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-2022-7bit'.
  29
  30 ;; This file does not define the syntax for Latin-N character sets;
  31 ;; those are defined by the files latin-N.el.
  32
  33 ;;; Predefined categories.
  34
  35 ;; For each character set.
  36
  37 (define-category ?a "ASCII")
  38 (define-category ?l "Latin")
  39 (define-category ?t "Thai")
  40 (define-category ?g "Greek")
  41 (define-category ?b "Arabic")
  42 (define-category ?w "Hebrew")
  43 (define-category ?y "Cyrillic")
  44 (define-category ?k "Japanese katakana")
  45 (define-category ?r "Japanese roman")
  46 (define-category ?c "Chinese")
  47 (define-category ?j "Japanese")
  48 (define-category ?h "Korean")
  49 (define-category ?e "Ethiopic (Ge'ez)")
  50 (define-category ?v "Vietnamese")
  51 (define-category ?i "Indian")
  52 (define-category ?o "Lao")
  53 (define-category ?q "Tibetan")
  54
  55 ;; For each group (row) of 2-byte character sets.
  56
  57 (define-category ?A "Alpha numeric characters of 2-byte character sets")
  58 (define-category ?C "Chinese (Han) characters of 2-byte character sets")
  59 (define-category ?G "Greek characters of 2-byte characters sets")
  60 (define-category ?H "Japanese Hiragana characters of 2-byte character sets")
  61 (define-category ?K "Japanese Katakana characters of 2-byte character sets")
  62 (define-category ?N "Korean Hangul characters of 2-byte character sets")
  63 (define-category ?Y "Cyrillic character of 2-byte character sets")
  64 (define-category ?I "Indian Glyphs")
  65
  66 ;; For phonetic classifications.
  67
  68 (define-category ?0 "consonant")
  69 (define-category ?1 "base vowel")
  70 (define-category ?2 "upper diacritical mark (including upper vowel)")
  71 (define-category ?3 "lower diacritical mark (including lower vowel)")
  72 (define-category ?4 "tone mark")
  73 (define-category ?5 "vowel")
  74 (define-category ?6 "digit")
  75 (define-category ?7 "vowel-modifying diacritical mark")
  76 (define-category ?8 "vowel-signs")
  77 (define-category ?9 "semivowel lower")
  78
  79 ;; For filling.
  80 (define-category ?| "While filling, we can break a line at this character.")
  81
  82 ;; Keep the followings for `kinsoku' processing.  See comments in
  83 ;; kinsoku.el.
  84 (define-category ?> "A character which can't be placed at beginning of line.")
  85 (define-category ?< "A character which can't be placed at end of line.")
  86
  87 \f
  88 ;;; Setting syntax and category.
  89
  90 ;; ASCII
  91
  92 (let ((ch 32))
  93   (while (< ch 127)                     ; All ASCII characters have
  94     (modify-category-entry ch ?a)       ; the category `a' (ASCII)
  95     (modify-category-entry ch ?l)       ; and `l' (Latin).
  96     (setq ch (1+ ch))))
  97
  98 ;; Arabic character set
  99
 100 (let ((charsets '(arabic-iso8859-6
 101                   arabic-digit
 102                   arabic-1-column
 103                   arabic-2-column)))
 104   (while charsets
 105     (modify-syntax-entry (make-char (car charsets)) "w")
 106     (modify-category-entry (make-char (car charsets)) ?b)
 107     (setq charsets (cdr charsets))))
 108
 109 ;; Chinese character set (GB2312)
 110
 111 (modify-syntax-entry (make-char 'chinese-gb2312) "w")
 112 (modify-syntax-entry (make-char 'chinese-gb2312 33) "_")
 113 (modify-syntax-entry (make-char 'chinese-gb2312 34) "_")
 114 (modify-syntax-entry (make-char 'chinese-gb2312 41) "_")
 115 (modify-syntax-entry ?\\e$A!2\e(B "(\e$A!3\e(B")
 116 (modify-syntax-entry ?\\e$A!4\e(B "(\e$A!5\e(B")
 117 (modify-syntax-entry ?\\e$A!6\e(B "(\e$A!7\e(B")
 118 (modify-syntax-entry ?\\e$A!8\e(B "(\e$A!9\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 (modify-syntax-entry ?\\e$A!3\e(B ")\e$A!2\e(B")
 123 (modify-syntax-entry ?\\e$A!5\e(B ")\e$A!4\e(B")
 124 (modify-syntax-entry ?\\e$A!7\e(B ")\e$A!6\e(B")
 125 (modify-syntax-entry ?\\e$A!9\e(B ")\e$A!8\e(B")
 126 (modify-syntax-entry ?\\e$A!;\e(B ")\e$A!:\e(B")
 127 (modify-syntax-entry ?\\e$A!=\e(B ")\e$A!<\e(B")
 128 (modify-syntax-entry ?\\e$A!?\e(B ")\e$A!>\e(B")
 129
 130 (modify-category-entry (make-char 'chinese-gb2312) ?c)
 131 (modify-category-entry (make-char 'chinese-gb2312) ?\|)
 132 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 133 (modify-category-entry (make-char 'chinese-gb2312 36) ?H)
 134 (modify-category-entry (make-char 'chinese-gb2312 37) ?K)
 135 (modify-category-entry (make-char 'chinese-gb2312 38) ?G)
 136 (modify-category-entry (make-char 'chinese-gb2312 39) ?Y)
 137 (modify-category-entry (make-char 'chinese-gb2312 35) ?A)
 138 (let ((row 48))
 139   (while (< row 127)
 140     (modify-category-entry (make-char 'chinese-gb2312 row) ?C)
 141     (setq row (1+ row))))
 142
 143 ;; Chinese character set (BIG5)
 144
 145 (let ((generic-big5-1-char (make-char 'chinese-big5-1))
 146       (generic-big5-2-char (make-char 'chinese-big5-2)))
 147   (modify-syntax-entry generic-big5-1-char "w")
 148   (modify-syntax-entry generic-big5-2-char "w")
 149
 150   (modify-category-entry generic-big5-1-char ?c)
 151   (modify-category-entry generic-big5-2-char ?c)
 152
 153   (modify-category-entry generic-big5-1-char ?C)
 154   (modify-category-entry generic-big5-2-char ?C)
 155
 156   (modify-category-entry generic-big5-1-char ?\|)
 157   (modify-category-entry generic-big5-2-char ?\|))
 158
 159
 160 ;; Chinese character set (CNS11643)
 161
 162 (let ((cns-list '(chinese-cns11643-1
 163                   chinese-cns11643-2
 164                   chinese-cns11643-3
 165                   chinese-cns11643-4
 166                   chinese-cns11643-5
 167                   chinese-cns11643-6
 168                   chinese-cns11643-7))
 169       generic-char)
 170   (while cns-list
 171     (setq generic-char (make-char (car cns-list)))
 172     (modify-syntax-entry generic-char "w")
 173     (modify-category-entry generic-char ?c)
 174     (modify-category-entry generic-char ?C)
 175     (modify-category-entry generic-char ?|)
 176     (setq cns-list (cdr cns-list))))
 177
 178 ;; Cyrillic character set (ISO-8859-5)
 179
 180 (modify-category-entry (make-char 'cyrillic-iso8859-5) ?y)
 181
 182 (let ((c 160))
 183   (while (< c 256)
 184     (modify-syntax-entry (make-char 'cyrillic-iso8859-5 c) "w")
 185     (setq c (1+ c))))
 186 (modify-syntax-entry ?\e,L-\e(B ".")
 187 (modify-syntax-entry ?\e,Lp\e(B ".")
 188 (modify-syntax-entry ?\e,L}\e(B ".")
 189
 190 ;; Devanagari character set
 191
 192 (let ((deflist  '(;; chars      syntax  category
 193                   ("\e$(5!!!"!#\e(B"      "w"     ?7) ; vowel-modifying diacritical mark
 194                                             ; chandrabindu, anuswar, visarga
 195                   ("\e$(5!$\e(B-\e$(5!2\e(B"        "w"     ?5) ; independent vowel
 196                   ("\e$(5!3\e(B-\e$(5!X\e(B"        "w"     ?0) ; consonant
 197                   ("\e$(5!Z\e(B-\e$(5!g\e(B"        "w"     ?8) ; matra
 198                   ("\e$(5!q\e(B-\e$(5!z\e(B"        "w"     ?6) ; digit
 199                   ))
 200       elm chars len syntax category to ch i)
 201   (while deflist
 202     (setq elm (car deflist))
 203     (setq chars (car elm)
 204           len (length chars)
 205           syntax (nth 1 elm)
 206           category (nth 2 elm)
 207           i 0)
 208     (while (< i len)
 209       (if (= (aref chars i) ?-)
 210           (setq i (1+ i)
 211                 to (sref chars i))
 212         (setq ch (sref chars i)
 213               to ch))
 214       (while (<= ch to)
 215         (modify-syntax-entry ch syntax)
 216         (modify-category-entry ch category)
 217         (setq ch (1+ ch)))
 218       (setq i (+ i (char-bytes to))))
 219     (setq deflist (cdr deflist))))
 220
 221 ;; Ethiopic character set
 222
 223 (modify-category-entry (make-char 'ethiopic) ?e)
 224
 225 ;; European character set (Latin-1,2,3,4,5)
 226
 227 (modify-category-entry (make-char 'latin-iso8859-1) ?l)
 228 (modify-category-entry (make-char 'latin-iso8859-2) ?l)
 229 (modify-category-entry (make-char 'latin-iso8859-3) ?l)
 230 (modify-category-entry (make-char 'latin-iso8859-4) ?l)
 231 (modify-category-entry (make-char 'latin-iso8859-9) ?l)
 232
 233 ;; Greek character set (ISO-8859-7)
 234
 235 (modify-category-entry (make-char 'greek-iso8859-7) ?g)
 236
 237 (let ((c 182))
 238   (while (< c 255)
 239     (modify-syntax-entry (make-char 'greek-iso8859-7 c) "w")
 240     (setq c (1+ c))))
 241 (modify-syntax-entry (make-char 'greek-iso8859-7 160) "w") ; NBSP
 242 (modify-syntax-entry ?\e,F7\e(B ".")
 243 (modify-syntax-entry ?\e,F;\e(B ".")
 244 (modify-syntax-entry ?\e,F=\e(B ".")
 245
 246 ;; Hebrew character set (ISO-8859-8)
 247
 248 (modify-category-entry (make-char 'hebrew-iso8859-8) ?w)
 249
 250 (let ((c 224))
 251   (while (< c 251)
 252     (modify-syntax-entry (make-char 'hebrew-iso8859-8 c) "w")
 253     (setq c (1+ c))))
 254 (modify-syntax-entry (make-char 'hebrew-iso8859-8 160) "w") ; NBSP
 255
 256 ;; Indian character set (IS 13194 and other Emacs original Indian charsets)
 257
 258 (modify-category-entry (make-char 'indian-is13194) ?i)
 259 (modify-category-entry (make-char 'indian-2-column) ?I)
 260 (modify-category-entry (make-char 'indian-1-column) ?I)
 261
 262 ;; Japanese character set (JISX0201-kana, JISX0201-roman, JISX0208, JISX0212)
 263
 264 (modify-category-entry (make-char 'katakana-jisx0201) ?k)
 265 (modify-category-entry (make-char 'latin-jisx0201) ?r)
 266 (modify-category-entry (make-char 'japanese-jisx0208) ?j)
 267 (modify-category-entry (make-char 'japanese-jisx0212) ?j)
 268 (modify-category-entry (make-char 'katakana-jisx0201) ?\|)
 269 (modify-category-entry (make-char 'japanese-jisx0208) ?\|)
 270 (modify-category-entry (make-char 'japanese-jisx0212) ?\|)
 271
 272 ;; JISX0208
 273 (modify-syntax-entry (make-char 'japanese-jisx0208) "w")
 274 (modify-syntax-entry (make-char 'japanese-jisx0208 33) "_")
 275 (modify-syntax-entry (make-char 'japanese-jisx0208 34) "_")
 276 (modify-syntax-entry (make-char 'japanese-jisx0208 40) "_")
 277 (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)))
 278   (while chars
 279     (modify-syntax-entry (car chars) "w")
 280     (setq chars (cdr chars))))
 281 (modify-syntax-entry ?\\e$B!J\e(B "(\e$B!K\e(B")
 282 (modify-syntax-entry ?\\e$B!N\e(B "(\e$B!O\e(B")
 283 (modify-syntax-entry ?\\e$B!P\e(B "(\e$B!Q\e(B")
 284 (modify-syntax-entry ?\\e$B!V\e(B "(\e$B!W\e(B")
 285 (modify-syntax-entry ?\\e$B!X\e(B "(\e$B!Y\e(B")
 286 (modify-syntax-entry ?\\e$B!K\e(B ")\e$B!J\e(B")
 287 (modify-syntax-entry ?\\e$B!O\e(B ")\e$B!N\e(B")
 288 (modify-syntax-entry ?\\e$B!Q\e(B ")\e$B!P\e(B")
 289 (modify-syntax-entry ?\\e$B!W\e(B ")\e$B!V\e(B")
 290 (modify-syntax-entry ?\\e$B!Y\e(B ")\e$B!X\e(B")
 291
 292 (modify-category-entry (make-char 'japanese-jisx0208 35) ?A)
 293 (modify-category-entry (make-char 'japanese-jisx0208 36) ?H)
 294 (modify-category-entry (make-char 'japanese-jisx0208 37) ?K)
 295 (modify-category-entry (make-char 'japanese-jisx0208 38) ?G)
 296 (modify-category-entry (make-char 'japanese-jisx0208 39) ?Y)
 297 (let ((row 48))
 298   (while (< row 127)
 299     (modify-category-entry (make-char 'japanese-jisx0208 row) ?C)
 300     (setq row (1+ row))))
 301 (modify-category-entry ?\e$B!<\e(B ?K)
 302 (let ((chars '(?\e$B!+\e(B ?\e$B!,\e(B)))
 303   (while chars
 304     (modify-category-entry (car chars) ?K)
 305     (modify-category-entry (car chars) ?H)
 306     (setq chars (cdr chars))))
 307 (let ((chars '(?\e$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)))
 308   (while chars
 309     (modify-category-entry (car chars) ?C)
 310     (setq chars (cdr chars))))
 311
 312 ;; JISX0212
 313 (modify-syntax-entry (make-char 'japanese-jisx0212) "w")
 314 (modify-syntax-entry (make-char 'japanese-jisx0212 33) "_")
 315 (modify-syntax-entry (make-char 'japanese-jisx0212 34) "_")
 316 (modify-syntax-entry (make-char 'japanese-jisx0212 35) "_")
 317
 318 (modify-category-entry (make-char 'japanese-jisx0212 ) ?C)
 319
 320 ;; JISX0201-Kana
 321 (modify-syntax-entry (make-char 'katakana-jisx0201) "w")
 322 (let ((chars '(?\e(I!\e(B ?\e(I"\e(B ?\e(I#\e(B ?\e(I$\e(B ?\e(I%\e(B)))
 323   (while chars
 324     (modify-syntax-entry (car chars) ".")
 325     (setq chars (cdr chars))))
 326
 327 ;; Korean character set (KSC5601)
 328
 329 (modify-syntax-entry (make-char 'korean-ksc5601) "w")
 330 (modify-syntax-entry (make-char 'korean-ksc5601 33) "_")
 331 (modify-syntax-entry (make-char 'korean-ksc5601 34) "_")
 332 (modify-syntax-entry (make-char 'korean-ksc5601 38) "_")
 333 (modify-syntax-entry (make-char 'korean-ksc5601 39) "_")
 334 (modify-syntax-entry (make-char 'korean-ksc5601 40) "_")
 335 (modify-syntax-entry (make-char 'korean-ksc5601 41) "_")
 336
 337 (modify-category-entry (make-char 'korean-ksc5601) ?h)
 338 (modify-category-entry (make-char 'korean-ksc5601 35) ?A)
 339 (modify-category-entry (make-char 'korean-ksc5601 37) ?G)
 340 (modify-category-entry (make-char 'korean-ksc5601 42) ?H)
 341 (modify-category-entry (make-char 'korean-ksc5601 43) ?K)
 342 (modify-category-entry (make-char 'korean-ksc5601 44) ?Y)
 343
 344 ;; Lao character set
 345
 346 (modify-category-entry (make-char 'lao) ?o)
 347
 348 (let ((deflist  '(;; chars      syntax  category
 349                   ("\e(1!\e(B-\e(1N\e(B"    "w"     ?0) ; consonant
 350                   ("\e(1PRS]`\e(B-\e(1d\e(B"        "w"     ?1) ; vowel base
 351                   ("\e(1QT\e(B-\e(1W[m\e(B" "w"     ?2) ; vowel upper
 352                   ("\e(1XY\e(B"           "w"     ?3) ; vowel lower
 353                   ("\e(1h\e(B-\e(1l\e(B"    "w"     ?4) ; tone mark
 354                   ("\e(1\\e(B"            "w"     ?9) ; semivowel lower
 355                   ("\e(1p\e(B-\e(1y\e(B"    "w"     ?0) ; digit and misc
 356                   ("\e(1Of\e(B"           "_"     ?0) ; symbol
 357                   ))
 358       elm chars len syntax category to ch i)
 359   (while deflist
 360     (setq elm (car deflist))
 361     (setq chars (car elm)
 362           len (length chars)
 363           syntax (nth 1 elm)
 364           category (nth 2 elm)
 365           i 0)
 366     (while (< i len)
 367       (if (= (aref chars i) ?-)
 368           (setq i (1+ i)
 369                 to (sref chars i))
 370         (setq ch (sref chars i)
 371               to ch))
 372       (while (<= ch to)
 373         (modify-syntax-entry ch syntax)
 374         (modify-category-entry ch category)
 375         (setq ch (1+ ch)))
 376       (setq i (+ i (char-bytes to))))
 377     (setq deflist (cdr deflist))))
 378
 379 ;; Thai character set (TIS620)
 380
 381 (modify-category-entry (make-char 'thai-tis620) ?t)
 382
 383 (let ((deflist  '(;; chars      syntax  category
 384                   ("\e,T!\e(B-\e,TCEG\e(B-\e,TN\e(B"  "w"     ?0) ; consonant
 385                   ("\e,TDFPRS`\e(B-\e,Te\e(B"       "w"     ?1) ; vowel base
 386                   ("\e,TQT\e(B-\e,TWgn\e(B" "w"     ?2) ; vowel upper
 387                   ("\e,TX\e(B-\e,TZ\e(B"    "w"     ?3) ; vowel lower
 388                   ("\e,Th\e(B-\e,Tm\e(B"    "w"     ?4) ; tone mark
 389                   ("\e,TOfp\e(B-\e,Ty\e(B"  "w"     ?0) ; digit and misc
 390                   ("\e,T_oz{\e(B" "_"     ?0) ; symbol
 391                   ))
 392       elm chars len syntax category to ch i)
 393   (while deflist
 394     (setq elm (car deflist))
 395     (setq chars (car elm)
 396           len (length chars)
 397           syntax (nth 1 elm)
 398           category (nth 2 elm)
 399           i 0)
 400     (while (< i len)
 401       (if (= (aref chars i) ?-)
 402           (setq i (1+ i)
 403                 to (sref chars i))
 404         (setq ch (sref chars i)
 405               to ch))
 406       (while (<= ch to)
 407         (modify-syntax-entry ch syntax)
 408         (modify-category-entry ch category)
 409         (setq ch (1+ ch)))
 410       (setq i (+ i (char-bytes to))))
 411     (setq deflist (cdr deflist))))
 412
 413 ;; Tibetan character set
 414
 415 (let ((row 33))
 416   (while (< row 38)
 417     (modify-category-entry (make-char 'tibetan row) ?q)
 418     (setq row (1+ row))))
 419
 420 (modify-category-entry (make-char 'tibetan-1-column 33) ?q)
 421
 422 (let ((deflist  '(;; chars             syntax category
 423                   ("\e$(7"!\e(B-\e$(7"J\e(B"                "w"     ?0) ; consonant
 424                   ("\e$(7#!\e(B-\e$(7#J#P#Q\e(B"          "w"     ?0) ;
 425                   ("\e$(7$!\e(B-\e$(7$e\e(B"              "w"     ?0) ;
 426                   ("\e$(7%!\e(B-\e$(7%u\e(B"              "w"     ?0) ;
 427                   ("\e$(7"S"["\"]"^"a\e(B"       "w"      ?2) ; upper vowel
 428                   ("\e$(7"_"c"d"g"h"i"j"k"l\e(B" "w"      ?2) ; upper modifier
 429                   ("\e$(7!I"Q"U"e!e!g\e(B"       "w"      ?3) ; lowel vowel/modifier
 430                   ("\e$(7!P\e(B-\e$(7!Y!Z\e(B-\e$(7!c\e(B"            "w"     ?6) ; digit
 431                   ("\e$(7!;!=\e(B-\e$(7!B!D"`\e(B"        "."     ?|) ; line-break char
 432                   ("\e$(8!;!=!?!@!A!D"`\e(B"            "."     ?|) ;
 433                   ("\e$(7!8!;!=\e(B-\e$(7!B!D"`!m!d\e(B"  "."     ?>) ; prohibition
 434                   ("\e$(8!;!=!?!@!A!D"`\e(B"            "."     ?>) ;
 435                   ("\e$(7!0\e(B-\e$(7!:!l#R#S"f\e(B"      "."     ?<) ; prohibition
 436                   ("\e$(7!C!E\e(B-\e$(7!H!J\e(B-\e$(7!O!f!h\e(B-\e$(7!k!n!o\e(B" "." ?q) ; others
 437                   ))
 438       elm chars len syntax category to ch i)
 439   (while deflist
 440     (setq elm (car deflist))
 441     (setq chars (car elm)
 442           len (length chars)
 443           syntax (nth 1 elm)
 444           category (nth 2 elm)
 445           i 0)
 446     (while (< i len)
 447       (if (= (aref chars i) ?-)
 448           (setq i (1+ i)
 449                 to (sref chars i))
 450         (setq ch (sref chars i)
 451               to ch))
 452       (while (<= ch to)
 453         (modify-syntax-entry ch syntax)
 454         (modify-category-entry ch category)
 455         (setq ch (1+ ch)))
 456       (setq i (+ i (char-bytes to))))
 457     (setq deflist (cdr deflist))))
 458
 459 ;; Vietnamese character set
 460
 461 (let ((lower (make-char 'vietnamese-viscii-lower))
 462       (upper (make-char 'vietnamese-viscii-upper)))
 463   (modify-syntax-entry lower "w")
 464   (modify-syntax-entry upper "w")
 465   (modify-category-entry lower ?v)
 466   (modify-category-entry upper ?v)
 467   (modify-category-entry lower ?l)      ; To make a word with
 468   (modify-category-entry upper ?l)      ; latin characters.
 469   )
 470
 471 \f
 472 ;;; Setting word boundary.
 473
 474 (setq word-combining-categories
 475       '((?l . ?l)))
 476
 477 (setq word-separating-categories        ;  (2-byte character sets)
 478       '((?A . ?K)                       ; Alpha numeric - Katakana
 479         (?A . ?C)                       ; Alpha numeric - Chinese
 480         (?H . ?A)                       ; Hiragana - Alpha numeric
 481         (?H . ?K)                       ; Hiragana - Katakana
 482         (?H . ?C)                       ; Hiragana - Chinese
 483         (?K . ?A)                       ; Katakana - Alpha numeric
 484         (?K . ?C)                       ; Katakana - Chinese
 485         (?C . ?A)                       ; Chinese - Alpha numeric
 486         (?C . ?K)                       ; Chinese - Katakana
 487         ))