Commit | Line | Data |
---|---|---|
ca4c9455 | 1 | /* Composite sequence support. |
aaef169d | 2 | Copyright (C) 2001, 2002, 2003, 2004, 2005, |
4e6835db GM |
3 | 2006, 2007 Free Software Foundation, Inc. |
4 | Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 | |
ce03bf76 KH |
5 | National Institute of Advanced Industrial Science and Technology (AIST) |
6 | Registration Number H14PRO021 | |
ca4c9455 KH |
7 | |
8 | This file is part of GNU Emacs. | |
9 | ||
10 | GNU Emacs is free software; you can redistribute it and/or modify | |
11 | it under the terms of the GNU General Public License as published by | |
12 | the Free Software Foundation; either version 2, or (at your option) | |
13 | any later version. | |
14 | ||
15 | GNU Emacs is distributed in the hope that it will be useful, | |
16 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
18 | GNU General Public License for more details. | |
19 | ||
20 | You should have received a copy of the GNU General Public License | |
21 | along with GNU Emacs; see the file COPYING. If not, write to | |
4fc5845f LK |
22 | the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
23 | Boston, MA 02110-1301, USA. */ | |
ca4c9455 KH |
24 | |
25 | #include <config.h> | |
26 | #include "lisp.h" | |
27 | #include "buffer.h" | |
28 | #include "charset.h" | |
29 | #include "intervals.h" | |
30 | ||
31 | /* Emacs uses special text property `composition' to support character | |
32 | composition. A sequence of characters that have the same (i.e. eq) | |
33 | `composition' property value is treated as a single composite | |
34 | sequence (we call it just `composition' here after). Characters in | |
35 | a composition are all composed somehow on the screen. | |
36 | ||
37 | The property value has this form when the composition is made: | |
38 | ((LENGTH . COMPONENTS) . MODIFICATION-FUNC) | |
39 | then turns to this form: | |
40 | (COMPOSITION-ID . (LENGTH COMPONENTS-VEC . MODIFICATION-FUNC)) | |
41 | when the composition is registered in composition_hash_table and | |
42 | composition_table. These rather peculiar structures were designed | |
43 | to make it easy to distinguish them quickly (we can do that by | |
44 | checking only the first element) and to extract LENGTH (from the | |
45 | former form) and COMPOSITION-ID (from the latter form). | |
46 | ||
47 | We register a composition when it is displayed, or when the width | |
48 | is required (for instance, to calculate columns). | |
49 | ||
50 | LENGTH -- Length of the composition. This information is used to | |
51 | check the validity of the composition. | |
52 | ||
53 | COMPONENTS -- Character, string, vector, list, or nil. | |
54 | ||
55 | If it is nil, characters in the text are composed relatively | |
56 | according to their metrics in font glyphs. | |
57 | ||
58 | If it is a character or a string, the character or characters | |
59 | in the string are composed relatively. | |
60 | ||
61 | If it is a vector or list of integers, the element is a | |
62 | character or an encoded composition rule. The characters are | |
63 | composed according to the rules. (2N)th elements are | |
64 | characters to be composed and (2N+1)th elements are | |
65 | composition rules to tell how to compose (2N+2)th element with | |
66 | the previously composed 2N glyphs. | |
67 | ||
68 | COMPONENTS-VEC -- Vector of integers. In relative composition, the | |
69 | elements are characters to be composed. In rule-base | |
70 | composition, the elements are characters or encoded | |
71 | composition rules. | |
72 | ||
73 | MODIFICATION-FUNC -- If non nil, it is a function to call when the | |
74 | composition gets invalid after a modification in a buffer. If | |
75 | it is nil, a function in `composition-function-table' of the | |
76 | first character in the sequence is called. | |
77 | ||
78 | COMPOSITION-ID --Identification number of the composition. It is | |
79 | used as an index to composition_table for the composition. | |
80 | ||
81 | When Emacs has to display a composition or has to know its | |
82 | displaying width, the function get_composition_id is called. It | |
83 | returns COMPOSITION-ID so that the caller can access the | |
84 | information about the composition through composition_table. If a | |
85 | COMPOSITION-ID has not yet been assigned to the composition, | |
86 | get_composition_id checks the validity of `composition' property, | |
87 | and, if valid, assigns a new ID, registers the information in | |
88 | composition_hash_table and composition_table, and changes the form | |
89 | of the property value. If the property is invalid, return -1 | |
90 | without changing the property value. | |
91 | ||
92 | We use two tables to keep information about composition; | |
93 | composition_hash_table and composition_table. | |
94 | ||
95 | The former is a hash table in which keys are COMPONENTS-VECs and | |
96 | values are the corresponding COMPOSITION-IDs. This hash table is | |
4abc7470 | 97 | weak, but as each key (COMPONENTS-VEC) is also kept as a value of the |
ca4c9455 | 98 | `composition' property, it won't be collected as garbage until all |
4abc7470 | 99 | bits of text that have the same COMPONENTS-VEC are deleted. |
ca4c9455 KH |
100 | |
101 | The latter is a table of pointers to `struct composition' indexed | |
4abc7470 | 102 | by COMPOSITION-ID. This structure keeps the other information (see |
ca4c9455 KH |
103 | composite.h). |
104 | ||
105 | In general, a text property holds information about individual | |
106 | characters. But, a `composition' property holds information about | |
4abc7470 | 107 | a sequence of characters (in this sense, it is like the `intangible' |
ca4c9455 | 108 | property). That means that we should not share the property value |
4abc7470 | 109 | in adjacent compositions -- we can't distinguish them if they have the |
ca4c9455 KH |
110 | same property. So, after any changes, we call |
111 | `update_compositions' and change a property of one of adjacent | |
112 | compositions to a copy of it. This function also runs a proper | |
113 | composition modification function to make a composition that gets | |
114 | invalid by the change valid again. | |
115 | ||
4abc7470 | 116 | As the value of the `composition' property holds information about a |
ca4c9455 | 117 | specific range of text, the value gets invalid if we change the |
4abc7470 | 118 | text in the range. We treat the `composition' property as always |
ca4c9455 KH |
119 | rear-nonsticky (currently by setting default-text-properties to |
120 | (rear-nonsticky (composition))) and we never make properties of | |
121 | adjacent compositions identical. Thus, any such changes make the | |
4abc7470 | 122 | range just shorter. So, we can check the validity of the `composition' |
ca4c9455 KH |
123 | property by comparing LENGTH information with the actual length of |
124 | the composition. | |
125 | ||
126 | */ | |
127 | ||
128 | ||
129 | Lisp_Object Qcomposition; | |
130 | ||
131 | /* Table of pointers to the structure `composition' indexed by | |
132 | COMPOSITION-ID. This structure is for storing information about | |
133 | each composition except for COMPONENTS-VEC. */ | |
134 | struct composition **composition_table; | |
135 | ||
136 | /* The current size of `composition_table'. */ | |
137 | static int composition_table_size; | |
138 | ||
139 | /* Number of compositions currently made. */ | |
140 | int n_compositions; | |
141 | ||
142 | /* Hash table for compositions. The key is COMPONENTS-VEC of | |
143 | `composition' property. The value is the corresponding | |
144 | COMPOSITION-ID. */ | |
145 | Lisp_Object composition_hash_table; | |
146 | ||
147 | /* Function to call to adjust composition. */ | |
148 | Lisp_Object Vcompose_chars_after_function; | |
149 | ||
40add26d KH |
150 | /* Char-table of patterns and functions to make a composition. */ |
151 | Lisp_Object Vcomposition_function_table; | |
152 | Lisp_Object Qcomposition_function_table; | |
153 | ||
ca4c9455 KH |
154 | /* Temporary variable used in macros COMPOSITION_XXX. */ |
155 | Lisp_Object composition_temp; | |
156 | \f | |
157 | /* Return how many columns C will occupy on the screen. It always | |
158 | returns 1 for control characters and 8-bit characters because those | |
159 | are just ignored in a composition. */ | |
160 | #define CHAR_WIDTH(c) \ | |
161 | (SINGLE_BYTE_CHAR_P (c) ? 1 : CHARSET_WIDTH (CHAR_CHARSET (c))) | |
162 | ||
ca4c9455 KH |
163 | /* Return COMPOSITION-ID of a composition at buffer position |
164 | CHARPOS/BYTEPOS and length NCHARS. The `composition' property of | |
165 | the sequence is PROP. STRING, if non-nil, is a string that | |
166 | contains the composition instead of the current buffer. | |
167 | ||
168 | If the composition is invalid, return -1. */ | |
169 | ||
170 | int | |
171 | get_composition_id (charpos, bytepos, nchars, prop, string) | |
172 | int charpos, bytepos, nchars; | |
173 | Lisp_Object prop, string; | |
174 | { | |
175 | Lisp_Object id, length, components, key, *key_contents; | |
176 | int glyph_len; | |
177 | struct Lisp_Hash_Table *hash_table = XHASH_TABLE (composition_hash_table); | |
178 | int hash_index; | |
179 | unsigned hash_code; | |
180 | struct composition *cmp; | |
181 | int i, ch; | |
182 | ||
183 | /* PROP should be | |
184 | Form-A: ((LENGTH . COMPONENTS) . MODIFICATION-FUNC) | |
185 | or | |
186 | Form-B: (COMPOSITION-ID . (LENGTH COMPONENTS-VEC . MODIFICATION-FUNC)) | |
187 | */ | |
188 | if (nchars == 0 || !CONSP (prop)) | |
189 | goto invalid_composition; | |
190 | ||
191 | id = XCAR (prop); | |
192 | if (INTEGERP (id)) | |
193 | { | |
194 | /* PROP should be Form-B. */ | |
195 | if (XINT (id) < 0 || XINT (id) >= n_compositions) | |
196 | goto invalid_composition; | |
197 | return XINT (id); | |
198 | } | |
199 | ||
200 | /* PROP should be Form-A. | |
201 | Thus, ID should be (LENGTH . COMPONENTS). */ | |
202 | if (!CONSP (id)) | |
203 | goto invalid_composition; | |
204 | length = XCAR (id); | |
205 | if (!INTEGERP (length) || XINT (length) != nchars) | |
206 | goto invalid_composition; | |
207 | ||
208 | components = XCDR (id); | |
209 | ||
210 | /* Check if the same composition has already been registered or not | |
211 | by consulting composition_hash_table. The key for this table is | |
212 | COMPONENTS (converted to a vector COMPONENTS-VEC) or, if it is | |
213 | nil, vector of characters in the composition range. */ | |
214 | if (INTEGERP (components)) | |
215 | key = Fmake_vector (make_number (1), components); | |
216 | else if (STRINGP (components) || CONSP (components)) | |
217 | key = Fvconcat (1, &components); | |
218 | else if (VECTORP (components)) | |
219 | key = components; | |
220 | else if (NILP (components)) | |
221 | { | |
222 | key = Fmake_vector (make_number (nchars), Qnil); | |
223 | if (STRINGP (string)) | |
224 | for (i = 0; i < nchars; i++) | |
225 | { | |
226 | FETCH_STRING_CHAR_ADVANCE (ch, string, charpos, bytepos); | |
227 | XVECTOR (key)->contents[i] = make_number (ch); | |
228 | } | |
229 | else | |
230 | for (i = 0; i < nchars; i++) | |
231 | { | |
232 | FETCH_CHAR_ADVANCE (ch, charpos, bytepos); | |
233 | XVECTOR (key)->contents[i] = make_number (ch); | |
234 | } | |
235 | } | |
236 | else | |
237 | goto invalid_composition; | |
238 | ||
239 | hash_index = hash_lookup (hash_table, key, &hash_code); | |
240 | if (hash_index >= 0) | |
241 | { | |
242 | /* We have already registered the same composition. Change PROP | |
243 | from Form-A above to Form-B while replacing COMPONENTS with | |
244 | COMPONENTS-VEC stored in the hash table. We can directly | |
245 | modify the cons cell of PROP because it is not shared. */ | |
246 | key = HASH_KEY (hash_table, hash_index); | |
247 | id = HASH_VALUE (hash_table, hash_index); | |
f3fbd155 KR |
248 | XSETCAR (prop, id); |
249 | XSETCDR (prop, Fcons (make_number (nchars), Fcons (key, XCDR (prop)))); | |
ca4c9455 KH |
250 | return XINT (id); |
251 | } | |
252 | ||
253 | /* This composition is a new one. We must register it. */ | |
177c0ea7 | 254 | |
ca4c9455 KH |
255 | /* Check if we have sufficient memory to store this information. */ |
256 | if (composition_table_size == 0) | |
257 | { | |
258 | composition_table_size = 256; | |
259 | composition_table | |
260 | = (struct composition **) xmalloc (sizeof (composition_table[0]) | |
261 | * composition_table_size); | |
262 | } | |
263 | else if (composition_table_size <= n_compositions) | |
264 | { | |
265 | composition_table_size += 256; | |
266 | composition_table | |
267 | = (struct composition **) xrealloc (composition_table, | |
268 | sizeof (composition_table[0]) | |
269 | * composition_table_size); | |
270 | } | |
271 | ||
272 | key_contents = XVECTOR (key)->contents; | |
273 | ||
274 | /* Check if the contents of COMPONENTS are valid if COMPONENTS is a | |
275 | vector or a list. It should be a sequence of: | |
276 | char1 rule1 char2 rule2 char3 ... ruleN charN+1 */ | |
277 | if (VECTORP (components) || CONSP (components)) | |
278 | { | |
279 | int len = XVECTOR (key)->size; | |
280 | ||
281 | /* The number of elements should be odd. */ | |
282 | if ((len % 2) == 0) | |
283 | goto invalid_composition; | |
284 | /* All elements should be integers (character or encoded | |
285 | composition rule). */ | |
286 | for (i = 0; i < len; i++) | |
287 | { | |
288 | if (!INTEGERP (key_contents[i])) | |
289 | goto invalid_composition; | |
290 | } | |
291 | } | |
292 | ||
293 | /* Change PROP from Form-A above to Form-B. We can directly modify | |
294 | the cons cell of PROP because it is not shared. */ | |
295 | XSETFASTINT (id, n_compositions); | |
f3fbd155 KR |
296 | XSETCAR (prop, id); |
297 | XSETCDR (prop, Fcons (make_number (nchars), Fcons (key, XCDR (prop)))); | |
ca4c9455 KH |
298 | |
299 | /* Register the composition in composition_hash_table. */ | |
300 | hash_index = hash_put (hash_table, key, id, hash_code); | |
301 | ||
302 | /* Register the composition in composition_table. */ | |
303 | cmp = (struct composition *) xmalloc (sizeof (struct composition)); | |
304 | ||
305 | cmp->method = (NILP (components) | |
306 | ? COMPOSITION_RELATIVE | |
307 | : ((INTEGERP (components) || STRINGP (components)) | |
308 | ? COMPOSITION_WITH_ALTCHARS | |
309 | : COMPOSITION_WITH_RULE_ALTCHARS)); | |
310 | cmp->hash_index = hash_index; | |
311 | glyph_len = (cmp->method == COMPOSITION_WITH_RULE_ALTCHARS | |
312 | ? (XVECTOR (key)->size + 1) / 2 | |
313 | : XVECTOR (key)->size); | |
314 | cmp->glyph_len = glyph_len; | |
315 | cmp->offsets = (short *) xmalloc (sizeof (short) * glyph_len * 2); | |
316 | cmp->font = NULL; | |
317 | ||
318 | /* Calculate the width of overall glyphs of the composition. */ | |
319 | if (cmp->method != COMPOSITION_WITH_RULE_ALTCHARS) | |
320 | { | |
321 | /* Relative composition. */ | |
322 | cmp->width = 0; | |
323 | for (i = 0; i < glyph_len; i++) | |
324 | { | |
325 | int this_width; | |
326 | ch = XINT (key_contents[i]); | |
327 | this_width = CHAR_WIDTH (ch); | |
328 | if (cmp->width < this_width) | |
329 | cmp->width = this_width; | |
330 | } | |
331 | } | |
332 | else | |
333 | { | |
334 | /* Rule-base composition. */ | |
335 | float leftmost = 0.0, rightmost; | |
336 | ||
337 | ch = XINT (key_contents[0]); | |
338 | rightmost = CHAR_WIDTH (ch); | |
339 | ||
340 | for (i = 1; i < glyph_len; i += 2) | |
341 | { | |
342 | int rule, gref, nref; | |
343 | int this_width; | |
344 | float this_left; | |
345 | ||
346 | rule = XINT (key_contents[i]); | |
347 | ch = XINT (key_contents[i + 1]); | |
348 | this_width = CHAR_WIDTH (ch); | |
349 | ||
350 | /* A composition rule is specified by an integer value | |
351 | that encodes global and new reference points (GREF and | |
352 | NREF). GREF and NREF are specified by numbers as | |
353 | below: | |
354 | 0---1---2 -- ascent | |
355 | | | | |
356 | | | | |
357 | | | | |
358 | 9--10--11 -- center | |
359 | | | | |
360 | ---3---4---5--- baseline | |
361 | | | | |
362 | 6---7---8 -- descent | |
363 | */ | |
364 | COMPOSITION_DECODE_RULE (rule, gref, nref); | |
365 | this_left = (leftmost | |
366 | + (gref % 3) * (rightmost - leftmost) / 2.0 | |
367 | - (nref % 3) * this_width / 2.0); | |
368 | ||
369 | if (this_left < leftmost) | |
370 | leftmost = this_left; | |
371 | if (this_left + this_width > rightmost) | |
372 | rightmost = this_left + this_width; | |
373 | } | |
374 | ||
375 | cmp->width = rightmost - leftmost; | |
376 | if (cmp->width < (rightmost - leftmost)) | |
377 | /* To get a ceiling integer value. */ | |
378 | cmp->width++; | |
379 | } | |
380 | ||
381 | composition_table[n_compositions] = cmp; | |
382 | ||
383 | return n_compositions++; | |
384 | ||
385 | invalid_composition: | |
386 | /* Would it be better to remove this `composition' property? */ | |
387 | return -1; | |
388 | } | |
389 | ||
390 | \f | |
391 | /* Find a composition at or nearest to position POS of OBJECT (buffer | |
392 | or string). | |
393 | ||
394 | OBJECT defaults to the current buffer. If there's a composition at | |
395 | POS, set *START and *END to the start and end of the sequence, | |
396 | *PROP to the `composition' property, and return 1. | |
397 | ||
398 | If there's no composition at POS and LIMIT is negative, return 0. | |
399 | ||
400 | Otherwise, search for a composition forward (LIMIT > POS) or | |
401 | backward (LIMIT < POS). In this case, LIMIT bounds the search. | |
402 | ||
403 | If a composition is found, set *START, *END, and *PROP as above, | |
404 | and return 1, else return 0. | |
405 | ||
406 | This doesn't check the validity of composition. */ | |
407 | ||
408 | int | |
409 | find_composition (pos, limit, start, end, prop, object) | |
410 | int pos, limit, *start, *end; | |
411 | Lisp_Object *prop, object; | |
412 | { | |
413 | Lisp_Object val; | |
414 | ||
415 | if (get_property_and_range (pos, Qcomposition, prop, start, end, object)) | |
416 | return 1; | |
417 | ||
418 | if (limit < 0 || limit == pos) | |
419 | return 0; | |
420 | ||
421 | if (limit > pos) /* search forward */ | |
d279f620 KH |
422 | { |
423 | val = Fnext_single_property_change (make_number (pos), Qcomposition, | |
424 | object, make_number (limit)); | |
425 | pos = XINT (val); | |
426 | if (pos == limit) | |
427 | return 0; | |
428 | } | |
ca4c9455 | 429 | else /* search backward */ |
d279f620 KH |
430 | { |
431 | if (get_property_and_range (pos - 1, Qcomposition, prop, start, end, | |
432 | object)) | |
433 | return 1; | |
434 | val = Fprevious_single_property_change (make_number (pos), Qcomposition, | |
435 | object, make_number (limit)); | |
436 | pos = XINT (val); | |
437 | if (pos == limit) | |
438 | return 0; | |
439 | pos--; | |
440 | } | |
ca4c9455 KH |
441 | get_property_and_range (pos, Qcomposition, prop, start, end, object); |
442 | return 1; | |
443 | } | |
444 | ||
445 | /* Run a proper function to adjust the composition sitting between | |
446 | FROM and TO with property PROP. */ | |
447 | ||
448 | static void | |
449 | run_composition_function (from, to, prop) | |
450 | int from, to; | |
451 | Lisp_Object prop; | |
452 | { | |
7d019510 | 453 | Lisp_Object func; |
ca4c9455 KH |
454 | int start, end; |
455 | ||
456 | func = COMPOSITION_MODIFICATION_FUNC (prop); | |
457 | /* If an invalid composition precedes or follows, try to make them | |
458 | valid too. */ | |
459 | if (from > BEGV | |
460 | && find_composition (from - 1, -1, &start, &end, &prop, Qnil) | |
461 | && !COMPOSITION_VALID_P (start, end, prop)) | |
462 | from = start; | |
463 | if (to < ZV | |
464 | && find_composition (to, -1, &start, &end, &prop, Qnil) | |
465 | && !COMPOSITION_VALID_P (start, end, prop)) | |
466 | to = end; | |
775b3d2d | 467 | if (!NILP (Ffboundp (func))) |
ca4c9455 | 468 | call2 (func, make_number (from), make_number (to)); |
09654086 | 469 | else if (!NILP (Ffboundp (Vcompose_chars_after_function))) |
40add26d KH |
470 | call3 (Vcompose_chars_after_function, |
471 | make_number (from), make_number (to), Qnil); | |
ca4c9455 KH |
472 | } |
473 | ||
474 | /* Make invalid compositions adjacent to or inside FROM and TO valid. | |
475 | CHECK_MASK is bitwise `or' of mask bits defined by macros | |
476 | CHECK_XXX (see the comment in composite.h). | |
477 | ||
478 | This function is called when a buffer text is changed. If the | |
479 | change is deletion, FROM == TO. Otherwise, FROM < TO. */ | |
480 | ||
481 | void | |
482 | update_compositions (from, to, check_mask) | |
1e26d5ef | 483 | int from, to, check_mask; |
ca4c9455 | 484 | { |
7d019510 | 485 | Lisp_Object prop; |
ca4c9455 KH |
486 | int start, end; |
487 | ||
6c1aa7f1 GM |
488 | if (inhibit_modification_hooks) |
489 | return; | |
177c0ea7 | 490 | |
d3f40cbd KH |
491 | /* If FROM and TO are not in a valid range, do nothing. */ |
492 | if (! (BEGV <= from && from <= to && to <= ZV)) | |
493 | return; | |
494 | ||
ca4c9455 KH |
495 | if (check_mask & CHECK_HEAD) |
496 | { | |
497 | /* FROM should be at composition boundary. But, insertion or | |
498 | deletion will make two compositions adjacent and | |
499 | indistinguishable when they have same (eq) property. To | |
500 | avoid it, in such a case, we change the property of the | |
501 | latter to the copy of it. */ | |
502 | if (from > BEGV | |
553d3164 KH |
503 | && find_composition (from - 1, -1, &start, &end, &prop, Qnil) |
504 | && COMPOSITION_VALID_P (start, end, prop)) | |
ca4c9455 KH |
505 | { |
506 | if (from < end) | |
507 | Fput_text_property (make_number (from), make_number (end), | |
508 | Qcomposition, | |
509 | Fcons (XCAR (prop), XCDR (prop)), Qnil); | |
510 | run_composition_function (start, end, prop); | |
511 | from = end; | |
512 | } | |
dd33cc56 | 513 | else if (from < ZV |
553d3164 KH |
514 | && find_composition (from, -1, &start, &from, &prop, Qnil) |
515 | && COMPOSITION_VALID_P (start, end, prop)) | |
ca4c9455 KH |
516 | run_composition_function (start, from, prop); |
517 | } | |
518 | ||
519 | if (check_mask & CHECK_INSIDE) | |
520 | { | |
521 | /* In this case, we are sure that (check & CHECK_TAIL) is also | |
522 | nonzero. Thus, here we should check only compositions before | |
523 | (to - 1). */ | |
524 | while (from < to - 1 | |
525 | && find_composition (from, to, &start, &from, &prop, Qnil) | |
553d3164 | 526 | && COMPOSITION_VALID_P (start, end, prop) |
ca4c9455 KH |
527 | && from < to - 1) |
528 | run_composition_function (start, from, prop); | |
529 | } | |
530 | ||
531 | if (check_mask & CHECK_TAIL) | |
532 | { | |
533 | if (from < to | |
553d3164 KH |
534 | && find_composition (to - 1, -1, &start, &end, &prop, Qnil) |
535 | && COMPOSITION_VALID_P (start, end, prop)) | |
ca4c9455 KH |
536 | { |
537 | /* TO should be also at composition boundary. But, | |
538 | insertion or deletion will make two compositions adjacent | |
539 | and indistinguishable when they have same (eq) property. | |
540 | To avoid it, in such a case, we change the property of | |
541 | the former to the copy of it. */ | |
542 | if (to < end) | |
543 | Fput_text_property (make_number (start), make_number (to), | |
544 | Qcomposition, | |
545 | Fcons (XCAR (prop), XCDR (prop)), Qnil); | |
546 | run_composition_function (start, end, prop); | |
547 | } | |
548 | else if (to < ZV | |
553d3164 KH |
549 | && find_composition (to, -1, &start, &end, &prop, Qnil) |
550 | && COMPOSITION_VALID_P (start, end, prop)) | |
ca4c9455 KH |
551 | run_composition_function (start, end, prop); |
552 | } | |
553 | } | |
554 | ||
c1361885 KH |
555 | |
556 | /* Modify composition property values in LIST destructively. LIST is | |
557 | a list as returned from text_property_list. Change values to the | |
558 | top-level copies of them so that none of them are `eq'. */ | |
559 | ||
560 | void | |
561 | make_composition_value_copy (list) | |
562 | Lisp_Object list; | |
563 | { | |
564 | Lisp_Object plist, val; | |
565 | ||
566 | for (; CONSP (list); list = XCDR (list)) | |
567 | { | |
568 | plist = XCAR (XCDR (XCDR (XCAR (list)))); | |
569 | while (CONSP (plist) && CONSP (XCDR (plist))) | |
570 | { | |
571 | if (EQ (XCAR (plist), Qcomposition) | |
572 | && (val = XCAR (XCDR (plist)), CONSP (val))) | |
f3fbd155 | 573 | XSETCAR (XCDR (plist), Fcons (XCAR (val), XCDR (val))); |
c1361885 KH |
574 | plist = XCDR (XCDR (plist)); |
575 | } | |
576 | } | |
577 | } | |
578 | ||
579 | ||
ca4c9455 KH |
580 | /* Make text in the region between START and END a composition that |
581 | has COMPONENTS and MODIFICATION-FUNC. | |
582 | ||
583 | If STRING is non-nil, then operate on characters contained between | |
584 | indices START and END in STRING. */ | |
585 | ||
586 | void | |
587 | compose_text (start, end, components, modification_func, string) | |
588 | int start, end; | |
589 | Lisp_Object components, modification_func, string; | |
590 | { | |
591 | Lisp_Object prop; | |
592 | ||
593 | prop = Fcons (Fcons (make_number (end - start), components), | |
594 | modification_func); | |
595 | Fput_text_property (make_number (start), make_number (end), | |
596 | Qcomposition, prop, string); | |
597 | } | |
598 | ||
599 | \f | |
600 | /* Emacs Lisp APIs. */ | |
601 | ||
602 | DEFUN ("compose-region-internal", Fcompose_region_internal, | |
603 | Scompose_region_internal, 2, 4, 0, | |
335c5470 PJ |
604 | doc: /* Internal use only. |
605 | ||
606 | Compose text in the region between START and END. | |
607 | Optional 3rd and 4th arguments are COMPONENTS and MODIFICATION-FUNC | |
9baf7418 | 608 | for the composition. See `compose-region' for more detail. */) |
d562f8ab JB |
609 | (start, end, components, modification_func) |
610 | Lisp_Object start, end, components, modification_func; | |
ca4c9455 KH |
611 | { |
612 | validate_region (&start, &end); | |
613 | if (!NILP (components) | |
614 | && !INTEGERP (components) | |
615 | && !CONSP (components) | |
616 | && !STRINGP (components)) | |
b7826503 | 617 | CHECK_VECTOR (components); |
ca4c9455 | 618 | |
d562f8ab | 619 | compose_text (XINT (start), XINT (end), components, modification_func, Qnil); |
ca4c9455 KH |
620 | return Qnil; |
621 | } | |
622 | ||
623 | DEFUN ("compose-string-internal", Fcompose_string_internal, | |
624 | Scompose_string_internal, 3, 5, 0, | |
335c5470 PJ |
625 | doc: /* Internal use only. |
626 | ||
627 | Compose text between indices START and END of STRING. | |
628 | Optional 4th and 5th arguments are COMPONENTS and MODIFICATION-FUNC | |
9baf7418 | 629 | for the composition. See `compose-string' for more detail. */) |
d562f8ab JB |
630 | (string, start, end, components, modification_func) |
631 | Lisp_Object string, start, end, components, modification_func; | |
ca4c9455 | 632 | { |
b7826503 PJ |
633 | CHECK_STRING (string); |
634 | CHECK_NUMBER (start); | |
635 | CHECK_NUMBER (end); | |
ca4c9455 KH |
636 | |
637 | if (XINT (start) < 0 || | |
638 | XINT (start) > XINT (end) | |
d5db4077 | 639 | || XINT (end) > SCHARS (string)) |
ca4c9455 KH |
640 | args_out_of_range (start, end); |
641 | ||
d562f8ab | 642 | compose_text (XINT (start), XINT (end), components, modification_func, string); |
ca4c9455 KH |
643 | return string; |
644 | } | |
645 | ||
646 | DEFUN ("find-composition-internal", Ffind_composition_internal, | |
177c0ea7 | 647 | Sfind_composition_internal, 4, 4, 0, |
335c5470 PJ |
648 | doc: /* Internal use only. |
649 | ||
650 | Return information about composition at or nearest to position POS. | |
651 | See `find-composition' for more detail. */) | |
652 | (pos, limit, string, detail_p) | |
ca4c9455 KH |
653 | Lisp_Object pos, limit, string, detail_p; |
654 | { | |
655 | Lisp_Object prop, tail; | |
656 | int start, end; | |
657 | int id; | |
658 | ||
b7826503 | 659 | CHECK_NUMBER_COERCE_MARKER (pos); |
ca4c9455 KH |
660 | start = XINT (pos); |
661 | if (!NILP (limit)) | |
662 | { | |
b7826503 | 663 | CHECK_NUMBER_COERCE_MARKER (limit); |
ca4c9455 KH |
664 | end = XINT (limit); |
665 | } | |
666 | else | |
667 | end = -1; | |
177c0ea7 | 668 | |
ca4c9455 | 669 | if (!NILP (string)) |
e3b3e327 | 670 | { |
b7826503 | 671 | CHECK_STRING (string); |
d5db4077 | 672 | if (XINT (pos) < 0 || XINT (pos) > SCHARS (string)) |
e3b3e327 GM |
673 | args_out_of_range (string, pos); |
674 | } | |
675 | else | |
676 | { | |
fa9090b8 | 677 | if (XINT (pos) < BEGV || XINT (pos) > ZV) |
e3b3e327 GM |
678 | args_out_of_range (Fcurrent_buffer (), pos); |
679 | } | |
ca4c9455 KH |
680 | |
681 | if (!find_composition (start, end, &start, &end, &prop, string)) | |
682 | return Qnil; | |
683 | if (!COMPOSITION_VALID_P (start, end, prop)) | |
684 | return Fcons (make_number (start), Fcons (make_number (end), | |
685 | Fcons (Qnil, Qnil))); | |
686 | if (NILP (detail_p)) | |
687 | return Fcons (make_number (start), Fcons (make_number (end), | |
688 | Fcons (Qt, Qnil))); | |
689 | ||
690 | if (COMPOSITION_REGISTERD_P (prop)) | |
691 | id = COMPOSITION_ID (prop); | |
692 | else | |
693 | { | |
694 | int start_byte = (NILP (string) | |
695 | ? CHAR_TO_BYTE (start) | |
696 | : string_char_to_byte (string, start)); | |
697 | id = get_composition_id (start, start_byte, end - start, prop, string); | |
698 | } | |
699 | ||
700 | if (id >= 0) | |
701 | { | |
702 | Lisp_Object components, relative_p, mod_func; | |
703 | enum composition_method method = COMPOSITION_METHOD (prop); | |
704 | int width = composition_table[id]->width; | |
705 | ||
706 | components = Fcopy_sequence (COMPOSITION_COMPONENTS (prop)); | |
707 | relative_p = (method == COMPOSITION_WITH_RULE_ALTCHARS | |
708 | ? Qnil : Qt); | |
709 | mod_func = COMPOSITION_MODIFICATION_FUNC (prop); | |
710 | tail = Fcons (components, | |
711 | Fcons (relative_p, | |
712 | Fcons (mod_func, | |
713 | Fcons (make_number (width), Qnil)))); | |
714 | } | |
715 | else | |
716 | tail = Qnil; | |
717 | ||
718 | return Fcons (make_number (start), Fcons (make_number (end), tail)); | |
719 | } | |
720 | ||
721 | \f | |
722 | void | |
723 | syms_of_composite () | |
724 | { | |
725 | Qcomposition = intern ("composition"); | |
726 | staticpro (&Qcomposition); | |
727 | ||
728 | /* Make a hash table for composition. */ | |
729 | { | |
09654086 | 730 | Lisp_Object args[6]; |
ca4c9455 | 731 | extern Lisp_Object QCsize; |
177c0ea7 | 732 | |
ca4c9455 KH |
733 | args[0] = QCtest; |
734 | args[1] = Qequal; | |
c7211bc5 KH |
735 | /* We used to make the hash table weak so that unreferenced |
736 | compostions can be garbage-collected. But, usually once | |
737 | created compositions are repeatedly used in an Emacs session, | |
738 | and thus it's not worth to save memory in such a way. So, we | |
739 | make the table not weak. */ | |
ca4c9455 | 740 | args[2] = QCweakness; |
1ace9906 | 741 | args[3] = Qnil; |
ca4c9455 KH |
742 | args[4] = QCsize; |
743 | args[5] = make_number (311); | |
09654086 | 744 | composition_hash_table = Fmake_hash_table (6, args); |
ca4c9455 KH |
745 | staticpro (&composition_hash_table); |
746 | } | |
747 | ||
748 | /* Text property `composition' should be nonsticky by default. */ | |
749 | Vtext_property_default_nonsticky | |
750 | = Fcons (Fcons (Qcomposition, Qt), Vtext_property_default_nonsticky); | |
751 | ||
752 | DEFVAR_LISP ("compose-chars-after-function", &Vcompose_chars_after_function, | |
335c5470 PJ |
753 | doc: /* Function to adjust composition of buffer text. |
754 | ||
755 | The function is called with three arguments FROM, TO, and OBJECT. | |
756 | FROM and TO specify the range of text of which composition should be | |
757 | adjusted. OBJECT, if non-nil, is a string that contains the text. | |
758 | ||
759 | This function is called after a text with `composition' property is | |
760 | inserted or deleted to keep `composition' property of buffer text | |
761 | valid. | |
762 | ||
763 | The default value is the function `compose-chars-after'. */); | |
ca4c9455 KH |
764 | Vcompose_chars_after_function = intern ("compose-chars-after"); |
765 | ||
40add26d KH |
766 | Qcomposition_function_table = intern ("composition-function-table"); |
767 | staticpro (&Qcomposition_function_table); | |
768 | ||
769 | /* Intern this now in case it isn't already done. | |
770 | Setting this variable twice is harmless. | |
771 | But don't staticpro it here--that is done in alloc.c. */ | |
772 | Qchar_table_extra_slots = intern ("char-table-extra-slots"); | |
773 | ||
774 | Fput (Qcomposition_function_table, Qchar_table_extra_slots, make_number (0)); | |
775 | ||
776 | DEFVAR_LISP ("composition-function-table", &Vcomposition_function_table, | |
335c5470 PJ |
777 | doc: /* Char table of patterns and functions to make a composition. |
778 | ||
779 | Each element is nil or an alist of PATTERNs vs FUNCs, where PATTERNs | |
780 | are regular expressions and FUNCs are functions. FUNC is responsible | |
781 | for composing text matching the corresponding PATTERN. FUNC is called | |
782 | with three arguments FROM, TO, and PATTERN. See the function | |
783 | `compose-chars-after' for more detail. | |
784 | ||
785 | This table is looked up by the first character of a composition when | |
786 | the composition gets invalid after a change in a buffer. */); | |
40add26d KH |
787 | Vcomposition_function_table |
788 | = Fmake_char_table (Qcomposition_function_table, Qnil); | |
789 | ||
ca4c9455 KH |
790 | defsubr (&Scompose_region_internal); |
791 | defsubr (&Scompose_string_internal); | |
792 | defsubr (&Sfind_composition_internal); | |
793 | } | |
ab5796a9 MB |
794 | |
795 | /* arch-tag: 79cefaf8-ca48-4eed-97e5-d5afb290d272 | |
796 | (do not change this comment) */ |