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 ;; `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
  50 ;; For each group (row) of 2-byte character sets.
  51
  52 (define-category ?A "Alpha numeric characters of 2-byte character sets")
  53 (define-category ?C "Chinese (Han) characters of 2-byte character sets")
  54 (define-category ?G "Greek characters of 2-byte characters sets")
  55 (define-category ?H "Japanese Hiragana characters of 2-byte character sets")
  56 (define-category ?K "Japanese Katakana characters of 2-byte character sets")
  57 (define-category ?N "Korean Hangul characters of 2-byte character sets")
  58 (define-category ?Y "Cyrillic character of 2-byte character sets")
  59 (define-category ?I "Indian Glyphs")
  60
  61 ;; For phonetic classifications.
  62
  63 (define-category ?0 "consonant")
  64 (define-category ?1 "base vowel")
  65 (define-category ?2 "upper diacritical mark (including upper vowel)")
  66 (define-category ?3 "lower diacritical mark (including lower vowel)")
  67 (define-category ?4 "tone mark")
  68 (define-category ?5 "vowel")
  69 (define-category ?6 "digit")
  70 (define-category ?7 "vowel-modifying diacritical mark")
  71 (define-category ?8 "vowel-signs.")
  72
  73 ;; For filling.
  74 (define-category ?| "While filling, we can break a line at this character.")
  75
  76 ;; Keep the followings for `kinsoku' processing.  See comments in
  77 ;; kinsoku.el.
  78 (define-category ?> "A character which can't be placed at beginning of line.")
  79 (define-category ?< "A character which can't be placed at end of line.")
  80
  81 \f
  82 ;;; Setting syntax and category.
  83
  84 ;; ASCII
  85
  86 (let ((ch 32))
  87   (while (< ch 127)                     ; All ASCII characters have
  88     (modify-category-entry ch ?a)       ; the category `a' (ASCII)
  89     (modify-category-entry ch ?l)       ; and `l' (Latin).
  90     (setq ch (1+ ch))))
  91
  92 ;; Arabic character set
  93
  94 (let ((charsets '(arabic-iso8859-6
  95                   arabic-digit
  96                   arabic-1-column
  97                   arabic-2-column)))
  98   (while charsets
  99     (modify-syntax-entry (make-char (car charsets)) "w")
 100     (modify-category-entry (make-char (car charsets)) ?b)
 101     (setq charsets (cdr charsets))))
 102
 103 ;; Chinese character set (GB2312)
 104
 105 (modify-syntax-entry (make-char 'chinese-gb2312) "w")
 106 (modify-syntax-entry (make-char 'chinese-gb2312 33) "_")
 107 (modify-syntax-entry (make-char 'chinese-gb2312 34) "_")
 108 (modify-syntax-entry (make-char 'chinese-gb2312 41) "_")
 109 (modify-syntax-entry ?\\e$A!2\e(B "(\e$A!3\e(B")
 110 (modify-syntax-entry ?\\e$A!4\e(B "(\e$A!5\e(B")
 111 (modify-syntax-entry ?\\e$A!6\e(B "(\e$A!7\e(B")
 112 (modify-syntax-entry ?\\e$A!8\e(B "(\e$A!9\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$A!>\e(B "(\e$A!?\e(B")
 116 (modify-syntax-entry ?\\e$A!3\e(B ")\e$A!2\e(B")
 117 (modify-syntax-entry ?\\e$A!5\e(B ")\e$A!4\e(B")
 118 (modify-syntax-entry ?\\e$A!7\e(B ")\e$A!6\e(B")
 119 (modify-syntax-entry ?\\e$A!9\e(B ")\e$A!8\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 (modify-syntax-entry ?\\e$A!?\e(B ")\e$A!>\e(B")
 123
 124 (modify-category-entry (make-char 'chinese-gb2312) ?c)
 125 (modify-category-entry (make-char 'chinese-gb2312) ?\|)
 126 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 127 (modify-category-entry (make-char 'chinese-gb2312 36) ?H)
 128 (modify-category-entry (make-char 'chinese-gb2312 37) ?K)
 129 (modify-category-entry (make-char 'chinese-gb2312 38) ?G)
 130 (modify-category-entry (make-char 'chinese-gb2312 39) ?Y)
 131 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 132 (let ((row 48))
 133   (while (< row 127)
 134     (modify-category-entry (make-char 'chinese-gb2312 row) ?C)
 135     (setq row (1+ row))))
 136
 137 ;; Chinese character set (BIG5)
 138
 139 (let ((generic-big5-1-char (make-char 'chinese-big5-1))
 140       (generic-big5-2-char (make-char 'chinese-big5-2)))
 141   (modify-syntax-entry generic-big5-1-char "w")
 142   (modify-syntax-entry generic-big5-2-char "w")
 143
 144   (modify-category-entry generic-big5-1-char ?c)
 145   (modify-category-entry generic-big5-2-char ?c)
 146
 147   (modify-category-entry generic-big5-1-char ?C)
 148   (modify-category-entry generic-big5-2-char ?C)
 149
 150   (modify-category-entry generic-big5-1-char ?\|)
 151   (modify-category-entry generic-big5-2-char ?\|))
 152
 153
 154 ;; Chinese character set (CNS11643)
 155
 156 (let ((cns-list '(chinese-cns11643-1
 157                   chinese-cns11643-2
 158                   chinese-cns11643-3
 159                   chinese-cns11643-4
 160                   chinese-cns11643-5
 161                   chinese-cns11643-6
 162                   chinese-cns11643-7))
 163       generic-char)
 164   (while cns-list
 165     (setq generic-char (make-char (car cns-list)))
 166     (modify-syntax-entry generic-char "w")
 167     (modify-category-entry generic-char ?c)
 168     (modify-category-entry generic-char ?C)
 169     (modify-category-entry generic-char ?|)
 170     (setq cns-list (cdr cns-list))))
 171
 172 ;; Cyrillic character set (ISO-8859-5)
 173
 174 (modify-category-entry (make-char 'cyrillic-iso8859-5) ?y)
 175
 176 (let ((c 160))
 177   (while (< c 256)
 178     (modify-syntax-entry (make-char 'cyrillic-iso8859-5 c) "w")
 179     (setq c (1+ c))))
 180 (modify-syntax-entry ?\e,L-\e(B ".")
 181 (modify-syntax-entry ?\e,Lp\e(B ".")
 182 (modify-syntax-entry ?\e,L}\e(B ".")
 183
 184 ;; Ethiopic character set
 185
 186 (modify-category-entry (make-char 'ethiopic) ?e)
 187
 188 ;; European character set (Latin-1,2,3,4,5)
 189
 190 (modify-category-entry (make-char 'latin-iso8859-1) ?l)
 191 (modify-category-entry (make-char 'latin-iso8859-2) ?l)
 192 (modify-category-entry (make-char 'latin-iso8859-3) ?l)
 193 (modify-category-entry (make-char 'latin-iso8859-4) ?l)
 194 (modify-category-entry (make-char 'latin-iso8859-9) ?l)
 195
 196 ;; ISO-8859-1 (Latin-1)
 197 (let ((c 64))
 198   (while (< c 128)                      ; from '\e,A@\e(B' to '\e,A\7f\e(B'
 199     (modify-syntax-entry (make-char 'latin-iso8859-1 c) "w")
 200     (setq c (1+ c)))
 201   (modify-syntax-entry (make-char 'latin-iso8859-1 32) "w") ; NBSP
 202   (modify-syntax-entry ?\e,AW\e(B "_")
 203   (modify-syntax-entry ?\e,Aw\e(B "_")
 204   )
 205
 206 ;; ISO-8859-2 (Latin-2)
 207 (let ((c 190))
 208   (while (< c 255)
 209     (modify-syntax-entry (make-char 'latin-iso8859-2 c) "w")
 210     (setq c (1+ c))))
 211 (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)))
 212   (while chars
 213     (modify-syntax-entry (car chars) "w")
 214     (setq chars (cdr chars))))
 215 (modify-syntax-entry (make-char 'latin-iso8859-2 160) "w") ; NBSP
 216 (modify-syntax-entry ?\e,BW\e(B ".")
 217 (modify-syntax-entry ?\e,Bw\e(B ".")
 218
 219 ;; Greek character set (ISO-8859-7)
 220
 221 (modify-category-entry (make-char 'greek-iso8859-7) ?g)
 222
 223 (let ((c 182))
 224   (while (< c 255)
 225     (modify-syntax-entry (make-char 'greek-iso8859-7 c) "w")
 226     (setq c (1+ c))))
 227 (modify-syntax-entry (make-char 'greek-iso8859-7 160) "w") ; NBSP
 228 (modify-syntax-entry ?\e,F7\e(B ".")
 229 (modify-syntax-entry ?\e,F;\e(B ".")
 230 (modify-syntax-entry ?\e,F=\e(B ".")
 231
 232 ;; Hebrew character set (ISO-8859-8)
 233
 234 (modify-category-entry (make-char 'hebrew-iso8859-8) ?w)
 235
 236 (let ((c 224))
 237   (while (< c 251)
 238     (modify-syntax-entry (make-char 'hebrew-iso8859-8 c) "w")
 239     (setq c (1+ c))))
 240 (modify-syntax-entry (make-char 'hebrew-iso8859-8 160) "w") ; NBSP
 241
 242 ;; Indian character set (IS 13194 and other Emacs original Indian charsets)
 243
 244 (modify-category-entry (make-char 'indian-is13194) ?i)
 245 (modify-category-entry (make-char 'indian-2-column) ?I)
 246 (modify-category-entry (make-char 'indian-1-column) ?I)
 247
 248 ;; Japanese character set (JISX0201-kana, JISX0201-roman, JISX0208, JISX0212)
 249
 250 (modify-category-entry (make-char 'katakana-jisx0201) ?k)
 251 (modify-category-entry (make-char 'latin-jisx0201) ?r)
 252 (modify-category-entry (make-char 'japanese-jisx0208) ?j)
 253 (modify-category-entry (make-char 'japanese-jisx0212) ?j)
 254 (modify-category-entry (make-char 'japanese-jisx0208) ?\|)
 255
 256 ;; JISX0208
 257 (modify-syntax-entry (make-char 'japanese-jisx0208) "w")
 258 (modify-syntax-entry (make-char 'japanese-jisx0208 33) "_")
 259 (modify-syntax-entry (make-char 'japanese-jisx0208 34) "_")
 260 (modify-syntax-entry (make-char 'japanese-jisx0208 40) "_")
 261 (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)))
 262   (while chars
 263     (modify-syntax-entry (car chars) "w")
 264     (setq chars (cdr chars))))
 265 (modify-syntax-entry ?\\e$B!J\e(B "(\e$B!K\e(B")
 266 (modify-syntax-entry ?\\e$B!N\e(B "(\e$B!O\e(B")
 267 (modify-syntax-entry ?\\e$B!P\e(B "(\e$B!Q\e(B")
 268 (modify-syntax-entry ?\\e$B!V\e(B "(\e$B!W\e(B")
 269 (modify-syntax-entry ?\\e$B!X\e(B "(\e$B!Y\e(B")
 270 (modify-syntax-entry ?\\e$B!K\e(B ")\e$B!J\e(B")
 271 (modify-syntax-entry ?\\e$B!O\e(B ")\e$B!N\e(B")
 272 (modify-syntax-entry ?\\e$B!Q\e(B ")\e$B!P\e(B")
 273 (modify-syntax-entry ?\\e$B!W\e(B ")\e$B!V\e(B")
 274 (modify-syntax-entry ?\\e$B!Y\e(B ")\e$B!X\e(B")
 275
 276 (modify-category-entry (make-char 'japanese-jisx0208 35) ?A)
 277 (modify-category-entry (make-char 'japanese-jisx0208 36) ?H)
 278 (modify-category-entry (make-char 'japanese-jisx0208 37) ?K)
 279 (modify-category-entry (make-char 'japanese-jisx0208 38) ?G)
 280 (modify-category-entry (make-char 'japanese-jisx0208 39) ?Y)
 281 (let ((row 48))
 282   (while (< row 127)
 283     (modify-category-entry (make-char 'japanese-jisx0208 row) ?C)
 284     (setq row (1+ row))))
 285 (let ((chars '(?\e$B!<\e(B ?\e$B!+\e(B ?\e$B!,\e(B)))
 286   (while chars
 287     (modify-category-entry (car chars) ?K)
 288     (modify-category-entry (car chars) ?H)
 289     (setq chars (cdr chars))))
 290 (let ((chars '(?\e$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)))
 291   (while chars
 292     (modify-category-entry (car chars) ?C)
 293     (setq chars (cdr chars))))
 294
 295 ;; JISX0212
 296 (modify-syntax-entry (make-char 'japanese-jisx0212) "w")
 297 (modify-syntax-entry (make-char 'japanese-jisx0212 33) "_")
 298 (modify-syntax-entry (make-char 'japanese-jisx0212 34) "_")
 299 (modify-syntax-entry (make-char 'japanese-jisx0212 35) "_")
 300
 301 (modify-category-entry (make-char 'japanese-jisx0212 ) ?C)
 302
 303 ;; JISX0201-Kana
 304 (modify-syntax-entry (make-char 'katakana-jisx0201) "w")
 305 (let ((chars '(?\e(I!\e(B ?\e(I"\e(B ?\e(I#\e(B ?\e(I$\e(B ?\e(I%\e(B)))
 306   (while chars
 307     (modify-syntax-entry (car chars) ".")
 308     (setq chars (cdr chars))))
 309
 310 ;; Korean character set (KSC5601)
 311
 312 (modify-syntax-entry (make-char 'korean-ksc5601) "w")
 313 (modify-syntax-entry (make-char 'korean-ksc5601 33) "_")
 314 (modify-syntax-entry (make-char 'korean-ksc5601 34) "_")
 315 (modify-syntax-entry (make-char 'korean-ksc5601 38) "_")
 316 (modify-syntax-entry (make-char 'korean-ksc5601 39) "_")
 317 (modify-syntax-entry (make-char 'korean-ksc5601 40) "_")
 318 (modify-syntax-entry (make-char 'korean-ksc5601 41) "_")
 319
 320 (modify-category-entry (make-char 'korean-ksc5601) ?h)
 321 (modify-category-entry (make-char 'korean-ksc5601 35) ?A)
 322 (modify-category-entry (make-char 'korean-ksc5601 37) ?G)
 323 (modify-category-entry (make-char 'korean-ksc5601 42) ?H)
 324 (modify-category-entry (make-char 'korean-ksc5601 43) ?K)
 325 (modify-category-entry (make-char 'korean-ksc5601 44) ?Y)
 326
 327 ;; Thai character set (TIS620)
 328
 329 (modify-category-entry (make-char 'thai-tis620) ?t)
 330
 331 (let ((deflist  '(;; chars      syntax  category
 332                   ("\e,T!\e(B-\e,TCEG\e(B-\e,TN\e(B"  "w"     ?0) ; consonant
 333                   ("\e,TDFPRS`\e(B-\e,Te\e(B"       "w"     ?1) ; vowel base
 334                   ("\e,TQT\e(B-\e,TWgn\e(B" "w"     ?2) ; vowel upper
 335                   ("\e,TX\e(B-\e,TZ\e(B"    "w"     ?3) ; vowel lower
 336                   ("\e,Th\e(B-\e,Tm\e(B"    "w"     ?4) ; tone mark
 337                   ("\e,TOfp\e(B-\e,Ty\e(B"  "w"     ?0) ; digit and misc
 338                   ("\e,T_oz{\e(B" "_"     ?0) ; symbol
 339                   ))
 340       elm chars len syntax category to ch i)
 341   (while deflist
 342     (setq elm (car deflist))
 343     (setq chars (car elm)
 344           len (length chars)
 345           syntax (nth 1 elm)
 346           category (nth 2 elm)
 347           i 0)
 348     (while (< i len)
 349       (if (= (aref chars i) ?-)
 350           (setq i (1+ i)
 351                 to (sref chars i))
 352         (setq ch (sref chars i)
 353               to ch))
 354       (while (<= ch to)
 355         (modify-syntax-entry ch syntax)
 356         (modify-category-entry ch category)
 357         (setq ch (1+ ch)))
 358       (setq i (+ i (char-bytes to))))
 359     (setq deflist (cdr deflist))))
 360
 361 ;; Vietnamese character set
 362
 363 (let ((lower (make-char 'vietnamese-viscii-lower))
 364       (upper (make-char 'vietnamese-viscii-upper)))
 365   (modify-syntax-entry lower "w")
 366   (modify-syntax-entry upper "w")
 367   (modify-category-entry lower ?v)
 368   (modify-category-entry upper ?v)
 369   (modify-category-entry lower ?l)      ; To make a word with
 370   (modify-category-entry upper ?l)      ; latin characters.
 371   )
 372
 373 \f
 374 ;;; Setting word boundary.
 375
 376 (setq word-combining-categories
 377       '((?l . ?l)))
 378
 379 (setq word-separating-categories        ;  (2-byte character sets)
 380       '((?A . ?K)                       ; Alpha numeric - Katakana
 381         (?A . ?C)                       ; Alpha numeric - Chinese
 382         (?H . ?A)                       ; Hiragana - Alpha numeric
 383         (?H . ?K)                       ; Hiragana - Katakana
 384         (?H . ?C)                       ; Hiragana - Chinese
 385         (?K . ?A)                       ; Katakana - Alpha numeric
 386         (?K . ?C)                       ; Katakana - Chinese
 387         (?C . ?A)                       ; Chinese - Alpha numeric
 388         (?C . ?K)                       ; Chinese - Katakana
 389         ))