Commit | Line | Data |
---|---|---|
ca1d1d23 | 1 | /* String search routines for GNU Emacs. |
3a22ee35 | 2 | Copyright (C) 1985, 1986, 1987, 1993, 1994 Free Software Foundation, Inc. |
ca1d1d23 JB |
3 | |
4 | This file is part of GNU Emacs. | |
5 | ||
6 | GNU Emacs is free software; you can redistribute it and/or modify | |
7 | it under the terms of the GNU General Public License as published by | |
8 | the Free Software Foundation; either version 1, or (at your option) | |
9 | any later version. | |
10 | ||
11 | GNU Emacs is distributed in the hope that it will be useful, | |
12 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
14 | GNU General Public License for more details. | |
15 | ||
16 | You should have received a copy of the GNU General Public License | |
17 | along with GNU Emacs; see the file COPYING. If not, write to | |
18 | the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ | |
19 | ||
20 | ||
18160b98 | 21 | #include <config.h> |
ca1d1d23 JB |
22 | #include "lisp.h" |
23 | #include "syntax.h" | |
24 | #include "buffer.h" | |
25 | #include "commands.h" | |
9ac0d9e0 | 26 | #include "blockinput.h" |
4746118a | 27 | |
ca1d1d23 JB |
28 | #include <sys/types.h> |
29 | #include "regex.h" | |
30 | ||
31 | #define max(a, b) ((a) > (b) ? (a) : (b)) | |
32 | #define min(a, b) ((a) < (b) ? (a) : (b)) | |
33 | ||
34 | /* We compile regexps into this buffer and then use it for searching. */ | |
35 | ||
36 | struct re_pattern_buffer searchbuf; | |
37 | ||
38 | char search_fastmap[0400]; | |
39 | ||
40 | /* Last regexp we compiled */ | |
41 | ||
42 | Lisp_Object last_regexp; | |
43 | ||
4746118a JB |
44 | /* Every call to re_match, etc., must pass &search_regs as the regs |
45 | argument unless you can show it is unnecessary (i.e., if re_match | |
46 | is certainly going to be called again before region-around-match | |
47 | can be called). | |
48 | ||
49 | Since the registers are now dynamically allocated, we need to make | |
50 | sure not to refer to the Nth register before checking that it has | |
1113d9db JB |
51 | been allocated by checking search_regs.num_regs. |
52 | ||
53 | The regex code keeps track of whether it has allocated the search | |
54 | buffer using bits in searchbuf. This means that whenever you | |
55 | compile a new pattern, it completely forgets whether it has | |
56 | allocated any registers, and will allocate new registers the next | |
57 | time you call a searching or matching function. Therefore, we need | |
58 | to call re_set_registers after compiling a new pattern or after | |
59 | setting the match registers, so that the regex functions will be | |
60 | able to free or re-allocate it properly. */ | |
ca1d1d23 JB |
61 | static struct re_registers search_regs; |
62 | ||
daa37602 JB |
63 | /* The buffer in which the last search was performed, or |
64 | Qt if the last search was done in a string; | |
65 | Qnil if no searching has been done yet. */ | |
66 | static Lisp_Object last_thing_searched; | |
ca1d1d23 JB |
67 | |
68 | /* error condition signalled when regexp compile_pattern fails */ | |
69 | ||
70 | Lisp_Object Qinvalid_regexp; | |
71 | ||
ca325161 RS |
72 | static void set_search_regs (); |
73 | ||
ca1d1d23 JB |
74 | static void |
75 | matcher_overflow () | |
76 | { | |
77 | error ("Stack overflow in regexp matcher"); | |
78 | } | |
79 | ||
80 | #ifdef __STDC__ | |
81 | #define CONST const | |
82 | #else | |
83 | #define CONST | |
84 | #endif | |
85 | ||
86 | /* Compile a regexp and signal a Lisp error if anything goes wrong. */ | |
87 | ||
1113d9db | 88 | compile_pattern (pattern, bufp, regp, translate) |
ca1d1d23 JB |
89 | Lisp_Object pattern; |
90 | struct re_pattern_buffer *bufp; | |
1113d9db | 91 | struct re_registers *regp; |
ca1d1d23 JB |
92 | char *translate; |
93 | { | |
94 | CONST char *val; | |
95 | Lisp_Object dummy; | |
96 | ||
97 | if (EQ (pattern, last_regexp) | |
98 | && translate == bufp->translate) | |
99 | return; | |
1113d9db | 100 | |
ca1d1d23 JB |
101 | last_regexp = Qnil; |
102 | bufp->translate = translate; | |
9ac0d9e0 | 103 | BLOCK_INPUT; |
b90d9e80 RS |
104 | val = (CONST char *) re_compile_pattern ((char *) XSTRING (pattern)->data, |
105 | XSTRING (pattern)->size, bufp); | |
9ac0d9e0 | 106 | UNBLOCK_INPUT; |
ca1d1d23 JB |
107 | if (val) |
108 | { | |
109 | dummy = build_string (val); | |
110 | while (1) | |
111 | Fsignal (Qinvalid_regexp, Fcons (dummy, Qnil)); | |
112 | } | |
1113d9db | 113 | |
ca1d1d23 | 114 | last_regexp = pattern; |
1113d9db JB |
115 | |
116 | /* Advise the searching functions about the space we have allocated | |
117 | for register data. */ | |
9ac0d9e0 | 118 | BLOCK_INPUT; |
ebb9e16f JB |
119 | if (regp) |
120 | re_set_registers (bufp, regp, regp->num_regs, regp->start, regp->end); | |
9ac0d9e0 | 121 | UNBLOCK_INPUT; |
1113d9db | 122 | |
ca1d1d23 JB |
123 | return; |
124 | } | |
125 | ||
126 | /* Error condition used for failing searches */ | |
127 | Lisp_Object Qsearch_failed; | |
128 | ||
129 | Lisp_Object | |
130 | signal_failure (arg) | |
131 | Lisp_Object arg; | |
132 | { | |
133 | Fsignal (Qsearch_failed, Fcons (arg, Qnil)); | |
134 | return Qnil; | |
135 | } | |
136 | \f | |
137 | DEFUN ("looking-at", Flooking_at, Slooking_at, 1, 1, 0, | |
e065a56e JB |
138 | "Return t if text after point matches regular expression PAT.\n\ |
139 | This function modifies the match data that `match-beginning',\n\ | |
140 | `match-end' and `match-data' access; save and restore the match\n\ | |
fe99283d | 141 | data if you want to preserve them.") |
ca1d1d23 JB |
142 | (string) |
143 | Lisp_Object string; | |
144 | { | |
145 | Lisp_Object val; | |
146 | unsigned char *p1, *p2; | |
147 | int s1, s2; | |
148 | register int i; | |
149 | ||
150 | CHECK_STRING (string, 0); | |
1113d9db | 151 | compile_pattern (string, &searchbuf, &search_regs, |
ca1d1d23 JB |
152 | !NILP (current_buffer->case_fold_search) ? DOWNCASE_TABLE : 0); |
153 | ||
154 | immediate_quit = 1; | |
155 | QUIT; /* Do a pending quit right away, to avoid paradoxical behavior */ | |
156 | ||
157 | /* Get pointers and sizes of the two strings | |
158 | that make up the visible portion of the buffer. */ | |
159 | ||
160 | p1 = BEGV_ADDR; | |
161 | s1 = GPT - BEGV; | |
162 | p2 = GAP_END_ADDR; | |
163 | s2 = ZV - GPT; | |
164 | if (s1 < 0) | |
165 | { | |
166 | p2 = p1; | |
167 | s2 = ZV - BEGV; | |
168 | s1 = 0; | |
169 | } | |
170 | if (s2 < 0) | |
171 | { | |
172 | s1 = ZV - BEGV; | |
173 | s2 = 0; | |
174 | } | |
175 | ||
176 | i = re_match_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2, | |
177 | point - BEGV, &search_regs, | |
178 | ZV - BEGV); | |
179 | if (i == -2) | |
180 | matcher_overflow (); | |
181 | ||
182 | val = (0 <= i ? Qt : Qnil); | |
4746118a | 183 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
184 | if (search_regs.start[i] >= 0) |
185 | { | |
186 | search_regs.start[i] += BEGV; | |
187 | search_regs.end[i] += BEGV; | |
188 | } | |
daa37602 | 189 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
190 | immediate_quit = 0; |
191 | return val; | |
192 | } | |
193 | ||
194 | DEFUN ("string-match", Fstring_match, Sstring_match, 2, 3, 0, | |
195 | "Return index of start of first match for REGEXP in STRING, or nil.\n\ | |
196 | If third arg START is non-nil, start search at that index in STRING.\n\ | |
197 | For index of first char beyond the match, do (match-end 0).\n\ | |
198 | `match-end' and `match-beginning' also give indices of substrings\n\ | |
199 | matched by parenthesis constructs in the pattern.") | |
200 | (regexp, string, start) | |
201 | Lisp_Object regexp, string, start; | |
202 | { | |
203 | int val; | |
204 | int s; | |
205 | ||
206 | CHECK_STRING (regexp, 0); | |
207 | CHECK_STRING (string, 1); | |
208 | ||
209 | if (NILP (start)) | |
210 | s = 0; | |
211 | else | |
212 | { | |
213 | int len = XSTRING (string)->size; | |
214 | ||
215 | CHECK_NUMBER (start, 2); | |
216 | s = XINT (start); | |
217 | if (s < 0 && -s <= len) | |
218 | s = len - s; | |
219 | else if (0 > s || s > len) | |
220 | args_out_of_range (string, start); | |
221 | } | |
222 | ||
1113d9db | 223 | compile_pattern (regexp, &searchbuf, &search_regs, |
ca1d1d23 JB |
224 | !NILP (current_buffer->case_fold_search) ? DOWNCASE_TABLE : 0); |
225 | immediate_quit = 1; | |
226 | val = re_search (&searchbuf, (char *) XSTRING (string)->data, | |
227 | XSTRING (string)->size, s, XSTRING (string)->size - s, | |
228 | &search_regs); | |
229 | immediate_quit = 0; | |
daa37602 | 230 | last_thing_searched = Qt; |
ca1d1d23 JB |
231 | if (val == -2) |
232 | matcher_overflow (); | |
233 | if (val < 0) return Qnil; | |
234 | return make_number (val); | |
235 | } | |
e59a8453 RS |
236 | |
237 | /* Match REGEXP against STRING, searching all of STRING, | |
238 | and return the index of the match, or negative on failure. | |
239 | This does not clobber the match data. */ | |
240 | ||
241 | int | |
242 | fast_string_match (regexp, string) | |
243 | Lisp_Object regexp, string; | |
244 | { | |
245 | int val; | |
246 | ||
247 | compile_pattern (regexp, &searchbuf, 0, 0); | |
248 | immediate_quit = 1; | |
249 | val = re_search (&searchbuf, (char *) XSTRING (string)->data, | |
250 | XSTRING (string)->size, 0, XSTRING (string)->size, | |
251 | 0); | |
252 | immediate_quit = 0; | |
253 | return val; | |
254 | } | |
ca1d1d23 | 255 | \f |
ffd56f97 JB |
256 | /* Search for COUNT instances of the character TARGET, starting at START. |
257 | If COUNT is negative, search backwards. | |
258 | ||
259 | If we find COUNT instances, set *SHORTAGE to zero, and return the | |
5bfe95c9 RS |
260 | position after the COUNTth match. Note that for reverse motion |
261 | this is not the same as the usual convention for Emacs motion commands. | |
ffd56f97 JB |
262 | |
263 | If we don't find COUNT instances before reaching the end of the | |
264 | buffer (or the beginning, if scanning backwards), set *SHORTAGE to | |
265 | the number of TARGETs left unfound, and return the end of the | |
087a5f81 | 266 | buffer we bumped up against. |
ffd56f97 | 267 | |
087a5f81 RS |
268 | If ALLOW_QUIT is non-zero, set immediate_quit. That's good to do |
269 | except when inside redisplay. */ | |
270 | ||
271 | scan_buffer (target, start, count, shortage, allow_quit) | |
ffd56f97 JB |
272 | int *shortage, start; |
273 | register int count, target; | |
087a5f81 | 274 | int allow_quit; |
ca1d1d23 | 275 | { |
ffd56f97 JB |
276 | int limit = ((count > 0) ? ZV - 1 : BEGV); |
277 | int direction = ((count > 0) ? 1 : -1); | |
278 | ||
279 | register unsigned char *cursor; | |
ca1d1d23 | 280 | unsigned char *base; |
ffd56f97 JB |
281 | |
282 | register int ceiling; | |
283 | register unsigned char *ceiling_addr; | |
ca1d1d23 JB |
284 | |
285 | if (shortage != 0) | |
286 | *shortage = 0; | |
287 | ||
087a5f81 | 288 | immediate_quit = allow_quit; |
ca1d1d23 | 289 | |
ffd56f97 JB |
290 | if (count > 0) |
291 | while (start != limit + 1) | |
ca1d1d23 | 292 | { |
ffd56f97 JB |
293 | ceiling = BUFFER_CEILING_OF (start); |
294 | ceiling = min (limit, ceiling); | |
295 | ceiling_addr = &FETCH_CHAR (ceiling) + 1; | |
296 | base = (cursor = &FETCH_CHAR (start)); | |
ca1d1d23 JB |
297 | while (1) |
298 | { | |
ffd56f97 | 299 | while (*cursor != target && ++cursor != ceiling_addr) |
ca1d1d23 | 300 | ; |
ffd56f97 | 301 | if (cursor != ceiling_addr) |
ca1d1d23 | 302 | { |
ffd56f97 | 303 | if (--count == 0) |
ca1d1d23 JB |
304 | { |
305 | immediate_quit = 0; | |
ffd56f97 | 306 | return (start + cursor - base + 1); |
ca1d1d23 JB |
307 | } |
308 | else | |
ffd56f97 | 309 | if (++cursor == ceiling_addr) |
ca1d1d23 JB |
310 | break; |
311 | } | |
312 | else | |
313 | break; | |
314 | } | |
ffd56f97 | 315 | start += cursor - base; |
ca1d1d23 JB |
316 | } |
317 | else | |
318 | { | |
ffd56f97 JB |
319 | start--; /* first character we scan */ |
320 | while (start > limit - 1) | |
321 | { /* we WILL scan under start */ | |
322 | ceiling = BUFFER_FLOOR_OF (start); | |
323 | ceiling = max (limit, ceiling); | |
324 | ceiling_addr = &FETCH_CHAR (ceiling) - 1; | |
325 | base = (cursor = &FETCH_CHAR (start)); | |
ca1d1d23 JB |
326 | cursor++; |
327 | while (1) | |
328 | { | |
ffd56f97 | 329 | while (--cursor != ceiling_addr && *cursor != target) |
ca1d1d23 | 330 | ; |
ffd56f97 | 331 | if (cursor != ceiling_addr) |
ca1d1d23 | 332 | { |
ffd56f97 | 333 | if (++count == 0) |
ca1d1d23 JB |
334 | { |
335 | immediate_quit = 0; | |
ffd56f97 | 336 | return (start + cursor - base + 1); |
ca1d1d23 JB |
337 | } |
338 | } | |
339 | else | |
340 | break; | |
341 | } | |
ffd56f97 | 342 | start += cursor - base; |
ca1d1d23 JB |
343 | } |
344 | } | |
345 | immediate_quit = 0; | |
346 | if (shortage != 0) | |
ffd56f97 JB |
347 | *shortage = count * direction; |
348 | return (start + ((direction == 1 ? 0 : 1))); | |
ca1d1d23 JB |
349 | } |
350 | ||
351 | int | |
352 | find_next_newline (from, cnt) | |
353 | register int from, cnt; | |
354 | { | |
087a5f81 | 355 | return scan_buffer ('\n', from, cnt, (int *) 0, 1); |
ca1d1d23 JB |
356 | } |
357 | \f | |
c1dc99a1 JB |
358 | Lisp_Object skip_chars (); |
359 | ||
ca1d1d23 | 360 | DEFUN ("skip-chars-forward", Fskip_chars_forward, Sskip_chars_forward, 1, 2, 0, |
3acb9a69 RS |
361 | "Move point forward, stopping before a char not in STRING, or at pos LIM.\n\ |
362 | STRING is like the inside of a `[...]' in a regular expression\n\ | |
ca1d1d23 JB |
363 | except that `]' is never special and `\\' quotes `^', `-' or `\\'.\n\ |
364 | Thus, with arg \"a-zA-Z\", this skips letters stopping before first nonletter.\n\ | |
c1dc99a1 JB |
365 | With arg \"^a-zA-Z\", skips nonletters stopping before first letter.\n\ |
366 | Returns the distance traveled, either zero or positive.") | |
ca1d1d23 JB |
367 | (string, lim) |
368 | Lisp_Object string, lim; | |
369 | { | |
17431c60 | 370 | return skip_chars (1, 0, string, lim); |
ca1d1d23 JB |
371 | } |
372 | ||
373 | DEFUN ("skip-chars-backward", Fskip_chars_backward, Sskip_chars_backward, 1, 2, 0, | |
3acb9a69 | 374 | "Move point backward, stopping after a char not in STRING, or at pos LIM.\n\ |
c1dc99a1 JB |
375 | See `skip-chars-forward' for details.\n\ |
376 | Returns the distance traveled, either zero or negative.") | |
ca1d1d23 JB |
377 | (string, lim) |
378 | Lisp_Object string, lim; | |
379 | { | |
17431c60 RS |
380 | return skip_chars (0, 0, string, lim); |
381 | } | |
382 | ||
383 | DEFUN ("skip-syntax-forward", Fskip_syntax_forward, Sskip_syntax_forward, 1, 2, 0, | |
384 | "Move point forward across chars in specified syntax classes.\n\ | |
385 | SYNTAX is a string of syntax code characters.\n\ | |
386 | Stop before a char whose syntax is not in SYNTAX, or at position LIM.\n\ | |
387 | If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.\n\ | |
388 | This function returns the distance traveled, either zero or positive.") | |
389 | (syntax, lim) | |
390 | Lisp_Object syntax, lim; | |
391 | { | |
392 | return skip_chars (1, 1, syntax, lim); | |
393 | } | |
394 | ||
395 | DEFUN ("skip-syntax-backward", Fskip_syntax_backward, Sskip_syntax_backward, 1, 2, 0, | |
396 | "Move point backward across chars in specified syntax classes.\n\ | |
397 | SYNTAX is a string of syntax code characters.\n\ | |
398 | Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.\n\ | |
399 | If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.\n\ | |
400 | This function returns the distance traveled, either zero or negative.") | |
401 | (syntax, lim) | |
402 | Lisp_Object syntax, lim; | |
403 | { | |
404 | return skip_chars (0, 1, syntax, lim); | |
ca1d1d23 JB |
405 | } |
406 | ||
c1dc99a1 | 407 | Lisp_Object |
17431c60 RS |
408 | skip_chars (forwardp, syntaxp, string, lim) |
409 | int forwardp, syntaxp; | |
ca1d1d23 JB |
410 | Lisp_Object string, lim; |
411 | { | |
412 | register unsigned char *p, *pend; | |
413 | register unsigned char c; | |
414 | unsigned char fastmap[0400]; | |
415 | int negate = 0; | |
416 | register int i; | |
417 | ||
418 | CHECK_STRING (string, 0); | |
419 | ||
420 | if (NILP (lim)) | |
421 | XSET (lim, Lisp_Int, forwardp ? ZV : BEGV); | |
422 | else | |
423 | CHECK_NUMBER_COERCE_MARKER (lim, 1); | |
424 | ||
ca1d1d23 | 425 | /* In any case, don't allow scan outside bounds of buffer. */ |
c5241910 RS |
426 | /* jla turned this off, for no known reason. |
427 | bfox turned the ZV part on, and rms turned the | |
428 | BEGV part back on. */ | |
429 | if (XINT (lim) > ZV) | |
ca1d1d23 | 430 | XFASTINT (lim) = ZV; |
c5241910 | 431 | if (XINT (lim) < BEGV) |
ca1d1d23 | 432 | XFASTINT (lim) = BEGV; |
ca1d1d23 JB |
433 | |
434 | p = XSTRING (string)->data; | |
435 | pend = p + XSTRING (string)->size; | |
436 | bzero (fastmap, sizeof fastmap); | |
437 | ||
438 | if (p != pend && *p == '^') | |
439 | { | |
440 | negate = 1; p++; | |
441 | } | |
442 | ||
17431c60 RS |
443 | /* Find the characters specified and set their elements of fastmap. |
444 | If syntaxp, each character counts as itself. | |
445 | Otherwise, handle backslashes and ranges specially */ | |
ca1d1d23 JB |
446 | |
447 | while (p != pend) | |
448 | { | |
449 | c = *p++; | |
17431c60 RS |
450 | if (syntaxp) |
451 | fastmap[c] = 1; | |
452 | else | |
ca1d1d23 | 453 | { |
17431c60 | 454 | if (c == '\\') |
ca1d1d23 | 455 | { |
17431c60 RS |
456 | if (p == pend) break; |
457 | c = *p++; | |
458 | } | |
459 | if (p != pend && *p == '-') | |
460 | { | |
461 | p++; | |
462 | if (p == pend) break; | |
463 | while (c <= *p) | |
464 | { | |
465 | fastmap[c] = 1; | |
466 | c++; | |
467 | } | |
468 | p++; | |
ca1d1d23 | 469 | } |
17431c60 RS |
470 | else |
471 | fastmap[c] = 1; | |
ca1d1d23 | 472 | } |
ca1d1d23 JB |
473 | } |
474 | ||
9239c6c1 RS |
475 | if (syntaxp && fastmap['-'] != 0) |
476 | fastmap[' '] = 1; | |
477 | ||
ca1d1d23 JB |
478 | /* If ^ was the first character, complement the fastmap. */ |
479 | ||
480 | if (negate) | |
481 | for (i = 0; i < sizeof fastmap; i++) | |
482 | fastmap[i] ^= 1; | |
483 | ||
c1dc99a1 JB |
484 | { |
485 | int start_point = point; | |
486 | ||
487 | immediate_quit = 1; | |
17431c60 | 488 | if (syntaxp) |
c1dc99a1 | 489 | { |
17431c60 RS |
490 | |
491 | if (forwardp) | |
492 | { | |
493 | while (point < XINT (lim) | |
494 | && fastmap[(unsigned char) syntax_code_spec[(int) SYNTAX (FETCH_CHAR (point))]]) | |
495 | SET_PT (point + 1); | |
496 | } | |
497 | else | |
498 | { | |
499 | while (point > XINT (lim) | |
500 | && fastmap[(unsigned char) syntax_code_spec[(int) SYNTAX (FETCH_CHAR (point - 1))]]) | |
501 | SET_PT (point - 1); | |
502 | } | |
c1dc99a1 JB |
503 | } |
504 | else | |
505 | { | |
17431c60 RS |
506 | if (forwardp) |
507 | { | |
508 | while (point < XINT (lim) && fastmap[FETCH_CHAR (point)]) | |
509 | SET_PT (point + 1); | |
510 | } | |
511 | else | |
512 | { | |
513 | while (point > XINT (lim) && fastmap[FETCH_CHAR (point - 1)]) | |
514 | SET_PT (point - 1); | |
515 | } | |
c1dc99a1 JB |
516 | } |
517 | immediate_quit = 0; | |
518 | ||
519 | return make_number (point - start_point); | |
520 | } | |
ca1d1d23 JB |
521 | } |
522 | \f | |
523 | /* Subroutines of Lisp buffer search functions. */ | |
524 | ||
525 | static Lisp_Object | |
526 | search_command (string, bound, noerror, count, direction, RE) | |
527 | Lisp_Object string, bound, noerror, count; | |
528 | int direction; | |
529 | int RE; | |
530 | { | |
531 | register int np; | |
532 | int lim; | |
533 | int n = direction; | |
534 | ||
535 | if (!NILP (count)) | |
536 | { | |
537 | CHECK_NUMBER (count, 3); | |
538 | n *= XINT (count); | |
539 | } | |
540 | ||
541 | CHECK_STRING (string, 0); | |
542 | if (NILP (bound)) | |
543 | lim = n > 0 ? ZV : BEGV; | |
544 | else | |
545 | { | |
546 | CHECK_NUMBER_COERCE_MARKER (bound, 1); | |
547 | lim = XINT (bound); | |
548 | if (n > 0 ? lim < point : lim > point) | |
549 | error ("Invalid search bound (wrong side of point)"); | |
550 | if (lim > ZV) | |
551 | lim = ZV; | |
552 | if (lim < BEGV) | |
553 | lim = BEGV; | |
554 | } | |
555 | ||
556 | np = search_buffer (string, point, lim, n, RE, | |
557 | (!NILP (current_buffer->case_fold_search) | |
558 | ? XSTRING (current_buffer->case_canon_table)->data : 0), | |
559 | (!NILP (current_buffer->case_fold_search) | |
560 | ? XSTRING (current_buffer->case_eqv_table)->data : 0)); | |
561 | if (np <= 0) | |
562 | { | |
563 | if (NILP (noerror)) | |
564 | return signal_failure (string); | |
565 | if (!EQ (noerror, Qt)) | |
566 | { | |
567 | if (lim < BEGV || lim > ZV) | |
568 | abort (); | |
a5f217b8 RS |
569 | SET_PT (lim); |
570 | return Qnil; | |
571 | #if 0 /* This would be clean, but maybe programs depend on | |
572 | a value of nil here. */ | |
481399bf | 573 | np = lim; |
a5f217b8 | 574 | #endif |
ca1d1d23 | 575 | } |
481399bf RS |
576 | else |
577 | return Qnil; | |
ca1d1d23 JB |
578 | } |
579 | ||
580 | if (np < BEGV || np > ZV) | |
581 | abort (); | |
582 | ||
583 | SET_PT (np); | |
584 | ||
585 | return make_number (np); | |
586 | } | |
587 | \f | |
ca325161 | 588 | /* Search for the n'th occurrence of STRING in the current buffer, |
ca1d1d23 JB |
589 | starting at position POS and stopping at position LIM, |
590 | treating PAT as a literal string if RE is false or as | |
591 | a regular expression if RE is true. | |
592 | ||
593 | If N is positive, searching is forward and LIM must be greater than POS. | |
594 | If N is negative, searching is backward and LIM must be less than POS. | |
595 | ||
596 | Returns -x if only N-x occurrences found (x > 0), | |
597 | or else the position at the beginning of the Nth occurrence | |
598 | (if searching backward) or the end (if searching forward). */ | |
599 | ||
600 | search_buffer (string, pos, lim, n, RE, trt, inverse_trt) | |
601 | Lisp_Object string; | |
602 | int pos; | |
603 | int lim; | |
604 | int n; | |
605 | int RE; | |
606 | register unsigned char *trt; | |
607 | register unsigned char *inverse_trt; | |
608 | { | |
609 | int len = XSTRING (string)->size; | |
610 | unsigned char *base_pat = XSTRING (string)->data; | |
611 | register int *BM_tab; | |
612 | int *BM_tab_base; | |
613 | register int direction = ((n > 0) ? 1 : -1); | |
614 | register int dirlen; | |
615 | int infinity, limit, k, stride_for_teases; | |
616 | register unsigned char *pat, *cursor, *p_limit; | |
617 | register int i, j; | |
618 | unsigned char *p1, *p2; | |
619 | int s1, s2; | |
620 | ||
621 | /* Null string is found at starting position. */ | |
3f57a499 | 622 | if (len == 0) |
ca325161 RS |
623 | { |
624 | set_search_regs (pos, 0); | |
625 | return pos; | |
626 | } | |
3f57a499 RS |
627 | |
628 | /* Searching 0 times means don't move. */ | |
629 | if (n == 0) | |
ca1d1d23 JB |
630 | return pos; |
631 | ||
632 | if (RE) | |
1113d9db | 633 | compile_pattern (string, &searchbuf, &search_regs, (char *) trt); |
ca1d1d23 JB |
634 | |
635 | if (RE /* Here we detect whether the */ | |
636 | /* generality of an RE search is */ | |
637 | /* really needed. */ | |
638 | /* first item is "exact match" */ | |
4746118a | 639 | && *(searchbuf.buffer) == (char) RE_EXACTN_VALUE |
ca1d1d23 JB |
640 | && searchbuf.buffer[1] + 2 == searchbuf.used) /*first is ONLY item */ |
641 | { | |
642 | RE = 0; /* can do straight (non RE) search */ | |
643 | pat = (base_pat = (unsigned char *) searchbuf.buffer + 2); | |
644 | /* trt already applied */ | |
645 | len = searchbuf.used - 2; | |
646 | } | |
647 | else if (!RE) | |
648 | { | |
649 | pat = (unsigned char *) alloca (len); | |
650 | ||
651 | for (i = len; i--;) /* Copy the pattern; apply trt */ | |
652 | *pat++ = (((int) trt) ? trt [*base_pat++] : *base_pat++); | |
653 | pat -= len; base_pat = pat; | |
654 | } | |
655 | ||
656 | if (RE) | |
657 | { | |
658 | immediate_quit = 1; /* Quit immediately if user types ^G, | |
659 | because letting this function finish | |
660 | can take too long. */ | |
661 | QUIT; /* Do a pending quit right away, | |
662 | to avoid paradoxical behavior */ | |
663 | /* Get pointers and sizes of the two strings | |
664 | that make up the visible portion of the buffer. */ | |
665 | ||
666 | p1 = BEGV_ADDR; | |
667 | s1 = GPT - BEGV; | |
668 | p2 = GAP_END_ADDR; | |
669 | s2 = ZV - GPT; | |
670 | if (s1 < 0) | |
671 | { | |
672 | p2 = p1; | |
673 | s2 = ZV - BEGV; | |
674 | s1 = 0; | |
675 | } | |
676 | if (s2 < 0) | |
677 | { | |
678 | s1 = ZV - BEGV; | |
679 | s2 = 0; | |
680 | } | |
681 | while (n < 0) | |
682 | { | |
42db823b | 683 | int val; |
42db823b RS |
684 | val = re_search_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2, |
685 | pos - BEGV, lim - pos, &search_regs, | |
686 | /* Don't allow match past current point */ | |
687 | pos - BEGV); | |
ca1d1d23 JB |
688 | if (val == -2) |
689 | matcher_overflow (); | |
690 | if (val >= 0) | |
691 | { | |
692 | j = BEGV; | |
4746118a | 693 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
694 | if (search_regs.start[i] >= 0) |
695 | { | |
696 | search_regs.start[i] += j; | |
697 | search_regs.end[i] += j; | |
698 | } | |
daa37602 | 699 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
700 | /* Set pos to the new position. */ |
701 | pos = search_regs.start[0]; | |
702 | } | |
703 | else | |
704 | { | |
705 | immediate_quit = 0; | |
706 | return (n); | |
707 | } | |
708 | n++; | |
709 | } | |
710 | while (n > 0) | |
711 | { | |
42db823b | 712 | int val; |
42db823b RS |
713 | val = re_search_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2, |
714 | pos - BEGV, lim - pos, &search_regs, | |
715 | lim - BEGV); | |
ca1d1d23 JB |
716 | if (val == -2) |
717 | matcher_overflow (); | |
718 | if (val >= 0) | |
719 | { | |
720 | j = BEGV; | |
4746118a | 721 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
722 | if (search_regs.start[i] >= 0) |
723 | { | |
724 | search_regs.start[i] += j; | |
725 | search_regs.end[i] += j; | |
726 | } | |
daa37602 | 727 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
728 | pos = search_regs.end[0]; |
729 | } | |
730 | else | |
731 | { | |
732 | immediate_quit = 0; | |
733 | return (0 - n); | |
734 | } | |
735 | n--; | |
736 | } | |
737 | immediate_quit = 0; | |
738 | return (pos); | |
739 | } | |
740 | else /* non-RE case */ | |
741 | { | |
742 | #ifdef C_ALLOCA | |
743 | int BM_tab_space[0400]; | |
744 | BM_tab = &BM_tab_space[0]; | |
745 | #else | |
746 | BM_tab = (int *) alloca (0400 * sizeof (int)); | |
747 | #endif | |
748 | /* The general approach is that we are going to maintain that we know */ | |
749 | /* the first (closest to the present position, in whatever direction */ | |
750 | /* we're searching) character that could possibly be the last */ | |
751 | /* (furthest from present position) character of a valid match. We */ | |
752 | /* advance the state of our knowledge by looking at that character */ | |
753 | /* and seeing whether it indeed matches the last character of the */ | |
754 | /* pattern. If it does, we take a closer look. If it does not, we */ | |
755 | /* move our pointer (to putative last characters) as far as is */ | |
756 | /* logically possible. This amount of movement, which I call a */ | |
757 | /* stride, will be the length of the pattern if the actual character */ | |
758 | /* appears nowhere in the pattern, otherwise it will be the distance */ | |
759 | /* from the last occurrence of that character to the end of the */ | |
760 | /* pattern. */ | |
761 | /* As a coding trick, an enormous stride is coded into the table for */ | |
762 | /* characters that match the last character. This allows use of only */ | |
763 | /* a single test, a test for having gone past the end of the */ | |
764 | /* permissible match region, to test for both possible matches (when */ | |
765 | /* the stride goes past the end immediately) and failure to */ | |
766 | /* match (where you get nudged past the end one stride at a time). */ | |
767 | ||
768 | /* Here we make a "mickey mouse" BM table. The stride of the search */ | |
769 | /* is determined only by the last character of the putative match. */ | |
770 | /* If that character does not match, we will stride the proper */ | |
771 | /* distance to propose a match that superimposes it on the last */ | |
772 | /* instance of a character that matches it (per trt), or misses */ | |
773 | /* it entirely if there is none. */ | |
774 | ||
775 | dirlen = len * direction; | |
776 | infinity = dirlen - (lim + pos + len + len) * direction; | |
777 | if (direction < 0) | |
778 | pat = (base_pat += len - 1); | |
779 | BM_tab_base = BM_tab; | |
780 | BM_tab += 0400; | |
781 | j = dirlen; /* to get it in a register */ | |
782 | /* A character that does not appear in the pattern induces a */ | |
783 | /* stride equal to the pattern length. */ | |
784 | while (BM_tab_base != BM_tab) | |
785 | { | |
786 | *--BM_tab = j; | |
787 | *--BM_tab = j; | |
788 | *--BM_tab = j; | |
789 | *--BM_tab = j; | |
790 | } | |
791 | i = 0; | |
792 | while (i != infinity) | |
793 | { | |
794 | j = pat[i]; i += direction; | |
795 | if (i == dirlen) i = infinity; | |
796 | if ((int) trt) | |
797 | { | |
798 | k = (j = trt[j]); | |
799 | if (i == infinity) | |
800 | stride_for_teases = BM_tab[j]; | |
801 | BM_tab[j] = dirlen - i; | |
802 | /* A translation table is accompanied by its inverse -- see */ | |
803 | /* comment following downcase_table for details */ | |
804 | while ((j = inverse_trt[j]) != k) | |
805 | BM_tab[j] = dirlen - i; | |
806 | } | |
807 | else | |
808 | { | |
809 | if (i == infinity) | |
810 | stride_for_teases = BM_tab[j]; | |
811 | BM_tab[j] = dirlen - i; | |
812 | } | |
813 | /* stride_for_teases tells how much to stride if we get a */ | |
814 | /* match on the far character but are subsequently */ | |
815 | /* disappointed, by recording what the stride would have been */ | |
816 | /* for that character if the last character had been */ | |
817 | /* different. */ | |
818 | } | |
819 | infinity = dirlen - infinity; | |
820 | pos += dirlen - ((direction > 0) ? direction : 0); | |
821 | /* loop invariant - pos points at where last char (first char if reverse) | |
822 | of pattern would align in a possible match. */ | |
823 | while (n != 0) | |
824 | { | |
b2c71fb4 KH |
825 | /* It's been reported that some (broken) compiler thinks that |
826 | Boolean expressions in an arithmetic context are unsigned. | |
827 | Using an explicit ?1:0 prevents this. */ | |
828 | if ((lim - pos - ((direction > 0) ? 1 : 0)) * direction < 0) | |
ca1d1d23 JB |
829 | return (n * (0 - direction)); |
830 | /* First we do the part we can by pointers (maybe nothing) */ | |
831 | QUIT; | |
832 | pat = base_pat; | |
833 | limit = pos - dirlen + direction; | |
834 | limit = ((direction > 0) | |
835 | ? BUFFER_CEILING_OF (limit) | |
836 | : BUFFER_FLOOR_OF (limit)); | |
837 | /* LIMIT is now the last (not beyond-last!) value | |
838 | POS can take on without hitting edge of buffer or the gap. */ | |
839 | limit = ((direction > 0) | |
840 | ? min (lim - 1, min (limit, pos + 20000)) | |
841 | : max (lim, max (limit, pos - 20000))); | |
842 | if ((limit - pos) * direction > 20) | |
843 | { | |
844 | p_limit = &FETCH_CHAR (limit); | |
845 | p2 = (cursor = &FETCH_CHAR (pos)); | |
846 | /* In this loop, pos + cursor - p2 is the surrogate for pos */ | |
847 | while (1) /* use one cursor setting as long as i can */ | |
848 | { | |
849 | if (direction > 0) /* worth duplicating */ | |
850 | { | |
851 | /* Use signed comparison if appropriate | |
852 | to make cursor+infinity sure to be > p_limit. | |
853 | Assuming that the buffer lies in a range of addresses | |
854 | that are all "positive" (as ints) or all "negative", | |
855 | either kind of comparison will work as long | |
856 | as we don't step by infinity. So pick the kind | |
857 | that works when we do step by infinity. */ | |
858 | if ((int) (p_limit + infinity) > (int) p_limit) | |
859 | while ((int) cursor <= (int) p_limit) | |
860 | cursor += BM_tab[*cursor]; | |
861 | else | |
862 | while ((unsigned int) cursor <= (unsigned int) p_limit) | |
863 | cursor += BM_tab[*cursor]; | |
864 | } | |
865 | else | |
866 | { | |
867 | if ((int) (p_limit + infinity) < (int) p_limit) | |
868 | while ((int) cursor >= (int) p_limit) | |
869 | cursor += BM_tab[*cursor]; | |
870 | else | |
871 | while ((unsigned int) cursor >= (unsigned int) p_limit) | |
872 | cursor += BM_tab[*cursor]; | |
873 | } | |
874 | /* If you are here, cursor is beyond the end of the searched region. */ | |
875 | /* This can happen if you match on the far character of the pattern, */ | |
876 | /* because the "stride" of that character is infinity, a number able */ | |
877 | /* to throw you well beyond the end of the search. It can also */ | |
878 | /* happen if you fail to match within the permitted region and would */ | |
879 | /* otherwise try a character beyond that region */ | |
880 | if ((cursor - p_limit) * direction <= len) | |
881 | break; /* a small overrun is genuine */ | |
882 | cursor -= infinity; /* large overrun = hit */ | |
883 | i = dirlen - direction; | |
884 | if ((int) trt) | |
885 | { | |
886 | while ((i -= direction) + direction != 0) | |
887 | if (pat[i] != trt[*(cursor -= direction)]) | |
888 | break; | |
889 | } | |
890 | else | |
891 | { | |
892 | while ((i -= direction) + direction != 0) | |
893 | if (pat[i] != *(cursor -= direction)) | |
894 | break; | |
895 | } | |
896 | cursor += dirlen - i - direction; /* fix cursor */ | |
897 | if (i + direction == 0) | |
898 | { | |
899 | cursor -= direction; | |
1113d9db | 900 | |
ca325161 RS |
901 | set_search_regs (pos + cursor - p2 + ((direction > 0) |
902 | ? 1 - len : 0), | |
903 | len); | |
904 | ||
ca1d1d23 JB |
905 | if ((n -= direction) != 0) |
906 | cursor += dirlen; /* to resume search */ | |
907 | else | |
908 | return ((direction > 0) | |
909 | ? search_regs.end[0] : search_regs.start[0]); | |
910 | } | |
911 | else | |
912 | cursor += stride_for_teases; /* <sigh> we lose - */ | |
913 | } | |
914 | pos += cursor - p2; | |
915 | } | |
916 | else | |
917 | /* Now we'll pick up a clump that has to be done the hard */ | |
918 | /* way because it covers a discontinuity */ | |
919 | { | |
920 | limit = ((direction > 0) | |
921 | ? BUFFER_CEILING_OF (pos - dirlen + 1) | |
922 | : BUFFER_FLOOR_OF (pos - dirlen - 1)); | |
923 | limit = ((direction > 0) | |
924 | ? min (limit + len, lim - 1) | |
925 | : max (limit - len, lim)); | |
926 | /* LIMIT is now the last value POS can have | |
927 | and still be valid for a possible match. */ | |
928 | while (1) | |
929 | { | |
930 | /* This loop can be coded for space rather than */ | |
931 | /* speed because it will usually run only once. */ | |
932 | /* (the reach is at most len + 21, and typically */ | |
933 | /* does not exceed len) */ | |
934 | while ((limit - pos) * direction >= 0) | |
935 | pos += BM_tab[FETCH_CHAR(pos)]; | |
936 | /* now run the same tests to distinguish going off the */ | |
eb8c3be9 | 937 | /* end, a match or a phony match. */ |
ca1d1d23 JB |
938 | if ((pos - limit) * direction <= len) |
939 | break; /* ran off the end */ | |
940 | /* Found what might be a match. | |
941 | Set POS back to last (first if reverse) char pos. */ | |
942 | pos -= infinity; | |
943 | i = dirlen - direction; | |
944 | while ((i -= direction) + direction != 0) | |
945 | { | |
946 | pos -= direction; | |
947 | if (pat[i] != (((int) trt) | |
948 | ? trt[FETCH_CHAR(pos)] | |
949 | : FETCH_CHAR (pos))) | |
950 | break; | |
951 | } | |
952 | /* Above loop has moved POS part or all the way | |
953 | back to the first char pos (last char pos if reverse). | |
954 | Set it once again at the last (first if reverse) char. */ | |
955 | pos += dirlen - i- direction; | |
956 | if (i + direction == 0) | |
957 | { | |
958 | pos -= direction; | |
1113d9db | 959 | |
ca325161 RS |
960 | set_search_regs (pos + ((direction > 0) ? 1 - len : 0), |
961 | len); | |
962 | ||
ca1d1d23 JB |
963 | if ((n -= direction) != 0) |
964 | pos += dirlen; /* to resume search */ | |
965 | else | |
966 | return ((direction > 0) | |
967 | ? search_regs.end[0] : search_regs.start[0]); | |
968 | } | |
969 | else | |
970 | pos += stride_for_teases; | |
971 | } | |
972 | } | |
973 | /* We have done one clump. Can we continue? */ | |
974 | if ((lim - pos) * direction < 0) | |
975 | return ((0 - n) * direction); | |
976 | } | |
977 | return pos; | |
978 | } | |
979 | } | |
ca325161 RS |
980 | |
981 | /* Record beginning BEG and end BEG + LEN | |
982 | for a match just found in the current buffer. */ | |
983 | ||
984 | static void | |
985 | set_search_regs (beg, len) | |
986 | int beg, len; | |
987 | { | |
988 | /* Make sure we have registers in which to store | |
989 | the match position. */ | |
990 | if (search_regs.num_regs == 0) | |
991 | { | |
992 | regoff_t *starts, *ends; | |
993 | ||
994 | starts = (regoff_t *) xmalloc (2 * sizeof (regoff_t)); | |
995 | ends = (regoff_t *) xmalloc (2 * sizeof (regoff_t)); | |
996 | BLOCK_INPUT; | |
997 | re_set_registers (&searchbuf, | |
998 | &search_regs, | |
999 | 2, starts, ends); | |
1000 | UNBLOCK_INPUT; | |
1001 | } | |
1002 | ||
1003 | search_regs.start[0] = beg; | |
1004 | search_regs.end[0] = beg + len; | |
1005 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); | |
1006 | } | |
ca1d1d23 JB |
1007 | \f |
1008 | /* Given a string of words separated by word delimiters, | |
1009 | compute a regexp that matches those exact words | |
1010 | separated by arbitrary punctuation. */ | |
1011 | ||
1012 | static Lisp_Object | |
1013 | wordify (string) | |
1014 | Lisp_Object string; | |
1015 | { | |
1016 | register unsigned char *p, *o; | |
1017 | register int i, len, punct_count = 0, word_count = 0; | |
1018 | Lisp_Object val; | |
1019 | ||
1020 | CHECK_STRING (string, 0); | |
1021 | p = XSTRING (string)->data; | |
1022 | len = XSTRING (string)->size; | |
1023 | ||
1024 | for (i = 0; i < len; i++) | |
1025 | if (SYNTAX (p[i]) != Sword) | |
1026 | { | |
1027 | punct_count++; | |
1028 | if (i > 0 && SYNTAX (p[i-1]) == Sword) word_count++; | |
1029 | } | |
1030 | if (SYNTAX (p[len-1]) == Sword) word_count++; | |
1031 | if (!word_count) return build_string (""); | |
1032 | ||
1033 | val = make_string (p, len - punct_count + 5 * (word_count - 1) + 4); | |
1034 | ||
1035 | o = XSTRING (val)->data; | |
1036 | *o++ = '\\'; | |
1037 | *o++ = 'b'; | |
1038 | ||
1039 | for (i = 0; i < len; i++) | |
1040 | if (SYNTAX (p[i]) == Sword) | |
1041 | *o++ = p[i]; | |
1042 | else if (i > 0 && SYNTAX (p[i-1]) == Sword && --word_count) | |
1043 | { | |
1044 | *o++ = '\\'; | |
1045 | *o++ = 'W'; | |
1046 | *o++ = '\\'; | |
1047 | *o++ = 'W'; | |
1048 | *o++ = '*'; | |
1049 | } | |
1050 | ||
1051 | *o++ = '\\'; | |
1052 | *o++ = 'b'; | |
1053 | ||
1054 | return val; | |
1055 | } | |
1056 | \f | |
1057 | DEFUN ("search-backward", Fsearch_backward, Ssearch_backward, 1, 4, | |
1058 | "sSearch backward: ", | |
1059 | "Search backward from point for STRING.\n\ | |
1060 | Set point to the beginning of the occurrence found, and return point.\n\ | |
1061 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1062 | The match found must not extend before that position.\n\ | |
1063 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1064 | If not nil and not t, position at limit of search and return nil.\n\ | |
1065 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1066 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
1067 | (string, bound, noerror, count) | |
1068 | Lisp_Object string, bound, noerror, count; | |
1069 | { | |
1070 | return search_command (string, bound, noerror, count, -1, 0); | |
1071 | } | |
1072 | ||
1073 | DEFUN ("search-forward", Fsearch_forward, Ssearch_forward, 1, 4, "sSearch: ", | |
1074 | "Search forward from point for STRING.\n\ | |
1075 | Set point to the end of the occurrence found, and return point.\n\ | |
1076 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1077 | The match found must not extend after that position. nil is equivalent\n\ | |
1078 | to (point-max).\n\ | |
1079 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1080 | If not nil and not t, move to limit of search and return nil.\n\ | |
1081 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1082 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
1083 | (string, bound, noerror, count) | |
1084 | Lisp_Object string, bound, noerror, count; | |
1085 | { | |
1086 | return search_command (string, bound, noerror, count, 1, 0); | |
1087 | } | |
1088 | ||
1089 | DEFUN ("word-search-backward", Fword_search_backward, Sword_search_backward, 1, 4, | |
1090 | "sWord search backward: ", | |
1091 | "Search backward from point for STRING, ignoring differences in punctuation.\n\ | |
1092 | Set point to the beginning of the occurrence found, and return point.\n\ | |
1093 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1094 | The match found must not extend before that position.\n\ | |
1095 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1096 | If not nil and not t, move to limit of search and return nil.\n\ | |
1097 | Optional fourth argument is repeat count--search for successive occurrences.") | |
1098 | (string, bound, noerror, count) | |
1099 | Lisp_Object string, bound, noerror, count; | |
1100 | { | |
1101 | return search_command (wordify (string), bound, noerror, count, -1, 1); | |
1102 | } | |
1103 | ||
1104 | DEFUN ("word-search-forward", Fword_search_forward, Sword_search_forward, 1, 4, | |
1105 | "sWord search: ", | |
1106 | "Search forward from point for STRING, ignoring differences in punctuation.\n\ | |
1107 | Set point to the end of the occurrence found, and return point.\n\ | |
1108 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1109 | The match found must not extend after that position.\n\ | |
1110 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1111 | If not nil and not t, move to limit of search and return nil.\n\ | |
1112 | Optional fourth argument is repeat count--search for successive occurrences.") | |
1113 | (string, bound, noerror, count) | |
1114 | Lisp_Object string, bound, noerror, count; | |
1115 | { | |
1116 | return search_command (wordify (string), bound, noerror, count, 1, 1); | |
1117 | } | |
1118 | ||
1119 | DEFUN ("re-search-backward", Fre_search_backward, Sre_search_backward, 1, 4, | |
1120 | "sRE search backward: ", | |
1121 | "Search backward from point for match for regular expression REGEXP.\n\ | |
1122 | Set point to the beginning of the match, and return point.\n\ | |
1123 | The match found is the one starting last in the buffer\n\ | |
19c0a730 | 1124 | and yet ending before the origin of the search.\n\ |
ca1d1d23 JB |
1125 | An optional second argument bounds the search; it is a buffer position.\n\ |
1126 | The match found must start at or after that position.\n\ | |
1127 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1128 | If not nil and not t, move to limit of search and return nil.\n\ | |
1129 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1130 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
19c0a730 KH |
1131 | (regexp, bound, noerror, count) |
1132 | Lisp_Object regexp, bound, noerror, count; | |
ca1d1d23 | 1133 | { |
19c0a730 | 1134 | return search_command (regexp, bound, noerror, count, -1, 1); |
ca1d1d23 JB |
1135 | } |
1136 | ||
1137 | DEFUN ("re-search-forward", Fre_search_forward, Sre_search_forward, 1, 4, | |
1138 | "sRE search: ", | |
1139 | "Search forward from point for regular expression REGEXP.\n\ | |
1140 | Set point to the end of the occurrence found, and return point.\n\ | |
1141 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1142 | The match found must not extend after that position.\n\ | |
1143 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1144 | If not nil and not t, move to limit of search and return nil.\n\ | |
1145 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1146 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
19c0a730 KH |
1147 | (regexp, bound, noerror, count) |
1148 | Lisp_Object regexp, bound, noerror, count; | |
ca1d1d23 | 1149 | { |
19c0a730 | 1150 | return search_command (regexp, bound, noerror, count, 1, 1); |
ca1d1d23 JB |
1151 | } |
1152 | \f | |
1153 | DEFUN ("replace-match", Freplace_match, Sreplace_match, 1, 3, 0, | |
1154 | "Replace text matched by last search with NEWTEXT.\n\ | |
1155 | If second arg FIXEDCASE is non-nil, do not alter case of replacement text.\n\ | |
5b9cf4b2 RS |
1156 | Otherwise maybe capitalize the whole text, or maybe just word initials,\n\ |
1157 | based on the replaced text.\n\ | |
1158 | If the replaced text has only capital letters\n\ | |
1159 | and has at least one multiletter word, convert NEWTEXT to all caps.\n\ | |
1160 | If the replaced text has at least one word starting with a capital letter,\n\ | |
1161 | then capitalize each word in NEWTEXT.\n\n\ | |
ca1d1d23 JB |
1162 | If third arg LITERAL is non-nil, insert NEWTEXT literally.\n\ |
1163 | Otherwise treat `\\' as special:\n\ | |
1164 | `\\&' in NEWTEXT means substitute original matched text.\n\ | |
1165 | `\\N' means substitute what matched the Nth `\\(...\\)'.\n\ | |
1166 | If Nth parens didn't match, substitute nothing.\n\ | |
1167 | `\\\\' means insert one `\\'.\n\ | |
1113d9db | 1168 | FIXEDCASE and LITERAL are optional arguments.\n\ |
ca1d1d23 | 1169 | Leaves point at end of replacement text.") |
16fdc568 BF |
1170 | (newtext, fixedcase, literal) |
1171 | Lisp_Object newtext, fixedcase, literal; | |
ca1d1d23 JB |
1172 | { |
1173 | enum { nochange, all_caps, cap_initial } case_action; | |
1174 | register int pos, last; | |
1175 | int some_multiletter_word; | |
97832bd0 | 1176 | int some_lowercase; |
c4d460ce | 1177 | int some_lowercase_initial; |
ca1d1d23 JB |
1178 | register int c, prevc; |
1179 | int inslen; | |
1180 | ||
16fdc568 | 1181 | CHECK_STRING (newtext, 0); |
ca1d1d23 JB |
1182 | |
1183 | case_action = nochange; /* We tried an initialization */ | |
1184 | /* but some C compilers blew it */ | |
4746118a JB |
1185 | |
1186 | if (search_regs.num_regs <= 0) | |
1187 | error ("replace-match called before any match found"); | |
1188 | ||
ca1d1d23 JB |
1189 | if (search_regs.start[0] < BEGV |
1190 | || search_regs.start[0] > search_regs.end[0] | |
1191 | || search_regs.end[0] > ZV) | |
97832bd0 RS |
1192 | args_out_of_range (make_number (search_regs.start[0]), |
1193 | make_number (search_regs.end[0])); | |
ca1d1d23 JB |
1194 | |
1195 | if (NILP (fixedcase)) | |
1196 | { | |
1197 | /* Decide how to casify by examining the matched text. */ | |
1198 | ||
1199 | last = search_regs.end[0]; | |
1200 | prevc = '\n'; | |
1201 | case_action = all_caps; | |
1202 | ||
1203 | /* some_multiletter_word is set nonzero if any original word | |
1204 | is more than one letter long. */ | |
1205 | some_multiletter_word = 0; | |
97832bd0 | 1206 | some_lowercase = 0; |
c4d460ce | 1207 | some_lowercase_initial = 0; |
ca1d1d23 JB |
1208 | |
1209 | for (pos = search_regs.start[0]; pos < last; pos++) | |
1210 | { | |
1211 | c = FETCH_CHAR (pos); | |
1212 | if (LOWERCASEP (c)) | |
1213 | { | |
1214 | /* Cannot be all caps if any original char is lower case */ | |
1215 | ||
97832bd0 | 1216 | some_lowercase = 1; |
ca1d1d23 | 1217 | if (SYNTAX (prevc) != Sword) |
c4d460ce | 1218 | some_lowercase_initial = 1; |
ca1d1d23 JB |
1219 | else |
1220 | some_multiletter_word = 1; | |
1221 | } | |
1222 | else if (!NOCASEP (c)) | |
1223 | { | |
97832bd0 | 1224 | if (SYNTAX (prevc) != Sword) |
c4d460ce | 1225 | ; |
97832bd0 | 1226 | else |
ca1d1d23 JB |
1227 | some_multiletter_word = 1; |
1228 | } | |
1229 | ||
1230 | prevc = c; | |
1231 | } | |
1232 | ||
97832bd0 RS |
1233 | /* Convert to all caps if the old text is all caps |
1234 | and has at least one multiletter word. */ | |
1235 | if (! some_lowercase && some_multiletter_word) | |
1236 | case_action = all_caps; | |
c4d460ce | 1237 | /* Capitalize each word, if the old text has all capitalized words. */ |
026e8330 RS |
1238 | /* We used to insist on some_multiletter_word here, |
1239 | but that screwed query replacing x with y, acting on X. | |
1240 | Even what we have now is more strict than what 19.22 had. */ | |
1241 | else if (!some_lowercase_initial) | |
ca1d1d23 | 1242 | case_action = cap_initial; |
97832bd0 RS |
1243 | else |
1244 | case_action = nochange; | |
ca1d1d23 JB |
1245 | } |
1246 | ||
9a76659d JB |
1247 | /* We insert the replacement text before the old text, and then |
1248 | delete the original text. This means that markers at the | |
1249 | beginning or end of the original will float to the corresponding | |
1250 | position in the replacement. */ | |
1251 | SET_PT (search_regs.start[0]); | |
ca1d1d23 | 1252 | if (!NILP (literal)) |
16fdc568 | 1253 | Finsert_and_inherit (1, &newtext); |
ca1d1d23 JB |
1254 | else |
1255 | { | |
1256 | struct gcpro gcpro1; | |
16fdc568 | 1257 | GCPRO1 (newtext); |
ca1d1d23 | 1258 | |
16fdc568 | 1259 | for (pos = 0; pos < XSTRING (newtext)->size; pos++) |
ca1d1d23 | 1260 | { |
9a76659d JB |
1261 | int offset = point - search_regs.start[0]; |
1262 | ||
16fdc568 | 1263 | c = XSTRING (newtext)->data[pos]; |
ca1d1d23 JB |
1264 | if (c == '\\') |
1265 | { | |
16fdc568 | 1266 | c = XSTRING (newtext)->data[++pos]; |
ca1d1d23 | 1267 | if (c == '&') |
9a76659d JB |
1268 | Finsert_buffer_substring |
1269 | (Fcurrent_buffer (), | |
1270 | make_number (search_regs.start[0] + offset), | |
1271 | make_number (search_regs.end[0] + offset)); | |
4746118a | 1272 | else if (c >= '1' && c <= search_regs.num_regs + '0') |
ca1d1d23 JB |
1273 | { |
1274 | if (search_regs.start[c - '0'] >= 1) | |
9a76659d JB |
1275 | Finsert_buffer_substring |
1276 | (Fcurrent_buffer (), | |
1277 | make_number (search_regs.start[c - '0'] + offset), | |
1278 | make_number (search_regs.end[c - '0'] + offset)); | |
ca1d1d23 JB |
1279 | } |
1280 | else | |
1281 | insert_char (c); | |
1282 | } | |
1283 | else | |
1284 | insert_char (c); | |
1285 | } | |
1286 | UNGCPRO; | |
1287 | } | |
1288 | ||
9a76659d JB |
1289 | inslen = point - (search_regs.start[0]); |
1290 | del_range (search_regs.start[0] + inslen, search_regs.end[0] + inslen); | |
ca1d1d23 JB |
1291 | |
1292 | if (case_action == all_caps) | |
1293 | Fupcase_region (make_number (point - inslen), make_number (point)); | |
1294 | else if (case_action == cap_initial) | |
1295 | upcase_initials_region (make_number (point - inslen), make_number (point)); | |
1296 | return Qnil; | |
1297 | } | |
1298 | \f | |
1299 | static Lisp_Object | |
1300 | match_limit (num, beginningp) | |
1301 | Lisp_Object num; | |
1302 | int beginningp; | |
1303 | { | |
1304 | register int n; | |
1305 | ||
1306 | CHECK_NUMBER (num, 0); | |
1307 | n = XINT (num); | |
4746118a JB |
1308 | if (n < 0 || n >= search_regs.num_regs) |
1309 | args_out_of_range (num, make_number (search_regs.num_regs)); | |
1310 | if (search_regs.num_regs <= 0 | |
1311 | || search_regs.start[n] < 0) | |
ca1d1d23 JB |
1312 | return Qnil; |
1313 | return (make_number ((beginningp) ? search_regs.start[n] | |
1314 | : search_regs.end[n])); | |
1315 | } | |
1316 | ||
1317 | DEFUN ("match-beginning", Fmatch_beginning, Smatch_beginning, 1, 1, 0, | |
1318 | "Return position of start of text matched by last search.\n\ | |
16fdc568 BF |
1319 | NUM specifies which parenthesized expression in the last regexp.\n\ |
1320 | Value is nil if NUMth pair didn't match, or there were less than NUM pairs.\n\ | |
ca1d1d23 JB |
1321 | Zero means the entire text matched by the whole regexp or whole string.") |
1322 | (num) | |
1323 | Lisp_Object num; | |
1324 | { | |
1325 | return match_limit (num, 1); | |
1326 | } | |
1327 | ||
1328 | DEFUN ("match-end", Fmatch_end, Smatch_end, 1, 1, 0, | |
1329 | "Return position of end of text matched by last search.\n\ | |
1330 | ARG, a number, specifies which parenthesized expression in the last regexp.\n\ | |
1331 | Value is nil if ARGth pair didn't match, or there were less than ARG pairs.\n\ | |
1332 | Zero means the entire text matched by the whole regexp or whole string.") | |
1333 | (num) | |
1334 | Lisp_Object num; | |
1335 | { | |
1336 | return match_limit (num, 0); | |
1337 | } | |
1338 | ||
1339 | DEFUN ("match-data", Fmatch_data, Smatch_data, 0, 0, 0, | |
1340 | "Return a list containing all info on what the last search matched.\n\ | |
1341 | Element 2N is `(match-beginning N)'; element 2N + 1 is `(match-end N)'.\n\ | |
1342 | All the elements are markers or nil (nil if the Nth pair didn't match)\n\ | |
1343 | if the last match was on a buffer; integers or nil if a string was matched.\n\ | |
1344 | Use `store-match-data' to reinstate the data in this list.") | |
1345 | () | |
1346 | { | |
4746118a | 1347 | Lisp_Object *data; |
ca1d1d23 JB |
1348 | int i, len; |
1349 | ||
daa37602 JB |
1350 | if (NILP (last_thing_searched)) |
1351 | error ("match-data called before any match found"); | |
1352 | ||
4746118a JB |
1353 | data = (Lisp_Object *) alloca ((2 * search_regs.num_regs) |
1354 | * sizeof (Lisp_Object)); | |
1355 | ||
ca1d1d23 | 1356 | len = -1; |
4746118a | 1357 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
1358 | { |
1359 | int start = search_regs.start[i]; | |
1360 | if (start >= 0) | |
1361 | { | |
daa37602 | 1362 | if (EQ (last_thing_searched, Qt)) |
ca1d1d23 JB |
1363 | { |
1364 | XFASTINT (data[2 * i]) = start; | |
1365 | XFASTINT (data[2 * i + 1]) = search_regs.end[i]; | |
1366 | } | |
daa37602 | 1367 | else if (XTYPE (last_thing_searched) == Lisp_Buffer) |
ca1d1d23 JB |
1368 | { |
1369 | data[2 * i] = Fmake_marker (); | |
daa37602 JB |
1370 | Fset_marker (data[2 * i], |
1371 | make_number (start), | |
1372 | last_thing_searched); | |
ca1d1d23 JB |
1373 | data[2 * i + 1] = Fmake_marker (); |
1374 | Fset_marker (data[2 * i + 1], | |
daa37602 JB |
1375 | make_number (search_regs.end[i]), |
1376 | last_thing_searched); | |
ca1d1d23 | 1377 | } |
daa37602 JB |
1378 | else |
1379 | /* last_thing_searched must always be Qt, a buffer, or Qnil. */ | |
1380 | abort (); | |
1381 | ||
ca1d1d23 JB |
1382 | len = i; |
1383 | } | |
1384 | else | |
1385 | data[2 * i] = data [2 * i + 1] = Qnil; | |
1386 | } | |
1387 | return Flist (2 * len + 2, data); | |
1388 | } | |
1389 | ||
1390 | ||
1391 | DEFUN ("store-match-data", Fstore_match_data, Sstore_match_data, 1, 1, 0, | |
1392 | "Set internal data on last search match from elements of LIST.\n\ | |
1393 | LIST should have been created by calling `match-data' previously.") | |
1394 | (list) | |
1395 | register Lisp_Object list; | |
1396 | { | |
1397 | register int i; | |
1398 | register Lisp_Object marker; | |
1399 | ||
1400 | if (!CONSP (list) && !NILP (list)) | |
b37902c8 | 1401 | list = wrong_type_argument (Qconsp, list); |
ca1d1d23 | 1402 | |
daa37602 JB |
1403 | /* Unless we find a marker with a buffer in LIST, assume that this |
1404 | match data came from a string. */ | |
1405 | last_thing_searched = Qt; | |
1406 | ||
4746118a JB |
1407 | /* Allocate registers if they don't already exist. */ |
1408 | { | |
d084e942 | 1409 | int length = XFASTINT (Flength (list)) / 2; |
4746118a JB |
1410 | |
1411 | if (length > search_regs.num_regs) | |
1412 | { | |
1113d9db JB |
1413 | if (search_regs.num_regs == 0) |
1414 | { | |
1415 | search_regs.start | |
1416 | = (regoff_t *) xmalloc (length * sizeof (regoff_t)); | |
1417 | search_regs.end | |
1418 | = (regoff_t *) xmalloc (length * sizeof (regoff_t)); | |
1419 | } | |
4746118a | 1420 | else |
1113d9db JB |
1421 | { |
1422 | search_regs.start | |
1423 | = (regoff_t *) xrealloc (search_regs.start, | |
1424 | length * sizeof (regoff_t)); | |
1425 | search_regs.end | |
1426 | = (regoff_t *) xrealloc (search_regs.end, | |
1427 | length * sizeof (regoff_t)); | |
1428 | } | |
4746118a | 1429 | |
9ac0d9e0 | 1430 | BLOCK_INPUT; |
1113d9db JB |
1431 | re_set_registers (&searchbuf, &search_regs, length, |
1432 | search_regs.start, search_regs.end); | |
9ac0d9e0 | 1433 | UNBLOCK_INPUT; |
4746118a JB |
1434 | } |
1435 | } | |
1436 | ||
1437 | for (i = 0; i < search_regs.num_regs; i++) | |
ca1d1d23 JB |
1438 | { |
1439 | marker = Fcar (list); | |
1440 | if (NILP (marker)) | |
1441 | { | |
1442 | search_regs.start[i] = -1; | |
1443 | list = Fcdr (list); | |
1444 | } | |
1445 | else | |
1446 | { | |
daa37602 JB |
1447 | if (XTYPE (marker) == Lisp_Marker) |
1448 | { | |
1449 | if (XMARKER (marker)->buffer == 0) | |
1450 | XFASTINT (marker) = 0; | |
1451 | else | |
1452 | XSET (last_thing_searched, Lisp_Buffer, | |
1453 | XMARKER (marker)->buffer); | |
1454 | } | |
ca1d1d23 JB |
1455 | |
1456 | CHECK_NUMBER_COERCE_MARKER (marker, 0); | |
1457 | search_regs.start[i] = XINT (marker); | |
1458 | list = Fcdr (list); | |
1459 | ||
1460 | marker = Fcar (list); | |
1461 | if (XTYPE (marker) == Lisp_Marker | |
1462 | && XMARKER (marker)->buffer == 0) | |
1463 | XFASTINT (marker) = 0; | |
1464 | ||
1465 | CHECK_NUMBER_COERCE_MARKER (marker, 0); | |
1466 | search_regs.end[i] = XINT (marker); | |
1467 | } | |
1468 | list = Fcdr (list); | |
1469 | } | |
1470 | ||
1471 | return Qnil; | |
1472 | } | |
1473 | ||
1474 | /* Quote a string to inactivate reg-expr chars */ | |
1475 | ||
1476 | DEFUN ("regexp-quote", Fregexp_quote, Sregexp_quote, 1, 1, 0, | |
1477 | "Return a regexp string which matches exactly STRING and nothing else.") | |
1478 | (str) | |
1479 | Lisp_Object str; | |
1480 | { | |
1481 | register unsigned char *in, *out, *end; | |
1482 | register unsigned char *temp; | |
1483 | ||
1484 | CHECK_STRING (str, 0); | |
1485 | ||
1486 | temp = (unsigned char *) alloca (XSTRING (str)->size * 2); | |
1487 | ||
1488 | /* Now copy the data into the new string, inserting escapes. */ | |
1489 | ||
1490 | in = XSTRING (str)->data; | |
1491 | end = in + XSTRING (str)->size; | |
1492 | out = temp; | |
1493 | ||
1494 | for (; in != end; in++) | |
1495 | { | |
1496 | if (*in == '[' || *in == ']' | |
1497 | || *in == '*' || *in == '.' || *in == '\\' | |
1498 | || *in == '?' || *in == '+' | |
1499 | || *in == '^' || *in == '$') | |
1500 | *out++ = '\\'; | |
1501 | *out++ = *in; | |
1502 | } | |
1503 | ||
1504 | return make_string (temp, out - temp); | |
1505 | } | |
1506 | \f | |
1507 | syms_of_search () | |
1508 | { | |
1509 | register int i; | |
1510 | ||
1511 | searchbuf.allocated = 100; | |
8c0e7b73 | 1512 | searchbuf.buffer = (unsigned char *) malloc (searchbuf.allocated); |
ca1d1d23 JB |
1513 | searchbuf.fastmap = search_fastmap; |
1514 | ||
1515 | Qsearch_failed = intern ("search-failed"); | |
1516 | staticpro (&Qsearch_failed); | |
1517 | Qinvalid_regexp = intern ("invalid-regexp"); | |
1518 | staticpro (&Qinvalid_regexp); | |
1519 | ||
1520 | Fput (Qsearch_failed, Qerror_conditions, | |
1521 | Fcons (Qsearch_failed, Fcons (Qerror, Qnil))); | |
1522 | Fput (Qsearch_failed, Qerror_message, | |
1523 | build_string ("Search failed")); | |
1524 | ||
1525 | Fput (Qinvalid_regexp, Qerror_conditions, | |
1526 | Fcons (Qinvalid_regexp, Fcons (Qerror, Qnil))); | |
1527 | Fput (Qinvalid_regexp, Qerror_message, | |
1528 | build_string ("Invalid regexp")); | |
1529 | ||
1530 | last_regexp = Qnil; | |
1531 | staticpro (&last_regexp); | |
1532 | ||
daa37602 JB |
1533 | last_thing_searched = Qnil; |
1534 | staticpro (&last_thing_searched); | |
1535 | ||
ca1d1d23 JB |
1536 | defsubr (&Sstring_match); |
1537 | defsubr (&Slooking_at); | |
1538 | defsubr (&Sskip_chars_forward); | |
1539 | defsubr (&Sskip_chars_backward); | |
17431c60 RS |
1540 | defsubr (&Sskip_syntax_forward); |
1541 | defsubr (&Sskip_syntax_backward); | |
ca1d1d23 JB |
1542 | defsubr (&Ssearch_forward); |
1543 | defsubr (&Ssearch_backward); | |
1544 | defsubr (&Sword_search_forward); | |
1545 | defsubr (&Sword_search_backward); | |
1546 | defsubr (&Sre_search_forward); | |
1547 | defsubr (&Sre_search_backward); | |
1548 | defsubr (&Sreplace_match); | |
1549 | defsubr (&Smatch_beginning); | |
1550 | defsubr (&Smatch_end); | |
1551 | defsubr (&Smatch_data); | |
1552 | defsubr (&Sstore_match_data); | |
1553 | defsubr (&Sregexp_quote); | |
1554 | } |