
 265 (modify-syntax-entry ?\\e$B!N\e(B "(\e$B!O\e(B")
 266 (modify-syntax-entry ?\\e$B!P\e(B "(\e$B!Q\e(B")
 267 (modify-syntax-entry ?\\e$B!V\e(B "(\e$B!W\e(B")
 268 (modify-syntax-entry ?\\e$B!X\e(B "(\e$B!Y\e(B")
 269 (modify-syntax-entry ?\\e$B!K\e(B ")\e$B!J\e(B")
 270 (modify-syntax-entry ?\\e$B!O\e(B ")\e$B!N\e(B")
 271 (modify-syntax-entry ?\\e$B!Q\e(B ")\e$B!P\e(B")
 272 (modify-syntax-entry ?\\e$B!W\e(B ")\e$B!V\e(B")
 273 (modify-syntax-entry ?\\e$B!Y\e(B ")\e$B!X\e(B")
 274
 275 (modify-category-entry (make-char 'japanese-jisx0208 35) ?A)
 276 (modify-category-entry (make-char 'japanese-jisx0208 36) ?H)
 277 (modify-category-entry (make-char 'japanese-jisx0208 37) ?K)
 278 (modify-category-entry (make-char 'japanese-jisx0208 38) ?G)
 279 (modify-category-entry (make-char 'japanese-jisx0208 39) ?Y)
 280 (let ((row 48))
 281   (while (< row 127)
 282     (modify-category-entry (make-char 'japanese-jisx0208 row) ?C)
 283     (setq row (1+ row))))
 284 (let ((chars '(?\e$B!<\e(B ?\e$B!+\e(B ?\e$B!,\e(B)))
 285   (while chars
 286     (modify-category-entry (car chars) ?K)
 287     (modify-category-entry (car chars) ?H)
 288     (setq chars (cdr chars))))
 289 (let ((chars '(?\e$B!3\e(B ?\e$B!4\e(B ?\e$B!5\e(B ?\e$B!6\e(B ?\e$B!7\e(B ?\e$B!8\e(B ?\e$B!9\e(B ?\e$B!:\e(B ?\e$B!;\e(B)))
 290   (while chars
 291     (modify-category-entry (car chars) ?C)
 292     (setq chars (cdr chars))))
 293
 294 ;; JISX0212
 295 (modify-syntax-entry (make-char 'japanese-jisx0212) "w")
 296 (modify-syntax-entry (make-char 'japanese-jisx0212 33) "_")
 297 (modify-syntax-entry (make-char 'japanese-jisx0212 34) "_")
 298 (modify-syntax-entry (make-char 'japanese-jisx0212 35) "_")
 299
 300 (modify-category-entry (make-char 'japanese-jisx0212 ) ?C)
 301
 302 ;; JISX0201-Kana
 303 (modify-syntax-entry (make-char 'katakana-jisx0201) "w")
 304 (let ((chars '(?\e(I!\e(B ?\e(I"\e(B ?\e(I#\e(B ?\e(I$\e(B ?\e(I%\e(B)))
 305   (while chars
 306     (modify-syntax-entry (car chars) ".")
 307     (setq chars (cdr chars))))
 308
 309 ;; Korean character set (KSC5601)
 310
 311 (modify-syntax-entry (make-char 'korean-ksc5601) "w")
 312 (modify-syntax-entry (make-char 'korean-ksc5601 33) "_")
 313 (modify-syntax-entry (make-char 'korean-ksc5601 34) "_")
 314 (modify-syntax-entry (make-char 'korean-ksc5601 38) "_")
 315 (modify-syntax-entry (make-char 'korean-ksc5601 39) "_")
 316 (modify-syntax-entry (make-char 'korean-ksc5601 40) "_")
 317 (modify-syntax-entry (make-char 'korean-ksc5601 41) "_")
 318
 319 (modify-category-entry (make-char 'korean-ksc5601) ?h)
 320 (modify-category-entry (make-char 'korean-ksc5601 35) ?A)
 321 (modify-category-entry (make-char 'korean-ksc5601 37) ?G)
 322 (modify-category-entry (make-char 'korean-ksc5601 42) ?H)
 323 (modify-category-entry (make-char 'korean-ksc5601 43) ?K)
 324 (modify-category-entry (make-char 'korean-ksc5601 44) ?Y)
 325
 326 ;; Thai character set (TIS620)
 327
 328 (modify-category-entry (make-char 'thai-tis620) ?t)
 329
 330 (let ((deflist  '(;; chars      syntax  category
 331                   ("\e,T!\e(B-\e,TCEG\e(B-\e,TN\e(B"  "w"     ?0) ; consonant
 332                   ("\e,TDFPRS`\e(B-\e,Te\e(B"       "w"     ?1) ; vowel base
 333                   ("\e,TQT\e(B-\e,TWgn\e(B" "w"     ?2) ; vowel upper
 334                   ("\e,TX\e(B-\e,TZ\e(B"    "w"     ?3) ; vowel lower
 335                   ("\e,Th\e(B-\e,Tm\e(B"    "w"     ?4) ; tone mark
 336                   ("\e,TOfp\e(B-\e,Ty\e(B"  "w"     ?0) ; digit and misc
 337                   ("\e,T_oz{\e(B" "_"     ?0) ; symbol
 338                   ))
 339       elm chars len syntax category to ch i)
 340   (while deflist
 341     (setq elm (car deflist))
 342     (setq chars (car elm)
 343           len (length chars)
 344           syntax (nth 1 elm)
 345           category (nth 2 elm)
 346           i 0)
 347     (while (< i len)
 348       (if (= (aref chars i) ?-)
 349           (setq i (1+ i)
 350                 to (sref chars i))
 351         (setq ch (sref chars i)
 352               to ch))
 353       (while (<= ch to)
 354         (modify-syntax-entry ch syntax)
 355         (modify-category-entry ch category)
 356         (setq ch (1+ ch)))
 357       (setq i (+ i (char-bytes to))))
 358     (setq deflist (cdr deflist))))
 359
 360 ;; Vietnamese character set
 361
 362 (let ((lower (make-char 'vietnamese-viscii-lower))
 363       (upper (make-char 'vietnamese-viscii-upper)))
 364   (modify-syntax-entry lower "w")
 365   (modify-syntax-entry upper "w")
 366   (modify-category-entry lower ?v)
 367   (modify-category-entry upper ?v)
 368   (modify-category-entry lower ?l)      ; To make a word with
 369   (modify-category-entry upper ?l)      ; latin characters.
 370   )
 371
 372 \f
 373 ;;; Setting word boundary.
 374
 375 (setq word-combining-categories
 376       '((?l . ?l)))
 377
 378 (setq word-separating-categories        ;  (2-byte character sets)
 379       '((?A . ?K)                       ; Alpha numeric - Katakana
 380         (?A . ?C)                       ; Alpha numeric - Chinese
 381         (?H . ?A)                       ; Hiragana - Alpha numeric
 382         (?H . ?K)                       ; Hiragana - Katakana
 383         (?H . ?C)                       ; Hiragana - Chinese
 384         (?K . ?A)                       ; Katakana - Alpha numeric
 385         (?K . ?C)                       ; Katakana - Chinese
 386         (?C . ?A)                       ; Chinese - Alpha numeric
 387         (?C . ?K)                       ; Chinese - Katakana
 388         ))