|
428
|
1 /* XEmacs routines to deal with char tables.
|
|
|
2 Copyright (C) 1992, 1995 Free Software Foundation, Inc.
|
|
|
3 Copyright (C) 1995 Sun Microsystems, Inc.
|
|
|
4 Copyright (C) 1995, 1996 Ben Wing.
|
|
|
5 Copyright (C) 1995, 1997, 1999 Electrotechnical Laboratory, JAPAN.
|
|
|
6 Licensed to the Free Software Foundation.
|
|
|
7
|
|
|
8 This file is part of XEmacs.
|
|
|
9
|
|
|
10 XEmacs is free software; you can redistribute it and/or modify it
|
|
|
11 under the terms of the GNU General Public License as published by the
|
|
|
12 Free Software Foundation; either version 2, or (at your option) any
|
|
|
13 later version.
|
|
|
14
|
|
|
15 XEmacs is distributed in the hope that it will be useful, but WITHOUT
|
|
|
16 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
|
17 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
|
18 for more details.
|
|
|
19
|
|
|
20 You should have received a copy of the GNU General Public License
|
|
|
21 along with XEmacs; see the file COPYING. If not, write to
|
|
|
22 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
23 Boston, MA 02111-1307, USA. */
|
|
|
24
|
|
|
25 /* Synched up with: Mule 2.3. Not synched with FSF.
|
|
|
26
|
|
|
27 This file was written independently of the FSF implementation,
|
|
|
28 and is not compatible. */
|
|
|
29
|
|
|
30 /* Authorship:
|
|
|
31
|
|
|
32 Ben Wing: wrote, for 19.13 (Mule). Some category table stuff
|
|
|
33 loosely based on the original Mule.
|
|
|
34 Jareth Hein: fixed a couple of bugs in the implementation, and
|
|
|
35 added regex support for categories with check_category_at
|
|
|
36 */
|
|
|
37
|
|
|
38 #include <config.h>
|
|
|
39 #include "lisp.h"
|
|
|
40
|
|
|
41 #include "buffer.h"
|
|
|
42 #include "chartab.h"
|
|
|
43 #include "syntax.h"
|
|
|
44
|
|
|
45 Lisp_Object Qchar_tablep, Qchar_table;
|
|
|
46
|
|
|
47 Lisp_Object Vall_syntax_tables;
|
|
|
48
|
|
|
49 #ifdef MULE
|
|
|
50 Lisp_Object Qcategory_table_p;
|
|
|
51 Lisp_Object Qcategory_designator_p;
|
|
|
52 Lisp_Object Qcategory_table_value_p;
|
|
|
53
|
|
|
54 Lisp_Object Vstandard_category_table;
|
|
|
55
|
|
|
56 /* Variables to determine word boundary. */
|
|
|
57 Lisp_Object Vword_combining_categories, Vword_separating_categories;
|
|
|
58 #endif /* MULE */
|
|
|
59
|
|
|
60
|
|
|
61 /* A char table maps from ranges of characters to values.
|
|
|
62
|
|
|
63 Implementing a general data structure that maps from arbitrary
|
|
|
64 ranges of numbers to values is tricky to do efficiently. As it
|
|
|
65 happens, it should suffice (and is usually more convenient, anyway)
|
|
|
66 when dealing with characters to restrict the sorts of ranges that
|
|
|
67 can be assigned values, as follows:
|
|
|
68
|
|
|
69 1) All characters.
|
|
|
70 2) All characters in a charset.
|
|
|
71 3) All characters in a particular row of a charset, where a "row"
|
|
|
72 means all characters with the same first byte.
|
|
|
73 4) A particular character in a charset.
|
|
|
74
|
|
|
75 We use char tables to generalize the 256-element vectors now
|
|
|
76 littering the Emacs code.
|
|
|
77
|
|
|
78 Possible uses (all should be converted at some point):
|
|
|
79
|
|
|
80 1) category tables
|
|
|
81 2) syntax tables
|
|
|
82 3) display tables
|
|
|
83 4) case tables
|
|
|
84 5) keyboard-translate-table?
|
|
|
85
|
|
|
86 We provide an
|
|
|
87 abstract type to generalize the Emacs vectors and Mule
|
|
|
88 vectors-of-vectors goo.
|
|
|
89 */
|
|
|
90
|
|
|
91 /************************************************************************/
|
|
|
92 /* Char Table object */
|
|
|
93 /************************************************************************/
|
|
|
94
|
|
|
95 #ifdef MULE
|
|
|
96
|
|
|
97 static Lisp_Object
|
|
|
98 mark_char_table_entry (Lisp_Object obj)
|
|
|
99 {
|
|
440
|
100 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (obj);
|
|
428
|
101 int i;
|
|
|
102
|
|
|
103 for (i = 0; i < 96; i++)
|
|
|
104 {
|
|
|
105 mark_object (cte->level2[i]);
|
|
|
106 }
|
|
|
107 return Qnil;
|
|
|
108 }
|
|
|
109
|
|
|
110 static int
|
|
|
111 char_table_entry_equal (Lisp_Object obj1, Lisp_Object obj2, int depth)
|
|
|
112 {
|
|
440
|
113 Lisp_Char_Table_Entry *cte1 = XCHAR_TABLE_ENTRY (obj1);
|
|
|
114 Lisp_Char_Table_Entry *cte2 = XCHAR_TABLE_ENTRY (obj2);
|
|
428
|
115 int i;
|
|
|
116
|
|
|
117 for (i = 0; i < 96; i++)
|
|
|
118 if (!internal_equal (cte1->level2[i], cte2->level2[i], depth + 1))
|
|
|
119 return 0;
|
|
|
120
|
|
|
121 return 1;
|
|
|
122 }
|
|
|
123
|
|
665
|
124 static Hashcode
|
|
428
|
125 char_table_entry_hash (Lisp_Object obj, int depth)
|
|
|
126 {
|
|
440
|
127 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (obj);
|
|
428
|
128
|
|
|
129 return internal_array_hash (cte->level2, 96, depth);
|
|
|
130 }
|
|
|
131
|
|
|
132 static const struct lrecord_description char_table_entry_description[] = {
|
|
440
|
133 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Char_Table_Entry, level2), 96 },
|
|
428
|
134 { XD_END }
|
|
|
135 };
|
|
|
136
|
|
|
137 DEFINE_LRECORD_IMPLEMENTATION ("char-table-entry", char_table_entry,
|
|
|
138 mark_char_table_entry, internal_object_printer,
|
|
|
139 0, char_table_entry_equal,
|
|
|
140 char_table_entry_hash,
|
|
|
141 char_table_entry_description,
|
|
440
|
142 Lisp_Char_Table_Entry);
|
|
428
|
143 #endif /* MULE */
|
|
|
144
|
|
|
145 static Lisp_Object
|
|
|
146 mark_char_table (Lisp_Object obj)
|
|
|
147 {
|
|
440
|
148 Lisp_Char_Table *ct = XCHAR_TABLE (obj);
|
|
428
|
149 int i;
|
|
|
150
|
|
|
151 for (i = 0; i < NUM_ASCII_CHARS; i++)
|
|
|
152 mark_object (ct->ascii[i]);
|
|
|
153 #ifdef MULE
|
|
|
154 for (i = 0; i < NUM_LEADING_BYTES; i++)
|
|
|
155 mark_object (ct->level1[i]);
|
|
|
156 #endif
|
|
|
157 return ct->mirror_table;
|
|
|
158 }
|
|
|
159
|
|
|
160 /* WARNING: All functions of this nature need to be written extremely
|
|
|
161 carefully to avoid crashes during GC. Cf. prune_specifiers()
|
|
|
162 and prune_weak_hash_tables(). */
|
|
|
163
|
|
|
164 void
|
|
|
165 prune_syntax_tables (void)
|
|
|
166 {
|
|
|
167 Lisp_Object rest, prev = Qnil;
|
|
|
168
|
|
|
169 for (rest = Vall_syntax_tables;
|
|
|
170 !NILP (rest);
|
|
|
171 rest = XCHAR_TABLE (rest)->next_table)
|
|
|
172 {
|
|
|
173 if (! marked_p (rest))
|
|
|
174 {
|
|
|
175 /* This table is garbage. Remove it from the list. */
|
|
|
176 if (NILP (prev))
|
|
|
177 Vall_syntax_tables = XCHAR_TABLE (rest)->next_table;
|
|
|
178 else
|
|
|
179 XCHAR_TABLE (prev)->next_table =
|
|
|
180 XCHAR_TABLE (rest)->next_table;
|
|
|
181 }
|
|
|
182 }
|
|
|
183 }
|
|
|
184
|
|
|
185 static Lisp_Object
|
|
|
186 char_table_type_to_symbol (enum char_table_type type)
|
|
|
187 {
|
|
|
188 switch (type)
|
|
|
189 {
|
|
|
190 default: abort();
|
|
|
191 case CHAR_TABLE_TYPE_GENERIC: return Qgeneric;
|
|
|
192 case CHAR_TABLE_TYPE_SYNTAX: return Qsyntax;
|
|
|
193 case CHAR_TABLE_TYPE_DISPLAY: return Qdisplay;
|
|
|
194 case CHAR_TABLE_TYPE_CHAR: return Qchar;
|
|
|
195 #ifdef MULE
|
|
|
196 case CHAR_TABLE_TYPE_CATEGORY: return Qcategory;
|
|
|
197 #endif
|
|
|
198 }
|
|
|
199 }
|
|
|
200
|
|
|
201 static enum char_table_type
|
|
|
202 symbol_to_char_table_type (Lisp_Object symbol)
|
|
|
203 {
|
|
|
204 CHECK_SYMBOL (symbol);
|
|
|
205
|
|
|
206 if (EQ (symbol, Qgeneric)) return CHAR_TABLE_TYPE_GENERIC;
|
|
|
207 if (EQ (symbol, Qsyntax)) return CHAR_TABLE_TYPE_SYNTAX;
|
|
|
208 if (EQ (symbol, Qdisplay)) return CHAR_TABLE_TYPE_DISPLAY;
|
|
|
209 if (EQ (symbol, Qchar)) return CHAR_TABLE_TYPE_CHAR;
|
|
|
210 #ifdef MULE
|
|
|
211 if (EQ (symbol, Qcategory)) return CHAR_TABLE_TYPE_CATEGORY;
|
|
|
212 #endif
|
|
|
213
|
|
563
|
214 invalid_constant ("Unrecognized char table type", symbol);
|
|
428
|
215 return CHAR_TABLE_TYPE_GENERIC; /* not reached */
|
|
|
216 }
|
|
|
217
|
|
|
218 static void
|
|
|
219 print_chartab_range (Emchar first, Emchar last, Lisp_Object val,
|
|
|
220 Lisp_Object printcharfun)
|
|
|
221 {
|
|
|
222 if (first != last)
|
|
|
223 {
|
|
|
224 write_c_string (" (", printcharfun);
|
|
|
225 print_internal (make_char (first), printcharfun, 0);
|
|
|
226 write_c_string (" ", printcharfun);
|
|
|
227 print_internal (make_char (last), printcharfun, 0);
|
|
|
228 write_c_string (") ", printcharfun);
|
|
|
229 }
|
|
|
230 else
|
|
|
231 {
|
|
|
232 write_c_string (" ", printcharfun);
|
|
|
233 print_internal (make_char (first), printcharfun, 0);
|
|
|
234 write_c_string (" ", printcharfun);
|
|
|
235 }
|
|
|
236 print_internal (val, printcharfun, 1);
|
|
|
237 }
|
|
|
238
|
|
|
239 #ifdef MULE
|
|
|
240
|
|
|
241 static void
|
|
|
242 print_chartab_charset_row (Lisp_Object charset,
|
|
|
243 int row,
|
|
440
|
244 Lisp_Char_Table_Entry *cte,
|
|
428
|
245 Lisp_Object printcharfun)
|
|
|
246 {
|
|
|
247 int i;
|
|
|
248 Lisp_Object cat = Qunbound;
|
|
|
249 int first = -1;
|
|
|
250
|
|
|
251 for (i = 32; i < 128; i++)
|
|
|
252 {
|
|
|
253 Lisp_Object pam = cte->level2[i - 32];
|
|
|
254
|
|
|
255 if (first == -1)
|
|
|
256 {
|
|
|
257 first = i;
|
|
|
258 cat = pam;
|
|
|
259 continue;
|
|
|
260 }
|
|
|
261
|
|
|
262 if (!EQ (cat, pam))
|
|
|
263 {
|
|
|
264 if (row == -1)
|
|
|
265 print_chartab_range (MAKE_CHAR (charset, first, 0),
|
|
|
266 MAKE_CHAR (charset, i - 1, 0),
|
|
|
267 cat, printcharfun);
|
|
|
268 else
|
|
|
269 print_chartab_range (MAKE_CHAR (charset, row, first),
|
|
|
270 MAKE_CHAR (charset, row, i - 1),
|
|
|
271 cat, printcharfun);
|
|
|
272 first = -1;
|
|
|
273 i--;
|
|
|
274 }
|
|
|
275 }
|
|
|
276
|
|
|
277 if (first != -1)
|
|
|
278 {
|
|
|
279 if (row == -1)
|
|
|
280 print_chartab_range (MAKE_CHAR (charset, first, 0),
|
|
|
281 MAKE_CHAR (charset, i - 1, 0),
|
|
|
282 cat, printcharfun);
|
|
|
283 else
|
|
|
284 print_chartab_range (MAKE_CHAR (charset, row, first),
|
|
|
285 MAKE_CHAR (charset, row, i - 1),
|
|
|
286 cat, printcharfun);
|
|
|
287 }
|
|
|
288 }
|
|
|
289
|
|
|
290 static void
|
|
|
291 print_chartab_two_byte_charset (Lisp_Object charset,
|
|
440
|
292 Lisp_Char_Table_Entry *cte,
|
|
428
|
293 Lisp_Object printcharfun)
|
|
|
294 {
|
|
|
295 int i;
|
|
|
296
|
|
|
297 for (i = 32; i < 128; i++)
|
|
|
298 {
|
|
|
299 Lisp_Object jen = cte->level2[i - 32];
|
|
|
300
|
|
|
301 if (!CHAR_TABLE_ENTRYP (jen))
|
|
|
302 {
|
|
|
303 char buf[100];
|
|
|
304
|
|
|
305 write_c_string (" [", printcharfun);
|
|
|
306 print_internal (XCHARSET_NAME (charset), printcharfun, 0);
|
|
|
307 sprintf (buf, " %d] ", i);
|
|
|
308 write_c_string (buf, printcharfun);
|
|
|
309 print_internal (jen, printcharfun, 0);
|
|
|
310 }
|
|
|
311 else
|
|
|
312 print_chartab_charset_row (charset, i, XCHAR_TABLE_ENTRY (jen),
|
|
|
313 printcharfun);
|
|
|
314 }
|
|
|
315 }
|
|
|
316
|
|
|
317 #endif /* MULE */
|
|
|
318
|
|
|
319 static void
|
|
|
320 print_char_table (Lisp_Object obj, Lisp_Object printcharfun, int escapeflag)
|
|
|
321 {
|
|
440
|
322 Lisp_Char_Table *ct = XCHAR_TABLE (obj);
|
|
428
|
323 char buf[200];
|
|
|
324
|
|
|
325 sprintf (buf, "#s(char-table type %s data (",
|
|
|
326 string_data (symbol_name (XSYMBOL
|
|
|
327 (char_table_type_to_symbol (ct->type)))));
|
|
|
328 write_c_string (buf, printcharfun);
|
|
|
329
|
|
|
330 /* Now write out the ASCII/Control-1 stuff. */
|
|
|
331 {
|
|
|
332 int i;
|
|
|
333 int first = -1;
|
|
|
334 Lisp_Object val = Qunbound;
|
|
|
335
|
|
|
336 for (i = 0; i < NUM_ASCII_CHARS; i++)
|
|
|
337 {
|
|
|
338 if (first == -1)
|
|
|
339 {
|
|
|
340 first = i;
|
|
|
341 val = ct->ascii[i];
|
|
|
342 continue;
|
|
|
343 }
|
|
|
344
|
|
|
345 if (!EQ (ct->ascii[i], val))
|
|
|
346 {
|
|
|
347 print_chartab_range (first, i - 1, val, printcharfun);
|
|
|
348 first = -1;
|
|
|
349 i--;
|
|
|
350 }
|
|
|
351 }
|
|
|
352
|
|
|
353 if (first != -1)
|
|
|
354 print_chartab_range (first, i - 1, val, printcharfun);
|
|
|
355 }
|
|
|
356
|
|
|
357 #ifdef MULE
|
|
|
358 {
|
|
|
359 int i;
|
|
|
360
|
|
|
361 for (i = MIN_LEADING_BYTE; i < MIN_LEADING_BYTE + NUM_LEADING_BYTES;
|
|
|
362 i++)
|
|
|
363 {
|
|
|
364 Lisp_Object ann = ct->level1[i - MIN_LEADING_BYTE];
|
|
|
365 Lisp_Object charset = CHARSET_BY_LEADING_BYTE (i);
|
|
|
366
|
|
|
367 if (!CHARSETP (charset) || i == LEADING_BYTE_ASCII
|
|
|
368 || i == LEADING_BYTE_CONTROL_1)
|
|
|
369 continue;
|
|
|
370 if (!CHAR_TABLE_ENTRYP (ann))
|
|
|
371 {
|
|
|
372 write_c_string (" ", printcharfun);
|
|
|
373 print_internal (XCHARSET_NAME (charset),
|
|
|
374 printcharfun, 0);
|
|
|
375 write_c_string (" ", printcharfun);
|
|
|
376 print_internal (ann, printcharfun, 0);
|
|
|
377 }
|
|
|
378 else
|
|
|
379 {
|
|
440
|
380 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (ann);
|
|
428
|
381 if (XCHARSET_DIMENSION (charset) == 1)
|
|
|
382 print_chartab_charset_row (charset, -1, cte, printcharfun);
|
|
|
383 else
|
|
|
384 print_chartab_two_byte_charset (charset, cte, printcharfun);
|
|
|
385 }
|
|
|
386 }
|
|
|
387 }
|
|
|
388 #endif /* MULE */
|
|
|
389
|
|
|
390 write_c_string ("))", printcharfun);
|
|
|
391 }
|
|
|
392
|
|
|
393 static int
|
|
|
394 char_table_equal (Lisp_Object obj1, Lisp_Object obj2, int depth)
|
|
|
395 {
|
|
440
|
396 Lisp_Char_Table *ct1 = XCHAR_TABLE (obj1);
|
|
|
397 Lisp_Char_Table *ct2 = XCHAR_TABLE (obj2);
|
|
428
|
398 int i;
|
|
|
399
|
|
|
400 if (CHAR_TABLE_TYPE (ct1) != CHAR_TABLE_TYPE (ct2))
|
|
|
401 return 0;
|
|
|
402
|
|
|
403 for (i = 0; i < NUM_ASCII_CHARS; i++)
|
|
|
404 if (!internal_equal (ct1->ascii[i], ct2->ascii[i], depth + 1))
|
|
|
405 return 0;
|
|
|
406
|
|
|
407 #ifdef MULE
|
|
|
408 for (i = 0; i < NUM_LEADING_BYTES; i++)
|
|
|
409 if (!internal_equal (ct1->level1[i], ct2->level1[i], depth + 1))
|
|
|
410 return 0;
|
|
|
411 #endif /* MULE */
|
|
|
412
|
|
|
413 return 1;
|
|
|
414 }
|
|
|
415
|
|
665
|
416 static Hashcode
|
|
428
|
417 char_table_hash (Lisp_Object obj, int depth)
|
|
|
418 {
|
|
440
|
419 Lisp_Char_Table *ct = XCHAR_TABLE (obj);
|
|
665
|
420 Hashcode hashval = internal_array_hash (ct->ascii, NUM_ASCII_CHARS,
|
|
647
|
421 depth);
|
|
428
|
422 #ifdef MULE
|
|
|
423 hashval = HASH2 (hashval,
|
|
|
424 internal_array_hash (ct->level1, NUM_LEADING_BYTES, depth));
|
|
|
425 #endif /* MULE */
|
|
|
426 return hashval;
|
|
|
427 }
|
|
|
428
|
|
|
429 static const struct lrecord_description char_table_description[] = {
|
|
440
|
430 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Char_Table, ascii), NUM_ASCII_CHARS },
|
|
428
|
431 #ifdef MULE
|
|
440
|
432 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Char_Table, level1), NUM_LEADING_BYTES },
|
|
428
|
433 #endif
|
|
440
|
434 { XD_LISP_OBJECT, offsetof (Lisp_Char_Table, mirror_table) },
|
|
|
435 { XD_LO_LINK, offsetof (Lisp_Char_Table, next_table) },
|
|
428
|
436 { XD_END }
|
|
|
437 };
|
|
|
438
|
|
|
439 DEFINE_LRECORD_IMPLEMENTATION ("char-table", char_table,
|
|
|
440 mark_char_table, print_char_table, 0,
|
|
|
441 char_table_equal, char_table_hash,
|
|
|
442 char_table_description,
|
|
440
|
443 Lisp_Char_Table);
|
|
428
|
444
|
|
|
445 DEFUN ("char-table-p", Fchar_table_p, 1, 1, 0, /*
|
|
|
446 Return non-nil if OBJECT is a char table.
|
|
|
447
|
|
|
448 A char table is a table that maps characters (or ranges of characters)
|
|
|
449 to values. Char tables are specialized for characters, only allowing
|
|
|
450 particular sorts of ranges to be assigned values. Although this
|
|
|
451 loses in generality, it makes for extremely fast (constant-time)
|
|
|
452 lookups, and thus is feasible for applications that do an extremely
|
|
|
453 large number of lookups (e.g. scanning a buffer for a character in
|
|
|
454 a particular syntax, where a lookup in the syntax table must occur
|
|
|
455 once per character).
|
|
|
456
|
|
|
457 When Mule support exists, the types of ranges that can be assigned
|
|
|
458 values are
|
|
|
459
|
|
|
460 -- all characters
|
|
|
461 -- an entire charset
|
|
|
462 -- a single row in a two-octet charset
|
|
|
463 -- a single character
|
|
|
464
|
|
|
465 When Mule support is not present, the types of ranges that can be
|
|
|
466 assigned values are
|
|
|
467
|
|
|
468 -- all characters
|
|
|
469 -- a single character
|
|
|
470
|
|
444
|
471 To create a char table, use `make-char-table'.
|
|
|
472 To modify a char table, use `put-char-table' or `remove-char-table'.
|
|
|
473 To retrieve the value for a particular character, use `get-char-table'.
|
|
|
474 See also `map-char-table', `clear-char-table', `copy-char-table',
|
|
|
475 `valid-char-table-type-p', `char-table-type-list',
|
|
|
476 `valid-char-table-value-p', and `check-char-table-value'.
|
|
428
|
477 */
|
|
|
478 (object))
|
|
|
479 {
|
|
|
480 return CHAR_TABLEP (object) ? Qt : Qnil;
|
|
|
481 }
|
|
|
482
|
|
|
483 DEFUN ("char-table-type-list", Fchar_table_type_list, 0, 0, 0, /*
|
|
|
484 Return a list of the recognized char table types.
|
|
|
485 See `valid-char-table-type-p'.
|
|
|
486 */
|
|
|
487 ())
|
|
|
488 {
|
|
|
489 #ifdef MULE
|
|
|
490 return list5 (Qchar, Qcategory, Qdisplay, Qgeneric, Qsyntax);
|
|
|
491 #else
|
|
|
492 return list4 (Qchar, Qdisplay, Qgeneric, Qsyntax);
|
|
|
493 #endif
|
|
|
494 }
|
|
|
495
|
|
|
496 DEFUN ("valid-char-table-type-p", Fvalid_char_table_type_p, 1, 1, 0, /*
|
|
|
497 Return t if TYPE if a recognized char table type.
|
|
|
498
|
|
|
499 Each char table type is used for a different purpose and allows different
|
|
|
500 sorts of values. The different char table types are
|
|
|
501
|
|
|
502 `category'
|
|
|
503 Used for category tables, which specify the regexp categories
|
|
|
504 that a character is in. The valid values are nil or a
|
|
|
505 bit vector of 95 elements. Higher-level Lisp functions are
|
|
|
506 provided for working with category tables. Currently categories
|
|
|
507 and category tables only exist when Mule support is present.
|
|
|
508 `char'
|
|
|
509 A generalized char table, for mapping from one character to
|
|
|
510 another. Used for case tables, syntax matching tables,
|
|
|
511 `keyboard-translate-table', etc. The valid values are characters.
|
|
|
512 `generic'
|
|
|
513 An even more generalized char table, for mapping from a
|
|
|
514 character to anything.
|
|
|
515 `display'
|
|
|
516 Used for display tables, which specify how a particular character
|
|
|
517 is to appear when displayed. #### Not yet implemented.
|
|
|
518 `syntax'
|
|
|
519 Used for syntax tables, which specify the syntax of a particular
|
|
|
520 character. Higher-level Lisp functions are provided for
|
|
|
521 working with syntax tables. The valid values are integers.
|
|
|
522
|
|
|
523 */
|
|
|
524 (type))
|
|
|
525 {
|
|
|
526 return (EQ (type, Qchar) ||
|
|
|
527 #ifdef MULE
|
|
|
528 EQ (type, Qcategory) ||
|
|
|
529 #endif
|
|
|
530 EQ (type, Qdisplay) ||
|
|
|
531 EQ (type, Qgeneric) ||
|
|
|
532 EQ (type, Qsyntax)) ? Qt : Qnil;
|
|
|
533 }
|
|
|
534
|
|
|
535 DEFUN ("char-table-type", Fchar_table_type, 1, 1, 0, /*
|
|
444
|
536 Return the type of CHAR-TABLE.
|
|
428
|
537 See `valid-char-table-type-p'.
|
|
|
538 */
|
|
444
|
539 (char_table))
|
|
428
|
540 {
|
|
444
|
541 CHECK_CHAR_TABLE (char_table);
|
|
|
542 return char_table_type_to_symbol (XCHAR_TABLE (char_table)->type);
|
|
428
|
543 }
|
|
|
544
|
|
|
545 void
|
|
440
|
546 fill_char_table (Lisp_Char_Table *ct, Lisp_Object value)
|
|
428
|
547 {
|
|
|
548 int i;
|
|
|
549
|
|
|
550 for (i = 0; i < NUM_ASCII_CHARS; i++)
|
|
|
551 ct->ascii[i] = value;
|
|
|
552 #ifdef MULE
|
|
|
553 for (i = 0; i < NUM_LEADING_BYTES; i++)
|
|
|
554 ct->level1[i] = value;
|
|
|
555 #endif /* MULE */
|
|
|
556
|
|
|
557 if (ct->type == CHAR_TABLE_TYPE_SYNTAX)
|
|
|
558 update_syntax_table (ct);
|
|
|
559 }
|
|
|
560
|
|
|
561 DEFUN ("reset-char-table", Freset_char_table, 1, 1, 0, /*
|
|
444
|
562 Reset CHAR-TABLE to its default state.
|
|
428
|
563 */
|
|
444
|
564 (char_table))
|
|
428
|
565 {
|
|
440
|
566 Lisp_Char_Table *ct;
|
|
428
|
567
|
|
444
|
568 CHECK_CHAR_TABLE (char_table);
|
|
|
569 ct = XCHAR_TABLE (char_table);
|
|
428
|
570
|
|
|
571 switch (ct->type)
|
|
|
572 {
|
|
|
573 case CHAR_TABLE_TYPE_CHAR:
|
|
|
574 fill_char_table (ct, make_char (0));
|
|
|
575 break;
|
|
|
576 case CHAR_TABLE_TYPE_DISPLAY:
|
|
|
577 case CHAR_TABLE_TYPE_GENERIC:
|
|
|
578 #ifdef MULE
|
|
|
579 case CHAR_TABLE_TYPE_CATEGORY:
|
|
|
580 #endif /* MULE */
|
|
|
581 fill_char_table (ct, Qnil);
|
|
|
582 break;
|
|
|
583
|
|
|
584 case CHAR_TABLE_TYPE_SYNTAX:
|
|
|
585 fill_char_table (ct, make_int (Sinherit));
|
|
|
586 break;
|
|
|
587
|
|
|
588 default:
|
|
|
589 abort ();
|
|
|
590 }
|
|
|
591
|
|
|
592 return Qnil;
|
|
|
593 }
|
|
|
594
|
|
|
595 DEFUN ("make-char-table", Fmake_char_table, 1, 1, 0, /*
|
|
|
596 Return a new, empty char table of type TYPE.
|
|
|
597 Currently recognized types are 'char, 'category, 'display, 'generic,
|
|
|
598 and 'syntax. See `valid-char-table-type-p'.
|
|
|
599 */
|
|
|
600 (type))
|
|
|
601 {
|
|
440
|
602 Lisp_Char_Table *ct;
|
|
428
|
603 Lisp_Object obj;
|
|
|
604 enum char_table_type ty = symbol_to_char_table_type (type);
|
|
|
605
|
|
440
|
606 ct = alloc_lcrecord_type (Lisp_Char_Table, &lrecord_char_table);
|
|
428
|
607 ct->type = ty;
|
|
|
608 if (ty == CHAR_TABLE_TYPE_SYNTAX)
|
|
|
609 {
|
|
|
610 ct->mirror_table = Fmake_char_table (Qgeneric);
|
|
|
611 fill_char_table (XCHAR_TABLE (ct->mirror_table),
|
|
|
612 make_int (Spunct));
|
|
|
613 }
|
|
|
614 else
|
|
|
615 ct->mirror_table = Qnil;
|
|
|
616 ct->next_table = Qnil;
|
|
|
617 XSETCHAR_TABLE (obj, ct);
|
|
|
618 if (ty == CHAR_TABLE_TYPE_SYNTAX)
|
|
|
619 {
|
|
|
620 ct->next_table = Vall_syntax_tables;
|
|
|
621 Vall_syntax_tables = obj;
|
|
|
622 }
|
|
|
623 Freset_char_table (obj);
|
|
|
624 return obj;
|
|
|
625 }
|
|
|
626
|
|
|
627 #ifdef MULE
|
|
|
628
|
|
|
629 static Lisp_Object
|
|
|
630 make_char_table_entry (Lisp_Object initval)
|
|
|
631 {
|
|
|
632 Lisp_Object obj;
|
|
|
633 int i;
|
|
440
|
634 Lisp_Char_Table_Entry *cte =
|
|
|
635 alloc_lcrecord_type (Lisp_Char_Table_Entry, &lrecord_char_table_entry);
|
|
428
|
636
|
|
|
637 for (i = 0; i < 96; i++)
|
|
|
638 cte->level2[i] = initval;
|
|
|
639
|
|
|
640 XSETCHAR_TABLE_ENTRY (obj, cte);
|
|
|
641 return obj;
|
|
|
642 }
|
|
|
643
|
|
|
644 static Lisp_Object
|
|
|
645 copy_char_table_entry (Lisp_Object entry)
|
|
|
646 {
|
|
440
|
647 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (entry);
|
|
428
|
648 Lisp_Object obj;
|
|
|
649 int i;
|
|
440
|
650 Lisp_Char_Table_Entry *ctenew =
|
|
|
651 alloc_lcrecord_type (Lisp_Char_Table_Entry, &lrecord_char_table_entry);
|
|
428
|
652
|
|
|
653 for (i = 0; i < 96; i++)
|
|
|
654 {
|
|
|
655 Lisp_Object new = cte->level2[i];
|
|
|
656 if (CHAR_TABLE_ENTRYP (new))
|
|
|
657 ctenew->level2[i] = copy_char_table_entry (new);
|
|
|
658 else
|
|
|
659 ctenew->level2[i] = new;
|
|
|
660 }
|
|
|
661
|
|
|
662 XSETCHAR_TABLE_ENTRY (obj, ctenew);
|
|
|
663 return obj;
|
|
|
664 }
|
|
|
665
|
|
|
666 #endif /* MULE */
|
|
|
667
|
|
|
668 DEFUN ("copy-char-table", Fcopy_char_table, 1, 1, 0, /*
|
|
444
|
669 Return a new char table which is a copy of CHAR-TABLE.
|
|
428
|
670 It will contain the same values for the same characters and ranges
|
|
444
|
671 as CHAR-TABLE. The values will not themselves be copied.
|
|
428
|
672 */
|
|
444
|
673 (char_table))
|
|
428
|
674 {
|
|
440
|
675 Lisp_Char_Table *ct, *ctnew;
|
|
428
|
676 Lisp_Object obj;
|
|
|
677 int i;
|
|
|
678
|
|
444
|
679 CHECK_CHAR_TABLE (char_table);
|
|
|
680 ct = XCHAR_TABLE (char_table);
|
|
440
|
681 ctnew = alloc_lcrecord_type (Lisp_Char_Table, &lrecord_char_table);
|
|
428
|
682 ctnew->type = ct->type;
|
|
|
683
|
|
|
684 for (i = 0; i < NUM_ASCII_CHARS; i++)
|
|
|
685 {
|
|
|
686 Lisp_Object new = ct->ascii[i];
|
|
|
687 #ifdef MULE
|
|
|
688 assert (! (CHAR_TABLE_ENTRYP (new)));
|
|
|
689 #endif /* MULE */
|
|
|
690 ctnew->ascii[i] = new;
|
|
|
691 }
|
|
|
692
|
|
|
693 #ifdef MULE
|
|
|
694
|
|
|
695 for (i = 0; i < NUM_LEADING_BYTES; i++)
|
|
|
696 {
|
|
|
697 Lisp_Object new = ct->level1[i];
|
|
|
698 if (CHAR_TABLE_ENTRYP (new))
|
|
|
699 ctnew->level1[i] = copy_char_table_entry (new);
|
|
|
700 else
|
|
|
701 ctnew->level1[i] = new;
|
|
|
702 }
|
|
|
703
|
|
|
704 #endif /* MULE */
|
|
|
705
|
|
|
706 if (CHAR_TABLEP (ct->mirror_table))
|
|
|
707 ctnew->mirror_table = Fcopy_char_table (ct->mirror_table);
|
|
|
708 else
|
|
|
709 ctnew->mirror_table = ct->mirror_table;
|
|
|
710 ctnew->next_table = Qnil;
|
|
|
711 XSETCHAR_TABLE (obj, ctnew);
|
|
|
712 if (ctnew->type == CHAR_TABLE_TYPE_SYNTAX)
|
|
|
713 {
|
|
|
714 ctnew->next_table = Vall_syntax_tables;
|
|
|
715 Vall_syntax_tables = obj;
|
|
|
716 }
|
|
|
717 return obj;
|
|
|
718 }
|
|
|
719
|
|
|
720 static void
|
|
|
721 decode_char_table_range (Lisp_Object range, struct chartab_range *outrange)
|
|
|
722 {
|
|
|
723 if (EQ (range, Qt))
|
|
|
724 outrange->type = CHARTAB_RANGE_ALL;
|
|
|
725 else if (CHAR_OR_CHAR_INTP (range))
|
|
|
726 {
|
|
|
727 outrange->type = CHARTAB_RANGE_CHAR;
|
|
|
728 outrange->ch = XCHAR_OR_CHAR_INT (range);
|
|
|
729 }
|
|
|
730 #ifndef MULE
|
|
|
731 else
|
|
563
|
732 sferror ("Range must be t or a character", range);
|
|
428
|
733 #else /* MULE */
|
|
|
734 else if (VECTORP (range))
|
|
|
735 {
|
|
440
|
736 Lisp_Vector *vec = XVECTOR (range);
|
|
428
|
737 Lisp_Object *elts = vector_data (vec);
|
|
|
738 if (vector_length (vec) != 2)
|
|
563
|
739 sferror ("Length of charset row vector must be 2",
|
|
428
|
740 range);
|
|
|
741 outrange->type = CHARTAB_RANGE_ROW;
|
|
|
742 outrange->charset = Fget_charset (elts[0]);
|
|
|
743 CHECK_INT (elts[1]);
|
|
|
744 outrange->row = XINT (elts[1]);
|
|
|
745 switch (XCHARSET_TYPE (outrange->charset))
|
|
|
746 {
|
|
|
747 case CHARSET_TYPE_94:
|
|
|
748 case CHARSET_TYPE_96:
|
|
563
|
749 sferror ("Charset in row vector must be multi-byte",
|
|
428
|
750 outrange->charset);
|
|
|
751 case CHARSET_TYPE_94X94:
|
|
|
752 check_int_range (outrange->row, 33, 126);
|
|
|
753 break;
|
|
|
754 case CHARSET_TYPE_96X96:
|
|
|
755 check_int_range (outrange->row, 32, 127);
|
|
|
756 break;
|
|
|
757 default:
|
|
|
758 abort ();
|
|
|
759 }
|
|
|
760 }
|
|
|
761 else
|
|
|
762 {
|
|
|
763 if (!CHARSETP (range) && !SYMBOLP (range))
|
|
563
|
764 sferror
|
|
428
|
765 ("Char table range must be t, charset, char, or vector", range);
|
|
|
766 outrange->type = CHARTAB_RANGE_CHARSET;
|
|
|
767 outrange->charset = Fget_charset (range);
|
|
|
768 }
|
|
|
769 #endif /* MULE */
|
|
|
770 }
|
|
|
771
|
|
|
772 #ifdef MULE
|
|
|
773
|
|
|
774 /* called from CHAR_TABLE_VALUE(). */
|
|
|
775 Lisp_Object
|
|
440
|
776 get_non_ascii_char_table_value (Lisp_Char_Table *ct, int leading_byte,
|
|
428
|
777 Emchar c)
|
|
|
778 {
|
|
|
779 Lisp_Object val;
|
|
|
780 Lisp_Object charset = CHARSET_BY_LEADING_BYTE (leading_byte);
|
|
|
781 int byte1, byte2;
|
|
|
782
|
|
|
783 BREAKUP_CHAR_1_UNSAFE (c, charset, byte1, byte2);
|
|
|
784 val = ct->level1[leading_byte - MIN_LEADING_BYTE];
|
|
|
785 if (CHAR_TABLE_ENTRYP (val))
|
|
|
786 {
|
|
440
|
787 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (val);
|
|
428
|
788 val = cte->level2[byte1 - 32];
|
|
|
789 if (CHAR_TABLE_ENTRYP (val))
|
|
|
790 {
|
|
|
791 cte = XCHAR_TABLE_ENTRY (val);
|
|
|
792 assert (byte2 >= 32);
|
|
|
793 val = cte->level2[byte2 - 32];
|
|
|
794 assert (!CHAR_TABLE_ENTRYP (val));
|
|
|
795 }
|
|
|
796 }
|
|
|
797
|
|
|
798 return val;
|
|
|
799 }
|
|
|
800
|
|
|
801 #endif /* MULE */
|
|
|
802
|
|
|
803 Lisp_Object
|
|
440
|
804 get_char_table (Emchar ch, Lisp_Char_Table *ct)
|
|
428
|
805 {
|
|
|
806 #ifdef MULE
|
|
|
807 {
|
|
|
808 Lisp_Object charset;
|
|
|
809 int byte1, byte2;
|
|
|
810 Lisp_Object val;
|
|
|
811
|
|
|
812 BREAKUP_CHAR (ch, charset, byte1, byte2);
|
|
|
813
|
|
|
814 if (EQ (charset, Vcharset_ascii))
|
|
|
815 val = ct->ascii[byte1];
|
|
|
816 else if (EQ (charset, Vcharset_control_1))
|
|
|
817 val = ct->ascii[byte1 + 128];
|
|
|
818 else
|
|
|
819 {
|
|
|
820 int lb = XCHARSET_LEADING_BYTE (charset) - MIN_LEADING_BYTE;
|
|
|
821 val = ct->level1[lb];
|
|
|
822 if (CHAR_TABLE_ENTRYP (val))
|
|
|
823 {
|
|
440
|
824 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (val);
|
|
428
|
825 val = cte->level2[byte1 - 32];
|
|
|
826 if (CHAR_TABLE_ENTRYP (val))
|
|
|
827 {
|
|
|
828 cte = XCHAR_TABLE_ENTRY (val);
|
|
|
829 assert (byte2 >= 32);
|
|
|
830 val = cte->level2[byte2 - 32];
|
|
|
831 assert (!CHAR_TABLE_ENTRYP (val));
|
|
|
832 }
|
|
|
833 }
|
|
|
834 }
|
|
|
835
|
|
|
836 return val;
|
|
|
837 }
|
|
|
838 #else /* not MULE */
|
|
|
839 return ct->ascii[(unsigned char)ch];
|
|
|
840 #endif /* not MULE */
|
|
|
841 }
|
|
|
842
|
|
|
843
|
|
|
844 DEFUN ("get-char-table", Fget_char_table, 2, 2, 0, /*
|
|
444
|
845 Find value for CHARACTER in CHAR-TABLE.
|
|
428
|
846 */
|
|
444
|
847 (character, char_table))
|
|
428
|
848 {
|
|
444
|
849 CHECK_CHAR_TABLE (char_table);
|
|
|
850 CHECK_CHAR_COERCE_INT (character);
|
|
428
|
851
|
|
444
|
852 return get_char_table (XCHAR (character), XCHAR_TABLE (char_table));
|
|
428
|
853 }
|
|
|
854
|
|
|
855 DEFUN ("get-range-char-table", Fget_range_char_table, 2, 3, 0, /*
|
|
444
|
856 Find value for a range in CHAR-TABLE.
|
|
428
|
857 If there is more than one value, return MULTI (defaults to nil).
|
|
|
858 */
|
|
444
|
859 (range, char_table, multi))
|
|
428
|
860 {
|
|
440
|
861 Lisp_Char_Table *ct;
|
|
428
|
862 struct chartab_range rainj;
|
|
|
863
|
|
|
864 if (CHAR_OR_CHAR_INTP (range))
|
|
444
|
865 return Fget_char_table (range, char_table);
|
|
|
866 CHECK_CHAR_TABLE (char_table);
|
|
|
867 ct = XCHAR_TABLE (char_table);
|
|
428
|
868
|
|
|
869 decode_char_table_range (range, &rainj);
|
|
|
870 switch (rainj.type)
|
|
|
871 {
|
|
|
872 case CHARTAB_RANGE_ALL:
|
|
|
873 {
|
|
|
874 int i;
|
|
|
875 Lisp_Object first = ct->ascii[0];
|
|
|
876
|
|
|
877 for (i = 1; i < NUM_ASCII_CHARS; i++)
|
|
|
878 if (!EQ (first, ct->ascii[i]))
|
|
|
879 return multi;
|
|
|
880
|
|
|
881 #ifdef MULE
|
|
|
882 for (i = MIN_LEADING_BYTE; i < MIN_LEADING_BYTE + NUM_LEADING_BYTES;
|
|
|
883 i++)
|
|
|
884 {
|
|
|
885 if (!CHARSETP (CHARSET_BY_LEADING_BYTE (i))
|
|
|
886 || i == LEADING_BYTE_ASCII
|
|
|
887 || i == LEADING_BYTE_CONTROL_1)
|
|
|
888 continue;
|
|
|
889 if (!EQ (first, ct->level1[i - MIN_LEADING_BYTE]))
|
|
|
890 return multi;
|
|
|
891 }
|
|
|
892 #endif /* MULE */
|
|
|
893
|
|
|
894 return first;
|
|
|
895 }
|
|
|
896
|
|
|
897 #ifdef MULE
|
|
|
898 case CHARTAB_RANGE_CHARSET:
|
|
|
899 if (EQ (rainj.charset, Vcharset_ascii))
|
|
|
900 {
|
|
|
901 int i;
|
|
|
902 Lisp_Object first = ct->ascii[0];
|
|
|
903
|
|
|
904 for (i = 1; i < 128; i++)
|
|
|
905 if (!EQ (first, ct->ascii[i]))
|
|
|
906 return multi;
|
|
|
907 return first;
|
|
|
908 }
|
|
|
909
|
|
|
910 if (EQ (rainj.charset, Vcharset_control_1))
|
|
|
911 {
|
|
|
912 int i;
|
|
|
913 Lisp_Object first = ct->ascii[128];
|
|
|
914
|
|
|
915 for (i = 129; i < 160; i++)
|
|
|
916 if (!EQ (first, ct->ascii[i]))
|
|
|
917 return multi;
|
|
|
918 return first;
|
|
|
919 }
|
|
|
920
|
|
|
921 {
|
|
|
922 Lisp_Object val = ct->level1[XCHARSET_LEADING_BYTE (rainj.charset) -
|
|
|
923 MIN_LEADING_BYTE];
|
|
|
924 if (CHAR_TABLE_ENTRYP (val))
|
|
|
925 return multi;
|
|
|
926 return val;
|
|
|
927 }
|
|
|
928
|
|
|
929 case CHARTAB_RANGE_ROW:
|
|
|
930 {
|
|
|
931 Lisp_Object val = ct->level1[XCHARSET_LEADING_BYTE (rainj.charset) -
|
|
|
932 MIN_LEADING_BYTE];
|
|
|
933 if (!CHAR_TABLE_ENTRYP (val))
|
|
|
934 return val;
|
|
|
935 val = XCHAR_TABLE_ENTRY (val)->level2[rainj.row - 32];
|
|
|
936 if (CHAR_TABLE_ENTRYP (val))
|
|
|
937 return multi;
|
|
|
938 return val;
|
|
|
939 }
|
|
|
940 #endif /* not MULE */
|
|
|
941
|
|
|
942 default:
|
|
|
943 abort ();
|
|
|
944 }
|
|
|
945
|
|
|
946 return Qnil; /* not reached */
|
|
|
947 }
|
|
|
948
|
|
|
949 static int
|
|
|
950 check_valid_char_table_value (Lisp_Object value, enum char_table_type type,
|
|
578
|
951 Error_Behavior errb)
|
|
428
|
952 {
|
|
|
953 switch (type)
|
|
|
954 {
|
|
|
955 case CHAR_TABLE_TYPE_SYNTAX:
|
|
|
956 if (!ERRB_EQ (errb, ERROR_ME))
|
|
|
957 return INTP (value) || (CONSP (value) && INTP (XCAR (value))
|
|
|
958 && CHAR_OR_CHAR_INTP (XCDR (value)));
|
|
|
959 if (CONSP (value))
|
|
|
960 {
|
|
|
961 Lisp_Object cdr = XCDR (value);
|
|
|
962 CHECK_INT (XCAR (value));
|
|
|
963 CHECK_CHAR_COERCE_INT (cdr);
|
|
|
964 }
|
|
|
965 else
|
|
|
966 CHECK_INT (value);
|
|
|
967 break;
|
|
|
968
|
|
|
969 #ifdef MULE
|
|
|
970 case CHAR_TABLE_TYPE_CATEGORY:
|
|
|
971 if (!ERRB_EQ (errb, ERROR_ME))
|
|
|
972 return CATEGORY_TABLE_VALUEP (value);
|
|
|
973 CHECK_CATEGORY_TABLE_VALUE (value);
|
|
|
974 break;
|
|
|
975 #endif /* MULE */
|
|
|
976
|
|
|
977 case CHAR_TABLE_TYPE_GENERIC:
|
|
|
978 return 1;
|
|
|
979
|
|
|
980 case CHAR_TABLE_TYPE_DISPLAY:
|
|
|
981 /* #### fix this */
|
|
563
|
982 maybe_signal_error (Qunimplemented,
|
|
|
983 "Display char tables not yet implemented",
|
|
|
984 value, Qchar_table, errb);
|
|
428
|
985 return 0;
|
|
|
986
|
|
|
987 case CHAR_TABLE_TYPE_CHAR:
|
|
|
988 if (!ERRB_EQ (errb, ERROR_ME))
|
|
|
989 return CHAR_OR_CHAR_INTP (value);
|
|
|
990 CHECK_CHAR_COERCE_INT (value);
|
|
|
991 break;
|
|
|
992
|
|
|
993 default:
|
|
|
994 abort ();
|
|
|
995 }
|
|
|
996
|
|
|
997 return 0; /* not reached */
|
|
|
998 }
|
|
|
999
|
|
|
1000 static Lisp_Object
|
|
|
1001 canonicalize_char_table_value (Lisp_Object value, enum char_table_type type)
|
|
|
1002 {
|
|
|
1003 switch (type)
|
|
|
1004 {
|
|
|
1005 case CHAR_TABLE_TYPE_SYNTAX:
|
|
|
1006 if (CONSP (value))
|
|
|
1007 {
|
|
|
1008 Lisp_Object car = XCAR (value);
|
|
|
1009 Lisp_Object cdr = XCDR (value);
|
|
|
1010 CHECK_CHAR_COERCE_INT (cdr);
|
|
|
1011 return Fcons (car, cdr);
|
|
|
1012 }
|
|
|
1013 break;
|
|
|
1014 case CHAR_TABLE_TYPE_CHAR:
|
|
|
1015 CHECK_CHAR_COERCE_INT (value);
|
|
|
1016 break;
|
|
|
1017 default:
|
|
|
1018 break;
|
|
|
1019 }
|
|
|
1020 return value;
|
|
|
1021 }
|
|
|
1022
|
|
|
1023 DEFUN ("valid-char-table-value-p", Fvalid_char_table_value_p, 2, 2, 0, /*
|
|
|
1024 Return non-nil if VALUE is a valid value for CHAR-TABLE-TYPE.
|
|
|
1025 */
|
|
|
1026 (value, char_table_type))
|
|
|
1027 {
|
|
|
1028 enum char_table_type type = symbol_to_char_table_type (char_table_type);
|
|
|
1029
|
|
|
1030 return check_valid_char_table_value (value, type, ERROR_ME_NOT) ? Qt : Qnil;
|
|
|
1031 }
|
|
|
1032
|
|
|
1033 DEFUN ("check-valid-char-table-value", Fcheck_valid_char_table_value, 2, 2, 0, /*
|
|
|
1034 Signal an error if VALUE is not a valid value for CHAR-TABLE-TYPE.
|
|
|
1035 */
|
|
|
1036 (value, char_table_type))
|
|
|
1037 {
|
|
|
1038 enum char_table_type type = symbol_to_char_table_type (char_table_type);
|
|
|
1039
|
|
|
1040 check_valid_char_table_value (value, type, ERROR_ME);
|
|
|
1041 return Qnil;
|
|
|
1042 }
|
|
|
1043
|
|
|
1044 /* Assign VAL to all characters in RANGE in char table CT. */
|
|
|
1045
|
|
|
1046 void
|
|
440
|
1047 put_char_table (Lisp_Char_Table *ct, struct chartab_range *range,
|
|
428
|
1048 Lisp_Object val)
|
|
|
1049 {
|
|
|
1050 switch (range->type)
|
|
|
1051 {
|
|
|
1052 case CHARTAB_RANGE_ALL:
|
|
|
1053 fill_char_table (ct, val);
|
|
|
1054 return; /* avoid the duplicate call to update_syntax_table() below,
|
|
|
1055 since fill_char_table() also did that. */
|
|
|
1056
|
|
|
1057 #ifdef MULE
|
|
|
1058 case CHARTAB_RANGE_CHARSET:
|
|
|
1059 if (EQ (range->charset, Vcharset_ascii))
|
|
|
1060 {
|
|
|
1061 int i;
|
|
|
1062 for (i = 0; i < 128; i++)
|
|
|
1063 ct->ascii[i] = val;
|
|
|
1064 }
|
|
|
1065 else if (EQ (range->charset, Vcharset_control_1))
|
|
|
1066 {
|
|
|
1067 int i;
|
|
|
1068 for (i = 128; i < 160; i++)
|
|
|
1069 ct->ascii[i] = val;
|
|
|
1070 }
|
|
|
1071 else
|
|
|
1072 {
|
|
|
1073 int lb = XCHARSET_LEADING_BYTE (range->charset) - MIN_LEADING_BYTE;
|
|
|
1074 ct->level1[lb] = val;
|
|
|
1075 }
|
|
|
1076 break;
|
|
|
1077
|
|
|
1078 case CHARTAB_RANGE_ROW:
|
|
|
1079 {
|
|
440
|
1080 Lisp_Char_Table_Entry *cte;
|
|
428
|
1081 int lb = XCHARSET_LEADING_BYTE (range->charset) - MIN_LEADING_BYTE;
|
|
|
1082 /* make sure that there is a separate entry for the row. */
|
|
|
1083 if (!CHAR_TABLE_ENTRYP (ct->level1[lb]))
|
|
|
1084 ct->level1[lb] = make_char_table_entry (ct->level1[lb]);
|
|
|
1085 cte = XCHAR_TABLE_ENTRY (ct->level1[lb]);
|
|
|
1086 cte->level2[range->row - 32] = val;
|
|
|
1087 }
|
|
|
1088 break;
|
|
|
1089 #endif /* MULE */
|
|
|
1090
|
|
|
1091 case CHARTAB_RANGE_CHAR:
|
|
|
1092 #ifdef MULE
|
|
|
1093 {
|
|
|
1094 Lisp_Object charset;
|
|
|
1095 int byte1, byte2;
|
|
|
1096
|
|
|
1097 BREAKUP_CHAR (range->ch, charset, byte1, byte2);
|
|
|
1098 if (EQ (charset, Vcharset_ascii))
|
|
|
1099 ct->ascii[byte1] = val;
|
|
|
1100 else if (EQ (charset, Vcharset_control_1))
|
|
|
1101 ct->ascii[byte1 + 128] = val;
|
|
|
1102 else
|
|
|
1103 {
|
|
440
|
1104 Lisp_Char_Table_Entry *cte;
|
|
428
|
1105 int lb = XCHARSET_LEADING_BYTE (charset) - MIN_LEADING_BYTE;
|
|
|
1106 /* make sure that there is a separate entry for the row. */
|
|
|
1107 if (!CHAR_TABLE_ENTRYP (ct->level1[lb]))
|
|
|
1108 ct->level1[lb] = make_char_table_entry (ct->level1[lb]);
|
|
|
1109 cte = XCHAR_TABLE_ENTRY (ct->level1[lb]);
|
|
|
1110 /* now CTE is a char table entry for the charset;
|
|
|
1111 each entry is for a single row (or character of
|
|
|
1112 a one-octet charset). */
|
|
|
1113 if (XCHARSET_DIMENSION (charset) == 1)
|
|
|
1114 cte->level2[byte1 - 32] = val;
|
|
|
1115 else
|
|
|
1116 {
|
|
|
1117 /* assigning to one character in a two-octet charset. */
|
|
|
1118 /* make sure that the charset row contains a separate
|
|
|
1119 entry for each character. */
|
|
|
1120 if (!CHAR_TABLE_ENTRYP (cte->level2[byte1 - 32]))
|
|
|
1121 cte->level2[byte1 - 32] =
|
|
|
1122 make_char_table_entry (cte->level2[byte1 - 32]);
|
|
|
1123 cte = XCHAR_TABLE_ENTRY (cte->level2[byte1 - 32]);
|
|
|
1124 cte->level2[byte2 - 32] = val;
|
|
|
1125 }
|
|
|
1126 }
|
|
|
1127 }
|
|
|
1128 #else /* not MULE */
|
|
|
1129 ct->ascii[(unsigned char) (range->ch)] = val;
|
|
|
1130 break;
|
|
|
1131 #endif /* not MULE */
|
|
|
1132 }
|
|
|
1133
|
|
|
1134 if (ct->type == CHAR_TABLE_TYPE_SYNTAX)
|
|
|
1135 update_syntax_table (ct);
|
|
|
1136 }
|
|
|
1137
|
|
|
1138 DEFUN ("put-char-table", Fput_char_table, 3, 3, 0, /*
|
|
444
|
1139 Set the value for chars in RANGE to be VALUE in CHAR-TABLE.
|
|
428
|
1140
|
|
|
1141 RANGE specifies one or more characters to be affected and should be
|
|
|
1142 one of the following:
|
|
|
1143
|
|
|
1144 -- t (all characters are affected)
|
|
|
1145 -- A charset (only allowed when Mule support is present)
|
|
|
1146 -- A vector of two elements: a two-octet charset and a row number
|
|
|
1147 (only allowed when Mule support is present)
|
|
|
1148 -- A single character
|
|
|
1149
|
|
444
|
1150 VALUE must be a value appropriate for the type of CHAR-TABLE.
|
|
428
|
1151 See `valid-char-table-type-p'.
|
|
|
1152 */
|
|
444
|
1153 (range, value, char_table))
|
|
428
|
1154 {
|
|
440
|
1155 Lisp_Char_Table *ct;
|
|
428
|
1156 struct chartab_range rainj;
|
|
|
1157
|
|
444
|
1158 CHECK_CHAR_TABLE (char_table);
|
|
|
1159 ct = XCHAR_TABLE (char_table);
|
|
|
1160 check_valid_char_table_value (value, ct->type, ERROR_ME);
|
|
428
|
1161 decode_char_table_range (range, &rainj);
|
|
444
|
1162 value = canonicalize_char_table_value (value, ct->type);
|
|
|
1163 put_char_table (ct, &rainj, value);
|
|
428
|
1164 return Qnil;
|
|
|
1165 }
|
|
|
1166
|
|
|
1167 /* Map FN over the ASCII chars in CT. */
|
|
|
1168
|
|
|
1169 static int
|
|
440
|
1170 map_over_charset_ascii (Lisp_Char_Table *ct,
|
|
428
|
1171 int (*fn) (struct chartab_range *range,
|
|
|
1172 Lisp_Object val, void *arg),
|
|
|
1173 void *arg)
|
|
|
1174 {
|
|
|
1175 struct chartab_range rainj;
|
|
|
1176 int i, retval;
|
|
|
1177 int start = 0;
|
|
|
1178 #ifdef MULE
|
|
|
1179 int stop = 128;
|
|
|
1180 #else
|
|
|
1181 int stop = 256;
|
|
|
1182 #endif
|
|
|
1183
|
|
|
1184 rainj.type = CHARTAB_RANGE_CHAR;
|
|
|
1185
|
|
|
1186 for (i = start, retval = 0; i < stop && retval == 0; i++)
|
|
|
1187 {
|
|
|
1188 rainj.ch = (Emchar) i;
|
|
|
1189 retval = (fn) (&rainj, ct->ascii[i], arg);
|
|
|
1190 }
|
|
|
1191
|
|
|
1192 return retval;
|
|
|
1193 }
|
|
|
1194
|
|
|
1195 #ifdef MULE
|
|
|
1196
|
|
|
1197 /* Map FN over the Control-1 chars in CT. */
|
|
|
1198
|
|
|
1199 static int
|
|
440
|
1200 map_over_charset_control_1 (Lisp_Char_Table *ct,
|
|
428
|
1201 int (*fn) (struct chartab_range *range,
|
|
|
1202 Lisp_Object val, void *arg),
|
|
|
1203 void *arg)
|
|
|
1204 {
|
|
|
1205 struct chartab_range rainj;
|
|
|
1206 int i, retval;
|
|
|
1207 int start = 128;
|
|
|
1208 int stop = start + 32;
|
|
|
1209
|
|
|
1210 rainj.type = CHARTAB_RANGE_CHAR;
|
|
|
1211
|
|
|
1212 for (i = start, retval = 0; i < stop && retval == 0; i++)
|
|
|
1213 {
|
|
|
1214 rainj.ch = (Emchar) (i);
|
|
|
1215 retval = (fn) (&rainj, ct->ascii[i], arg);
|
|
|
1216 }
|
|
|
1217
|
|
|
1218 return retval;
|
|
|
1219 }
|
|
|
1220
|
|
|
1221 /* Map FN over the row ROW of two-byte charset CHARSET.
|
|
|
1222 There must be a separate value for that row in the char table.
|
|
|
1223 CTE specifies the char table entry for CHARSET. */
|
|
|
1224
|
|
|
1225 static int
|
|
440
|
1226 map_over_charset_row (Lisp_Char_Table_Entry *cte,
|
|
428
|
1227 Lisp_Object charset, int row,
|
|
|
1228 int (*fn) (struct chartab_range *range,
|
|
|
1229 Lisp_Object val, void *arg),
|
|
|
1230 void *arg)
|
|
|
1231 {
|
|
|
1232 Lisp_Object val = cte->level2[row - 32];
|
|
|
1233
|
|
|
1234 if (!CHAR_TABLE_ENTRYP (val))
|
|
|
1235 {
|
|
|
1236 struct chartab_range rainj;
|
|
|
1237
|
|
|
1238 rainj.type = CHARTAB_RANGE_ROW;
|
|
|
1239 rainj.charset = charset;
|
|
|
1240 rainj.row = row;
|
|
|
1241 return (fn) (&rainj, val, arg);
|
|
|
1242 }
|
|
|
1243 else
|
|
|
1244 {
|
|
|
1245 struct chartab_range rainj;
|
|
|
1246 int i, retval;
|
|
|
1247 int charset94_p = (XCHARSET_CHARS (charset) == 94);
|
|
|
1248 int start = charset94_p ? 33 : 32;
|
|
|
1249 int stop = charset94_p ? 127 : 128;
|
|
|
1250
|
|
|
1251 cte = XCHAR_TABLE_ENTRY (val);
|
|
|
1252
|
|
|
1253 rainj.type = CHARTAB_RANGE_CHAR;
|
|
|
1254
|
|
|
1255 for (i = start, retval = 0; i < stop && retval == 0; i++)
|
|
|
1256 {
|
|
|
1257 rainj.ch = MAKE_CHAR (charset, row, i);
|
|
|
1258 retval = (fn) (&rainj, cte->level2[i - 32], arg);
|
|
|
1259 }
|
|
|
1260 return retval;
|
|
|
1261 }
|
|
|
1262 }
|
|
|
1263
|
|
|
1264
|
|
|
1265 static int
|
|
440
|
1266 map_over_other_charset (Lisp_Char_Table *ct, int lb,
|
|
428
|
1267 int (*fn) (struct chartab_range *range,
|
|
|
1268 Lisp_Object val, void *arg),
|
|
|
1269 void *arg)
|
|
|
1270 {
|
|
|
1271 Lisp_Object val = ct->level1[lb - MIN_LEADING_BYTE];
|
|
|
1272 Lisp_Object charset = CHARSET_BY_LEADING_BYTE (lb);
|
|
|
1273
|
|
|
1274 if (!CHARSETP (charset)
|
|
|
1275 || lb == LEADING_BYTE_ASCII
|
|
|
1276 || lb == LEADING_BYTE_CONTROL_1)
|
|
|
1277 return 0;
|
|
|
1278
|
|
|
1279 if (!CHAR_TABLE_ENTRYP (val))
|
|
|
1280 {
|
|
|
1281 struct chartab_range rainj;
|
|
|
1282
|
|
|
1283 rainj.type = CHARTAB_RANGE_CHARSET;
|
|
|
1284 rainj.charset = charset;
|
|
|
1285 return (fn) (&rainj, val, arg);
|
|
|
1286 }
|
|
|
1287
|
|
|
1288 {
|
|
440
|
1289 Lisp_Char_Table_Entry *cte = XCHAR_TABLE_ENTRY (val);
|
|
428
|
1290 int charset94_p = (XCHARSET_CHARS (charset) == 94);
|
|
|
1291 int start = charset94_p ? 33 : 32;
|
|
|
1292 int stop = charset94_p ? 127 : 128;
|
|
|
1293 int i, retval;
|
|
|
1294
|
|
|
1295 if (XCHARSET_DIMENSION (charset) == 1)
|
|
|
1296 {
|
|
|
1297 struct chartab_range rainj;
|
|
|
1298 rainj.type = CHARTAB_RANGE_CHAR;
|
|
|
1299
|
|
|
1300 for (i = start, retval = 0; i < stop && retval == 0; i++)
|
|
|
1301 {
|
|
|
1302 rainj.ch = MAKE_CHAR (charset, i, 0);
|
|
|
1303 retval = (fn) (&rainj, cte->level2[i - 32], arg);
|
|
|
1304 }
|
|
|
1305 }
|
|
|
1306 else
|
|
|
1307 {
|
|
|
1308 for (i = start, retval = 0; i < stop && retval == 0; i++)
|
|
|
1309 retval = map_over_charset_row (cte, charset, i, fn, arg);
|
|
|
1310 }
|
|
|
1311
|
|
|
1312 return retval;
|
|
|
1313 }
|
|
|
1314 }
|
|
|
1315
|
|
|
1316 #endif /* MULE */
|
|
|
1317
|
|
|
1318 /* Map FN (with client data ARG) over range RANGE in char table CT.
|
|
|
1319 Mapping stops the first time FN returns non-zero, and that value
|
|
|
1320 becomes the return value of map_char_table(). */
|
|
|
1321
|
|
|
1322 int
|
|
440
|
1323 map_char_table (Lisp_Char_Table *ct,
|
|
428
|
1324 struct chartab_range *range,
|
|
|
1325 int (*fn) (struct chartab_range *range,
|
|
|
1326 Lisp_Object val, void *arg),
|
|
|
1327 void *arg)
|
|
|
1328 {
|
|
|
1329 switch (range->type)
|
|
|
1330 {
|
|
|
1331 case CHARTAB_RANGE_ALL:
|
|
|
1332 {
|
|
|
1333 int retval;
|
|
|
1334
|
|
|
1335 retval = map_over_charset_ascii (ct, fn, arg);
|
|
|
1336 if (retval)
|
|
|
1337 return retval;
|
|
|
1338 #ifdef MULE
|
|
|
1339 retval = map_over_charset_control_1 (ct, fn, arg);
|
|
|
1340 if (retval)
|
|
|
1341 return retval;
|
|
|
1342 {
|
|
|
1343 int i;
|
|
|
1344 int start = MIN_LEADING_BYTE;
|
|
|
1345 int stop = start + NUM_LEADING_BYTES;
|
|
|
1346
|
|
|
1347 for (i = start, retval = 0; i < stop && retval == 0; i++)
|
|
|
1348 {
|
|
771
|
1349 if (i != LEADING_BYTE_ASCII && i != LEADING_BYTE_CONTROL_1)
|
|
|
1350 retval = map_over_other_charset (ct, i, fn, arg);
|
|
428
|
1351 }
|
|
|
1352 }
|
|
|
1353 #endif /* MULE */
|
|
|
1354 return retval;
|
|
|
1355 }
|
|
|
1356
|
|
|
1357 #ifdef MULE
|
|
|
1358 case CHARTAB_RANGE_CHARSET:
|
|
|
1359 return map_over_other_charset (ct,
|
|
|
1360 XCHARSET_LEADING_BYTE (range->charset),
|
|
|
1361 fn, arg);
|
|
|
1362
|
|
|
1363 case CHARTAB_RANGE_ROW:
|
|
|
1364 {
|
|
771
|
1365 Lisp_Object val = ct->level1[XCHARSET_LEADING_BYTE (range->charset) -
|
|
|
1366 MIN_LEADING_BYTE];
|
|
428
|
1367 if (!CHAR_TABLE_ENTRYP (val))
|
|
|
1368 {
|
|
|
1369 struct chartab_range rainj;
|
|
|
1370
|
|
|
1371 rainj.type = CHARTAB_RANGE_ROW;
|
|
|
1372 rainj.charset = range->charset;
|
|
|
1373 rainj.row = range->row;
|
|
|
1374 return (fn) (&rainj, val, arg);
|
|
|
1375 }
|
|
|
1376 else
|
|
|
1377 return map_over_charset_row (XCHAR_TABLE_ENTRY (val),
|
|
|
1378 range->charset, range->row,
|
|
|
1379 fn, arg);
|
|
|
1380 }
|
|
|
1381 #endif /* MULE */
|
|
|
1382
|
|
|
1383 case CHARTAB_RANGE_CHAR:
|
|
|
1384 {
|
|
|
1385 Emchar ch = range->ch;
|
|
|
1386 Lisp_Object val = CHAR_TABLE_VALUE_UNSAFE (ct, ch);
|
|
|
1387 struct chartab_range rainj;
|
|
|
1388
|
|
|
1389 rainj.type = CHARTAB_RANGE_CHAR;
|
|
|
1390 rainj.ch = ch;
|
|
|
1391 return (fn) (&rainj, val, arg);
|
|
|
1392 }
|
|
|
1393
|
|
|
1394 default:
|
|
|
1395 abort ();
|
|
|
1396 }
|
|
|
1397
|
|
|
1398 return 0;
|
|
|
1399 }
|
|
|
1400
|
|
|
1401 struct slow_map_char_table_arg
|
|
|
1402 {
|
|
|
1403 Lisp_Object function;
|
|
|
1404 Lisp_Object retval;
|
|
|
1405 };
|
|
|
1406
|
|
|
1407 static int
|
|
|
1408 slow_map_char_table_fun (struct chartab_range *range,
|
|
|
1409 Lisp_Object val, void *arg)
|
|
|
1410 {
|
|
|
1411 Lisp_Object ranjarg = Qnil;
|
|
|
1412 struct slow_map_char_table_arg *closure =
|
|
|
1413 (struct slow_map_char_table_arg *) arg;
|
|
|
1414
|
|
|
1415 switch (range->type)
|
|
|
1416 {
|
|
|
1417 case CHARTAB_RANGE_ALL:
|
|
|
1418 ranjarg = Qt;
|
|
|
1419 break;
|
|
|
1420
|
|
|
1421 #ifdef MULE
|
|
|
1422 case CHARTAB_RANGE_CHARSET:
|
|
|
1423 ranjarg = XCHARSET_NAME (range->charset);
|
|
|
1424 break;
|
|
|
1425
|
|
|
1426 case CHARTAB_RANGE_ROW:
|
|
|
1427 ranjarg = vector2 (XCHARSET_NAME (range->charset),
|
|
|
1428 make_int (range->row));
|
|
|
1429 break;
|
|
|
1430 #endif /* MULE */
|
|
|
1431 case CHARTAB_RANGE_CHAR:
|
|
|
1432 ranjarg = make_char (range->ch);
|
|
|
1433 break;
|
|
|
1434 default:
|
|
|
1435 abort ();
|
|
|
1436 }
|
|
|
1437
|
|
|
1438 closure->retval = call2 (closure->function, ranjarg, val);
|
|
|
1439 return !NILP (closure->retval);
|
|
|
1440 }
|
|
|
1441
|
|
|
1442 DEFUN ("map-char-table", Fmap_char_table, 2, 3, 0, /*
|
|
444
|
1443 Map FUNCTION over entries in CHAR-TABLE, calling it with two args,
|
|
428
|
1444 each key and value in the table.
|
|
|
1445
|
|
|
1446 RANGE specifies a subrange to map over and is in the same format as
|
|
|
1447 the RANGE argument to `put-range-table'. If omitted or t, it defaults to
|
|
|
1448 the entire table.
|
|
|
1449 */
|
|
444
|
1450 (function, char_table, range))
|
|
428
|
1451 {
|
|
440
|
1452 Lisp_Char_Table *ct;
|
|
428
|
1453 struct slow_map_char_table_arg slarg;
|
|
|
1454 struct gcpro gcpro1, gcpro2;
|
|
|
1455 struct chartab_range rainj;
|
|
|
1456
|
|
444
|
1457 CHECK_CHAR_TABLE (char_table);
|
|
|
1458 ct = XCHAR_TABLE (char_table);
|
|
428
|
1459 if (NILP (range))
|
|
|
1460 range = Qt;
|
|
|
1461 decode_char_table_range (range, &rainj);
|
|
|
1462 slarg.function = function;
|
|
|
1463 slarg.retval = Qnil;
|
|
|
1464 GCPRO2 (slarg.function, slarg.retval);
|
|
|
1465 map_char_table (ct, &rainj, slow_map_char_table_fun, &slarg);
|
|
|
1466 UNGCPRO;
|
|
|
1467
|
|
|
1468 return slarg.retval;
|
|
|
1469 }
|
|
|
1470
|
|
|
1471
|
|
|
1472
|
|
|
1473 /************************************************************************/
|
|
|
1474 /* Char table read syntax */
|
|
|
1475 /************************************************************************/
|
|
|
1476
|
|
|
1477 static int
|
|
|
1478 chartab_type_validate (Lisp_Object keyword, Lisp_Object value,
|
|
578
|
1479 Error_Behavior errb)
|
|
428
|
1480 {
|
|
|
1481 /* #### should deal with ERRB */
|
|
|
1482 symbol_to_char_table_type (value);
|
|
|
1483 return 1;
|
|
|
1484 }
|
|
|
1485
|
|
|
1486 static int
|
|
|
1487 chartab_data_validate (Lisp_Object keyword, Lisp_Object value,
|
|
578
|
1488 Error_Behavior errb)
|
|
428
|
1489 {
|
|
|
1490 Lisp_Object rest;
|
|
|
1491
|
|
|
1492 /* #### should deal with ERRB */
|
|
|
1493 EXTERNAL_LIST_LOOP (rest, value)
|
|
|
1494 {
|
|
|
1495 Lisp_Object range = XCAR (rest);
|
|
|
1496 struct chartab_range dummy;
|
|
|
1497
|
|
|
1498 rest = XCDR (rest);
|
|
|
1499 if (!CONSP (rest))
|
|
563
|
1500 signal_error (Qlist_formation_error, "Invalid list format", value);
|
|
428
|
1501 if (CONSP (range))
|
|
|
1502 {
|
|
|
1503 if (!CONSP (XCDR (range))
|
|
|
1504 || !NILP (XCDR (XCDR (range))))
|
|
563
|
1505 sferror ("Invalid range format", range);
|
|
428
|
1506 decode_char_table_range (XCAR (range), &dummy);
|
|
|
1507 decode_char_table_range (XCAR (XCDR (range)), &dummy);
|
|
|
1508 }
|
|
|
1509 else
|
|
|
1510 decode_char_table_range (range, &dummy);
|
|
|
1511 }
|
|
|
1512
|
|
|
1513 return 1;
|
|
|
1514 }
|
|
|
1515
|
|
|
1516 static Lisp_Object
|
|
|
1517 chartab_instantiate (Lisp_Object data)
|
|
|
1518 {
|
|
|
1519 Lisp_Object chartab;
|
|
|
1520 Lisp_Object type = Qgeneric;
|
|
|
1521 Lisp_Object dataval = Qnil;
|
|
|
1522
|
|
|
1523 while (!NILP (data))
|
|
|
1524 {
|
|
|
1525 Lisp_Object keyw = Fcar (data);
|
|
|
1526 Lisp_Object valw;
|
|
|
1527
|
|
|
1528 data = Fcdr (data);
|
|
|
1529 valw = Fcar (data);
|
|
|
1530 data = Fcdr (data);
|
|
|
1531 if (EQ (keyw, Qtype))
|
|
|
1532 type = valw;
|
|
|
1533 else if (EQ (keyw, Qdata))
|
|
|
1534 dataval = valw;
|
|
|
1535 }
|
|
|
1536
|
|
|
1537 chartab = Fmake_char_table (type);
|
|
|
1538
|
|
|
1539 data = dataval;
|
|
|
1540 while (!NILP (data))
|
|
|
1541 {
|
|
|
1542 Lisp_Object range = Fcar (data);
|
|
|
1543 Lisp_Object val = Fcar (Fcdr (data));
|
|
|
1544
|
|
|
1545 data = Fcdr (Fcdr (data));
|
|
|
1546 if (CONSP (range))
|
|
|
1547 {
|
|
|
1548 if (CHAR_OR_CHAR_INTP (XCAR (range)))
|
|
|
1549 {
|
|
|
1550 Emchar first = XCHAR_OR_CHAR_INT (Fcar (range));
|
|
|
1551 Emchar last = XCHAR_OR_CHAR_INT (Fcar (Fcdr (range)));
|
|
|
1552 Emchar i;
|
|
|
1553
|
|
|
1554 for (i = first; i <= last; i++)
|
|
|
1555 Fput_char_table (make_char (i), val, chartab);
|
|
|
1556 }
|
|
|
1557 else
|
|
|
1558 abort ();
|
|
|
1559 }
|
|
|
1560 else
|
|
|
1561 Fput_char_table (range, val, chartab);
|
|
|
1562 }
|
|
|
1563
|
|
|
1564 return chartab;
|
|
|
1565 }
|
|
|
1566
|
|
|
1567 #ifdef MULE
|
|
|
1568
|
|
|
1569
|
|
|
1570 /************************************************************************/
|
|
|
1571 /* Category Tables, specifically */
|
|
|
1572 /************************************************************************/
|
|
|
1573
|
|
|
1574 DEFUN ("category-table-p", Fcategory_table_p, 1, 1, 0, /*
|
|
444
|
1575 Return t if OBJECT is a category table.
|
|
428
|
1576 A category table is a type of char table used for keeping track of
|
|
|
1577 categories. Categories are used for classifying characters for use
|
|
|
1578 in regexps -- you can refer to a category rather than having to use
|
|
|
1579 a complicated [] expression (and category lookups are significantly
|
|
|
1580 faster).
|
|
|
1581
|
|
|
1582 There are 95 different categories available, one for each printable
|
|
|
1583 character (including space) in the ASCII charset. Each category
|
|
|
1584 is designated by one such character, called a "category designator".
|
|
|
1585 They are specified in a regexp using the syntax "\\cX", where X is
|
|
|
1586 a category designator.
|
|
|
1587
|
|
|
1588 A category table specifies, for each character, the categories that
|
|
|
1589 the character is in. Note that a character can be in more than one
|
|
|
1590 category. More specifically, a category table maps from a character
|
|
|
1591 to either the value nil (meaning the character is in no categories)
|
|
|
1592 or a 95-element bit vector, specifying for each of the 95 categories
|
|
|
1593 whether the character is in that category.
|
|
|
1594
|
|
|
1595 Special Lisp functions are provided that abstract this, so you do not
|
|
|
1596 have to directly manipulate bit vectors.
|
|
|
1597 */
|
|
444
|
1598 (object))
|
|
428
|
1599 {
|
|
444
|
1600 return (CHAR_TABLEP (object) &&
|
|
|
1601 XCHAR_TABLE_TYPE (object) == CHAR_TABLE_TYPE_CATEGORY) ?
|
|
428
|
1602 Qt : Qnil;
|
|
|
1603 }
|
|
|
1604
|
|
|
1605 static Lisp_Object
|
|
444
|
1606 check_category_table (Lisp_Object object, Lisp_Object default_)
|
|
428
|
1607 {
|
|
444
|
1608 if (NILP (object))
|
|
|
1609 object = default_;
|
|
|
1610 while (NILP (Fcategory_table_p (object)))
|
|
|
1611 object = wrong_type_argument (Qcategory_table_p, object);
|
|
|
1612 return object;
|
|
428
|
1613 }
|
|
|
1614
|
|
|
1615 int
|
|
|
1616 check_category_char (Emchar ch, Lisp_Object table,
|
|
647
|
1617 int designator, int not_p)
|
|
428
|
1618 {
|
|
|
1619 REGISTER Lisp_Object temp;
|
|
440
|
1620 Lisp_Char_Table *ctbl;
|
|
428
|
1621 #ifdef ERROR_CHECK_TYPECHECK
|
|
|
1622 if (NILP (Fcategory_table_p (table)))
|
|
563
|
1623 wtaerror ("Expected category table", table);
|
|
428
|
1624 #endif
|
|
|
1625 ctbl = XCHAR_TABLE (table);
|
|
|
1626 temp = get_char_table (ch, ctbl);
|
|
|
1627 if (NILP (temp))
|
|
458
|
1628 return not_p;
|
|
428
|
1629
|
|
|
1630 designator -= ' ';
|
|
458
|
1631 return bit_vector_bit (XBIT_VECTOR (temp), designator) ? !not_p : not_p;
|
|
428
|
1632 }
|
|
|
1633
|
|
|
1634 DEFUN ("check-category-at", Fcheck_category_at, 2, 4, 0, /*
|
|
444
|
1635 Return t if category of the character at POSITION includes DESIGNATOR.
|
|
|
1636 Optional third arg BUFFER specifies which buffer to use, and defaults
|
|
|
1637 to the current buffer.
|
|
|
1638 Optional fourth arg CATEGORY-TABLE specifies the category table to
|
|
|
1639 use, and defaults to BUFFER's category table.
|
|
428
|
1640 */
|
|
444
|
1641 (position, designator, buffer, category_table))
|
|
428
|
1642 {
|
|
|
1643 Lisp_Object ctbl;
|
|
|
1644 Emchar ch;
|
|
647
|
1645 int des;
|
|
428
|
1646 struct buffer *buf = decode_buffer (buffer, 0);
|
|
|
1647
|
|
444
|
1648 CHECK_INT (position);
|
|
428
|
1649 CHECK_CATEGORY_DESIGNATOR (designator);
|
|
|
1650 des = XCHAR (designator);
|
|
|
1651 ctbl = check_category_table (category_table, Vstandard_category_table);
|
|
444
|
1652 ch = BUF_FETCH_CHAR (buf, XINT (position));
|
|
428
|
1653 return check_category_char (ch, ctbl, des, 0) ? Qt : Qnil;
|
|
|
1654 }
|
|
|
1655
|
|
|
1656 DEFUN ("char-in-category-p", Fchar_in_category_p, 2, 3, 0, /*
|
|
444
|
1657 Return t if category of CHARACTER includes DESIGNATOR, else nil.
|
|
|
1658 Optional third arg CATEGORY-TABLE specifies the category table to use,
|
|
|
1659 and defaults to the standard category table.
|
|
428
|
1660 */
|
|
444
|
1661 (character, designator, category_table))
|
|
428
|
1662 {
|
|
|
1663 Lisp_Object ctbl;
|
|
|
1664 Emchar ch;
|
|
647
|
1665 int des;
|
|
428
|
1666
|
|
|
1667 CHECK_CATEGORY_DESIGNATOR (designator);
|
|
|
1668 des = XCHAR (designator);
|
|
444
|
1669 CHECK_CHAR (character);
|
|
|
1670 ch = XCHAR (character);
|
|
428
|
1671 ctbl = check_category_table (category_table, Vstandard_category_table);
|
|
|
1672 return check_category_char (ch, ctbl, des, 0) ? Qt : Qnil;
|
|
|
1673 }
|
|
|
1674
|
|
|
1675 DEFUN ("category-table", Fcategory_table, 0, 1, 0, /*
|
|
444
|
1676 Return BUFFER's current category table.
|
|
|
1677 BUFFER defaults to the current buffer.
|
|
428
|
1678 */
|
|
|
1679 (buffer))
|
|
|
1680 {
|
|
|
1681 return decode_buffer (buffer, 0)->category_table;
|
|
|
1682 }
|
|
|
1683
|
|
|
1684 DEFUN ("standard-category-table", Fstandard_category_table, 0, 0, 0, /*
|
|
|
1685 Return the standard category table.
|
|
|
1686 This is the one used for new buffers.
|
|
|
1687 */
|
|
|
1688 ())
|
|
|
1689 {
|
|
|
1690 return Vstandard_category_table;
|
|
|
1691 }
|
|
|
1692
|
|
|
1693 DEFUN ("copy-category-table", Fcopy_category_table, 0, 1, 0, /*
|
|
444
|
1694 Return a new category table which is a copy of CATEGORY-TABLE.
|
|
|
1695 CATEGORY-TABLE defaults to the standard category table.
|
|
428
|
1696 */
|
|
444
|
1697 (category_table))
|
|
428
|
1698 {
|
|
|
1699 if (NILP (Vstandard_category_table))
|
|
|
1700 return Fmake_char_table (Qcategory);
|
|
|
1701
|
|
444
|
1702 category_table =
|
|
|
1703 check_category_table (category_table, Vstandard_category_table);
|
|
|
1704 return Fcopy_char_table (category_table);
|
|
428
|
1705 }
|
|
|
1706
|
|
|
1707 DEFUN ("set-category-table", Fset_category_table, 1, 2, 0, /*
|
|
444
|
1708 Select CATEGORY-TABLE as the new category table for BUFFER.
|
|
428
|
1709 BUFFER defaults to the current buffer if omitted.
|
|
|
1710 */
|
|
444
|
1711 (category_table, buffer))
|
|
428
|
1712 {
|
|
|
1713 struct buffer *buf = decode_buffer (buffer, 0);
|
|
444
|
1714 category_table = check_category_table (category_table, Qnil);
|
|
|
1715 buf->category_table = category_table;
|
|
428
|
1716 /* Indicate that this buffer now has a specified category table. */
|
|
|
1717 buf->local_var_flags |= XINT (buffer_local_flags.category_table);
|
|
444
|
1718 return category_table;
|
|
428
|
1719 }
|
|
|
1720
|
|
|
1721 DEFUN ("category-designator-p", Fcategory_designator_p, 1, 1, 0, /*
|
|
444
|
1722 Return t if OBJECT is a category designator (a char in the range ' ' to '~').
|
|
428
|
1723 */
|
|
444
|
1724 (object))
|
|
428
|
1725 {
|
|
444
|
1726 return CATEGORY_DESIGNATORP (object) ? Qt : Qnil;
|
|
428
|
1727 }
|
|
|
1728
|
|
|
1729 DEFUN ("category-table-value-p", Fcategory_table_value_p, 1, 1, 0, /*
|
|
444
|
1730 Return t if OBJECT is a category table value.
|
|
428
|
1731 Valid values are nil or a bit vector of size 95.
|
|
|
1732 */
|
|
444
|
1733 (object))
|
|
428
|
1734 {
|
|
444
|
1735 return CATEGORY_TABLE_VALUEP (object) ? Qt : Qnil;
|
|
428
|
1736 }
|
|
|
1737
|
|
|
1738
|
|
|
1739 #define CATEGORYP(x) \
|
|
|
1740 (CHARP (x) && XCHAR (x) >= 0x20 && XCHAR (x) <= 0x7E)
|
|
|
1741
|
|
|
1742 #define CATEGORY_SET(c) \
|
|
|
1743 (get_char_table(c, XCHAR_TABLE(current_buffer->category_table)))
|
|
|
1744
|
|
|
1745 /* Return 1 if CATEGORY_SET contains CATEGORY, else return 0.
|
|
|
1746 The faster version of `!NILP (Faref (category_set, category))'. */
|
|
|
1747 #define CATEGORY_MEMBER(category, category_set) \
|
|
|
1748 (bit_vector_bit(XBIT_VECTOR (category_set), category - 32))
|
|
|
1749
|
|
|
1750 /* Return 1 if there is a word boundary between two word-constituent
|
|
|
1751 characters C1 and C2 if they appear in this order, else return 0.
|
|
|
1752 Use the macro WORD_BOUNDARY_P instead of calling this function
|
|
|
1753 directly. */
|
|
|
1754
|
|
|
1755 int word_boundary_p (Emchar c1, Emchar c2);
|
|
|
1756 int
|
|
|
1757 word_boundary_p (Emchar c1, Emchar c2)
|
|
|
1758 {
|
|
|
1759 Lisp_Object category_set1, category_set2;
|
|
|
1760 Lisp_Object tail;
|
|
|
1761 int default_result;
|
|
|
1762
|
|
|
1763 #if 0
|
|
|
1764 if (COMPOSITE_CHAR_P (c1))
|
|
|
1765 c1 = cmpchar_component (c1, 0, 1);
|
|
|
1766 if (COMPOSITE_CHAR_P (c2))
|
|
|
1767 c2 = cmpchar_component (c2, 0, 1);
|
|
|
1768 #endif
|
|
|
1769
|
|
|
1770 if (EQ (CHAR_CHARSET (c1), CHAR_CHARSET (c2)))
|
|
|
1771 {
|
|
|
1772 tail = Vword_separating_categories;
|
|
|
1773 default_result = 0;
|
|
|
1774 }
|
|
|
1775 else
|
|
|
1776 {
|
|
|
1777 tail = Vword_combining_categories;
|
|
|
1778 default_result = 1;
|
|
|
1779 }
|
|
|
1780
|
|
|
1781 category_set1 = CATEGORY_SET (c1);
|
|
|
1782 if (NILP (category_set1))
|
|
|
1783 return default_result;
|
|
|
1784 category_set2 = CATEGORY_SET (c2);
|
|
|
1785 if (NILP (category_set2))
|
|
|
1786 return default_result;
|
|
|
1787
|
|
|
1788 for (; CONSP (tail); tail = XCONS (tail)->cdr)
|
|
|
1789 {
|
|
|
1790 Lisp_Object elt = XCONS(tail)->car;
|
|
|
1791
|
|
|
1792 if (CONSP (elt)
|
|
|
1793 && CATEGORYP (XCONS (elt)->car)
|
|
|
1794 && CATEGORYP (XCONS (elt)->cdr)
|
|
|
1795 && CATEGORY_MEMBER (XCHAR (XCONS (elt)->car), category_set1)
|
|
|
1796 && CATEGORY_MEMBER (XCHAR (XCONS (elt)->cdr), category_set2))
|
|
|
1797 return !default_result;
|
|
|
1798 }
|
|
|
1799 return default_result;
|
|
|
1800 }
|
|
|
1801 #endif /* MULE */
|
|
|
1802
|
|
|
1803
|
|
|
1804 void
|
|
|
1805 syms_of_chartab (void)
|
|
|
1806 {
|
|
442
|
1807 INIT_LRECORD_IMPLEMENTATION (char_table);
|
|
|
1808
|
|
428
|
1809 #ifdef MULE
|
|
442
|
1810 INIT_LRECORD_IMPLEMENTATION (char_table_entry);
|
|
|
1811
|
|
563
|
1812 DEFSYMBOL (Qcategory_table_p);
|
|
|
1813 DEFSYMBOL (Qcategory_designator_p);
|
|
|
1814 DEFSYMBOL (Qcategory_table_value_p);
|
|
428
|
1815 #endif /* MULE */
|
|
|
1816
|
|
563
|
1817 DEFSYMBOL (Qchar_table);
|
|
|
1818 DEFSYMBOL_MULTIWORD_PREDICATE (Qchar_tablep);
|
|
428
|
1819
|
|
|
1820 DEFSUBR (Fchar_table_p);
|
|
|
1821 DEFSUBR (Fchar_table_type_list);
|
|
|
1822 DEFSUBR (Fvalid_char_table_type_p);
|
|
|
1823 DEFSUBR (Fchar_table_type);
|
|
|
1824 DEFSUBR (Freset_char_table);
|
|
|
1825 DEFSUBR (Fmake_char_table);
|
|
|
1826 DEFSUBR (Fcopy_char_table);
|
|
|
1827 DEFSUBR (Fget_char_table);
|
|
|
1828 DEFSUBR (Fget_range_char_table);
|
|
|
1829 DEFSUBR (Fvalid_char_table_value_p);
|
|
|
1830 DEFSUBR (Fcheck_valid_char_table_value);
|
|
|
1831 DEFSUBR (Fput_char_table);
|
|
|
1832 DEFSUBR (Fmap_char_table);
|
|
|
1833
|
|
|
1834 #ifdef MULE
|
|
|
1835 DEFSUBR (Fcategory_table_p);
|
|
|
1836 DEFSUBR (Fcategory_table);
|
|
|
1837 DEFSUBR (Fstandard_category_table);
|
|
|
1838 DEFSUBR (Fcopy_category_table);
|
|
|
1839 DEFSUBR (Fset_category_table);
|
|
|
1840 DEFSUBR (Fcheck_category_at);
|
|
|
1841 DEFSUBR (Fchar_in_category_p);
|
|
|
1842 DEFSUBR (Fcategory_designator_p);
|
|
|
1843 DEFSUBR (Fcategory_table_value_p);
|
|
|
1844 #endif /* MULE */
|
|
|
1845
|
|
|
1846 }
|
|
|
1847
|
|
|
1848 void
|
|
|
1849 vars_of_chartab (void)
|
|
|
1850 {
|
|
|
1851 /* DO NOT staticpro this. It works just like Vweak_hash_tables. */
|
|
|
1852 Vall_syntax_tables = Qnil;
|
|
452
|
1853 dump_add_weak_object_chain (&Vall_syntax_tables);
|
|
428
|
1854 }
|
|
|
1855
|
|
|
1856 void
|
|
|
1857 structure_type_create_chartab (void)
|
|
|
1858 {
|
|
|
1859 struct structure_type *st;
|
|
|
1860
|
|
|
1861 st = define_structure_type (Qchar_table, 0, chartab_instantiate);
|
|
|
1862
|
|
|
1863 define_structure_type_keyword (st, Qtype, chartab_type_validate);
|
|
|
1864 define_structure_type_keyword (st, Qdata, chartab_data_validate);
|
|
|
1865 }
|
|
|
1866
|
|
|
1867 void
|
|
|
1868 complex_vars_of_chartab (void)
|
|
|
1869 {
|
|
|
1870 #ifdef MULE
|
|
|
1871 /* Set this now, so first buffer creation can refer to it. */
|
|
|
1872 /* Make it nil before calling copy-category-table
|
|
|
1873 so that copy-category-table will know not to try to copy from garbage */
|
|
|
1874 Vstandard_category_table = Qnil;
|
|
|
1875 Vstandard_category_table = Fcopy_category_table (Qnil);
|
|
|
1876 staticpro (&Vstandard_category_table);
|
|
|
1877
|
|
|
1878 DEFVAR_LISP ("word-combining-categories", &Vword_combining_categories /*
|
|
|
1879 List of pair (cons) of categories to determine word boundary.
|
|
|
1880
|
|
|
1881 Emacs treats a sequence of word constituent characters as a single
|
|
|
1882 word (i.e. finds no word boundary between them) iff they belongs to
|
|
|
1883 the same charset. But, exceptions are allowed in the following cases.
|
|
|
1884
|
|
444
|
1885 \(1) The case that characters are in different charsets is controlled
|
|
428
|
1886 by the variable `word-combining-categories'.
|
|
|
1887
|
|
|
1888 Emacs finds no word boundary between characters of different charsets
|
|
|
1889 if they have categories matching some element of this list.
|
|
|
1890
|
|
|
1891 More precisely, if an element of this list is a cons of category CAT1
|
|
|
1892 and CAT2, and a multibyte character C1 which has CAT1 is followed by
|
|
|
1893 C2 which has CAT2, there's no word boundary between C1 and C2.
|
|
|
1894
|
|
|
1895 For instance, to tell that ASCII characters and Latin-1 characters can
|
|
|
1896 form a single word, the element `(?l . ?l)' should be in this list
|
|
|
1897 because both characters have the category `l' (Latin characters).
|
|
|
1898
|
|
444
|
1899 \(2) The case that character are in the same charset is controlled by
|
|
428
|
1900 the variable `word-separating-categories'.
|
|
|
1901
|
|
|
1902 Emacs find a word boundary between characters of the same charset
|
|
|
1903 if they have categories matching some element of this list.
|
|
|
1904
|
|
|
1905 More precisely, if an element of this list is a cons of category CAT1
|
|
|
1906 and CAT2, and a multibyte character C1 which has CAT1 is followed by
|
|
|
1907 C2 which has CAT2, there's a word boundary between C1 and C2.
|
|
|
1908
|
|
|
1909 For instance, to tell that there's a word boundary between Japanese
|
|
|
1910 Hiragana and Japanese Kanji (both are in the same charset), the
|
|
|
1911 element `(?H . ?C) should be in this list.
|
|
|
1912 */ );
|
|
|
1913
|
|
|
1914 Vword_combining_categories = Qnil;
|
|
|
1915
|
|
|
1916 DEFVAR_LISP ("word-separating-categories", &Vword_separating_categories /*
|
|
|
1917 List of pair (cons) of categories to determine word boundary.
|
|
|
1918 See the documentation of the variable `word-combining-categories'.
|
|
|
1919 */ );
|
|
|
1920
|
|
|
1921 Vword_separating_categories = Qnil;
|
|
|
1922 #endif /* MULE */
|
|
|
1923 }
|
