src/category.c

   1 /* GNU Emacs routines to deal with category tables.
   2
   3 Copyright (C) 1998, 2001-2011  Free Software Foundation, Inc.
   4 Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   5   2005, 2006, 2007, 2008, 2009, 2010, 2011
   6   National Institute of Advanced Industrial Science and Technology (AIST)
   7   Registration Number H14PRO021
   8 Copyright (C) 2003
   9   National Institute of Advanced Industrial Science and Technology (AIST)
  10   Registration Number H13PRO009
  11
  12 This file is part of GNU Emacs.
  13
  14 GNU Emacs is free software: you can redistribute it and/or modify
  15 it under the terms of the GNU General Public License as published by
  16 the Free Software Foundation, either version 3 of the License, or
  17 (at your option) any later version.
  18
  19 GNU Emacs is distributed in the hope that it will be useful,
  20 but WITHOUT ANY WARRANTY; without even the implied warranty of
  21 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22 GNU General Public License for more details.
  23
  24 You should have received a copy of the GNU General Public License
  25 along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.  */
  26
  27
  28 /* Here we handle three objects: category, category set, and category
  29    table.  Read comments in the file category.h to understand them.  */
  30
  31 #include <config.h>
  32 #include <ctype.h>
  33 #include <setjmp.h>
  34 #include "lisp.h"
  35 #include "buffer.h"
  36 #include "character.h"
  37 #include "charset.h"
  38 #include "category.h"
  39 #include "keymap.h"
  40
  41 /* The version number of the latest category table.  Each category
  42    table has a unique version number.  It is assigned a new number
  43    also when it is modified.  When a regular expression is compiled
  44    into the struct re_pattern_buffer, the version number of the
  45    category table (of the current buffer) at that moment is also
  46    embedded in the structure.
  47
  48    For the moment, we are not using this feature.  */
  49 static int category_table_version;
  50
  51 Lisp_Object Qcategory_table, Qcategoryp, Qcategorysetp, Qcategory_table_p;
  52
  53 /* Temporary internal variable used in macro CHAR_HAS_CATEGORY.  */
  54 Lisp_Object _temp_category_set;
  55
  56 \f
  57 /* Category set staff.  */
  58
  59 static Lisp_Object hash_get_category_set (Lisp_Object, Lisp_Object);
  60
  61 static Lisp_Object
  62 hash_get_category_set (Lisp_Object table, Lisp_Object category_set)
  63 {
  64   Lisp_Object val;
  65   struct Lisp_Hash_Table *h;
  66   int i;
  67   unsigned hash;
  68
  69   if (NILP (XCHAR_TABLE (table)->extras[1]))
  70     XCHAR_TABLE (table)->extras[1]
  71       = make_hash_table (Qequal, make_number (DEFAULT_HASH_SIZE),
  72                          make_float (DEFAULT_REHASH_SIZE),
  73                          make_float (DEFAULT_REHASH_THRESHOLD),
  74                          Qnil, Qnil, Qnil);
  75   h = XHASH_TABLE (XCHAR_TABLE (table)->extras[1]);
  76   i = hash_lookup (h, category_set, &hash);
  77   if (i >= 0)
  78     return HASH_KEY (h, i);
  79   hash_put (h, category_set, Qnil, hash);
  80   return category_set;
  81 }
  82
  83
  84 DEFUN ("make-category-set", Fmake_category_set, Smake_category_set, 1, 1, 0,
  85        doc: /* Return a newly created category-set which contains CATEGORIES.
  86 CATEGORIES is a string of category mnemonics.
  87 The value is a bool-vector which has t at the indices corresponding to
  88 those categories.  */)
  89   (Lisp_Object categories)
  90 {
  91   Lisp_Object val;
  92   int len;
  93
  94   CHECK_STRING (categories);
  95   val = MAKE_CATEGORY_SET;
  96
  97   if (STRING_MULTIBYTE (categories))
  98     error ("Multibyte string in `make-category-set'");
  99
 100   len = SCHARS (categories);
 101   while (--len >= 0)
 102     {
 103       Lisp_Object category;
 104
 105       XSETFASTINT (category, SREF (categories, len));
 106       CHECK_CATEGORY (category);
 107       SET_CATEGORY_SET (val, category, Qt);
 108     }
 109   return val;
 110 }
 111
 112 \f
 113 /* Category staff.  */
 114
 115 Lisp_Object check_category_table (Lisp_Object table);
 116
 117 DEFUN ("define-category", Fdefine_category, Sdefine_category, 2, 3, 0,
 118        doc: /* Define CATEGORY as a category which is described by DOCSTRING.
 119 CATEGORY should be an ASCII printing character in the range ` ' to `~'.
 120 DOCSTRING is the documentation string of the category.  The first line
 121 should be a terse text (preferably less than 16 characters),
 122 and the rest lines should be the full description.
 123 The category is defined only in category table TABLE, which defaults to
 124 the current buffer's category table.  */)
 125   (Lisp_Object category, Lisp_Object docstring, Lisp_Object table)
 126 {
 127   CHECK_CATEGORY (category);
 128   CHECK_STRING (docstring);
 129   table = check_category_table (table);
 130
 131   if (!NILP (CATEGORY_DOCSTRING (table, XFASTINT (category))))
 132     error ("Category `%c' is already defined", XFASTINT (category));
 133   if (!NILP (Vpurify_flag))
 134     docstring = Fpurecopy (docstring);
 135   CATEGORY_DOCSTRING (table, XFASTINT (category)) = docstring;
 136
 137   return Qnil;
 138 }
 139
 140 DEFUN ("category-docstring", Fcategory_docstring, Scategory_docstring, 1, 2, 0,
 141        doc: /* Return the documentation string of CATEGORY, as defined in TABLE.
 142 TABLE should be a category table and defaults to the current buffer's
 143 category table.  */)
 144   (Lisp_Object category, Lisp_Object table)
 145 {
 146   CHECK_CATEGORY (category);
 147   table = check_category_table (table);
 148
 149   return CATEGORY_DOCSTRING (table, XFASTINT (category));
 150 }
 151
 152 DEFUN ("get-unused-category", Fget_unused_category, Sget_unused_category,
 153        0, 1, 0,
 154        doc: /* Return a category which is not yet defined in TABLE.
 155 If no category remains available, return nil.
 156 The optional argument TABLE specifies which category table to modify;
 157 it defaults to the current buffer's category table.  */)
 158   (Lisp_Object table)
 159 {
 160   int i;
 161
 162   table = check_category_table (table);
 163
 164   for (i = ' '; i <= '~'; i++)
 165     if (NILP (CATEGORY_DOCSTRING (table, i)))
 166       return make_number (i);
 167
 168   return Qnil;
 169 }
 170
 171 \f
 172 /* Category-table staff.  */
 173
 174 DEFUN ("category-table-p", Fcategory_table_p, Scategory_table_p, 1, 1, 0,
 175        doc: /* Return t if ARG is a category table.  */)
 176   (Lisp_Object arg)
 177 {
 178   if (CHAR_TABLE_P (arg)
 179       && EQ (XCHAR_TABLE (arg)->purpose, Qcategory_table))
 180     return Qt;
 181   return Qnil;
 182 }
 183
 184 /* If TABLE is nil, return the current category table.  If TABLE is
 185    not nil, check the validity of TABLE as a category table.  If
 186    valid, return TABLE itself, but if not valid, signal an error of
 187    wrong-type-argument.  */
 188
 189 Lisp_Object
 190 check_category_table (Lisp_Object table)
 191 {
 192   if (NILP (table))
 193     return current_buffer->category_table;
 194   CHECK_TYPE (!NILP (Fcategory_table_p (table)), Qcategory_table_p, table);
 195   return table;
 196 }
 197
 198 DEFUN ("category-table", Fcategory_table, Scategory_table, 0, 0, 0,
 199        doc: /* Return the current category table.
 200 This is the one specified by the current buffer.  */)
 201   (void)
 202 {
 203   return current_buffer->category_table;
 204 }
 205
 206 DEFUN ("standard-category-table", Fstandard_category_table,
 207    Sstandard_category_table, 0, 0, 0,
 208        doc: /* Return the standard category table.
 209 This is the one used for new buffers.  */)
 210   (void)
 211 {
 212   return Vstandard_category_table;
 213 }
 214
 215
 216 static void
 217 copy_category_entry (Lisp_Object table, Lisp_Object c, Lisp_Object val)
 218 {
 219   val = Fcopy_sequence (val);
 220   if (CONSP (c))
 221     char_table_set_range (table, XINT (XCAR (c)), XINT (XCDR (c)), val);
 222   else
 223     char_table_set (table, XINT (c), val);
 224 }
 225
 226 /* Return a copy of category table TABLE.  We can't simply use the
 227    function copy-sequence because no contents should be shared between
 228    the original and the copy.  This function is called recursively by
 229    binding TABLE to a sub char table.  */
 230
 231 Lisp_Object
 232 copy_category_table (Lisp_Object table)
 233 {
 234   table = copy_char_table (table);
 235
 236   if (! NILP (XCHAR_TABLE (table)->defalt))
 237     XCHAR_TABLE (table)->defalt
 238       = Fcopy_sequence (XCHAR_TABLE (table)->defalt);
 239   XCHAR_TABLE (table)->extras[0]
 240     = Fcopy_sequence (XCHAR_TABLE (table)->extras[0]);
 241   map_char_table (copy_category_entry, Qnil, table, table);
 242
 243   return table;
 244 }
 245
 246 DEFUN ("copy-category-table", Fcopy_category_table, Scopy_category_table,
 247        0, 1, 0,
 248        doc: /* Construct a new category table and return it.
 249 It is a copy of the TABLE, which defaults to the standard category table.  */)
 250   (Lisp_Object table)
 251 {
 252   if (!NILP (table))
 253     check_category_table (table);
 254   else
 255     table = Vstandard_category_table;
 256
 257   return copy_category_table (table);
 258 }
 259
 260 DEFUN ("make-category-table", Fmake_category_table, Smake_category_table,
 261        0, 0, 0,
 262        doc: /* Construct a new and empty category table and return it.  */)
 263   (void)
 264 {
 265   Lisp_Object val;
 266   int i;
 267
 268   val = Fmake_char_table (Qcategory_table, Qnil);
 269   XCHAR_TABLE (val)->defalt = MAKE_CATEGORY_SET;
 270   for (i = 0; i < (1 << CHARTAB_SIZE_BITS_0); i++)
 271     XCHAR_TABLE (val)->contents[i] = MAKE_CATEGORY_SET;
 272   Fset_char_table_extra_slot (val, make_number (0),
 273                               Fmake_vector (make_number (95), Qnil));
 274   return val;
 275 }
 276
 277 DEFUN ("set-category-table", Fset_category_table, Sset_category_table, 1, 1, 0,
 278        doc: /* Specify TABLE as the category table for the current buffer.
 279 Return TABLE.  */)
 280   (Lisp_Object table)
 281 {
 282   int idx;
 283   table = check_category_table (table);
 284   current_buffer->category_table = table;
 285   /* Indicate that this buffer now has a specified category table.  */
 286   idx = PER_BUFFER_VAR_IDX (category_table);
 287   SET_PER_BUFFER_VALUE_P (current_buffer, idx, 1);
 288   return table;
 289 }
 290
 291 \f
 292 Lisp_Object
 293 char_category_set (int c)
 294 {
 295   return CHAR_TABLE_REF (current_buffer->category_table, c);
 296 }
 297
 298 DEFUN ("char-category-set", Fchar_category_set, Schar_category_set, 1, 1, 0,
 299        doc: /* Return the category set of CHAR.
 300 usage: (char-category-set CHAR)  */)
 301   (Lisp_Object ch)
 302 {
 303   CHECK_NUMBER (ch);
 304   return CATEGORY_SET (XFASTINT (ch));
 305 }
 306
 307 DEFUN ("category-set-mnemonics", Fcategory_set_mnemonics,
 308        Scategory_set_mnemonics, 1, 1, 0,
 309        doc: /* Return a string containing mnemonics of the categories in CATEGORY-SET.
 310 CATEGORY-SET is a bool-vector, and the categories \"in\" it are those
 311 that are indexes where t occurs in the bool-vector.
 312 The return value is a string containing those same categories.  */)
 313   (Lisp_Object category_set)
 314 {
 315   int i, j;
 316   char str[96];
 317
 318   CHECK_CATEGORY_SET (category_set);
 319
 320   j = 0;
 321   for (i = 32; i < 127; i++)
 322     if (CATEGORY_MEMBER (i, category_set))
 323       str[j++] = i;
 324   str[j] = '\0';
 325
 326   return build_string (str);
 327 }
 328
 329 void
 330 set_category_set (Lisp_Object category_set, Lisp_Object category, Lisp_Object val)
 331 {
 332   do {
 333     int idx = XINT (category) / 8;
 334     unsigned char bits = 1 << (XINT (category) % 8);
 335
 336     if (NILP (val))
 337       XCATEGORY_SET (category_set)->data[idx] &= ~bits;
 338     else
 339       XCATEGORY_SET (category_set)->data[idx] |= bits;
 340   } while (0);
 341 }
 342
 343 DEFUN ("modify-category-entry", Fmodify_category_entry,
 344        Smodify_category_entry, 2, 4, 0,
 345        doc: /* Modify the category set of CHARACTER by adding CATEGORY to it.
 346 The category is changed only for table TABLE, which defaults to
 347 the current buffer's category table.
 348 CHARACTER can be either a single character or a cons representing the
 349 lower and upper ends of an inclusive character range to modify.
 350 If optional fourth argument RESET is non-nil,
 351 then delete CATEGORY from the category set instead of adding it.  */)
 352   (Lisp_Object character, Lisp_Object category, Lisp_Object table, Lisp_Object reset)
 353 {
 354   Lisp_Object set_value;        /* Actual value to be set in category sets.  */
 355   Lisp_Object category_set;
 356   int start, end;
 357   int from, to;
 358
 359   if (INTEGERP (character))
 360     {
 361       CHECK_CHARACTER (character);
 362       start = end = XFASTINT (character);
 363     }
 364   else
 365     {
 366       CHECK_CONS (character);
 367       CHECK_CHARACTER_CAR (character);
 368       CHECK_CHARACTER_CDR (character);
 369       start = XFASTINT (XCAR (character));
 370       end = XFASTINT (XCDR (character));
 371     }
 372
 373   CHECK_CATEGORY (category);
 374   table = check_category_table (table);
 375
 376   if (NILP (CATEGORY_DOCSTRING (table, XFASTINT (category))))
 377     error ("Undefined category: %c", XFASTINT (category));
 378
 379   set_value = NILP (reset) ? Qt : Qnil;
 380
 381   while (start <= end)
 382     {
 383       from = start, to = end;
 384       category_set = char_table_ref_and_range (table, start, &from, &to);
 385       if (CATEGORY_MEMBER (XFASTINT (category), category_set) != NILP (reset))
 386         {
 387           category_set = Fcopy_sequence (category_set);
 388           SET_CATEGORY_SET (category_set, category, set_value);
 389           category_set = hash_get_category_set (table, category_set);
 390           char_table_set_range (table, start, to, category_set);
 391         }
 392       start = to + 1;
 393     }
 394
 395   return Qnil;
 396 }
 397 \f
 398 /* Return 1 if there is a word boundary between two word-constituent
 399    characters C1 and C2 if they appear in this order, else return 0.
 400    Use the macro WORD_BOUNDARY_P instead of calling this function
 401    directly.  */
 402
 403 int
 404 word_boundary_p (int c1, int c2)
 405 {
 406   Lisp_Object category_set1, category_set2;
 407   Lisp_Object tail;
 408   int default_result;
 409
 410   if (EQ (CHAR_TABLE_REF (Vchar_script_table, c1),
 411           CHAR_TABLE_REF (Vchar_script_table, c2)))
 412     {
 413       tail = Vword_separating_categories;
 414       default_result = 0;
 415     }
 416   else
 417     {
 418       tail = Vword_combining_categories;
 419       default_result = 1;
 420     }
 421
 422   category_set1 = CATEGORY_SET (c1);
 423   if (NILP (category_set1))
 424     return default_result;
 425   category_set2 = CATEGORY_SET (c2);
 426   if (NILP (category_set2))
 427     return default_result;
 428
 429   for (; CONSP (tail); tail = XCDR (tail))
 430     {
 431       Lisp_Object elt = XCAR (tail);
 432
 433       if (CONSP (elt)
 434           && (NILP (XCAR (elt))
 435               || (CATEGORYP (XCAR (elt))
 436                   && CATEGORY_MEMBER (XFASTINT (XCAR (elt)), category_set1)
 437                   && ! CATEGORY_MEMBER (XFASTINT (XCAR (elt)), category_set2)))
 438           && (NILP (XCDR (elt))
 439               || (CATEGORYP (XCDR (elt))
 440                   && ! CATEGORY_MEMBER (XFASTINT (XCDR (elt)), category_set1)
 441                   && CATEGORY_MEMBER (XFASTINT (XCDR (elt)), category_set2))))
 442         return !default_result;
 443     }
 444   return default_result;
 445 }
 446
 447 \f
 448 void
 449 init_category_once (void)
 450 {
 451   /* This has to be done here, before we call Fmake_char_table.  */
 452   Qcategory_table = intern_c_string ("category-table");
 453   staticpro (&Qcategory_table);
 454
 455   /* Intern this now in case it isn't already done.
 456      Setting this variable twice is harmless.
 457      But don't staticpro it here--that is done in alloc.c.  */
 458   Qchar_table_extra_slots = intern_c_string ("char-table-extra-slots");
 459
 460   /* Now we are ready to set up this property, so we can
 461      create category tables.  */
 462   Fput (Qcategory_table, Qchar_table_extra_slots, make_number (2));
 463
 464   Vstandard_category_table = Fmake_char_table (Qcategory_table, Qnil);
 465   /* Set a category set which contains nothing to the default.  */
 466   XCHAR_TABLE (Vstandard_category_table)->defalt = MAKE_CATEGORY_SET;
 467   Fset_char_table_extra_slot (Vstandard_category_table, make_number (0),
 468                               Fmake_vector (make_number (95), Qnil));
 469 }
 470
 471 void
 472 syms_of_category (void)
 473 {
 474   Qcategoryp = intern_c_string ("categoryp");
 475   staticpro (&Qcategoryp);
 476   Qcategorysetp = intern_c_string ("categorysetp");
 477   staticpro (&Qcategorysetp);
 478   Qcategory_table_p = intern_c_string ("category-table-p");
 479   staticpro (&Qcategory_table_p);
 480
 481   DEFVAR_LISP ("word-combining-categories", Vword_combining_categories,
 482                doc: /* List of pair (cons) of categories to determine word boundary.
 483
 484 Emacs treats a sequence of word constituent characters as a single
 485 word (i.e. finds no word boundary between them) only if they belong to
 486 the same script.  But, exceptions are allowed in the following cases.
 487
 488 \(1) The case that characters are in different scripts is controlled
 489 by the variable `word-combining-categories'.
 490
 491 Emacs finds no word boundary between characters of different scripts
 492 if they have categories matching some element of this list.
 493
 494 More precisely, if an element of this list is a cons of category CAT1
 495 and CAT2, and a multibyte character C1 which has CAT1 is followed by
 496 C2 which has CAT2, there's no word boundary between C1 and C2.
 497
 498 For instance, to tell that Han characters followed by Hiragana
 499 characters can form a single word, the element `(?C . ?H)' should be
 500 in this list.
 501
 502 \(2) The case that character are in the same script is controlled by
 503 the variable `word-separating-categories'.
 504
 505 Emacs finds a word boundary between characters of the same script
 506 if they have categories matching some element of this list.
 507
 508 More precisely, if an element of this list is a cons of category CAT1
 509 and CAT2, and a multibyte character C1 which has CAT1 but not CAT2 is
 510 followed by C2 which has CAT2 but not CAT1, there's a word boundary
 511 between C1 and C2.
 512
 513 For instance, to tell that there's a word boundary between Hiragana
 514 and Katakana (both are in the same script `kana'),
 515 the element `(?H . ?K) should be in this list.  */);
 516
 517   Vword_combining_categories = Qnil;
 518
 519   DEFVAR_LISP ("word-separating-categories", Vword_separating_categories,
 520                doc: /* List of pair (cons) of categories to determine word boundary.
 521 See the documentation of the variable `word-combining-categories'.  */);
 522
 523   Vword_separating_categories = Qnil;
 524
 525   defsubr (&Smake_category_set);
 526   defsubr (&Sdefine_category);
 527   defsubr (&Scategory_docstring);
 528   defsubr (&Sget_unused_category);
 529   defsubr (&Scategory_table_p);
 530   defsubr (&Scategory_table);
 531   defsubr (&Sstandard_category_table);
 532   defsubr (&Scopy_category_table);
 533   defsubr (&Smake_category_table);
 534   defsubr (&Sset_category_table);
 535   defsubr (&Schar_category_set);
 536   defsubr (&Scategory_set_mnemonics);
 537   defsubr (&Smodify_category_entry);
 538
 539   category_table_version = 0;
 540 }
 541