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 the
  22 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  23 ;; Boston, MA 02111-1307, USA.
  24
  25 ;;; Commentary:
  26
  27 ;; This file contains multibyte characters.  Save this file always in
  28 ;; the coding system `iso-2022-7'.
  29
  30 ;;; Predefined categories.
  31
  32 ;; For each character set.
  33
  34 (define-category ?a "ASCII")
  35 (define-category ?l "Latin")
  36 (define-category ?t "Thai")
  37 (define-category ?g "Greek")
  38 (define-category ?b "Arabic")
  39 (define-category ?w "Hebrew")
  40 (define-category ?y "Cyrillic")
  41 (define-category ?k "Japanese katakana")
  42 (define-category ?r "Japanese roman")
  43 (define-category ?c "Chinese")
  44 (define-category ?j "Japanese")
  45 (define-category ?h "Korean")
  46 (define-category ?e "Ethiopic (Ge'ez)")
  47 (define-category ?v "Vietnamese")
  48 (define-category ?i "Indian")
  49 (define-category ?o "Lao")
  50
  51 ;; For each group (row) of 2-byte character sets.
  52
  53 (define-category ?A "Alpha numeric characters of 2-byte character sets")
  54 (define-category ?C "Chinese (Han) characters of 2-byte character sets")
  55 (define-category ?G "Greek characters of 2-byte characters sets")
  56 (define-category ?H "Japanese Hiragana characters of 2-byte character sets")
  57 (define-category ?K "Japanese Katakana characters of 2-byte character sets")
  58 (define-category ?N "Korean Hangul characters of 2-byte character sets")
  59 (define-category ?Y "Cyrillic character of 2-byte character sets")
  60 (define-category ?I "Indian Glyphs")
  61
  62 ;; For phonetic classifications.
  63
  64 (define-category ?0 "consonant")
  65 (define-category ?1 "base vowel")
  66 (define-category ?2 "upper diacritical mark (including upper vowel)")
  67 (define-category ?3 "lower diacritical mark (including lower vowel)")
  68 (define-category ?4 "tone mark")
  69 (define-category ?5 "vowel")
  70 (define-category ?6 "digit")
  71 (define-category ?7 "vowel-modifying diacritical mark")
  72 (define-category ?8 "vowel-signs")
  73 (define-category ?9 "semivowel lower")
  74
  75 ;; For filling.
  76 (define-category ?| "While filling, we can break a line at this character.")
  77
  78 ;; Keep the followings for `kinsoku' processing.  See comments in
  79 ;; kinsoku.el.
  80 (define-category ?> "A character which can't be placed at beginning of line.")
  81 (define-category ?< "A character which can't be placed at end of line.")
  82
  83 \f
  84 ;;; Setting syntax and category.
  85
  86 ;; ASCII
  87
  88 (let ((ch 32))
  89   (while (< ch 127)                     ; All ASCII characters have
  90     (modify-category-entry ch ?a)       ; the category `a' (ASCII)
  91     (modify-category-entry ch ?l)       ; and `l' (Latin).
  92     (setq ch (1+ ch))))
  93
  94 ;; Arabic character set
  95
  96 (let ((charsets '(arabic-iso8859-6
  97                   arabic-digit
  98                   arabic-1-column
  99                   arabic-2-column)))
 100   (while charsets
 101     (modify-syntax-entry (make-char (car charsets)) "w")
 102     (modify-category-entry (make-char (car charsets)) ?b)
 103     (setq charsets (cdr charsets))))
 104
 105 ;; Chinese character set (GB2312)
 106
 107 (modify-syntax-entry (make-char 'chinese-gb2312) "w")
 108 (modify-syntax-entry (make-char 'chinese-gb2312 33) "_")
 109 (modify-syntax-entry (make-char 'chinese-gb2312 34) "_")
 110 (modify-syntax-entry (make-char 'chinese-gb2312 41) "_")
 111 (modify-syntax-entry ?\\e$A!2\e(B "(\e$A!3\e(B")
 112 (modify-syntax-entry ?\\e$A!4\e(B "(\e$A!5\e(B")
 113 (modify-syntax-entry ?\\e$A!6\e(B "(\e$A!7\e(B")
 114 (modify-syntax-entry ?\\e$A!8\e(B "(\e$A!9\e(B")
 115 (modify-syntax-entry ?\\e$A!:\e(B "(\e$A!;\e(B")
 116 (modify-syntax-entry ?\\e$A!<\e(B "(\e$A!=\e(B")
 117 (modify-syntax-entry ?\\e$A!>\e(B "(\e$A!?\e(B")
 118 (modify-syntax-entry ?\\e$A!3\e(B ")\e$A!2\e(B")
 119 (modify-syntax-entry ?\\e$A!5\e(B ")\e$A!4\e(B")
 120 (modify-syntax-entry ?\\e$A!7\e(B ")\e$A!6\e(B")
 121 (modify-syntax-entry ?\\e$A!9\e(B ")\e$A!8\e(B")
 122 (modify-syntax-entry ?\\e$A!;\e(B ")\e$A!:\e(B")
 123 (modify-syntax-entry ?\\e$A!=\e(B ")\e$A!<\e(B")
 124 (modify-syntax-entry ?\\e$A!?\e(B ")\e$A!>\e(B")
 125
 126 (modify-category-entry (make-char 'chinese-gb2312) ?c)
 127 (modify-category-entry (make-char 'chinese-gb2312) ?\|)
 128 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 129 (modify-category-entry (make-char 'chinese-gb2312 36) ?H)
 130 (modify-category-entry (make-char 'chinese-gb2312 37) ?K)
 131 (modify-category-entry (make-char 'chinese-gb2312 38) ?G)
 132 (modify-category-entry (make-char 'chinese-gb2312 39) ?Y)
 133 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 134 (let ((row 48))
 135   (while (< row 127)
 136     (modify-category-entry (make-char 'chinese-gb2312 row) ?C)
 137     (setq row (1+ row))))
 138
 139 ;; Chinese character set (BIG5)
 140
 141 (let ((generic-big5-1-char (make-char 'chinese-big5-1))
 142       (generic-big5-2-char (make-char 'chinese-big5-2)))
 143   (modify-syntax-entry generic-big5-1-char "w")
 144   (modify-syntax-entry generic-big5-2-char "w")
 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 ?C)
 150   (modify-category-entry generic-big5-2-char ?C)
 151
 152   (modify-category-entry generic-big5-1-char ?\|)
 153   (modify-category-entry generic-big5-2-char ?\|))
 154
 155
 156 ;; Chinese character set (CNS11643)
 157
 158 (let ((cns-list '(chinese-cns11643-1
 159                   chinese-cns11643-2
 160                   chinese-cns11643-3
 161                   chinese-cns11643-4
 162                   chinese-cns11643-5
 163                   chinese-cns11643-6
 164                   chinese-cns11643-7))
 165       generic-char)
 166   (while cns-list
 167     (setq generic-char (make-char (car cns-list)))
 168     (modify-syntax-entry generic-char "w")
 169     (modify-category-entry generic-char ?c)
 170     (modify-category-entry generic-char ?C)
 171     (modify-category-entry generic-char ?|)
 172     (setq cns-list (cdr cns-list))))
 173
 174 ;; Cyrillic character set (ISO-8859-5)
 175
 176 (modify-category-entry (make-char 'cyrillic-iso8859-5) ?y)
 177
 178 (let ((c 160))
 179   (while (< c 256)
 180     (modify-syntax-entry (make-char 'cyrillic-iso8859-5 c) "w")
 181     (setq c (1+ c))))
 182 (modify-syntax-entry ?\e,L-\e(B ".")
 183 (modify-syntax-entry ?\e,Lp\e(B ".")
 184 (modify-syntax-entry ?\e,L}\e(B ".")
 185
 186 ;; Ethiopic character set
 187
 188 (modify-category-entry (make-char 'ethiopic) ?e)
 189
 190 ;; European character set (Latin-1,2,3,4,5)
 191
 192 (modify-category-entry (make-char 'latin-iso8859-1) ?l)
 193 (modify-category-entry (make-char 'latin-iso8859-2) ?l)
 194 (modify-category-entry (make-char 'latin-iso8859-3) ?l)
 195 (modify-category-entry (make-char 'latin-iso8859-4) ?l)
 196 (modify-category-entry (make-char 'latin-iso8859-9) ?l)
 197
 198 ;; ISO-8859-1 (Latin-1)
 199 (let ((c 64))
 200   (while (< c 128)                      ; from '\e,A@\e(B' to '\e,A\7f\e(B'
 201     (modify-syntax-entry (make-char 'latin-iso8859-1 c) "w")
 202     (setq c (1+ c)))
 203   (modify-syntax-entry (make-char 'latin-iso8859-1 32) "w") ; NBSP
 204   (modify-syntax-entry ?\e,AW\e(B "_")
 205   (modify-syntax-entry ?\e,Aw\e(B "_")
 206   )
 207
 208 ;; ISO-8859-2 (Latin-2)
 209 (let ((c 190))
 210   (while (< c 255)
 211     (modify-syntax-entry (make-char 'latin-iso8859-2 c) "w")
 212     (setq c (1+ c))))
 213 (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)))
 214   (while chars
 215     (modify-syntax-entry (car chars) "w")
 216     (setq chars (cdr chars))))
 217 (modify-syntax-entry (make-char 'latin-iso8859-2 160) "w") ; NBSP
 218 (modify-syntax-entry ?\e,BW\e(B ".")
 219 (modify-syntax-entry ?\e,Bw\e(B ".")
 220
 221 ;; Greek character set (ISO-8859-7)
 222
 223 (modify-category-entry (make-char 'greek-iso8859-7) ?g)
 224
 225 (let ((c 182))
 226   (while (< c 255)
 227     (modify-syntax-entry (make-char 'greek-iso8859-7 c) "w")
 228     (setq c (1+ c))))
 229 (modify-syntax-entry (make-char 'greek-iso8859-7 160) "w") ; NBSP
 230 (modify-syntax-entry ?\e,F7\e(B ".")
 231 (modify-syntax-entry ?\e,F;\e(B ".")
 232 (modify-syntax-entry ?\e,F=\e(B ".")
 233
 234 ;; Hebrew character set (ISO-8859-8)
 235
 236 (modify-category-entry (make-char 'hebrew-iso8859-8) ?w)
 237
 238 (let ((c 224))
 239   (while (< c 251)
 240     (modify-syntax-entry (make-char 'hebrew-iso8859-8 c) "w")
 241     (setq c (1+ c))))
 242 (modify-syntax-entry (make-char 'hebrew-iso8859-8 160) "w") ; NBSP
 243
 244 ;; Indian character set (IS 13194 and other Emacs original Indian charsets)
 245
 246 (modify-category-entry (make-char 'indian-is13194) ?i)
 247 (modify-category-entry (make-char 'indian-2-column) ?I)
 248 (modify-category-entry (make-char 'indian-1-column) ?I)
 249
 250 ;; Japanese character set (JISX0201-kana, JISX0201-roman, JISX0208, JISX0212)
 251
 252 (modify-category-entry (make-char 'katakana-jisx0201) ?k)
 253 (modify-category-entry (make-char 'latin-jisx0201) ?r)
 254 (modify-category-entry (make-char 'japanese-jisx0208) ?j)
 255 (modify-category-entry (make-char 'japanese-jisx0212) ?j)
 256 (modify-category-entry (make-char 'japanese-jisx0208) ?\|)
 257
 258 ;; JISX0208
 259 (modify-syntax-entry (make-char 'japanese-jisx0208) "w")
 260 (modify-syntax-entry (make-char 'japanese-jisx0208 33) "_")
 261 (modify-syntax-entry (make-char 'japanese-jisx0208 34) "_")
 262 (modify-syntax-entry (make-char 'japanese-jisx0208 40) "_")
 263 (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)))
 264   (while chars
 265     (modify-syntax-entry (car chars) "w")
 266     (setq chars (cdr chars))))
 267 (modify-syntax-entry ?\\e$B!J\e(B "(\e$B!K\e(B")
 268 (modify-syntax-entry ?\\e$B!N\e(B "(\e$B!O\e(B")
 269 (modify-syntax-entry ?\\e$B!P\e(B "(\e$B!Q\e(B")
 270 (modify-syntax-entry ?\\e$B!V\e(B "(\e$B!W\e(B")
 271 (modify-syntax-entry ?\\e$B!X\e(B "(\e$B!Y\e(B")
 272 (modify-syntax-entry ?\\e$B!K\e(B ")\e$B!J\e(B")
 273 (modify-syntax-entry ?\\e$B!O\e(B ")\e$B!N\e(B")
 274 (modify-syntax-entry ?\\e$B!Q\e(B ")\e$B!P\e(B")
 275 (modify-syntax-entry ?\\e$B!W\e(B ")\e$B!V\e(B")
 276 (modify-syntax-entry ?\\e$B!Y\e(B ")\e$B!X\e(B")
 277
 278 (modify-category-entry (make-char 'japanese-jisx0208 35) ?A)
 279 (modify-category-entry (make-char 'japanese-jisx0208 36) ?H)
 280 (modify-category-entry (make-char 'japanese-jisx0208 37) ?K)
 281 (modify-category-entry (make-char 'japanese-jisx0208 38) ?G)
 282 (modify-category-entry (make-char 'japanese-jisx0208 39) ?Y)
 283 (let ((row 48))
 284   (while (< row 127)
 285     (modify-category-entry (make-char 'japanese-jisx0208 row) ?C)
 286     (setq row (1+ row))))
 287 (let ((chars '(?\e$B!<\e(B ?\e$B!+\e(B ?\e$B!,\e(B)))
 288   (while chars
 289     (modify-category-entry (car chars) ?K)
 290     (modify-category-entry (car chars) ?H)
 291     (setq chars (cdr chars))))
 292 (let ((chars '(?\e$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)))
 293   (while chars
 294     (modify-category-entry (car chars) ?C)
 295     (setq chars (cdr chars))))
 296
 297 ;; JISX0212
 298 (modify-syntax-entry (make-char 'japanese-jisx0212) "w")
 299 (modify-syntax-entry (make-char 'japanese-jisx0212 33) "_")
 300 (modify-syntax-entry (make-char 'japanese-jisx0212 34) "_")
 301 (modify-syntax-entry (make-char 'japanese-jisx0212 35) "_")
 302
 303 (modify-category-entry (make-char 'japanese-jisx0212 ) ?C)
 304
 305 ;; JISX0201-Kana
 306 (modify-syntax-entry (make-char 'katakana-jisx0201) "w")
 307 (let ((chars '(?\e(I!\e(B ?\e(I"\e(B ?\e(I#\e(B ?\e(I$\e(B ?\e(I%\e(B)))
 308   (while chars
 309     (modify-syntax-entry (car chars) ".")
 310     (setq chars (cdr chars))))
 311
 312 ;; Korean character set (KSC5601)
 313
 314 (modify-syntax-entry (make-char 'korean-ksc5601) "w")
 315 (modify-syntax-entry (make-char 'korean-ksc5601 33) "_")
 316 (modify-syntax-entry (make-char 'korean-ksc5601 34) "_")
 317 (modify-syntax-entry (make-char 'korean-ksc5601 38) "_")
 318 (modify-syntax-entry (make-char 'korean-ksc5601 39) "_")
 319 (modify-syntax-entry (make-char 'korean-ksc5601 40) "_")
 320 (modify-syntax-entry (make-char 'korean-ksc5601 41) "_")
 321
 322 (modify-category-entry (make-char 'korean-ksc5601) ?h)
 323 (modify-category-entry (make-char 'korean-ksc5601 35) ?A)
 324 (modify-category-entry (make-char 'korean-ksc5601 37) ?G)
 325 (modify-category-entry (make-char 'korean-ksc5601 42) ?H)
 326 (modify-category-entry (make-char 'korean-ksc5601 43) ?K)
 327 (modify-category-entry (make-char 'korean-ksc5601 44) ?Y)
 328
 329 ;; Lao character set
 330
 331 (modify-category-entry (make-char 'lao) ?o)
 332
 333 (let ((deflist  '(;; chars      syntax  category
 334                   ("\e(1!\e(B-\e(1N\e(B"    "w"     ?0) ; consonant
 335                   ("\e(1PRS]`\e(B-\e(1d\e(B"        "w"     ?1) ; vowel base
 336                   ("\e(1QT\e(B-\e(1W[m\e(B" "w"     ?2) ; vowel upper
 337                   ("\e(1XY\e(B"           "w"     ?3) ; vowel lower
 338                   ("\e(1h\e(B-\e(1l\e(B"    "w"     ?4) ; tone mark
 339                   ("\e(1\\e(B"            "w"     ?9) ; semivowel lower
 340                   ("\e(1p\e(B-\e(1y\e(B"    "w"     ?0) ; digit and misc
 341                   ("\e(1Of\e(B"           "_"     ?0) ; symbol
 342                   ))
 343       elm chars len syntax category to ch i)
 344   (while deflist
 345     (setq elm (car deflist))
 346     (setq chars (car elm)
 347           len (length chars)
 348           syntax (nth 1 elm)
 349           category (nth 2 elm)
 350           i 0)
 351     (while (< i len)
 352       (if (= (aref chars i) ?-)
 353           (setq i (1+ i)
 354                 to (sref chars i))
 355         (setq ch (sref chars i)
 356               to ch))
 357       (while (<= ch to)
 358         (modify-syntax-entry ch syntax)
 359         (modify-category-entry ch category)
 360         (setq ch (1+ ch)))
 361       (setq i (+ i (char-bytes to))))
 362     (setq deflist (cdr deflist))))
 363
 364 ;; Thai character set (TIS620)
 365
 366 (modify-category-entry (make-char 'thai-tis620) ?t)
 367
 368 (let ((deflist  '(;; chars      syntax  category
 369                   ("\e,T!\e(B-\e,TCEG\e(B-\e,TN\e(B"  "w"     ?0) ; consonant
 370                   ("\e,TDFPRS`\e(B-\e,Te\e(B"       "w"     ?1) ; vowel base
 371                   ("\e,TQT\e(B-\e,TWgn\e(B" "w"     ?2) ; vowel upper
 372                   ("\e,TX\e(B-\e,TZ\e(B"    "w"     ?3) ; vowel lower
 373                   ("\e,Th\e(B-\e,Tm\e(B"    "w"     ?4) ; tone mark
 374                   ("\e,TOfp\e(B-\e,Ty\e(B"  "w"     ?0) ; digit and misc
 375                   ("\e,T_oz{\e(B" "_"     ?0) ; symbol
 376                   ))
 377       elm chars len syntax category to ch i)
 378   (while deflist
 379     (setq elm (car deflist))
 380     (setq chars (car elm)
 381           len (length chars)
 382           syntax (nth 1 elm)
 383           category (nth 2 elm)
 384           i 0)
 385     (while (< i len)
 386       (if (= (aref chars i) ?-)
 387           (setq i (1+ i)
 388                 to (sref chars i))
 389         (setq ch (sref chars i)
 390               to ch))
 391       (while (<= ch to)
 392         (modify-syntax-entry ch syntax)
 393         (modify-category-entry ch category)
 394         (setq ch (1+ ch)))
 395       (setq i (+ i (char-bytes to))))
 396     (setq deflist (cdr deflist))))
 397
 398 ;; Vietnamese character set
 399
 400 (let ((lower (make-char 'vietnamese-viscii-lower))
 401       (upper (make-char 'vietnamese-viscii-upper)))
 402   (modify-syntax-entry lower "w")
 403   (modify-syntax-entry upper "w")
 404   (modify-category-entry lower ?v)
 405   (modify-category-entry upper ?v)
 406   (modify-category-entry lower ?l)      ; To make a word with
 407   (modify-category-entry upper ?l)      ; latin characters.
 408   )
 409
 410 \f
 411 ;;; Setting word boundary.
 412
 413 (setq word-combining-categories
 414       '((?l . ?l)))
 415
 416 (setq word-separating-categories        ;  (2-byte character sets)
 417       '((?A . ?K)                       ; Alpha numeric - Katakana
 418         (?A . ?C)                       ; Alpha numeric - Chinese
 419         (?H . ?A)                       ; Hiragana - Alpha numeric
 420         (?H . ?K)                       ; Hiragana - Katakana
 421         (?H . ?C)                       ; Hiragana - Chinese
 422         (?K . ?A)                       ; Katakana - Alpha numeric
 423         (?K . ?C)                       ; Katakana - Chinese
 424         (?C . ?A)                       ; Chinese - Alpha numeric
 425         (?C . ?K)                       ; Chinese - Katakana
 426         ))