| 1 | ;; unidata-gen.el -- Create files containing character property data. |
| 2 | |
| 3 | ;; Copyright 2008-2013 (C) Free Software Foundation, Inc. |
| 4 | |
| 5 | ;; Copyright (C) 2005, 2006, 2007, 2008, 2009, 2010, 2011 |
| 6 | ;; National Institute of Advanced Industrial Science and Technology (AIST) |
| 7 | ;; Registration Number H13PRO009 |
| 8 | |
| 9 | ;; This file is part of GNU Emacs. |
| 10 | |
| 11 | ;; GNU Emacs is free software: you can redistribute it and/or modify |
| 12 | ;; it under the terms of the GNU General Public License as published by |
| 13 | ;; the Free Software Foundation, either version 3 of the License, or |
| 14 | ;; (at your option) any later version. |
| 15 | |
| 16 | ;; GNU Emacs is distributed in the hope that it will be useful, |
| 17 | ;; but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 18 | ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 19 | ;; GNU General Public License for more details. |
| 20 | |
| 21 | ;; You should have received a copy of the GNU General Public License |
| 22 | ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. |
| 23 | |
| 24 | ;;; Commentary: |
| 25 | |
| 26 | ;; SPECIAL NOTICE |
| 27 | ;; |
| 28 | ;; This file must be byte-compilable/loadable by `temacs' and also |
| 29 | ;; the entry function `unidata-gen-files' must be runnable by `temacs'. |
| 30 | |
| 31 | ;; FILES TO BE GENERATED |
| 32 | ;; |
| 33 | ;; The entry function `unidata-gen-files' generates these files in |
| 34 | ;; in directory specified by its dest-dir argument. |
| 35 | ;; |
| 36 | ;; charprop.el |
| 37 | ;; It contains a series of forms of this format: |
| 38 | ;; (define-char-code-property PROP FILE) |
| 39 | ;; where PROP is a symbol representing a character property |
| 40 | ;; (name, general-category, etc), and FILE is a name of one of |
| 41 | ;; the following files. |
| 42 | ;; |
| 43 | ;; uni-name.el, uni-category.el, uni-combining.el, uni-bidi.el, |
| 44 | ;; uni-decomposition.el, uni-decimal.el, uni-digit.el, uni-numeric.el, |
| 45 | ;; uni-mirrored.el, uni-old-name.el, uni-comment.el, uni-uppercase.el, |
| 46 | ;; uni-lowercase.el, uni-titlecase.el |
| 47 | ;; They contain one or more forms of this format: |
| 48 | ;; (define-char-code-property PROP CHAR-TABLE) |
| 49 | ;; where PROP is the same as above, and CHAR-TABLE is a |
| 50 | ;; char-table containing property values in a compressed format. |
| 51 | ;; |
| 52 | ;; When they are installed in .../lisp/international/, the file |
| 53 | ;; "charprop.el" is preloaded in loadup.el. The other files are |
| 54 | ;; automatically loaded when the Lisp functions |
| 55 | ;; `get-char-code-property' and `put-char-code-property', and C |
| 56 | ;; function uniprop_table are called. |
| 57 | ;; |
| 58 | ;; FORMAT OF A CHAR TABLE |
| 59 | ;; |
| 60 | ;; We want to make a file size containing a char-table small. We |
| 61 | ;; also want to load the file and get a property value fast. We |
| 62 | ;; also want to reduce the used memory after loading it. So, |
| 63 | ;; instead of naively storing a property value for each character in |
| 64 | ;; a char-table (and write it out into a file), we store compressed |
| 65 | ;; data in a char-table as below. |
| 66 | ;; |
| 67 | ;; If succeeding 128*N characters have the same property value, we |
| 68 | ;; store that value (or the encoded one) for them. Otherwise, |
| 69 | ;; compress values (or the encoded ones) for succeeding 128 |
| 70 | ;; characters into a single string and store it for those |
| 71 | ;; characters. The way of compression depends on a property. See |
| 72 | ;; the section "SIMPLE TABLE", "RUN-LENGTH TABLE", and "WORD-LIST |
| 73 | ;; TABLE". |
| 74 | |
| 75 | ;; The char table has five extra slots: |
| 76 | ;; 1st: property symbol |
| 77 | ;; 2nd: function to call to get a property value, |
| 78 | ;; or an index number of C function to decode the value, |
| 79 | ;; or nil if the value can be directly got from the table. |
| 80 | ;; 3nd: function to call to put a property value, |
| 81 | ;; or an index number of C function to encode the value, |
| 82 | ;; or nil if the value can be directly stored in the table. |
| 83 | ;; 4th: function to call to get a description of a property value, or nil |
| 84 | ;; 5th: data referred by the above functions |
| 85 | |
| 86 | ;; List of elements of this form: |
| 87 | ;; (CHAR-or-RANGE PROP1 PROP2 ... PROPn) |
| 88 | ;; CHAR-or-RANGE: a character code or a cons of character codes |
| 89 | ;; PROPn: string representing the nth property value |
| 90 | |
| 91 | (defvar unidata-list nil) |
| 92 | |
| 93 | ;; Name of the directory containing files of Unicode Character Database. |
| 94 | |
| 95 | ;; Dynamically bound in unidata-gen-files. |
| 96 | (defvar unidata-dir nil) |
| 97 | |
| 98 | (defun unidata-setup-list (unidata-text-file) |
| 99 | (let* ((table (list nil)) |
| 100 | (tail table) |
| 101 | (block-names '(("^<CJK Ideograph" . CJK\ IDEOGRAPH) |
| 102 | ("^<Hangul Syllable" . HANGUL\ SYLLABLE) |
| 103 | ("^<.*Surrogate" . nil) |
| 104 | ("^<.*Private Use" . PRIVATE\ USE))) |
| 105 | val char name) |
| 106 | (setq unidata-text-file (expand-file-name unidata-text-file unidata-dir)) |
| 107 | (or (file-readable-p unidata-text-file) |
| 108 | (error "File not readable: %s" unidata-text-file)) |
| 109 | (with-temp-buffer |
| 110 | ;; Insert a file of this format: |
| 111 | ;; (CHAR NAME CATEGORY ...) |
| 112 | ;; where CHAR is a character code, the following elements are strings |
| 113 | ;; representing character properties. |
| 114 | (insert-file-contents unidata-text-file) |
| 115 | (goto-char (point-min)) |
| 116 | (condition-case nil |
| 117 | (while t |
| 118 | (setq val (read (current-buffer)) |
| 119 | char (car val) |
| 120 | name (cadr val)) |
| 121 | |
| 122 | ;; Check this kind of block. |
| 123 | ;; 4E00;<CJK Ideograph, First>;Lo;0;L;;;;;N;;;;; |
| 124 | ;; 9FCB;<CJK Ideograph, Last>;Lo;0;L;;;;;N;;;;; |
| 125 | (if (and (= (aref name 0) ?<) |
| 126 | (string-match ", First>$" name)) |
| 127 | (let ((first char) |
| 128 | (l block-names) |
| 129 | block-name) |
| 130 | (setq val (read (current-buffer)) |
| 131 | char (car val) |
| 132 | block-name (cadr val) |
| 133 | name nil) |
| 134 | (while l |
| 135 | (if (string-match (caar l) block-name) |
| 136 | (setq name (cdar l) l nil) |
| 137 | (setq l (cdr l)))) |
| 138 | (if (not name) |
| 139 | ;; As this is a surrogate pair range, ignore it. |
| 140 | (setq val nil) |
| 141 | (setcar val (cons first char)) |
| 142 | (setcar (cdr val) name)))) |
| 143 | |
| 144 | (when val |
| 145 | (setcdr tail (list val)) |
| 146 | (setq tail (cdr tail)))) |
| 147 | (error nil))) |
| 148 | (setq unidata-list (cdr table)))) |
| 149 | |
| 150 | ;; Alist of this form: |
| 151 | ;; (PROP INDEX GENERATOR FILENAME DOCSTRING DESCRIBER DEFAULT VAL-LIST) |
| 152 | ;; PROP: character property |
| 153 | ;; INDEX: index to each element of unidata-list for PROP. |
| 154 | ;; It may be a function that generates an alist of character codes |
| 155 | ;; vs. the corresponding property values. |
| 156 | ;; GENERATOR: function to generate a char-table |
| 157 | ;; FILENAME: filename to store the char-table |
| 158 | ;; DOCSTRING: docstring for the property |
| 159 | ;; DESCRIBER: function to call to get a description string of property value |
| 160 | ;; DEFAULT: the default value of the property. It may have the form |
| 161 | ;; (VAL0 (FROM1 TO1 VAL1) ...) which indicates that the default |
| 162 | ;; value is VAL0 except for characters in the ranges specified by |
| 163 | ;; FROMn and TOn (inclusive). The default value of characters |
| 164 | ;; between FROMn and TOn is VALn. |
| 165 | ;; VAL-LIST: list of specially ordered property values |
| 166 | |
| 167 | (defconst unidata-prop-alist |
| 168 | '((name |
| 169 | 1 unidata-gen-table-name "uni-name.el" |
| 170 | "Unicode character name. |
| 171 | Property value is a string or nil. |
| 172 | The value nil stands for the default value \"null string\")." |
| 173 | nil |
| 174 | nil) |
| 175 | (general-category |
| 176 | 2 unidata-gen-table-symbol "uni-category.el" |
| 177 | "Unicode general category. |
| 178 | Property value is one of the following symbols: |
| 179 | Lu, Ll, Lt, Lm, Lo, Mn, Mc, Me, Nd, Nl, No, Pc, Pd, Ps, Pe, Pi, Pf, Po, |
| 180 | Sm, Sc, Sk, So, Zs, Zl, Zp, Cc, Cf, Cs, Co, Cn" |
| 181 | unidata-describe-general-category |
| 182 | Cn |
| 183 | ;; The order of elements must be in sync with unicode_category_t |
| 184 | ;; in src/character.h. |
| 185 | (Lu Ll Lt Lm Lo Mn Mc Me Nd Nl No Pc Pd Ps Pe Pi Pf Po |
| 186 | Sm Sc Sk So Zs Zl Zp Cc Cf Cs Co Cn)) |
| 187 | (canonical-combining-class |
| 188 | 3 unidata-gen-table-integer "uni-combining.el" |
| 189 | "Unicode canonical combining class. |
| 190 | Property value is an integer." |
| 191 | unidata-describe-canonical-combining-class |
| 192 | 0) |
| 193 | (bidi-class |
| 194 | 4 unidata-gen-table-symbol "uni-bidi.el" |
| 195 | "Unicode bidi class. |
| 196 | Property value is one of the following symbols: |
| 197 | L, LRE, LRO, LRI, R, AL, RLE, RLO, RLI, FSI, PDF, PDI, |
| 198 | EN, ES, ET, AN, CS, NSM, BN, B, S, WS, ON" |
| 199 | unidata-describe-bidi-class |
| 200 | ;; The assignment of default values to blocks of code points |
| 201 | ;; follows the file DerivedBidiClass.txt from the Unicode |
| 202 | ;; Character Database (UCD). |
| 203 | (L (#x0600 #x06FF AL) (#xFB50 #xFDFF AL) (#xFE70 #xFEFF AL) |
| 204 | (#x0590 #x05FF R) (#x07C0 #x08FF R) |
| 205 | (#xFB1D #xFB4F R) (#x10800 #x10FFF R) (#x1E800 #x1EFFF R)) |
| 206 | ;; The order of elements must be in sync with bidi_type_t in |
| 207 | ;; src/dispextern.h. |
| 208 | (L R EN AN BN B AL LRE LRO RLE RLO PDF LRI RLI FSI PDI |
| 209 | ES ET CS NSM S WS ON)) |
| 210 | (decomposition |
| 211 | 5 unidata-gen-table-decomposition "uni-decomposition.el" |
| 212 | "Unicode decomposition mapping. |
| 213 | Property value is a list of characters. The first element may be |
| 214 | one of these symbols representing compatibility formatting tag: |
| 215 | font, noBreak, initial, medial, final, isolated, circle, super, |
| 216 | sub, vertical, wide, narrow, small, square, fraction, compat" |
| 217 | unidata-describe-decomposition) |
| 218 | (decimal-digit-value |
| 219 | 6 unidata-gen-table-integer "uni-decimal.el" |
| 220 | "Unicode numeric value (decimal digit). |
| 221 | Property value is an integer 0..9, or nil. |
| 222 | The value nil stands for NaN \"Numeric_Value\".") |
| 223 | (digit-value |
| 224 | 7 unidata-gen-table-integer "uni-digit.el" |
| 225 | "Unicode numeric value (digit). |
| 226 | Property value is an integer 0..9, or nil. |
| 227 | The value nil stands for NaN \"Numeric_Value\".") |
| 228 | (numeric-value |
| 229 | 8 unidata-gen-table-numeric "uni-numeric.el" |
| 230 | "Unicode numeric value (numeric). |
| 231 | Property value is an integer, a floating point, or nil. |
| 232 | The value nil stands for NaN \"Numeric_Value\".") |
| 233 | (mirrored |
| 234 | 9 unidata-gen-table-symbol "uni-mirrored.el" |
| 235 | "Unicode bidi mirrored flag. |
| 236 | Property value is a symbol `Y' or `N'. See also the property `mirroring'." |
| 237 | nil |
| 238 | N) |
| 239 | (old-name |
| 240 | 10 unidata-gen-table-name "uni-old-name.el" |
| 241 | "Unicode old names as published in Unicode 1.0. |
| 242 | Property value is a string or nil. |
| 243 | The value nil stands for the default value \"null string\").") |
| 244 | (iso-10646-comment |
| 245 | 11 unidata-gen-table-name "uni-comment.el" |
| 246 | "Unicode ISO 10646 comment. |
| 247 | Property value is a string.") |
| 248 | (uppercase |
| 249 | 12 unidata-gen-table-character "uni-uppercase.el" |
| 250 | "Unicode simple uppercase mapping. |
| 251 | Property value is a character or nil. |
| 252 | The value nil means that the actual property value of a character |
| 253 | is the character itself." |
| 254 | string) |
| 255 | (lowercase |
| 256 | 13 unidata-gen-table-character "uni-lowercase.el" |
| 257 | "Unicode simple lowercase mapping. |
| 258 | Property value is a character or nil. |
| 259 | The value nil means that the actual property value of a character |
| 260 | is the character itself." |
| 261 | string) |
| 262 | (titlecase |
| 263 | 14 unidata-gen-table-character "uni-titlecase.el" |
| 264 | "Unicode simple titlecase mapping. |
| 265 | Property value is a character or nil. |
| 266 | The value nil means that the actual property value of a character |
| 267 | is the character itself." |
| 268 | string) |
| 269 | (mirroring |
| 270 | unidata-gen-mirroring-list unidata-gen-table-character "uni-mirrored.el" |
| 271 | "Unicode bidi-mirroring characters. |
| 272 | Property value is a character that has the corresponding mirroring image or nil. |
| 273 | The value nil means that the actual property value of a character |
| 274 | is the character itself."))) |
| 275 | |
| 276 | ;; Functions to access the above data. |
| 277 | (defsubst unidata-prop-index (prop) (nth 1 (assq prop unidata-prop-alist))) |
| 278 | (defsubst unidata-prop-generator (prop) (nth 2 (assq prop unidata-prop-alist))) |
| 279 | (defsubst unidata-prop-file (prop) (nth 3 (assq prop unidata-prop-alist))) |
| 280 | (defsubst unidata-prop-docstring (prop) (nth 4 (assq prop unidata-prop-alist))) |
| 281 | (defsubst unidata-prop-describer (prop) (nth 5 (assq prop unidata-prop-alist))) |
| 282 | (defsubst unidata-prop-default (prop) (nth 6 (assq prop unidata-prop-alist))) |
| 283 | (defsubst unidata-prop-val-list (prop) (nth 7 (assq prop unidata-prop-alist))) |
| 284 | |
| 285 | \f |
| 286 | ;; SIMPLE TABLE |
| 287 | ;; |
| 288 | ;; If the type of character property value is character, and the |
| 289 | ;; values of succeeding character codes are usually different, we use |
| 290 | ;; a char-table described here to store such values. |
| 291 | ;; |
| 292 | ;; A char-table divides character code space (#x0..#x3FFFFF) into |
| 293 | ;; #x8000 blocks (each block contains 128 characters). |
| 294 | |
| 295 | ;; If all characters of a block have no property, a char-table has the |
| 296 | ;; symbol nil for that block. Otherwise a char-table has a string of |
| 297 | ;; the following format for it. |
| 298 | ;; |
| 299 | ;; The first character of the string is ?\001. |
| 300 | ;; The second character of the string is FIRST-INDEX. |
| 301 | ;; The Nth (N > 1) character of the string is a property value of the |
| 302 | ;; character (BLOCK-HEAD + FIRST-INDEX + N - 2), where BLOCK-HEAD is |
| 303 | ;; the first character of the block. |
| 304 | ;; |
| 305 | ;; This kind of char-table has these extra slots: |
| 306 | ;; 1st: the property symbol |
| 307 | ;; 2nd: nil |
| 308 | ;; 3rd: 0 (corresponding to uniprop_encode_character in chartab.c) |
| 309 | ;; 4th to 5th: nil |
| 310 | |
| 311 | (defun unidata-gen-table-character (prop &rest ignore) |
| 312 | (let ((table (make-char-table 'char-code-property-table)) |
| 313 | (prop-idx (unidata-prop-index prop)) |
| 314 | (vec (make-vector 128 0)) |
| 315 | (tail unidata-list) |
| 316 | elt range val idx slot) |
| 317 | (if (functionp prop-idx) |
| 318 | (setq tail (funcall prop-idx) |
| 319 | prop-idx 1)) |
| 320 | (while tail |
| 321 | (setq elt (car tail) tail (cdr tail)) |
| 322 | (setq range (car elt) |
| 323 | val (nth prop-idx elt)) |
| 324 | (if (= (length val) 0) |
| 325 | (setq val nil) |
| 326 | (setq val (string-to-number val 16))) |
| 327 | (if (consp range) |
| 328 | (if val |
| 329 | (set-char-table-range table range val)) |
| 330 | (let* ((start (lsh (lsh range -7) 7)) |
| 331 | (limit (+ start 127)) |
| 332 | first-index last-index) |
| 333 | (fillarray vec 0) |
| 334 | (if val |
| 335 | (aset vec (setq last-index (setq first-index (- range start))) |
| 336 | val)) |
| 337 | (while (and (setq elt (car tail) range (car elt)) |
| 338 | (integerp range) |
| 339 | (<= range limit)) |
| 340 | (setq val (nth prop-idx elt)) |
| 341 | (when (> (length val) 0) |
| 342 | (aset vec (setq last-index (- range start)) |
| 343 | (string-to-number val 16)) |
| 344 | (or first-index |
| 345 | (setq first-index last-index))) |
| 346 | (setq tail (cdr tail))) |
| 347 | (when first-index |
| 348 | (let ((str (string 1 first-index)) |
| 349 | c) |
| 350 | (while (<= first-index last-index) |
| 351 | (setq str (format "%s%c" str (or (aref vec first-index) 0)) |
| 352 | first-index (1+ first-index))) |
| 353 | (set-char-table-range table (cons start limit) str)))))) |
| 354 | |
| 355 | (set-char-table-extra-slot table 0 prop) |
| 356 | (set-char-table-extra-slot table 2 0) |
| 357 | table)) |
| 358 | |
| 359 | |
| 360 | \f |
| 361 | ;; RUN-LENGTH TABLE |
| 362 | ;; |
| 363 | ;; If many characters of successive character codes have the same |
| 364 | ;; property value, we use a char-table described here to store the |
| 365 | ;; values. |
| 366 | ;; |
| 367 | ;; At first, instead of a value itself, we store an index number to |
| 368 | ;; the VAL-TABLE (5th extra slot) in the table. We call that index |
| 369 | ;; number as VAL-CODE here after. |
| 370 | ;; |
| 371 | ;; A char-table divides character code space (#x0..#x3FFFFF) into |
| 372 | ;; #x8000 blocks (each block contains 128 characters). |
| 373 | ;; |
| 374 | ;; If all characters of a block have the same value, a char-table has |
| 375 | ;; VAL-CODE for that block. Otherwise a char-table has a string of |
| 376 | ;; the following format for that block. |
| 377 | ;; |
| 378 | ;; The first character of the string is ?\002. |
| 379 | ;; The following characters has this form: |
| 380 | ;; ( VAL-CODE RUN-LENGTH ? ) + |
| 381 | ;; where: |
| 382 | ;; VAL-CODE (0..127): index into VAL-TABLE. |
| 383 | ;; RUN-LENGTH (130..255): |
| 384 | ;; (RUN-LENGTH - 128) specifies how many characters have the same |
| 385 | ;; value. If omitted, it means 1. |
| 386 | ;; |
| 387 | ;; This kind of char-table has these extra slots: |
| 388 | ;; 1st: the property symbol |
| 389 | ;; 2nd: 0 (corresponding to uniprop_decode_value in chartab.c) |
| 390 | ;; 3rd: 1..3 (corresponding to uniprop_encode_xxx in chartab.c) |
| 391 | ;; 4th: function or nil |
| 392 | ;; 5th: VAL-TABLE |
| 393 | |
| 394 | ;; Encode the character property value VAL into an integer value by |
| 395 | ;; VAL-LIST. By side effect, VAL-LIST is modified. |
| 396 | ;; VAL-LIST has this form: |
| 397 | ;; ((nil . 0) (VAL1 . 1) (VAL2 . 2) ...) |
| 398 | ;; If VAL is one of VALn, just return n. |
| 399 | ;; Otherwise, VAL-LIST is modified to this: |
| 400 | ;; ((nil . 0) (VAL1 . 1) (VAL2 . 2) ... (VAL . n+1)) |
| 401 | ;; |
| 402 | ;; WARN is an optional warning to display when the value list is |
| 403 | ;; extended, for property values that need to be in sync with other |
| 404 | ;; parts of Emacs; currently only used for bidi-class. |
| 405 | |
| 406 | (defun unidata-encode-val (val-list val &optional warn) |
| 407 | (let ((slot (assoc val val-list)) |
| 408 | val-code) |
| 409 | (if slot |
| 410 | (cdr slot) |
| 411 | (if warn (message warn val)) |
| 412 | (setq val-code (length val-list)) |
| 413 | (nconc val-list (list (cons val val-code))) |
| 414 | val-code))) |
| 415 | |
| 416 | ;; Generate a char-table for the character property PROP. |
| 417 | |
| 418 | (defun unidata-gen-table (prop val-func default-value val-list) |
| 419 | (let ((table (make-char-table 'char-code-property-table)) |
| 420 | (prop-idx (unidata-prop-index prop)) |
| 421 | (vec (make-vector 128 0)) |
| 422 | ;; When this warning is printed, there's a need to make the |
| 423 | ;; following changes: |
| 424 | ;; (1) update unidata-prop-alist with the new bidi-class values; |
| 425 | ;; (2) extend bidi_type_t enumeration on src/dispextern.h to |
| 426 | ;; include the new classes; |
| 427 | ;; (3) possibly update the assertion in bidi.c:bidi_check_type; and |
| 428 | ;; (4) possibly update the switch cases in |
| 429 | ;; bidi.c:bidi_get_type and bidi.c:bidi_get_category. |
| 430 | (bidi-warning "\ |
| 431 | ** Found new bidi-class '%s', please update bidi.c and dispextern.h") |
| 432 | tail elt range val val-code idx slot |
| 433 | prev-range-data) |
| 434 | (setq val-list (cons nil (copy-sequence val-list))) |
| 435 | (setq tail val-list val-code 0) |
| 436 | ;; Convert (nil A B ...) to ((nil . 0) (A . 1) (B . 2) ...) |
| 437 | (while tail |
| 438 | (setcar tail (cons (car tail) val-code)) |
| 439 | (setq tail (cdr tail) val-code (1+ val-code))) |
| 440 | (if (consp default-value) |
| 441 | (setq default-value (copy-sequence default-value)) |
| 442 | (setq default-value (list default-value))) |
| 443 | (setcar default-value |
| 444 | (unidata-encode-val val-list (car default-value))) |
| 445 | (set-char-table-range table t (car default-value)) |
| 446 | (set-char-table-range table nil (car default-value)) |
| 447 | (dolist (elm (cdr default-value)) |
| 448 | (setcar (nthcdr 2 elm) |
| 449 | (unidata-encode-val val-list (nth 2 elm))) |
| 450 | (set-char-table-range table (cons (car elm) (nth 1 elm)) (nth 2 elm))) |
| 451 | |
| 452 | (setq tail unidata-list) |
| 453 | (while tail |
| 454 | (setq elt (car tail) tail (cdr tail)) |
| 455 | (setq range (car elt) |
| 456 | val (funcall val-func (nth prop-idx elt))) |
| 457 | (setq val-code (if val (unidata-encode-val val-list val |
| 458 | (and (eq prop 'bidi-class) |
| 459 | bidi-warning)))) |
| 460 | (if (consp range) |
| 461 | (when val-code |
| 462 | (set-char-table-range table range val-code) |
| 463 | (let ((from (car range)) (to (cdr range))) |
| 464 | ;; If RANGE doesn't end at the char-table boundary (each |
| 465 | ;; 128 characters), we may have to carry over the data |
| 466 | ;; for the last several characters (at most 127 chars) |
| 467 | ;; to the next loop. In that case, set PREV-RANGE-DATA |
| 468 | ;; to ((FROM . TO) . VAL-CODE) where (FROM . TO) |
| 469 | ;; specifies the range of characters handled in the next |
| 470 | ;; loop. |
| 471 | (when (< (logand to #x7F) #x7F) |
| 472 | (if (< from (logand to #x1FFF80)) |
| 473 | (setq from (logand to #x1FFF80))) |
| 474 | (setq prev-range-data (cons (cons from to) val-code))))) |
| 475 | (let* ((start (lsh (lsh range -7) 7)) |
| 476 | (limit (+ start 127)) |
| 477 | str count new-val from to vcode) |
| 478 | (fillarray vec (car default-value)) |
| 479 | (dolist (elm (cdr default-value)) |
| 480 | (setq from (car elm) to (nth 1 elm)) |
| 481 | (when (and (<= from limit) |
| 482 | (or (>= from start) (>= to start))) |
| 483 | (setq from (max from start) |
| 484 | to (min to limit) |
| 485 | vcode (nth 2 elm)) |
| 486 | (while (<= from to) |
| 487 | (aset vec (- from start) vcode) |
| 488 | (setq from (1+ from))))) |
| 489 | ;; See the comment above. |
| 490 | (when (and prev-range-data |
| 491 | (>= (cdr (car prev-range-data)) start)) |
| 492 | (setq from (car (car prev-range-data)) |
| 493 | to (cdr (car prev-range-data)) |
| 494 | vcode (cdr prev-range-data)) |
| 495 | (while (<= from to) |
| 496 | (aset vec (- from start) vcode) |
| 497 | (setq from (1+ from)))) |
| 498 | (setq prev-range-data nil) |
| 499 | (if val-code |
| 500 | (aset vec (- range start) val-code)) |
| 501 | (while (and (setq elt (car tail) range (car elt)) |
| 502 | (integerp range) |
| 503 | (<= range limit)) |
| 504 | (setq new-val (funcall val-func (nth prop-idx elt))) |
| 505 | (if (not (eq val new-val)) |
| 506 | (setq val new-val |
| 507 | val-code (if val (unidata-encode-val |
| 508 | val-list val (and (eq prop 'bidi-class) |
| 509 | bidi-warning))))) |
| 510 | (if val-code |
| 511 | (aset vec (- range start) val-code)) |
| 512 | (setq tail (cdr tail))) |
| 513 | (setq str "\002" val-code -1 count 0) |
| 514 | (mapc #'(lambda (x) |
| 515 | (if (= val-code x) |
| 516 | (setq count (1+ count)) |
| 517 | (if (> count 2) |
| 518 | (setq str (concat str (string val-code |
| 519 | (+ count 128)))) |
| 520 | (if (= count 2) |
| 521 | (setq str (concat str (string val-code val-code))) |
| 522 | (if (= count 1) |
| 523 | (setq str (concat str (string val-code)))))) |
| 524 | (setq val-code x count 1))) |
| 525 | vec) |
| 526 | (if (= count 128) |
| 527 | (if val |
| 528 | (set-char-table-range table (cons start limit) val-code)) |
| 529 | (if (= val-code 0) |
| 530 | (set-char-table-range table (cons start limit) str) |
| 531 | (if (> count 2) |
| 532 | (setq str (concat str (string val-code (+ count 128)))) |
| 533 | (if (= count 2) |
| 534 | (setq str (concat str (string val-code val-code))) |
| 535 | (setq str (concat str (string val-code))))) |
| 536 | (set-char-table-range table (cons start limit) str)))))) |
| 537 | |
| 538 | (set-char-table-extra-slot table 0 prop) |
| 539 | (set-char-table-extra-slot table 4 (vconcat (mapcar 'car val-list))) |
| 540 | table)) |
| 541 | |
| 542 | (defun unidata-gen-table-symbol (prop default-value val-list) |
| 543 | (let ((table (unidata-gen-table prop |
| 544 | #'(lambda (x) (and (> (length x) 0) |
| 545 | (intern x))) |
| 546 | default-value val-list))) |
| 547 | (set-char-table-extra-slot table 1 0) |
| 548 | (set-char-table-extra-slot table 2 1) |
| 549 | table)) |
| 550 | |
| 551 | (defun unidata-gen-table-integer (prop default-value val-list) |
| 552 | (let ((table (unidata-gen-table prop |
| 553 | #'(lambda (x) (and (> (length x) 0) |
| 554 | (string-to-number x))) |
| 555 | default-value val-list))) |
| 556 | (set-char-table-extra-slot table 1 0) |
| 557 | (set-char-table-extra-slot table 2 1) |
| 558 | table)) |
| 559 | |
| 560 | (defun unidata-gen-table-numeric (prop default-value val-list) |
| 561 | (let ((table (unidata-gen-table prop |
| 562 | #'(lambda (x) |
| 563 | (if (string-match "/" x) |
| 564 | (/ (float (string-to-number x)) |
| 565 | (string-to-number |
| 566 | (substring x (match-end 0)))) |
| 567 | (if (> (length x) 0) |
| 568 | (string-to-number x)))) |
| 569 | default-value val-list))) |
| 570 | (set-char-table-extra-slot table 1 0) |
| 571 | (set-char-table-extra-slot table 2 2) |
| 572 | table)) |
| 573 | |
| 574 | \f |
| 575 | ;; WORD-LIST TABLE |
| 576 | |
| 577 | ;; If the table is for `name' property, each character in the string |
| 578 | ;; is one of these: |
| 579 | ;; DIFF-HEAD-CODE (0, 1, or 2): |
| 580 | ;; specifies how to decode the following characters. |
| 581 | ;; WORD-CODE (3..#x7FF excluding '-', '0'..'9', 'A'..'Z'): |
| 582 | ;; specifies an index number into WORD-TABLE (see below) |
| 583 | ;; Otherwise (' ', '-', '0'..'9', 'A'..'Z'): |
| 584 | ;; specifies a literal word. |
| 585 | ;; |
| 586 | ;; The 4th slots is a vector: |
| 587 | ;; [ WORD-TABLE BLOCK-NAME HANGUL-JAMO-TABLE ] |
| 588 | ;; WORD-TABLE is a vector of word symbols. |
| 589 | ;; BLOCK-NAME is a vector of name symbols for a block of characters. |
| 590 | ;; HANGUL-JAMO-TABLE is `unidata-name-jamo-name-table'. |
| 591 | |
| 592 | ;; Return the difference of symbol list L1 and L2 in this form: |
| 593 | ;; (DIFF-HEAD SYM1 SYM2 ...) |
| 594 | ;; DIFF-HEAD is ((SAME-HEAD-LENGTH * 16) + SAME-TAIL-LENGTH). |
| 595 | ;; Ex: If L1 is (a b c d e f) and L2 is (a g h e f), this function |
| 596 | ;; returns ((+ (* 1 16) 2) g h). |
| 597 | ;; It means that we can get L2 from L1 by prepending the first element |
| 598 | ;; of L1 and appending the last 2 elements of L1 to the list (g h). |
| 599 | ;; If L1 and L2 don't have common elements at the head and tail, |
| 600 | ;; set DIFF-HEAD to -1 and SYM1 ... to the elements of L2. |
| 601 | |
| 602 | (defun unidata-word-list-diff (l1 l2) |
| 603 | (let ((beg 0) |
| 604 | (end 0) |
| 605 | (len1 (length l1)) |
| 606 | (len2 (length l2)) |
| 607 | result) |
| 608 | (when (< len1 16) |
| 609 | (while (and l1 (eq (car l1) (car l2))) |
| 610 | (setq beg (1+ beg) |
| 611 | l1 (cdr l1) len1 (1- len1) l2 (cdr l2) len2 (1- len2))) |
| 612 | (while (and (< end len1) (< end len2) |
| 613 | (eq (nth (- len1 end 1) l1) (nth (- len2 end 1) l2))) |
| 614 | (setq end (1+ end)))) |
| 615 | (if (= (+ beg end) 0) |
| 616 | (setq result (list -1)) |
| 617 | (setq result (list (+ (* beg 16) (+ beg (- len1 end)))))) |
| 618 | (while (< end len2) |
| 619 | (setcdr result (cons (nth (- len2 end 1) l2) (cdr result))) |
| 620 | (setq end (1+ end))) |
| 621 | result)) |
| 622 | |
| 623 | ;; Return a compressed form of the vector VEC. Each element of VEC is |
| 624 | ;; a list of symbols of which names can be concatenated to form a |
| 625 | ;; character name. This function changes those elements into |
| 626 | ;; compressed forms by utilizing the fact that diff of consecutive |
| 627 | ;; elements is usually small. |
| 628 | |
| 629 | (defun unidata-word-list-compress (vec) |
| 630 | (let (last-elt last-idx diff-head tail elt val) |
| 631 | (dotimes (i 128) |
| 632 | (setq elt (aref vec i)) |
| 633 | (when elt |
| 634 | (if (null last-elt) |
| 635 | (setq diff-head -1 |
| 636 | val (cons 0 elt)) |
| 637 | (setq val (unidata-word-list-diff last-elt elt)) |
| 638 | (if (= (car val) -1) |
| 639 | (setq diff-head -1 |
| 640 | val (cons 0 (cdr val))) |
| 641 | (if (eq diff-head (car val)) |
| 642 | (setq val (cons 2 (cdr val))) |
| 643 | (setq diff-head (car val)) |
| 644 | (if (>= diff-head 0) |
| 645 | (setq val (cons 1 val)))))) |
| 646 | (aset vec i val) |
| 647 | (setq last-idx i last-elt elt))) |
| 648 | (if (not last-idx) |
| 649 | (setq vec nil) |
| 650 | (if (< last-idx 127) |
| 651 | (let ((shorter (make-vector (1+ last-idx) nil))) |
| 652 | (dotimes (i (1+ last-idx)) |
| 653 | (aset shorter i (aref vec i))) |
| 654 | (setq vec shorter)))) |
| 655 | vec)) |
| 656 | |
| 657 | ;; Encode the word index IDX into a characters code that can be |
| 658 | ;; embedded in a string. |
| 659 | |
| 660 | (defsubst unidata-encode-word (idx) |
| 661 | ;; Exclude 0, 1, 2. |
| 662 | (+ idx 3)) |
| 663 | |
| 664 | ;; Decode the character code CODE (that is embedded in a string) into |
| 665 | ;; the corresponding word name by looking up WORD-TABLE. |
| 666 | |
| 667 | (defsubst unidata-decode-word (code word-table) |
| 668 | (setq code (- code 3)) |
| 669 | (if (< code (length word-table)) |
| 670 | (aref word-table code))) |
| 671 | |
| 672 | ;; Table of short transliterated name symbols of Hangul Jamo divided |
| 673 | ;; into Choseong, Jungseong, and Jongseong. |
| 674 | |
| 675 | (defconst unidata-name-jamo-name-table |
| 676 | [[G GG N D DD R M B BB S SS nil J JJ C K T P H] |
| 677 | [A AE YA YAE EO E YEO YE O WA WAE OE YO U WEO WE WI YU EU YI I] |
| 678 | [G GG GS N NJ NH D L LG LM LB LS LT LP LH M B BS S SS NG J C K T P H]]) |
| 679 | |
| 680 | ;; Return a name of CHAR. VAL is the current value of (aref TABLE |
| 681 | ;; CHAR). |
| 682 | |
| 683 | (defun unidata-get-name (char val table) |
| 684 | (cond |
| 685 | ((stringp val) |
| 686 | (if (> (aref val 0) 0) |
| 687 | val |
| 688 | (let* ((first-char (lsh (lsh char -7) 7)) |
| 689 | (word-table (aref (char-table-extra-slot table 4) 0)) |
| 690 | (i 1) |
| 691 | (len (length val)) |
| 692 | (vec (make-vector 128 nil)) |
| 693 | (idx 0) |
| 694 | (case-fold-search nil) |
| 695 | c word-list tail-list last-list word diff-head) |
| 696 | (while (< i len) |
| 697 | (setq c (aref val i)) |
| 698 | (if (< c 3) |
| 699 | (progn |
| 700 | (if (or word-list tail-list) |
| 701 | (aset vec idx |
| 702 | (setq last-list (nconc word-list tail-list)))) |
| 703 | (setq i (1+ i) idx (1+ idx) |
| 704 | word-list nil tail-list nil) |
| 705 | (if (> c 0) |
| 706 | (let ((l last-list)) |
| 707 | (if (= c 1) |
| 708 | (setq diff-head |
| 709 | (prog1 (aref val i) (setq i (1+ i))))) |
| 710 | (setq tail-list (nthcdr (% diff-head 16) last-list)) |
| 711 | (dotimes (i (/ diff-head 16)) |
| 712 | (setq word-list (nconc word-list (list (car l))) |
| 713 | l (cdr l)))))) |
| 714 | (setq word-list |
| 715 | (nconc word-list |
| 716 | (list (symbol-name |
| 717 | (unidata-decode-word c word-table)))) |
| 718 | i (1+ i)))) |
| 719 | (if (or word-list tail-list) |
| 720 | (aset vec idx (nconc word-list tail-list))) |
| 721 | (setq val nil) |
| 722 | (dotimes (i 128) |
| 723 | (setq c (+ first-char i)) |
| 724 | (let ((name (aref vec i))) |
| 725 | (if name |
| 726 | (let ((tail (cdr (setq name (copy-sequence name)))) |
| 727 | elt) |
| 728 | (while tail |
| 729 | (setq elt (car tail)) |
| 730 | (or (string= elt "-") |
| 731 | (progn |
| 732 | (setcdr tail (cons elt (cdr tail))) |
| 733 | (setcar tail " "))) |
| 734 | (setq tail (cddr tail))) |
| 735 | (setq name (apply 'concat name)))) |
| 736 | (aset table c name) |
| 737 | (if (= c char) |
| 738 | (setq val name)))) |
| 739 | val))) |
| 740 | |
| 741 | ((and (integerp val) (> val 0)) |
| 742 | (let* ((symbol-table (aref (char-table-extra-slot table 4) 1)) |
| 743 | (sym (aref symbol-table (1- val)))) |
| 744 | (cond ((eq sym 'HANGUL\ SYLLABLE) |
| 745 | (let ((jamo-name-table (aref (char-table-extra-slot table 4) 2))) |
| 746 | ;; SIndex = S - SBase |
| 747 | (setq char (- char #xAC00)) |
| 748 | (let ( ;; LIndex = SIndex / NCount |
| 749 | (L (/ char 588)) |
| 750 | ;; VIndex = (SIndex % NCount) * TCount |
| 751 | (V (/ (% char 588) 28)) |
| 752 | ;; TIndex = SIndex % TCount |
| 753 | (T (% char 28))) |
| 754 | (format "HANGUL SYLLABLE %s%s%s" |
| 755 | ;; U+110B is nil in this table. |
| 756 | (or (aref (aref jamo-name-table 0) L) "") |
| 757 | (aref (aref jamo-name-table 1) V) |
| 758 | (if (= T 0) "" |
| 759 | (aref (aref jamo-name-table 2) (1- T))))))) |
| 760 | ((eq sym 'CJK\ IDEOGRAPH) |
| 761 | (format "%s-%04X" sym char)) |
| 762 | ((eq sym 'CJK\ COMPATIBILITY\ IDEOGRAPH) |
| 763 | (format "%s-%04X" sym char)) |
| 764 | ((eq sym 'VARIATION\ SELECTOR) |
| 765 | (format "%s-%d" sym (+ (- char #xe0100) 17)))))))) |
| 766 | |
| 767 | ;; Store VAL as the name of CHAR in TABLE. |
| 768 | |
| 769 | (defun unidata-put-name (char val table) |
| 770 | (let ((current-val (aref table char))) |
| 771 | (if (and (stringp current-val) (= (aref current-val 0) 0)) |
| 772 | (funcall (char-table-extra-slot table 1) char current-val table)) |
| 773 | (aset table char val))) |
| 774 | |
| 775 | (defun unidata-get-decomposition (char val table) |
| 776 | (cond |
| 777 | ((not val) |
| 778 | (list char)) |
| 779 | |
| 780 | ((consp val) |
| 781 | val) |
| 782 | |
| 783 | ((stringp val) |
| 784 | (if (> (aref val 0) 0) |
| 785 | val |
| 786 | (let* ((first-char (lsh (lsh char -7) 7)) |
| 787 | (word-table (char-table-extra-slot table 4)) |
| 788 | (i 1) |
| 789 | (len (length val)) |
| 790 | (vec (make-vector 128 nil)) |
| 791 | (idx 0) |
| 792 | (case-fold-search nil) |
| 793 | c word-list tail-list last-list word diff-head) |
| 794 | (while (< i len) |
| 795 | (setq c (aref val i)) |
| 796 | (if (< c 3) |
| 797 | (progn |
| 798 | (if (or word-list tail-list) |
| 799 | (aset vec idx |
| 800 | (setq last-list (nconc word-list tail-list)))) |
| 801 | (setq i (1+ i) idx (1+ idx) |
| 802 | word-list nil tail-list nil) |
| 803 | (if (> c 0) |
| 804 | (let ((l last-list)) |
| 805 | (if (= c 1) |
| 806 | (setq diff-head |
| 807 | (prog1 (aref val i) (setq i (1+ i))))) |
| 808 | (setq tail-list (nthcdr (% diff-head 16) last-list)) |
| 809 | (dotimes (i (/ diff-head 16)) |
| 810 | (setq word-list (nconc word-list (list (car l))) |
| 811 | l (cdr l)))))) |
| 812 | (setq word-list |
| 813 | (nconc word-list |
| 814 | (list (or (unidata-decode-word c word-table) c))) |
| 815 | i (1+ i)))) |
| 816 | (if (or word-list tail-list) |
| 817 | (aset vec idx (nconc word-list tail-list))) |
| 818 | (dotimes (i 128) |
| 819 | (aset table (+ first-char i) (aref vec i))) |
| 820 | (setq val (aref vec (- char first-char))) |
| 821 | (or val (list char))))) |
| 822 | |
| 823 | ;; Hangul syllable |
| 824 | ((and (eq val 0) (>= char #xAC00) (<= char #xD7A3)) |
| 825 | ;; SIndex = S (char) - SBase (#xAC00) |
| 826 | (setq char (- char #xAC00)) |
| 827 | (let (;; L = LBase + SIndex / NCount |
| 828 | (L (+ #x1100 (/ char 588))) |
| 829 | ;; V = VBase + (SIndex % NCount) * TCount |
| 830 | (V (+ #x1161 (/ (% char 588) 28))) |
| 831 | ;; LV = SBase + (SIndex / TCount) * TCount |
| 832 | (LV (+ #xAC00 (* (/ char 28) 28))) |
| 833 | ;; T = TBase + SIndex % TCount |
| 834 | (T (+ #x11A7 (% char 28)))) |
| 835 | (if (= T #x11A7) |
| 836 | (list L V) |
| 837 | (list LV T)))) |
| 838 | |
| 839 | )) |
| 840 | |
| 841 | ;; Store VAL as the decomposition information of CHAR in TABLE. |
| 842 | |
| 843 | (defun unidata-put-decomposition (char val table) |
| 844 | (let ((current-val (aref table char))) |
| 845 | (if (and (stringp current-val) (= (aref current-val 0) 0)) |
| 846 | (funcall (char-table-extra-slot table 1) char current-val table)) |
| 847 | (aset table char val))) |
| 848 | |
| 849 | ;; UnicodeData.txt contains these lines: |
| 850 | ;; 0000;<control>;Cc;0;BN;;;;;N;NULL;;;; |
| 851 | ;; ... |
| 852 | ;; 0020;SPACE;Zs;0;WS;;;;;N;;;;; |
| 853 | ;; ... |
| 854 | ;; The following command yields a file of about 96K bytes. |
| 855 | ;; % gawk -F ';' '{print $1,$2;}' < UnicodeData.txt | gzip > temp.gz |
| 856 | ;; With the following function, we can get a file of almost the same |
| 857 | ;; the size. |
| 858 | |
| 859 | ;; Generate a char-table for character names. |
| 860 | |
| 861 | (defun unidata-gen-table-word-list (prop val-func) |
| 862 | (let ((table (make-char-table 'char-code-property-table)) |
| 863 | (prop-idx (unidata-prop-index prop)) |
| 864 | (word-list (list nil)) |
| 865 | word-table |
| 866 | block-list block-word-table block-end |
| 867 | tail elt range val idx slot) |
| 868 | (setq tail unidata-list) |
| 869 | (setq block-end -1) |
| 870 | (while tail |
| 871 | (setq elt (car tail) tail (cdr tail)) |
| 872 | (setq range (car elt) |
| 873 | val (funcall val-func (nth prop-idx elt))) |
| 874 | ;; Treat the sequence of "CJK COMPATIBILITY IDEOGRAPH-XXXX" and |
| 875 | ;; "VARIATION SELECTOR-XXX" as a block. |
| 876 | (if (and (consp val) (eq prop 'name) |
| 877 | (or (and (eq (car val) 'CJK) |
| 878 | (eq (nth 1 val) 'COMPATIBILITY)) |
| 879 | (and (>= range #xe0100) |
| 880 | (eq (car val) 'VARIATION) |
| 881 | (eq (nth 1 val) 'SELECTOR)))) |
| 882 | (let ((first (car val)) |
| 883 | (second (nth 1 val)) |
| 884 | (start range)) |
| 885 | (while (and (setq elt (car tail) range (car elt) |
| 886 | val (funcall val-func (nth prop-idx elt))) |
| 887 | (consp val) |
| 888 | (eq first (car val)) |
| 889 | (eq second (nth 1 val))) |
| 890 | (setq block-end range |
| 891 | tail (cdr tail))) |
| 892 | (setq range (cons start block-end) |
| 893 | val (if (eq first 'CJK) 'CJK\ COMPATIBILITY\ IDEOGRAPH |
| 894 | 'VARIATION\ SELECTOR)))) |
| 895 | |
| 896 | (if (consp range) |
| 897 | (if val |
| 898 | (let ((slot (assq val block-list))) |
| 899 | (setq range (cons (car range) (cdr range))) |
| 900 | (setq block-end (cdr range)) |
| 901 | (if slot |
| 902 | (nconc slot (list range)) |
| 903 | (push (list val range) block-list)))) |
| 904 | (let* ((start (lsh (lsh range -7) 7)) |
| 905 | (limit (+ start 127)) |
| 906 | (first tail) |
| 907 | (vec (make-vector 128 nil)) |
| 908 | c name len) |
| 909 | (if (<= start block-end) |
| 910 | ;; START overlap with the previous block. |
| 911 | (aset table range (nth prop-idx elt)) |
| 912 | (if val |
| 913 | (aset vec (- range start) val)) |
| 914 | (while (and (setq elt (car tail) range (car elt)) |
| 915 | (integerp range) |
| 916 | (<= range limit)) |
| 917 | (setq val (funcall val-func (nth prop-idx elt))) |
| 918 | (if val |
| 919 | (aset vec (- range start) val)) |
| 920 | (setq tail (cdr tail))) |
| 921 | (setq vec (unidata-word-list-compress vec)) |
| 922 | (when vec |
| 923 | (dotimes (i (length vec)) |
| 924 | (dolist (elt (aref vec i)) |
| 925 | (if (symbolp elt) |
| 926 | (let ((slot (assq elt word-list))) |
| 927 | (if slot |
| 928 | (setcdr slot (1+ (cdr slot))) |
| 929 | (setcdr word-list |
| 930 | (cons (cons elt 1) (cdr word-list)))))))) |
| 931 | (set-char-table-range table (cons start limit) vec)))))) |
| 932 | (setq word-list (sort (cdr word-list) |
| 933 | #'(lambda (x y) (> (cdr x) (cdr y))))) |
| 934 | (setq tail word-list idx 0) |
| 935 | (while tail |
| 936 | (setcdr (car tail) (unidata-encode-word idx)) |
| 937 | (setq idx (1+ idx) tail (cdr tail))) |
| 938 | (setq word-table (make-vector (length word-list) nil)) |
| 939 | (setq idx 0) |
| 940 | (dolist (elt word-list) |
| 941 | (aset word-table idx (car elt)) |
| 942 | (setq idx (1+ idx))) |
| 943 | |
| 944 | (if (and (eq prop 'decomposition) |
| 945 | (> idx 32)) |
| 946 | (error "Too many symbols in decomposition data")) |
| 947 | |
| 948 | (dotimes (i (/ #x110000 128)) |
| 949 | (let* ((idx (* i 128)) |
| 950 | (vec (aref table idx))) |
| 951 | (when (vectorp vec) |
| 952 | (dotimes (i (length vec)) |
| 953 | (let ((tail (aref vec i)) |
| 954 | elt code) |
| 955 | (if (not tail) |
| 956 | (aset vec i "\0") |
| 957 | (while tail |
| 958 | (setq elt (car tail) |
| 959 | code (if (integerp elt) elt |
| 960 | (cdr (assq elt word-list)))) |
| 961 | (setcar tail (string code)) |
| 962 | (setq tail (cdr tail))) |
| 963 | (aset vec i (mapconcat 'identity (aref vec i) ""))))) |
| 964 | (set-char-table-range |
| 965 | table (cons idx (+ idx 127)) |
| 966 | (mapconcat 'identity vec ""))))) |
| 967 | |
| 968 | (setq block-word-table (make-vector (length block-list) nil)) |
| 969 | (setq idx 0) |
| 970 | (dolist (elt block-list) |
| 971 | (dolist (e (cdr elt)) |
| 972 | (set-char-table-range table e (1+ idx))) |
| 973 | (aset block-word-table idx (car elt)) |
| 974 | (setq idx (1+ idx))) |
| 975 | |
| 976 | (set-char-table-extra-slot table 0 prop) |
| 977 | (set-char-table-extra-slot table 4 (cons word-table block-word-table)) |
| 978 | table)) |
| 979 | |
| 980 | (defun unidata-split-name (str) |
| 981 | (if (symbolp str) |
| 982 | str |
| 983 | (let ((len (length str)) |
| 984 | (l nil) |
| 985 | (idx 0) |
| 986 | c) |
| 987 | (if (= len 0) |
| 988 | nil |
| 989 | (dotimes (i len) |
| 990 | (setq c (aref str i)) |
| 991 | (if (= c 32) |
| 992 | (setq l (cons (intern (substring str idx i)) l) |
| 993 | idx (1+ i)) |
| 994 | (if (and (= c ?-) (< idx i) |
| 995 | (< (1+ i) len) (/= (aref str (1+ i)) 32)) |
| 996 | (setq l (cons '- (cons (intern (substring str idx i)) l)) |
| 997 | idx (1+ i))))) |
| 998 | (nreverse (cons (intern (substring str idx)) l)))))) |
| 999 | |
| 1000 | (defun unidata--ensure-compiled (&rest funcs) |
| 1001 | (dolist (fun funcs) |
| 1002 | (or (byte-code-function-p (symbol-function fun)) |
| 1003 | (byte-compile fun)))) |
| 1004 | |
| 1005 | (defun unidata-gen-table-name (prop &rest ignore) |
| 1006 | (let* ((table (unidata-gen-table-word-list prop 'unidata-split-name)) |
| 1007 | (word-tables (char-table-extra-slot table 4))) |
| 1008 | (unidata--ensure-compiled 'unidata-get-name 'unidata-put-name) |
| 1009 | (set-char-table-extra-slot table 1 (symbol-function 'unidata-get-name)) |
| 1010 | (set-char-table-extra-slot table 2 (symbol-function 'unidata-put-name)) |
| 1011 | |
| 1012 | (if (eq prop 'name) |
| 1013 | (set-char-table-extra-slot table 4 |
| 1014 | (vector (car word-tables) |
| 1015 | (cdr word-tables) |
| 1016 | unidata-name-jamo-name-table)) |
| 1017 | (set-char-table-extra-slot table 4 |
| 1018 | (vector (car word-tables)))) |
| 1019 | table)) |
| 1020 | |
| 1021 | (defun unidata-split-decomposition (str) |
| 1022 | (if (symbolp str) |
| 1023 | str |
| 1024 | (let ((len (length str)) |
| 1025 | (l nil) |
| 1026 | (idx 0) |
| 1027 | c) |
| 1028 | (if (= len 0) |
| 1029 | nil |
| 1030 | (dotimes (i len) |
| 1031 | (setq c (aref str i)) |
| 1032 | (if (= c 32) |
| 1033 | (setq l (if (= (aref str idx) ?<) |
| 1034 | (cons (intern (substring str (1+ idx) (1- i))) l) |
| 1035 | (cons (string-to-number (substring str idx i) 16) l)) |
| 1036 | idx (1+ i)))) |
| 1037 | (if (= (aref str idx) ?<) |
| 1038 | (setq l (cons (intern (substring str (1+ idx) (1- len))) l)) |
| 1039 | (setq l (cons (string-to-number (substring str idx len) 16) l))) |
| 1040 | (nreverse l))))) |
| 1041 | |
| 1042 | |
| 1043 | (defun unidata-gen-table-decomposition (prop &rest ignore) |
| 1044 | (let* ((table (unidata-gen-table-word-list prop 'unidata-split-decomposition)) |
| 1045 | (word-tables (char-table-extra-slot table 4))) |
| 1046 | (unidata--ensure-compiled 'unidata-get-decomposition |
| 1047 | 'unidata-put-decomposition) |
| 1048 | (set-char-table-extra-slot table 1 |
| 1049 | (symbol-function 'unidata-get-decomposition)) |
| 1050 | (set-char-table-extra-slot table 2 |
| 1051 | (symbol-function 'unidata-put-decomposition)) |
| 1052 | (set-char-table-extra-slot table 4 (car word-tables)) |
| 1053 | table)) |
| 1054 | |
| 1055 | |
| 1056 | \f |
| 1057 | (defun unidata-describe-general-category (val) |
| 1058 | (cdr (assq val |
| 1059 | '((nil . "Uknown") |
| 1060 | (Lu . "Letter, Uppercase") |
| 1061 | (Ll . "Letter, Lowercase") |
| 1062 | (Lt . "Letter, Titlecase") |
| 1063 | (Lm . "Letter, Modifier") |
| 1064 | (Lo . "Letter, Other") |
| 1065 | (Mn . "Mark, Nonspacing") |
| 1066 | (Mc . "Mark, Spacing Combining") |
| 1067 | (Me . "Mark, Enclosing") |
| 1068 | (Nd . "Number, Decimal Digit") |
| 1069 | (Nl . "Number, Letter") |
| 1070 | (No . "Number, Other") |
| 1071 | (Pc . "Punctuation, Connector") |
| 1072 | (Pd . "Punctuation, Dash") |
| 1073 | (Ps . "Punctuation, Open") |
| 1074 | (Pe . "Punctuation, Close") |
| 1075 | (Pi . "Punctuation, Initial quote") |
| 1076 | (Pf . "Punctuation, Final quote") |
| 1077 | (Po . "Punctuation, Other") |
| 1078 | (Sm . "Symbol, Math") |
| 1079 | (Sc . "Symbol, Currency") |
| 1080 | (Sk . "Symbol, Modifier") |
| 1081 | (So . "Symbol, Other") |
| 1082 | (Zs . "Separator, Space") |
| 1083 | (Zl . "Separator, Line") |
| 1084 | (Zp . "Separator, Paragraph") |
| 1085 | (Cc . "Other, Control") |
| 1086 | (Cf . "Other, Format") |
| 1087 | (Cs . "Other, Surrogate") |
| 1088 | (Co . "Other, Private Use") |
| 1089 | (Cn . "Other, Not Assigned"))))) |
| 1090 | |
| 1091 | (defun unidata-describe-canonical-combining-class (val) |
| 1092 | (cdr (assq val |
| 1093 | '((0 . "Spacing, split, enclosing, reordrant, and Tibetan subjoined") |
| 1094 | (1 . "Overlays and interior") |
| 1095 | (7 . "Nuktas") |
| 1096 | (8 . "Hiragana/Katakana voicing marks") |
| 1097 | (9 . "Viramas") |
| 1098 | (10 . "Start of fixed position classes") |
| 1099 | (199 . "End of fixed position classes") |
| 1100 | (200 . "Below left attached") |
| 1101 | (202 . "Below attached") |
| 1102 | (204 . "Below right attached") |
| 1103 | (208 . "Left attached (reordrant around single base character)") |
| 1104 | (210 . "Right attached") |
| 1105 | (212 . "Above left attached") |
| 1106 | (214 . "Above attached") |
| 1107 | (216 . "Above right attached") |
| 1108 | (218 . "Below left") |
| 1109 | (220 . "Below") |
| 1110 | (222 . "Below right") |
| 1111 | (224 . "Left (reordrant around single base character)") |
| 1112 | (226 . "Right") |
| 1113 | (228 . "Above left") |
| 1114 | (230 . "Above") |
| 1115 | (232 . "Above right") |
| 1116 | (233 . "Double below") |
| 1117 | (234 . "Double above") |
| 1118 | (240 . "Below (iota subscript)"))))) |
| 1119 | |
| 1120 | (defun unidata-describe-bidi-class (val) |
| 1121 | (cdr (assq val |
| 1122 | '((L . "Left-to-Right") |
| 1123 | (LRE . "Left-to-Right Embedding") |
| 1124 | (LRO . "Left-to-Right Override") |
| 1125 | (R . "Right-to-Left") |
| 1126 | (AL . "Right-to-Left Arabic") |
| 1127 | (RLE . "Right-to-Left Embedding") |
| 1128 | (RLO . "Right-to-Left Override") |
| 1129 | (PDF . "Pop Directional Format") |
| 1130 | (EN . "European Number") |
| 1131 | (ES . "European Number Separator") |
| 1132 | (ET . "European Number Terminator") |
| 1133 | (AN . "Arabic Number") |
| 1134 | (CS . "Common Number Separator") |
| 1135 | (NSM . "Non-Spacing Mark") |
| 1136 | (BN . "Boundary Neutral") |
| 1137 | (B . "Paragraph Separator") |
| 1138 | (S . "Segment Separator") |
| 1139 | (WS . "Whitespace") |
| 1140 | (ON . "Other Neutrals"))))) |
| 1141 | |
| 1142 | (defun unidata-describe-decomposition (val) |
| 1143 | (mapconcat |
| 1144 | #'(lambda (x) |
| 1145 | (if (symbolp x) (symbol-name x) |
| 1146 | (concat (string ?') |
| 1147 | (compose-string (string x) 0 1 (string ?\t x ?\t)) |
| 1148 | (string ?')))) |
| 1149 | val " ")) |
| 1150 | |
| 1151 | (defun unidata-gen-mirroring-list () |
| 1152 | (let ((head (list nil)) |
| 1153 | tail) |
| 1154 | (with-temp-buffer |
| 1155 | (insert-file-contents (expand-file-name "BidiMirroring.txt" unidata-dir)) |
| 1156 | (goto-char (point-min)) |
| 1157 | (setq tail head) |
| 1158 | (while (re-search-forward "^\\([0-9A-F]+\\);\\s +\\([0-9A-F]+\\)" nil t) |
| 1159 | (let ((char (string-to-number (match-string 1) 16)) |
| 1160 | (mirror (match-string 2))) |
| 1161 | (setq tail (setcdr tail (list (list char mirror))))))) |
| 1162 | (cdr head))) |
| 1163 | |
| 1164 | ;; Verify if we can retrieve correct values from the generated |
| 1165 | ;; char-tables. |
| 1166 | |
| 1167 | (defun unidata-check () |
| 1168 | (dolist (elt unidata-prop-alist) |
| 1169 | (let* ((prop (car elt)) |
| 1170 | (index (unidata-prop-index prop)) |
| 1171 | (generator (unidata-prop-generator prop)) |
| 1172 | (table (progn |
| 1173 | (message "Generating %S table..." prop) |
| 1174 | (funcall generator prop))) |
| 1175 | (decoder (char-table-extra-slot table 1)) |
| 1176 | (check #x400)) |
| 1177 | (dolist (e unidata-list) |
| 1178 | (let ((char (car e)) |
| 1179 | (val1 (nth index e)) |
| 1180 | val2) |
| 1181 | (if (and (stringp val1) (= (length val1) 0)) |
| 1182 | (setq val1 nil)) |
| 1183 | (unless (consp char) |
| 1184 | (setq val2 (funcall decoder char (aref table char) table)) |
| 1185 | (if val1 |
| 1186 | (cond ((eq generator 'unidata-gen-table-symbol) |
| 1187 | (setq val1 (intern val1))) |
| 1188 | ((eq generator 'unidata-gen-table-integer) |
| 1189 | (setq val1 (string-to-number val1))) |
| 1190 | ((eq generator 'unidata-gen-table-character) |
| 1191 | (setq val1 (string-to-number val1 16))) |
| 1192 | ((eq generator 'unidata-gen-table-decomposition) |
| 1193 | (setq val1 (unidata-split-decomposition val1))))) |
| 1194 | (when (>= char check) |
| 1195 | (message "%S %04X" prop check) |
| 1196 | (setq check (+ check #x400))) |
| 1197 | (or (equal val1 val2) |
| 1198 | (insert (format "> %04X %S\n< %04X %S\n" |
| 1199 | char val1 char val2))) |
| 1200 | (sit-for 0))))))) |
| 1201 | |
| 1202 | ;; The entry function. It generates files described in the header |
| 1203 | ;; comment of this file. |
| 1204 | |
| 1205 | ;; Write files (charprop.el, uni-*.el) to dest-dir (default PWD), |
| 1206 | ;; using as input files from data-dir, and |
| 1207 | ;; unidata-text-file (default "unidata.txt" in PWD). |
| 1208 | (defun unidata-gen-files (&optional data-dir dest-dir unidata-text-file) |
| 1209 | (or data-dir |
| 1210 | (setq data-dir (pop command-line-args-left) |
| 1211 | dest-dir (or (pop command-line-args-left) default-directory) |
| 1212 | unidata-text-file (or (pop command-line-args-left) |
| 1213 | (expand-file-name "unidata.txt")))) |
| 1214 | (let ((coding-system-for-write 'utf-8-unix) |
| 1215 | (charprop-file (expand-file-name "charprop.el" dest-dir)) |
| 1216 | (unidata-dir data-dir)) |
| 1217 | (dolist (elt unidata-prop-alist) |
| 1218 | (let* ((prop (car elt)) |
| 1219 | (file (expand-file-name (unidata-prop-file prop) dest-dir))) |
| 1220 | (if (file-exists-p file) |
| 1221 | (delete-file file)))) |
| 1222 | (unidata-setup-list unidata-text-file) |
| 1223 | (with-temp-file charprop-file |
| 1224 | (insert ";; Automatically generated by unidata-gen.el.\n") |
| 1225 | (dolist (elt unidata-prop-alist) |
| 1226 | (let* ((prop (car elt)) |
| 1227 | (generator (unidata-prop-generator prop)) |
| 1228 | (file (expand-file-name (unidata-prop-file prop) dest-dir)) |
| 1229 | (basename (file-name-nondirectory file)) |
| 1230 | (docstring (unidata-prop-docstring prop)) |
| 1231 | (describer (unidata-prop-describer prop)) |
| 1232 | (default-value (unidata-prop-default prop)) |
| 1233 | (val-list (unidata-prop-val-list prop)) |
| 1234 | table) |
| 1235 | ;; Filename in this comment line is extracted by sed in |
| 1236 | ;; Makefile. |
| 1237 | (insert (format ";; FILE: %s\n" basename)) |
| 1238 | (insert (format "(define-char-code-property '%S %S\n %S)\n" |
| 1239 | prop basename docstring)) |
| 1240 | (with-temp-buffer |
| 1241 | (message "Generating %s..." file) |
| 1242 | (when (file-exists-p file) |
| 1243 | (insert-file-contents file) |
| 1244 | (goto-char (point-max)) |
| 1245 | (search-backward ";; Local Variables:")) |
| 1246 | (setq table (funcall generator prop default-value val-list)) |
| 1247 | (when describer |
| 1248 | (unless (subrp (symbol-function describer)) |
| 1249 | (unidata--ensure-compiled describer) |
| 1250 | (setq describer (symbol-function describer))) |
| 1251 | (set-char-table-extra-slot table 3 describer)) |
| 1252 | (if (bobp) |
| 1253 | (insert ";; Copyright (C) 1991-2013 Unicode, Inc. |
| 1254 | ;; This file was generated from the Unicode data files at |
| 1255 | ;; http://www.unicode.org/Public/UNIDATA/. |
| 1256 | ;; See lisp/international/README for the copyright and permission notice.\n")) |
| 1257 | (insert (format "(define-char-code-property '%S\n %S\n %S)\n" |
| 1258 | prop table docstring)) |
| 1259 | (if (eobp) |
| 1260 | (insert ";; Local Variables:\n" |
| 1261 | ";; coding: utf-8\n" |
| 1262 | ";; version-control: never\n" |
| 1263 | ";; no-byte-compile: t\n" |
| 1264 | ";; no-update-autoloads: t\n" |
| 1265 | ";; End:\n\n" |
| 1266 | (format ";; %s ends here\n" basename))) |
| 1267 | (write-file file) |
| 1268 | (message "Generating %s...done" file)))) |
| 1269 | (message "Writing %s..." charprop-file) |
| 1270 | (insert ";; Local Variables:\n" |
| 1271 | ";; coding: utf-8\n" |
| 1272 | ";; version-control: never\n" |
| 1273 | ";; no-byte-compile: t\n" |
| 1274 | ";; no-update-autoloads: t\n" |
| 1275 | ";; End:\n\n" |
| 1276 | (format ";; %s ends here\n" |
| 1277 | (file-name-nondirectory charprop-file)))))) |
| 1278 | |
| 1279 | \f |
| 1280 | |
| 1281 | ;;; unidata-gen.el ends here |