Commit | Line | Data |
---|---|---|
ca1d1d23 | 1 | /* String search routines for GNU Emacs. |
c6c5df7f | 2 | Copyright (C) 1985, 1986, 1987, 1993 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 | ||
72 | static void | |
73 | matcher_overflow () | |
74 | { | |
75 | error ("Stack overflow in regexp matcher"); | |
76 | } | |
77 | ||
78 | #ifdef __STDC__ | |
79 | #define CONST const | |
80 | #else | |
81 | #define CONST | |
82 | #endif | |
83 | ||
84 | /* Compile a regexp and signal a Lisp error if anything goes wrong. */ | |
85 | ||
1113d9db | 86 | compile_pattern (pattern, bufp, regp, translate) |
ca1d1d23 JB |
87 | Lisp_Object pattern; |
88 | struct re_pattern_buffer *bufp; | |
1113d9db | 89 | struct re_registers *regp; |
ca1d1d23 JB |
90 | char *translate; |
91 | { | |
92 | CONST char *val; | |
93 | Lisp_Object dummy; | |
94 | ||
95 | if (EQ (pattern, last_regexp) | |
96 | && translate == bufp->translate) | |
97 | return; | |
1113d9db | 98 | |
ca1d1d23 JB |
99 | last_regexp = Qnil; |
100 | bufp->translate = translate; | |
9ac0d9e0 | 101 | BLOCK_INPUT; |
b90d9e80 RS |
102 | val = (CONST char *) re_compile_pattern ((char *) XSTRING (pattern)->data, |
103 | XSTRING (pattern)->size, bufp); | |
9ac0d9e0 | 104 | UNBLOCK_INPUT; |
ca1d1d23 JB |
105 | if (val) |
106 | { | |
107 | dummy = build_string (val); | |
108 | while (1) | |
109 | Fsignal (Qinvalid_regexp, Fcons (dummy, Qnil)); | |
110 | } | |
1113d9db | 111 | |
ca1d1d23 | 112 | last_regexp = pattern; |
1113d9db JB |
113 | |
114 | /* Advise the searching functions about the space we have allocated | |
115 | for register data. */ | |
9ac0d9e0 | 116 | BLOCK_INPUT; |
ebb9e16f JB |
117 | if (regp) |
118 | re_set_registers (bufp, regp, regp->num_regs, regp->start, regp->end); | |
9ac0d9e0 | 119 | UNBLOCK_INPUT; |
1113d9db | 120 | |
ca1d1d23 JB |
121 | return; |
122 | } | |
123 | ||
124 | /* Error condition used for failing searches */ | |
125 | Lisp_Object Qsearch_failed; | |
126 | ||
127 | Lisp_Object | |
128 | signal_failure (arg) | |
129 | Lisp_Object arg; | |
130 | { | |
131 | Fsignal (Qsearch_failed, Fcons (arg, Qnil)); | |
132 | return Qnil; | |
133 | } | |
134 | \f | |
135 | DEFUN ("looking-at", Flooking_at, Slooking_at, 1, 1, 0, | |
e065a56e JB |
136 | "Return t if text after point matches regular expression PAT.\n\ |
137 | This function modifies the match data that `match-beginning',\n\ | |
138 | `match-end' and `match-data' access; save and restore the match\n\ | |
fe99283d | 139 | data if you want to preserve them.") |
ca1d1d23 JB |
140 | (string) |
141 | Lisp_Object string; | |
142 | { | |
143 | Lisp_Object val; | |
144 | unsigned char *p1, *p2; | |
145 | int s1, s2; | |
146 | register int i; | |
147 | ||
148 | CHECK_STRING (string, 0); | |
1113d9db | 149 | compile_pattern (string, &searchbuf, &search_regs, |
ca1d1d23 JB |
150 | !NILP (current_buffer->case_fold_search) ? DOWNCASE_TABLE : 0); |
151 | ||
152 | immediate_quit = 1; | |
153 | QUIT; /* Do a pending quit right away, to avoid paradoxical behavior */ | |
154 | ||
155 | /* Get pointers and sizes of the two strings | |
156 | that make up the visible portion of the buffer. */ | |
157 | ||
158 | p1 = BEGV_ADDR; | |
159 | s1 = GPT - BEGV; | |
160 | p2 = GAP_END_ADDR; | |
161 | s2 = ZV - GPT; | |
162 | if (s1 < 0) | |
163 | { | |
164 | p2 = p1; | |
165 | s2 = ZV - BEGV; | |
166 | s1 = 0; | |
167 | } | |
168 | if (s2 < 0) | |
169 | { | |
170 | s1 = ZV - BEGV; | |
171 | s2 = 0; | |
172 | } | |
173 | ||
174 | i = re_match_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2, | |
175 | point - BEGV, &search_regs, | |
176 | ZV - BEGV); | |
177 | if (i == -2) | |
178 | matcher_overflow (); | |
179 | ||
180 | val = (0 <= i ? Qt : Qnil); | |
4746118a | 181 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
182 | if (search_regs.start[i] >= 0) |
183 | { | |
184 | search_regs.start[i] += BEGV; | |
185 | search_regs.end[i] += BEGV; | |
186 | } | |
daa37602 | 187 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
188 | immediate_quit = 0; |
189 | return val; | |
190 | } | |
191 | ||
192 | DEFUN ("string-match", Fstring_match, Sstring_match, 2, 3, 0, | |
193 | "Return index of start of first match for REGEXP in STRING, or nil.\n\ | |
194 | If third arg START is non-nil, start search at that index in STRING.\n\ | |
195 | For index of first char beyond the match, do (match-end 0).\n\ | |
196 | `match-end' and `match-beginning' also give indices of substrings\n\ | |
197 | matched by parenthesis constructs in the pattern.") | |
198 | (regexp, string, start) | |
199 | Lisp_Object regexp, string, start; | |
200 | { | |
201 | int val; | |
202 | int s; | |
203 | ||
204 | CHECK_STRING (regexp, 0); | |
205 | CHECK_STRING (string, 1); | |
206 | ||
207 | if (NILP (start)) | |
208 | s = 0; | |
209 | else | |
210 | { | |
211 | int len = XSTRING (string)->size; | |
212 | ||
213 | CHECK_NUMBER (start, 2); | |
214 | s = XINT (start); | |
215 | if (s < 0 && -s <= len) | |
216 | s = len - s; | |
217 | else if (0 > s || s > len) | |
218 | args_out_of_range (string, start); | |
219 | } | |
220 | ||
1113d9db | 221 | compile_pattern (regexp, &searchbuf, &search_regs, |
ca1d1d23 JB |
222 | !NILP (current_buffer->case_fold_search) ? DOWNCASE_TABLE : 0); |
223 | immediate_quit = 1; | |
224 | val = re_search (&searchbuf, (char *) XSTRING (string)->data, | |
225 | XSTRING (string)->size, s, XSTRING (string)->size - s, | |
226 | &search_regs); | |
227 | immediate_quit = 0; | |
daa37602 | 228 | last_thing_searched = Qt; |
ca1d1d23 JB |
229 | if (val == -2) |
230 | matcher_overflow (); | |
231 | if (val < 0) return Qnil; | |
232 | return make_number (val); | |
233 | } | |
e59a8453 RS |
234 | |
235 | /* Match REGEXP against STRING, searching all of STRING, | |
236 | and return the index of the match, or negative on failure. | |
237 | This does not clobber the match data. */ | |
238 | ||
239 | int | |
240 | fast_string_match (regexp, string) | |
241 | Lisp_Object regexp, string; | |
242 | { | |
243 | int val; | |
244 | ||
245 | compile_pattern (regexp, &searchbuf, 0, 0); | |
246 | immediate_quit = 1; | |
247 | val = re_search (&searchbuf, (char *) XSTRING (string)->data, | |
248 | XSTRING (string)->size, 0, XSTRING (string)->size, | |
249 | 0); | |
250 | immediate_quit = 0; | |
251 | return val; | |
252 | } | |
ca1d1d23 | 253 | \f |
ffd56f97 JB |
254 | /* Search for COUNT instances of the character TARGET, starting at START. |
255 | If COUNT is negative, search backwards. | |
256 | ||
257 | If we find COUNT instances, set *SHORTAGE to zero, and return the | |
5bfe95c9 RS |
258 | position after the COUNTth match. Note that for reverse motion |
259 | this is not the same as the usual convention for Emacs motion commands. | |
ffd56f97 JB |
260 | |
261 | If we don't find COUNT instances before reaching the end of the | |
262 | buffer (or the beginning, if scanning backwards), set *SHORTAGE to | |
263 | the number of TARGETs left unfound, and return the end of the | |
264 | buffer we bumped up against. */ | |
265 | ||
266 | scan_buffer (target, start, count, shortage) | |
267 | int *shortage, start; | |
268 | register int count, target; | |
ca1d1d23 | 269 | { |
ffd56f97 JB |
270 | int limit = ((count > 0) ? ZV - 1 : BEGV); |
271 | int direction = ((count > 0) ? 1 : -1); | |
272 | ||
273 | register unsigned char *cursor; | |
ca1d1d23 | 274 | unsigned char *base; |
ffd56f97 JB |
275 | |
276 | register int ceiling; | |
277 | register unsigned char *ceiling_addr; | |
ca1d1d23 JB |
278 | |
279 | if (shortage != 0) | |
280 | *shortage = 0; | |
281 | ||
282 | immediate_quit = 1; | |
283 | ||
ffd56f97 JB |
284 | if (count > 0) |
285 | while (start != limit + 1) | |
ca1d1d23 | 286 | { |
ffd56f97 JB |
287 | ceiling = BUFFER_CEILING_OF (start); |
288 | ceiling = min (limit, ceiling); | |
289 | ceiling_addr = &FETCH_CHAR (ceiling) + 1; | |
290 | base = (cursor = &FETCH_CHAR (start)); | |
ca1d1d23 JB |
291 | while (1) |
292 | { | |
ffd56f97 | 293 | while (*cursor != target && ++cursor != ceiling_addr) |
ca1d1d23 | 294 | ; |
ffd56f97 | 295 | if (cursor != ceiling_addr) |
ca1d1d23 | 296 | { |
ffd56f97 | 297 | if (--count == 0) |
ca1d1d23 JB |
298 | { |
299 | immediate_quit = 0; | |
ffd56f97 | 300 | return (start + cursor - base + 1); |
ca1d1d23 JB |
301 | } |
302 | else | |
ffd56f97 | 303 | if (++cursor == ceiling_addr) |
ca1d1d23 JB |
304 | break; |
305 | } | |
306 | else | |
307 | break; | |
308 | } | |
ffd56f97 | 309 | start += cursor - base; |
ca1d1d23 JB |
310 | } |
311 | else | |
312 | { | |
ffd56f97 JB |
313 | start--; /* first character we scan */ |
314 | while (start > limit - 1) | |
315 | { /* we WILL scan under start */ | |
316 | ceiling = BUFFER_FLOOR_OF (start); | |
317 | ceiling = max (limit, ceiling); | |
318 | ceiling_addr = &FETCH_CHAR (ceiling) - 1; | |
319 | base = (cursor = &FETCH_CHAR (start)); | |
ca1d1d23 JB |
320 | cursor++; |
321 | while (1) | |
322 | { | |
ffd56f97 | 323 | while (--cursor != ceiling_addr && *cursor != target) |
ca1d1d23 | 324 | ; |
ffd56f97 | 325 | if (cursor != ceiling_addr) |
ca1d1d23 | 326 | { |
ffd56f97 | 327 | if (++count == 0) |
ca1d1d23 JB |
328 | { |
329 | immediate_quit = 0; | |
ffd56f97 | 330 | return (start + cursor - base + 1); |
ca1d1d23 JB |
331 | } |
332 | } | |
333 | else | |
334 | break; | |
335 | } | |
ffd56f97 | 336 | start += cursor - base; |
ca1d1d23 JB |
337 | } |
338 | } | |
339 | immediate_quit = 0; | |
340 | if (shortage != 0) | |
ffd56f97 JB |
341 | *shortage = count * direction; |
342 | return (start + ((direction == 1 ? 0 : 1))); | |
ca1d1d23 JB |
343 | } |
344 | ||
345 | int | |
346 | find_next_newline (from, cnt) | |
347 | register int from, cnt; | |
348 | { | |
349 | return (scan_buffer ('\n', from, cnt, (int *) 0)); | |
350 | } | |
351 | \f | |
c1dc99a1 JB |
352 | Lisp_Object skip_chars (); |
353 | ||
ca1d1d23 | 354 | DEFUN ("skip-chars-forward", Fskip_chars_forward, Sskip_chars_forward, 1, 2, 0, |
3acb9a69 RS |
355 | "Move point forward, stopping before a char not in STRING, or at pos LIM.\n\ |
356 | STRING is like the inside of a `[...]' in a regular expression\n\ | |
ca1d1d23 JB |
357 | except that `]' is never special and `\\' quotes `^', `-' or `\\'.\n\ |
358 | Thus, with arg \"a-zA-Z\", this skips letters stopping before first nonletter.\n\ | |
c1dc99a1 JB |
359 | With arg \"^a-zA-Z\", skips nonletters stopping before first letter.\n\ |
360 | Returns the distance traveled, either zero or positive.") | |
ca1d1d23 JB |
361 | (string, lim) |
362 | Lisp_Object string, lim; | |
363 | { | |
17431c60 | 364 | return skip_chars (1, 0, string, lim); |
ca1d1d23 JB |
365 | } |
366 | ||
367 | DEFUN ("skip-chars-backward", Fskip_chars_backward, Sskip_chars_backward, 1, 2, 0, | |
3acb9a69 | 368 | "Move point backward, stopping after a char not in STRING, or at pos LIM.\n\ |
c1dc99a1 JB |
369 | See `skip-chars-forward' for details.\n\ |
370 | Returns the distance traveled, either zero or negative.") | |
ca1d1d23 JB |
371 | (string, lim) |
372 | Lisp_Object string, lim; | |
373 | { | |
17431c60 RS |
374 | return skip_chars (0, 0, string, lim); |
375 | } | |
376 | ||
377 | DEFUN ("skip-syntax-forward", Fskip_syntax_forward, Sskip_syntax_forward, 1, 2, 0, | |
378 | "Move point forward across chars in specified syntax classes.\n\ | |
379 | SYNTAX is a string of syntax code characters.\n\ | |
380 | Stop before a char whose syntax is not in SYNTAX, or at position LIM.\n\ | |
381 | If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.\n\ | |
382 | This function returns the distance traveled, either zero or positive.") | |
383 | (syntax, lim) | |
384 | Lisp_Object syntax, lim; | |
385 | { | |
386 | return skip_chars (1, 1, syntax, lim); | |
387 | } | |
388 | ||
389 | DEFUN ("skip-syntax-backward", Fskip_syntax_backward, Sskip_syntax_backward, 1, 2, 0, | |
390 | "Move point backward across chars in specified syntax classes.\n\ | |
391 | SYNTAX is a string of syntax code characters.\n\ | |
392 | Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.\n\ | |
393 | If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.\n\ | |
394 | This function returns the distance traveled, either zero or negative.") | |
395 | (syntax, lim) | |
396 | Lisp_Object syntax, lim; | |
397 | { | |
398 | return skip_chars (0, 1, syntax, lim); | |
ca1d1d23 JB |
399 | } |
400 | ||
c1dc99a1 | 401 | Lisp_Object |
17431c60 RS |
402 | skip_chars (forwardp, syntaxp, string, lim) |
403 | int forwardp, syntaxp; | |
ca1d1d23 JB |
404 | Lisp_Object string, lim; |
405 | { | |
406 | register unsigned char *p, *pend; | |
407 | register unsigned char c; | |
408 | unsigned char fastmap[0400]; | |
409 | int negate = 0; | |
410 | register int i; | |
411 | ||
412 | CHECK_STRING (string, 0); | |
413 | ||
414 | if (NILP (lim)) | |
415 | XSET (lim, Lisp_Int, forwardp ? ZV : BEGV); | |
416 | else | |
417 | CHECK_NUMBER_COERCE_MARKER (lim, 1); | |
418 | ||
ca1d1d23 | 419 | /* In any case, don't allow scan outside bounds of buffer. */ |
c5241910 RS |
420 | /* jla turned this off, for no known reason. |
421 | bfox turned the ZV part on, and rms turned the | |
422 | BEGV part back on. */ | |
423 | if (XINT (lim) > ZV) | |
ca1d1d23 | 424 | XFASTINT (lim) = ZV; |
c5241910 | 425 | if (XINT (lim) < BEGV) |
ca1d1d23 | 426 | XFASTINT (lim) = BEGV; |
ca1d1d23 JB |
427 | |
428 | p = XSTRING (string)->data; | |
429 | pend = p + XSTRING (string)->size; | |
430 | bzero (fastmap, sizeof fastmap); | |
431 | ||
432 | if (p != pend && *p == '^') | |
433 | { | |
434 | negate = 1; p++; | |
435 | } | |
436 | ||
17431c60 RS |
437 | /* Find the characters specified and set their elements of fastmap. |
438 | If syntaxp, each character counts as itself. | |
439 | Otherwise, handle backslashes and ranges specially */ | |
ca1d1d23 JB |
440 | |
441 | while (p != pend) | |
442 | { | |
443 | c = *p++; | |
17431c60 RS |
444 | if (syntaxp) |
445 | fastmap[c] = 1; | |
446 | else | |
ca1d1d23 | 447 | { |
17431c60 | 448 | if (c == '\\') |
ca1d1d23 | 449 | { |
17431c60 RS |
450 | if (p == pend) break; |
451 | c = *p++; | |
452 | } | |
453 | if (p != pend && *p == '-') | |
454 | { | |
455 | p++; | |
456 | if (p == pend) break; | |
457 | while (c <= *p) | |
458 | { | |
459 | fastmap[c] = 1; | |
460 | c++; | |
461 | } | |
462 | p++; | |
ca1d1d23 | 463 | } |
17431c60 RS |
464 | else |
465 | fastmap[c] = 1; | |
ca1d1d23 | 466 | } |
ca1d1d23 JB |
467 | } |
468 | ||
469 | /* If ^ was the first character, complement the fastmap. */ | |
470 | ||
471 | if (negate) | |
472 | for (i = 0; i < sizeof fastmap; i++) | |
473 | fastmap[i] ^= 1; | |
474 | ||
c1dc99a1 JB |
475 | { |
476 | int start_point = point; | |
477 | ||
478 | immediate_quit = 1; | |
17431c60 | 479 | if (syntaxp) |
c1dc99a1 | 480 | { |
17431c60 RS |
481 | |
482 | if (forwardp) | |
483 | { | |
484 | while (point < XINT (lim) | |
485 | && fastmap[(unsigned char) syntax_code_spec[(int) SYNTAX (FETCH_CHAR (point))]]) | |
486 | SET_PT (point + 1); | |
487 | } | |
488 | else | |
489 | { | |
490 | while (point > XINT (lim) | |
491 | && fastmap[(unsigned char) syntax_code_spec[(int) SYNTAX (FETCH_CHAR (point - 1))]]) | |
492 | SET_PT (point - 1); | |
493 | } | |
c1dc99a1 JB |
494 | } |
495 | else | |
496 | { | |
17431c60 RS |
497 | if (forwardp) |
498 | { | |
499 | while (point < XINT (lim) && fastmap[FETCH_CHAR (point)]) | |
500 | SET_PT (point + 1); | |
501 | } | |
502 | else | |
503 | { | |
504 | while (point > XINT (lim) && fastmap[FETCH_CHAR (point - 1)]) | |
505 | SET_PT (point - 1); | |
506 | } | |
c1dc99a1 JB |
507 | } |
508 | immediate_quit = 0; | |
509 | ||
510 | return make_number (point - start_point); | |
511 | } | |
ca1d1d23 JB |
512 | } |
513 | \f | |
514 | /* Subroutines of Lisp buffer search functions. */ | |
515 | ||
516 | static Lisp_Object | |
517 | search_command (string, bound, noerror, count, direction, RE) | |
518 | Lisp_Object string, bound, noerror, count; | |
519 | int direction; | |
520 | int RE; | |
521 | { | |
522 | register int np; | |
523 | int lim; | |
524 | int n = direction; | |
525 | ||
526 | if (!NILP (count)) | |
527 | { | |
528 | CHECK_NUMBER (count, 3); | |
529 | n *= XINT (count); | |
530 | } | |
531 | ||
532 | CHECK_STRING (string, 0); | |
533 | if (NILP (bound)) | |
534 | lim = n > 0 ? ZV : BEGV; | |
535 | else | |
536 | { | |
537 | CHECK_NUMBER_COERCE_MARKER (bound, 1); | |
538 | lim = XINT (bound); | |
539 | if (n > 0 ? lim < point : lim > point) | |
540 | error ("Invalid search bound (wrong side of point)"); | |
541 | if (lim > ZV) | |
542 | lim = ZV; | |
543 | if (lim < BEGV) | |
544 | lim = BEGV; | |
545 | } | |
546 | ||
547 | np = search_buffer (string, point, lim, n, RE, | |
548 | (!NILP (current_buffer->case_fold_search) | |
549 | ? XSTRING (current_buffer->case_canon_table)->data : 0), | |
550 | (!NILP (current_buffer->case_fold_search) | |
551 | ? XSTRING (current_buffer->case_eqv_table)->data : 0)); | |
552 | if (np <= 0) | |
553 | { | |
554 | if (NILP (noerror)) | |
555 | return signal_failure (string); | |
556 | if (!EQ (noerror, Qt)) | |
557 | { | |
558 | if (lim < BEGV || lim > ZV) | |
559 | abort (); | |
a5f217b8 RS |
560 | SET_PT (lim); |
561 | return Qnil; | |
562 | #if 0 /* This would be clean, but maybe programs depend on | |
563 | a value of nil here. */ | |
481399bf | 564 | np = lim; |
a5f217b8 | 565 | #endif |
ca1d1d23 | 566 | } |
481399bf RS |
567 | else |
568 | return Qnil; | |
ca1d1d23 JB |
569 | } |
570 | ||
571 | if (np < BEGV || np > ZV) | |
572 | abort (); | |
573 | ||
574 | SET_PT (np); | |
575 | ||
576 | return make_number (np); | |
577 | } | |
578 | \f | |
579 | /* search for the n'th occurrence of STRING in the current buffer, | |
580 | starting at position POS and stopping at position LIM, | |
581 | treating PAT as a literal string if RE is false or as | |
582 | a regular expression if RE is true. | |
583 | ||
584 | If N is positive, searching is forward and LIM must be greater than POS. | |
585 | If N is negative, searching is backward and LIM must be less than POS. | |
586 | ||
587 | Returns -x if only N-x occurrences found (x > 0), | |
588 | or else the position at the beginning of the Nth occurrence | |
589 | (if searching backward) or the end (if searching forward). */ | |
590 | ||
591 | search_buffer (string, pos, lim, n, RE, trt, inverse_trt) | |
592 | Lisp_Object string; | |
593 | int pos; | |
594 | int lim; | |
595 | int n; | |
596 | int RE; | |
597 | register unsigned char *trt; | |
598 | register unsigned char *inverse_trt; | |
599 | { | |
600 | int len = XSTRING (string)->size; | |
601 | unsigned char *base_pat = XSTRING (string)->data; | |
602 | register int *BM_tab; | |
603 | int *BM_tab_base; | |
604 | register int direction = ((n > 0) ? 1 : -1); | |
605 | register int dirlen; | |
606 | int infinity, limit, k, stride_for_teases; | |
607 | register unsigned char *pat, *cursor, *p_limit; | |
608 | register int i, j; | |
609 | unsigned char *p1, *p2; | |
610 | int s1, s2; | |
611 | ||
612 | /* Null string is found at starting position. */ | |
3f57a499 RS |
613 | if (len == 0) |
614 | return pos; | |
615 | ||
616 | /* Searching 0 times means don't move. */ | |
617 | if (n == 0) | |
ca1d1d23 JB |
618 | return pos; |
619 | ||
620 | if (RE) | |
1113d9db | 621 | compile_pattern (string, &searchbuf, &search_regs, (char *) trt); |
ca1d1d23 JB |
622 | |
623 | if (RE /* Here we detect whether the */ | |
624 | /* generality of an RE search is */ | |
625 | /* really needed. */ | |
626 | /* first item is "exact match" */ | |
4746118a | 627 | && *(searchbuf.buffer) == (char) RE_EXACTN_VALUE |
ca1d1d23 JB |
628 | && searchbuf.buffer[1] + 2 == searchbuf.used) /*first is ONLY item */ |
629 | { | |
630 | RE = 0; /* can do straight (non RE) search */ | |
631 | pat = (base_pat = (unsigned char *) searchbuf.buffer + 2); | |
632 | /* trt already applied */ | |
633 | len = searchbuf.used - 2; | |
634 | } | |
635 | else if (!RE) | |
636 | { | |
637 | pat = (unsigned char *) alloca (len); | |
638 | ||
639 | for (i = len; i--;) /* Copy the pattern; apply trt */ | |
640 | *pat++ = (((int) trt) ? trt [*base_pat++] : *base_pat++); | |
641 | pat -= len; base_pat = pat; | |
642 | } | |
643 | ||
644 | if (RE) | |
645 | { | |
646 | immediate_quit = 1; /* Quit immediately if user types ^G, | |
647 | because letting this function finish | |
648 | can take too long. */ | |
649 | QUIT; /* Do a pending quit right away, | |
650 | to avoid paradoxical behavior */ | |
651 | /* Get pointers and sizes of the two strings | |
652 | that make up the visible portion of the buffer. */ | |
653 | ||
654 | p1 = BEGV_ADDR; | |
655 | s1 = GPT - BEGV; | |
656 | p2 = GAP_END_ADDR; | |
657 | s2 = ZV - GPT; | |
658 | if (s1 < 0) | |
659 | { | |
660 | p2 = p1; | |
661 | s2 = ZV - BEGV; | |
662 | s1 = 0; | |
663 | } | |
664 | if (s2 < 0) | |
665 | { | |
666 | s1 = ZV - BEGV; | |
667 | s2 = 0; | |
668 | } | |
669 | while (n < 0) | |
670 | { | |
42db823b | 671 | int val; |
42db823b RS |
672 | val = re_search_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2, |
673 | pos - BEGV, lim - pos, &search_regs, | |
674 | /* Don't allow match past current point */ | |
675 | pos - BEGV); | |
ca1d1d23 JB |
676 | if (val == -2) |
677 | matcher_overflow (); | |
678 | if (val >= 0) | |
679 | { | |
680 | j = BEGV; | |
4746118a | 681 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
682 | if (search_regs.start[i] >= 0) |
683 | { | |
684 | search_regs.start[i] += j; | |
685 | search_regs.end[i] += j; | |
686 | } | |
daa37602 | 687 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
688 | /* Set pos to the new position. */ |
689 | pos = search_regs.start[0]; | |
690 | } | |
691 | else | |
692 | { | |
693 | immediate_quit = 0; | |
694 | return (n); | |
695 | } | |
696 | n++; | |
697 | } | |
698 | while (n > 0) | |
699 | { | |
42db823b | 700 | int val; |
42db823b RS |
701 | val = re_search_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2, |
702 | pos - BEGV, lim - pos, &search_regs, | |
703 | lim - BEGV); | |
ca1d1d23 JB |
704 | if (val == -2) |
705 | matcher_overflow (); | |
706 | if (val >= 0) | |
707 | { | |
708 | j = BEGV; | |
4746118a | 709 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
710 | if (search_regs.start[i] >= 0) |
711 | { | |
712 | search_regs.start[i] += j; | |
713 | search_regs.end[i] += j; | |
714 | } | |
daa37602 | 715 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
716 | pos = search_regs.end[0]; |
717 | } | |
718 | else | |
719 | { | |
720 | immediate_quit = 0; | |
721 | return (0 - n); | |
722 | } | |
723 | n--; | |
724 | } | |
725 | immediate_quit = 0; | |
726 | return (pos); | |
727 | } | |
728 | else /* non-RE case */ | |
729 | { | |
730 | #ifdef C_ALLOCA | |
731 | int BM_tab_space[0400]; | |
732 | BM_tab = &BM_tab_space[0]; | |
733 | #else | |
734 | BM_tab = (int *) alloca (0400 * sizeof (int)); | |
735 | #endif | |
736 | /* The general approach is that we are going to maintain that we know */ | |
737 | /* the first (closest to the present position, in whatever direction */ | |
738 | /* we're searching) character that could possibly be the last */ | |
739 | /* (furthest from present position) character of a valid match. We */ | |
740 | /* advance the state of our knowledge by looking at that character */ | |
741 | /* and seeing whether it indeed matches the last character of the */ | |
742 | /* pattern. If it does, we take a closer look. If it does not, we */ | |
743 | /* move our pointer (to putative last characters) as far as is */ | |
744 | /* logically possible. This amount of movement, which I call a */ | |
745 | /* stride, will be the length of the pattern if the actual character */ | |
746 | /* appears nowhere in the pattern, otherwise it will be the distance */ | |
747 | /* from the last occurrence of that character to the end of the */ | |
748 | /* pattern. */ | |
749 | /* As a coding trick, an enormous stride is coded into the table for */ | |
750 | /* characters that match the last character. This allows use of only */ | |
751 | /* a single test, a test for having gone past the end of the */ | |
752 | /* permissible match region, to test for both possible matches (when */ | |
753 | /* the stride goes past the end immediately) and failure to */ | |
754 | /* match (where you get nudged past the end one stride at a time). */ | |
755 | ||
756 | /* Here we make a "mickey mouse" BM table. The stride of the search */ | |
757 | /* is determined only by the last character of the putative match. */ | |
758 | /* If that character does not match, we will stride the proper */ | |
759 | /* distance to propose a match that superimposes it on the last */ | |
760 | /* instance of a character that matches it (per trt), or misses */ | |
761 | /* it entirely if there is none. */ | |
762 | ||
763 | dirlen = len * direction; | |
764 | infinity = dirlen - (lim + pos + len + len) * direction; | |
765 | if (direction < 0) | |
766 | pat = (base_pat += len - 1); | |
767 | BM_tab_base = BM_tab; | |
768 | BM_tab += 0400; | |
769 | j = dirlen; /* to get it in a register */ | |
770 | /* A character that does not appear in the pattern induces a */ | |
771 | /* stride equal to the pattern length. */ | |
772 | while (BM_tab_base != BM_tab) | |
773 | { | |
774 | *--BM_tab = j; | |
775 | *--BM_tab = j; | |
776 | *--BM_tab = j; | |
777 | *--BM_tab = j; | |
778 | } | |
779 | i = 0; | |
780 | while (i != infinity) | |
781 | { | |
782 | j = pat[i]; i += direction; | |
783 | if (i == dirlen) i = infinity; | |
784 | if ((int) trt) | |
785 | { | |
786 | k = (j = trt[j]); | |
787 | if (i == infinity) | |
788 | stride_for_teases = BM_tab[j]; | |
789 | BM_tab[j] = dirlen - i; | |
790 | /* A translation table is accompanied by its inverse -- see */ | |
791 | /* comment following downcase_table for details */ | |
792 | while ((j = inverse_trt[j]) != k) | |
793 | BM_tab[j] = dirlen - i; | |
794 | } | |
795 | else | |
796 | { | |
797 | if (i == infinity) | |
798 | stride_for_teases = BM_tab[j]; | |
799 | BM_tab[j] = dirlen - i; | |
800 | } | |
801 | /* stride_for_teases tells how much to stride if we get a */ | |
802 | /* match on the far character but are subsequently */ | |
803 | /* disappointed, by recording what the stride would have been */ | |
804 | /* for that character if the last character had been */ | |
805 | /* different. */ | |
806 | } | |
807 | infinity = dirlen - infinity; | |
808 | pos += dirlen - ((direction > 0) ? direction : 0); | |
809 | /* loop invariant - pos points at where last char (first char if reverse) | |
810 | of pattern would align in a possible match. */ | |
811 | while (n != 0) | |
812 | { | |
813 | if ((lim - pos - (direction > 0)) * direction < 0) | |
814 | return (n * (0 - direction)); | |
815 | /* First we do the part we can by pointers (maybe nothing) */ | |
816 | QUIT; | |
817 | pat = base_pat; | |
818 | limit = pos - dirlen + direction; | |
819 | limit = ((direction > 0) | |
820 | ? BUFFER_CEILING_OF (limit) | |
821 | : BUFFER_FLOOR_OF (limit)); | |
822 | /* LIMIT is now the last (not beyond-last!) value | |
823 | POS can take on without hitting edge of buffer or the gap. */ | |
824 | limit = ((direction > 0) | |
825 | ? min (lim - 1, min (limit, pos + 20000)) | |
826 | : max (lim, max (limit, pos - 20000))); | |
827 | if ((limit - pos) * direction > 20) | |
828 | { | |
829 | p_limit = &FETCH_CHAR (limit); | |
830 | p2 = (cursor = &FETCH_CHAR (pos)); | |
831 | /* In this loop, pos + cursor - p2 is the surrogate for pos */ | |
832 | while (1) /* use one cursor setting as long as i can */ | |
833 | { | |
834 | if (direction > 0) /* worth duplicating */ | |
835 | { | |
836 | /* Use signed comparison if appropriate | |
837 | to make cursor+infinity sure to be > p_limit. | |
838 | Assuming that the buffer lies in a range of addresses | |
839 | that are all "positive" (as ints) or all "negative", | |
840 | either kind of comparison will work as long | |
841 | as we don't step by infinity. So pick the kind | |
842 | that works when we do step by infinity. */ | |
843 | if ((int) (p_limit + infinity) > (int) p_limit) | |
844 | while ((int) cursor <= (int) p_limit) | |
845 | cursor += BM_tab[*cursor]; | |
846 | else | |
847 | while ((unsigned int) cursor <= (unsigned int) p_limit) | |
848 | cursor += BM_tab[*cursor]; | |
849 | } | |
850 | else | |
851 | { | |
852 | if ((int) (p_limit + infinity) < (int) p_limit) | |
853 | while ((int) cursor >= (int) p_limit) | |
854 | cursor += BM_tab[*cursor]; | |
855 | else | |
856 | while ((unsigned int) cursor >= (unsigned int) p_limit) | |
857 | cursor += BM_tab[*cursor]; | |
858 | } | |
859 | /* If you are here, cursor is beyond the end of the searched region. */ | |
860 | /* This can happen if you match on the far character of the pattern, */ | |
861 | /* because the "stride" of that character is infinity, a number able */ | |
862 | /* to throw you well beyond the end of the search. It can also */ | |
863 | /* happen if you fail to match within the permitted region and would */ | |
864 | /* otherwise try a character beyond that region */ | |
865 | if ((cursor - p_limit) * direction <= len) | |
866 | break; /* a small overrun is genuine */ | |
867 | cursor -= infinity; /* large overrun = hit */ | |
868 | i = dirlen - direction; | |
869 | if ((int) trt) | |
870 | { | |
871 | while ((i -= direction) + direction != 0) | |
872 | if (pat[i] != trt[*(cursor -= direction)]) | |
873 | break; | |
874 | } | |
875 | else | |
876 | { | |
877 | while ((i -= direction) + direction != 0) | |
878 | if (pat[i] != *(cursor -= direction)) | |
879 | break; | |
880 | } | |
881 | cursor += dirlen - i - direction; /* fix cursor */ | |
882 | if (i + direction == 0) | |
883 | { | |
884 | cursor -= direction; | |
1113d9db JB |
885 | |
886 | /* Make sure we have registers in which to store | |
887 | the match position. */ | |
888 | if (search_regs.num_regs == 0) | |
889 | { | |
890 | regoff_t *starts, *ends; | |
891 | ||
892 | starts = | |
893 | (regoff_t *) xmalloc (2 * sizeof (regoff_t)); | |
894 | ends = | |
895 | (regoff_t *) xmalloc (2 * sizeof (regoff_t)); | |
9ac0d9e0 | 896 | BLOCK_INPUT; |
1113d9db JB |
897 | re_set_registers (&searchbuf, |
898 | &search_regs, | |
899 | 2, starts, ends); | |
9ac0d9e0 | 900 | UNBLOCK_INPUT; |
1113d9db JB |
901 | } |
902 | ||
ca1d1d23 JB |
903 | search_regs.start[0] |
904 | = pos + cursor - p2 + ((direction > 0) | |
905 | ? 1 - len : 0); | |
906 | search_regs.end[0] = len + search_regs.start[0]; | |
daa37602 | 907 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
908 | if ((n -= direction) != 0) |
909 | cursor += dirlen; /* to resume search */ | |
910 | else | |
911 | return ((direction > 0) | |
912 | ? search_regs.end[0] : search_regs.start[0]); | |
913 | } | |
914 | else | |
915 | cursor += stride_for_teases; /* <sigh> we lose - */ | |
916 | } | |
917 | pos += cursor - p2; | |
918 | } | |
919 | else | |
920 | /* Now we'll pick up a clump that has to be done the hard */ | |
921 | /* way because it covers a discontinuity */ | |
922 | { | |
923 | limit = ((direction > 0) | |
924 | ? BUFFER_CEILING_OF (pos - dirlen + 1) | |
925 | : BUFFER_FLOOR_OF (pos - dirlen - 1)); | |
926 | limit = ((direction > 0) | |
927 | ? min (limit + len, lim - 1) | |
928 | : max (limit - len, lim)); | |
929 | /* LIMIT is now the last value POS can have | |
930 | and still be valid for a possible match. */ | |
931 | while (1) | |
932 | { | |
933 | /* This loop can be coded for space rather than */ | |
934 | /* speed because it will usually run only once. */ | |
935 | /* (the reach is at most len + 21, and typically */ | |
936 | /* does not exceed len) */ | |
937 | while ((limit - pos) * direction >= 0) | |
938 | pos += BM_tab[FETCH_CHAR(pos)]; | |
939 | /* now run the same tests to distinguish going off the */ | |
eb8c3be9 | 940 | /* end, a match or a phony match. */ |
ca1d1d23 JB |
941 | if ((pos - limit) * direction <= len) |
942 | break; /* ran off the end */ | |
943 | /* Found what might be a match. | |
944 | Set POS back to last (first if reverse) char pos. */ | |
945 | pos -= infinity; | |
946 | i = dirlen - direction; | |
947 | while ((i -= direction) + direction != 0) | |
948 | { | |
949 | pos -= direction; | |
950 | if (pat[i] != (((int) trt) | |
951 | ? trt[FETCH_CHAR(pos)] | |
952 | : FETCH_CHAR (pos))) | |
953 | break; | |
954 | } | |
955 | /* Above loop has moved POS part or all the way | |
956 | back to the first char pos (last char pos if reverse). | |
957 | Set it once again at the last (first if reverse) char. */ | |
958 | pos += dirlen - i- direction; | |
959 | if (i + direction == 0) | |
960 | { | |
961 | pos -= direction; | |
1113d9db JB |
962 | |
963 | /* Make sure we have registers in which to store | |
964 | the match position. */ | |
965 | if (search_regs.num_regs == 0) | |
966 | { | |
967 | regoff_t *starts, *ends; | |
968 | ||
969 | starts = | |
970 | (regoff_t *) xmalloc (2 * sizeof (regoff_t)); | |
971 | ends = | |
972 | (regoff_t *) xmalloc (2 * sizeof (regoff_t)); | |
9ac0d9e0 | 973 | BLOCK_INPUT; |
1113d9db JB |
974 | re_set_registers (&searchbuf, |
975 | &search_regs, | |
976 | 2, starts, ends); | |
9ac0d9e0 | 977 | UNBLOCK_INPUT; |
1113d9db JB |
978 | } |
979 | ||
ca1d1d23 JB |
980 | search_regs.start[0] |
981 | = pos + ((direction > 0) ? 1 - len : 0); | |
982 | search_regs.end[0] = len + search_regs.start[0]; | |
daa37602 | 983 | XSET (last_thing_searched, Lisp_Buffer, current_buffer); |
ca1d1d23 JB |
984 | if ((n -= direction) != 0) |
985 | pos += dirlen; /* to resume search */ | |
986 | else | |
987 | return ((direction > 0) | |
988 | ? search_regs.end[0] : search_regs.start[0]); | |
989 | } | |
990 | else | |
991 | pos += stride_for_teases; | |
992 | } | |
993 | } | |
994 | /* We have done one clump. Can we continue? */ | |
995 | if ((lim - pos) * direction < 0) | |
996 | return ((0 - n) * direction); | |
997 | } | |
998 | return pos; | |
999 | } | |
1000 | } | |
1001 | \f | |
1002 | /* Given a string of words separated by word delimiters, | |
1003 | compute a regexp that matches those exact words | |
1004 | separated by arbitrary punctuation. */ | |
1005 | ||
1006 | static Lisp_Object | |
1007 | wordify (string) | |
1008 | Lisp_Object string; | |
1009 | { | |
1010 | register unsigned char *p, *o; | |
1011 | register int i, len, punct_count = 0, word_count = 0; | |
1012 | Lisp_Object val; | |
1013 | ||
1014 | CHECK_STRING (string, 0); | |
1015 | p = XSTRING (string)->data; | |
1016 | len = XSTRING (string)->size; | |
1017 | ||
1018 | for (i = 0; i < len; i++) | |
1019 | if (SYNTAX (p[i]) != Sword) | |
1020 | { | |
1021 | punct_count++; | |
1022 | if (i > 0 && SYNTAX (p[i-1]) == Sword) word_count++; | |
1023 | } | |
1024 | if (SYNTAX (p[len-1]) == Sword) word_count++; | |
1025 | if (!word_count) return build_string (""); | |
1026 | ||
1027 | val = make_string (p, len - punct_count + 5 * (word_count - 1) + 4); | |
1028 | ||
1029 | o = XSTRING (val)->data; | |
1030 | *o++ = '\\'; | |
1031 | *o++ = 'b'; | |
1032 | ||
1033 | for (i = 0; i < len; i++) | |
1034 | if (SYNTAX (p[i]) == Sword) | |
1035 | *o++ = p[i]; | |
1036 | else if (i > 0 && SYNTAX (p[i-1]) == Sword && --word_count) | |
1037 | { | |
1038 | *o++ = '\\'; | |
1039 | *o++ = 'W'; | |
1040 | *o++ = '\\'; | |
1041 | *o++ = 'W'; | |
1042 | *o++ = '*'; | |
1043 | } | |
1044 | ||
1045 | *o++ = '\\'; | |
1046 | *o++ = 'b'; | |
1047 | ||
1048 | return val; | |
1049 | } | |
1050 | \f | |
1051 | DEFUN ("search-backward", Fsearch_backward, Ssearch_backward, 1, 4, | |
1052 | "sSearch backward: ", | |
1053 | "Search backward from point for STRING.\n\ | |
1054 | Set point to the beginning of the occurrence found, and return point.\n\ | |
1055 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1056 | The match found must not extend before that position.\n\ | |
1057 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1058 | If not nil and not t, position at limit of search and return nil.\n\ | |
1059 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1060 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
1061 | (string, bound, noerror, count) | |
1062 | Lisp_Object string, bound, noerror, count; | |
1063 | { | |
1064 | return search_command (string, bound, noerror, count, -1, 0); | |
1065 | } | |
1066 | ||
1067 | DEFUN ("search-forward", Fsearch_forward, Ssearch_forward, 1, 4, "sSearch: ", | |
1068 | "Search forward from point for STRING.\n\ | |
1069 | Set point to the end of the occurrence found, and return point.\n\ | |
1070 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1071 | The match found must not extend after that position. nil is equivalent\n\ | |
1072 | to (point-max).\n\ | |
1073 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1074 | If not nil and not t, move to limit of search and return nil.\n\ | |
1075 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1076 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
1077 | (string, bound, noerror, count) | |
1078 | Lisp_Object string, bound, noerror, count; | |
1079 | { | |
1080 | return search_command (string, bound, noerror, count, 1, 0); | |
1081 | } | |
1082 | ||
1083 | DEFUN ("word-search-backward", Fword_search_backward, Sword_search_backward, 1, 4, | |
1084 | "sWord search backward: ", | |
1085 | "Search backward from point for STRING, ignoring differences in punctuation.\n\ | |
1086 | Set point to the beginning of the occurrence found, and return point.\n\ | |
1087 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1088 | The match found must not extend before that position.\n\ | |
1089 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1090 | If not nil and not t, move to limit of search and return nil.\n\ | |
1091 | Optional fourth argument is repeat count--search for successive occurrences.") | |
1092 | (string, bound, noerror, count) | |
1093 | Lisp_Object string, bound, noerror, count; | |
1094 | { | |
1095 | return search_command (wordify (string), bound, noerror, count, -1, 1); | |
1096 | } | |
1097 | ||
1098 | DEFUN ("word-search-forward", Fword_search_forward, Sword_search_forward, 1, 4, | |
1099 | "sWord search: ", | |
1100 | "Search forward from point for STRING, ignoring differences in punctuation.\n\ | |
1101 | Set point to the end of the occurrence found, and return point.\n\ | |
1102 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1103 | The match found must not extend after that position.\n\ | |
1104 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1105 | If not nil and not t, move to limit of search and return nil.\n\ | |
1106 | Optional fourth argument is repeat count--search for successive occurrences.") | |
1107 | (string, bound, noerror, count) | |
1108 | Lisp_Object string, bound, noerror, count; | |
1109 | { | |
1110 | return search_command (wordify (string), bound, noerror, count, 1, 1); | |
1111 | } | |
1112 | ||
1113 | DEFUN ("re-search-backward", Fre_search_backward, Sre_search_backward, 1, 4, | |
1114 | "sRE search backward: ", | |
1115 | "Search backward from point for match for regular expression REGEXP.\n\ | |
1116 | Set point to the beginning of the match, and return point.\n\ | |
1117 | The match found is the one starting last in the buffer\n\ | |
1118 | and yet ending before the place the origin of the search.\n\ | |
1119 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1120 | The match found must start at or after that position.\n\ | |
1121 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1122 | If not nil and not t, move to limit of search and return nil.\n\ | |
1123 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1124 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
1125 | (string, bound, noerror, count) | |
1126 | Lisp_Object string, bound, noerror, count; | |
1127 | { | |
1128 | return search_command (string, bound, noerror, count, -1, 1); | |
1129 | } | |
1130 | ||
1131 | DEFUN ("re-search-forward", Fre_search_forward, Sre_search_forward, 1, 4, | |
1132 | "sRE search: ", | |
1133 | "Search forward from point for regular expression REGEXP.\n\ | |
1134 | Set point to the end of the occurrence found, and return point.\n\ | |
1135 | An optional second argument bounds the search; it is a buffer position.\n\ | |
1136 | The match found must not extend after that position.\n\ | |
1137 | Optional third argument, if t, means if fail just return nil (no error).\n\ | |
1138 | If not nil and not t, move to limit of search and return nil.\n\ | |
1139 | Optional fourth argument is repeat count--search for successive occurrences.\n\ | |
1140 | See also the functions `match-beginning', `match-end' and `replace-match'.") | |
1141 | (string, bound, noerror, count) | |
1142 | Lisp_Object string, bound, noerror, count; | |
1143 | { | |
1144 | return search_command (string, bound, noerror, count, 1, 1); | |
1145 | } | |
1146 | \f | |
1147 | DEFUN ("replace-match", Freplace_match, Sreplace_match, 1, 3, 0, | |
1148 | "Replace text matched by last search with NEWTEXT.\n\ | |
1149 | If second arg FIXEDCASE is non-nil, do not alter case of replacement text.\n\ | |
1150 | Otherwise convert to all caps or cap initials, like replaced text.\n\ | |
1151 | If third arg LITERAL is non-nil, insert NEWTEXT literally.\n\ | |
1152 | Otherwise treat `\\' as special:\n\ | |
1153 | `\\&' in NEWTEXT means substitute original matched text.\n\ | |
1154 | `\\N' means substitute what matched the Nth `\\(...\\)'.\n\ | |
1155 | If Nth parens didn't match, substitute nothing.\n\ | |
1156 | `\\\\' means insert one `\\'.\n\ | |
1113d9db | 1157 | FIXEDCASE and LITERAL are optional arguments.\n\ |
ca1d1d23 | 1158 | Leaves point at end of replacement text.") |
16fdc568 BF |
1159 | (newtext, fixedcase, literal) |
1160 | Lisp_Object newtext, fixedcase, literal; | |
ca1d1d23 JB |
1161 | { |
1162 | enum { nochange, all_caps, cap_initial } case_action; | |
1163 | register int pos, last; | |
1164 | int some_multiletter_word; | |
97832bd0 RS |
1165 | int some_lowercase; |
1166 | int some_uppercase_initial; | |
ca1d1d23 JB |
1167 | register int c, prevc; |
1168 | int inslen; | |
1169 | ||
16fdc568 | 1170 | CHECK_STRING (newtext, 0); |
ca1d1d23 JB |
1171 | |
1172 | case_action = nochange; /* We tried an initialization */ | |
1173 | /* but some C compilers blew it */ | |
4746118a JB |
1174 | |
1175 | if (search_regs.num_regs <= 0) | |
1176 | error ("replace-match called before any match found"); | |
1177 | ||
ca1d1d23 JB |
1178 | if (search_regs.start[0] < BEGV |
1179 | || search_regs.start[0] > search_regs.end[0] | |
1180 | || search_regs.end[0] > ZV) | |
97832bd0 RS |
1181 | args_out_of_range (make_number (search_regs.start[0]), |
1182 | make_number (search_regs.end[0])); | |
ca1d1d23 JB |
1183 | |
1184 | if (NILP (fixedcase)) | |
1185 | { | |
1186 | /* Decide how to casify by examining the matched text. */ | |
1187 | ||
1188 | last = search_regs.end[0]; | |
1189 | prevc = '\n'; | |
1190 | case_action = all_caps; | |
1191 | ||
1192 | /* some_multiletter_word is set nonzero if any original word | |
1193 | is more than one letter long. */ | |
1194 | some_multiletter_word = 0; | |
97832bd0 RS |
1195 | some_lowercase = 0; |
1196 | some_uppercase_initial = 0; | |
ca1d1d23 JB |
1197 | |
1198 | for (pos = search_regs.start[0]; pos < last; pos++) | |
1199 | { | |
1200 | c = FETCH_CHAR (pos); | |
1201 | if (LOWERCASEP (c)) | |
1202 | { | |
1203 | /* Cannot be all caps if any original char is lower case */ | |
1204 | ||
97832bd0 | 1205 | some_lowercase = 1; |
ca1d1d23 | 1206 | if (SYNTAX (prevc) != Sword) |
97832bd0 | 1207 | ; |
ca1d1d23 JB |
1208 | else |
1209 | some_multiletter_word = 1; | |
1210 | } | |
1211 | else if (!NOCASEP (c)) | |
1212 | { | |
97832bd0 RS |
1213 | if (SYNTAX (prevc) != Sword) |
1214 | some_uppercase_initial = 1; | |
1215 | else | |
ca1d1d23 JB |
1216 | some_multiletter_word = 1; |
1217 | } | |
1218 | ||
1219 | prevc = c; | |
1220 | } | |
1221 | ||
97832bd0 RS |
1222 | /* Convert to all caps if the old text is all caps |
1223 | and has at least one multiletter word. */ | |
1224 | if (! some_lowercase && some_multiletter_word) | |
1225 | case_action = all_caps; | |
1226 | /* Capitalize each word, if the old text has a capitalized word. */ | |
1227 | else if (some_uppercase_initial) | |
ca1d1d23 | 1228 | case_action = cap_initial; |
97832bd0 RS |
1229 | else |
1230 | case_action = nochange; | |
ca1d1d23 JB |
1231 | } |
1232 | ||
9a76659d JB |
1233 | /* We insert the replacement text before the old text, and then |
1234 | delete the original text. This means that markers at the | |
1235 | beginning or end of the original will float to the corresponding | |
1236 | position in the replacement. */ | |
1237 | SET_PT (search_regs.start[0]); | |
ca1d1d23 | 1238 | if (!NILP (literal)) |
16fdc568 | 1239 | Finsert_and_inherit (1, &newtext); |
ca1d1d23 JB |
1240 | else |
1241 | { | |
1242 | struct gcpro gcpro1; | |
16fdc568 | 1243 | GCPRO1 (newtext); |
ca1d1d23 | 1244 | |
16fdc568 | 1245 | for (pos = 0; pos < XSTRING (newtext)->size; pos++) |
ca1d1d23 | 1246 | { |
9a76659d JB |
1247 | int offset = point - search_regs.start[0]; |
1248 | ||
16fdc568 | 1249 | c = XSTRING (newtext)->data[pos]; |
ca1d1d23 JB |
1250 | if (c == '\\') |
1251 | { | |
16fdc568 | 1252 | c = XSTRING (newtext)->data[++pos]; |
ca1d1d23 | 1253 | if (c == '&') |
9a76659d JB |
1254 | Finsert_buffer_substring |
1255 | (Fcurrent_buffer (), | |
1256 | make_number (search_regs.start[0] + offset), | |
1257 | make_number (search_regs.end[0] + offset)); | |
4746118a | 1258 | else if (c >= '1' && c <= search_regs.num_regs + '0') |
ca1d1d23 JB |
1259 | { |
1260 | if (search_regs.start[c - '0'] >= 1) | |
9a76659d JB |
1261 | Finsert_buffer_substring |
1262 | (Fcurrent_buffer (), | |
1263 | make_number (search_regs.start[c - '0'] + offset), | |
1264 | make_number (search_regs.end[c - '0'] + offset)); | |
ca1d1d23 JB |
1265 | } |
1266 | else | |
1267 | insert_char (c); | |
1268 | } | |
1269 | else | |
1270 | insert_char (c); | |
1271 | } | |
1272 | UNGCPRO; | |
1273 | } | |
1274 | ||
9a76659d JB |
1275 | inslen = point - (search_regs.start[0]); |
1276 | del_range (search_regs.start[0] + inslen, search_regs.end[0] + inslen); | |
ca1d1d23 JB |
1277 | |
1278 | if (case_action == all_caps) | |
1279 | Fupcase_region (make_number (point - inslen), make_number (point)); | |
1280 | else if (case_action == cap_initial) | |
1281 | upcase_initials_region (make_number (point - inslen), make_number (point)); | |
1282 | return Qnil; | |
1283 | } | |
1284 | \f | |
1285 | static Lisp_Object | |
1286 | match_limit (num, beginningp) | |
1287 | Lisp_Object num; | |
1288 | int beginningp; | |
1289 | { | |
1290 | register int n; | |
1291 | ||
1292 | CHECK_NUMBER (num, 0); | |
1293 | n = XINT (num); | |
4746118a JB |
1294 | if (n < 0 || n >= search_regs.num_regs) |
1295 | args_out_of_range (num, make_number (search_regs.num_regs)); | |
1296 | if (search_regs.num_regs <= 0 | |
1297 | || search_regs.start[n] < 0) | |
ca1d1d23 JB |
1298 | return Qnil; |
1299 | return (make_number ((beginningp) ? search_regs.start[n] | |
1300 | : search_regs.end[n])); | |
1301 | } | |
1302 | ||
1303 | DEFUN ("match-beginning", Fmatch_beginning, Smatch_beginning, 1, 1, 0, | |
1304 | "Return position of start of text matched by last search.\n\ | |
16fdc568 BF |
1305 | NUM specifies which parenthesized expression in the last regexp.\n\ |
1306 | Value is nil if NUMth pair didn't match, or there were less than NUM pairs.\n\ | |
ca1d1d23 JB |
1307 | Zero means the entire text matched by the whole regexp or whole string.") |
1308 | (num) | |
1309 | Lisp_Object num; | |
1310 | { | |
1311 | return match_limit (num, 1); | |
1312 | } | |
1313 | ||
1314 | DEFUN ("match-end", Fmatch_end, Smatch_end, 1, 1, 0, | |
1315 | "Return position of end of text matched by last search.\n\ | |
1316 | ARG, a number, specifies which parenthesized expression in the last regexp.\n\ | |
1317 | Value is nil if ARGth pair didn't match, or there were less than ARG pairs.\n\ | |
1318 | Zero means the entire text matched by the whole regexp or whole string.") | |
1319 | (num) | |
1320 | Lisp_Object num; | |
1321 | { | |
1322 | return match_limit (num, 0); | |
1323 | } | |
1324 | ||
1325 | DEFUN ("match-data", Fmatch_data, Smatch_data, 0, 0, 0, | |
1326 | "Return a list containing all info on what the last search matched.\n\ | |
1327 | Element 2N is `(match-beginning N)'; element 2N + 1 is `(match-end N)'.\n\ | |
1328 | All the elements are markers or nil (nil if the Nth pair didn't match)\n\ | |
1329 | if the last match was on a buffer; integers or nil if a string was matched.\n\ | |
1330 | Use `store-match-data' to reinstate the data in this list.") | |
1331 | () | |
1332 | { | |
4746118a | 1333 | Lisp_Object *data; |
ca1d1d23 JB |
1334 | int i, len; |
1335 | ||
daa37602 JB |
1336 | if (NILP (last_thing_searched)) |
1337 | error ("match-data called before any match found"); | |
1338 | ||
4746118a JB |
1339 | data = (Lisp_Object *) alloca ((2 * search_regs.num_regs) |
1340 | * sizeof (Lisp_Object)); | |
1341 | ||
ca1d1d23 | 1342 | len = -1; |
4746118a | 1343 | for (i = 0; i < search_regs.num_regs; i++) |
ca1d1d23 JB |
1344 | { |
1345 | int start = search_regs.start[i]; | |
1346 | if (start >= 0) | |
1347 | { | |
daa37602 | 1348 | if (EQ (last_thing_searched, Qt)) |
ca1d1d23 JB |
1349 | { |
1350 | XFASTINT (data[2 * i]) = start; | |
1351 | XFASTINT (data[2 * i + 1]) = search_regs.end[i]; | |
1352 | } | |
daa37602 | 1353 | else if (XTYPE (last_thing_searched) == Lisp_Buffer) |
ca1d1d23 JB |
1354 | { |
1355 | data[2 * i] = Fmake_marker (); | |
daa37602 JB |
1356 | Fset_marker (data[2 * i], |
1357 | make_number (start), | |
1358 | last_thing_searched); | |
ca1d1d23 JB |
1359 | data[2 * i + 1] = Fmake_marker (); |
1360 | Fset_marker (data[2 * i + 1], | |
daa37602 JB |
1361 | make_number (search_regs.end[i]), |
1362 | last_thing_searched); | |
ca1d1d23 | 1363 | } |
daa37602 JB |
1364 | else |
1365 | /* last_thing_searched must always be Qt, a buffer, or Qnil. */ | |
1366 | abort (); | |
1367 | ||
ca1d1d23 JB |
1368 | len = i; |
1369 | } | |
1370 | else | |
1371 | data[2 * i] = data [2 * i + 1] = Qnil; | |
1372 | } | |
1373 | return Flist (2 * len + 2, data); | |
1374 | } | |
1375 | ||
1376 | ||
1377 | DEFUN ("store-match-data", Fstore_match_data, Sstore_match_data, 1, 1, 0, | |
1378 | "Set internal data on last search match from elements of LIST.\n\ | |
1379 | LIST should have been created by calling `match-data' previously.") | |
1380 | (list) | |
1381 | register Lisp_Object list; | |
1382 | { | |
1383 | register int i; | |
1384 | register Lisp_Object marker; | |
1385 | ||
1386 | if (!CONSP (list) && !NILP (list)) | |
b37902c8 | 1387 | list = wrong_type_argument (Qconsp, list); |
ca1d1d23 | 1388 | |
daa37602 JB |
1389 | /* Unless we find a marker with a buffer in LIST, assume that this |
1390 | match data came from a string. */ | |
1391 | last_thing_searched = Qt; | |
1392 | ||
4746118a JB |
1393 | /* Allocate registers if they don't already exist. */ |
1394 | { | |
d084e942 | 1395 | int length = XFASTINT (Flength (list)) / 2; |
4746118a JB |
1396 | |
1397 | if (length > search_regs.num_regs) | |
1398 | { | |
1113d9db JB |
1399 | if (search_regs.num_regs == 0) |
1400 | { | |
1401 | search_regs.start | |
1402 | = (regoff_t *) xmalloc (length * sizeof (regoff_t)); | |
1403 | search_regs.end | |
1404 | = (regoff_t *) xmalloc (length * sizeof (regoff_t)); | |
1405 | } | |
4746118a | 1406 | else |
1113d9db JB |
1407 | { |
1408 | search_regs.start | |
1409 | = (regoff_t *) xrealloc (search_regs.start, | |
1410 | length * sizeof (regoff_t)); | |
1411 | search_regs.end | |
1412 | = (regoff_t *) xrealloc (search_regs.end, | |
1413 | length * sizeof (regoff_t)); | |
1414 | } | |
4746118a | 1415 | |
9ac0d9e0 | 1416 | BLOCK_INPUT; |
1113d9db JB |
1417 | re_set_registers (&searchbuf, &search_regs, length, |
1418 | search_regs.start, search_regs.end); | |
9ac0d9e0 | 1419 | UNBLOCK_INPUT; |
4746118a JB |
1420 | } |
1421 | } | |
1422 | ||
1423 | for (i = 0; i < search_regs.num_regs; i++) | |
ca1d1d23 JB |
1424 | { |
1425 | marker = Fcar (list); | |
1426 | if (NILP (marker)) | |
1427 | { | |
1428 | search_regs.start[i] = -1; | |
1429 | list = Fcdr (list); | |
1430 | } | |
1431 | else | |
1432 | { | |
daa37602 JB |
1433 | if (XTYPE (marker) == Lisp_Marker) |
1434 | { | |
1435 | if (XMARKER (marker)->buffer == 0) | |
1436 | XFASTINT (marker) = 0; | |
1437 | else | |
1438 | XSET (last_thing_searched, Lisp_Buffer, | |
1439 | XMARKER (marker)->buffer); | |
1440 | } | |
ca1d1d23 JB |
1441 | |
1442 | CHECK_NUMBER_COERCE_MARKER (marker, 0); | |
1443 | search_regs.start[i] = XINT (marker); | |
1444 | list = Fcdr (list); | |
1445 | ||
1446 | marker = Fcar (list); | |
1447 | if (XTYPE (marker) == Lisp_Marker | |
1448 | && XMARKER (marker)->buffer == 0) | |
1449 | XFASTINT (marker) = 0; | |
1450 | ||
1451 | CHECK_NUMBER_COERCE_MARKER (marker, 0); | |
1452 | search_regs.end[i] = XINT (marker); | |
1453 | } | |
1454 | list = Fcdr (list); | |
1455 | } | |
1456 | ||
1457 | return Qnil; | |
1458 | } | |
1459 | ||
1460 | /* Quote a string to inactivate reg-expr chars */ | |
1461 | ||
1462 | DEFUN ("regexp-quote", Fregexp_quote, Sregexp_quote, 1, 1, 0, | |
1463 | "Return a regexp string which matches exactly STRING and nothing else.") | |
1464 | (str) | |
1465 | Lisp_Object str; | |
1466 | { | |
1467 | register unsigned char *in, *out, *end; | |
1468 | register unsigned char *temp; | |
1469 | ||
1470 | CHECK_STRING (str, 0); | |
1471 | ||
1472 | temp = (unsigned char *) alloca (XSTRING (str)->size * 2); | |
1473 | ||
1474 | /* Now copy the data into the new string, inserting escapes. */ | |
1475 | ||
1476 | in = XSTRING (str)->data; | |
1477 | end = in + XSTRING (str)->size; | |
1478 | out = temp; | |
1479 | ||
1480 | for (; in != end; in++) | |
1481 | { | |
1482 | if (*in == '[' || *in == ']' | |
1483 | || *in == '*' || *in == '.' || *in == '\\' | |
1484 | || *in == '?' || *in == '+' | |
1485 | || *in == '^' || *in == '$') | |
1486 | *out++ = '\\'; | |
1487 | *out++ = *in; | |
1488 | } | |
1489 | ||
1490 | return make_string (temp, out - temp); | |
1491 | } | |
1492 | \f | |
1493 | syms_of_search () | |
1494 | { | |
1495 | register int i; | |
1496 | ||
1497 | searchbuf.allocated = 100; | |
8c0e7b73 | 1498 | searchbuf.buffer = (unsigned char *) malloc (searchbuf.allocated); |
ca1d1d23 JB |
1499 | searchbuf.fastmap = search_fastmap; |
1500 | ||
1501 | Qsearch_failed = intern ("search-failed"); | |
1502 | staticpro (&Qsearch_failed); | |
1503 | Qinvalid_regexp = intern ("invalid-regexp"); | |
1504 | staticpro (&Qinvalid_regexp); | |
1505 | ||
1506 | Fput (Qsearch_failed, Qerror_conditions, | |
1507 | Fcons (Qsearch_failed, Fcons (Qerror, Qnil))); | |
1508 | Fput (Qsearch_failed, Qerror_message, | |
1509 | build_string ("Search failed")); | |
1510 | ||
1511 | Fput (Qinvalid_regexp, Qerror_conditions, | |
1512 | Fcons (Qinvalid_regexp, Fcons (Qerror, Qnil))); | |
1513 | Fput (Qinvalid_regexp, Qerror_message, | |
1514 | build_string ("Invalid regexp")); | |
1515 | ||
1516 | last_regexp = Qnil; | |
1517 | staticpro (&last_regexp); | |
1518 | ||
daa37602 JB |
1519 | last_thing_searched = Qnil; |
1520 | staticpro (&last_thing_searched); | |
1521 | ||
ca1d1d23 JB |
1522 | defsubr (&Sstring_match); |
1523 | defsubr (&Slooking_at); | |
1524 | defsubr (&Sskip_chars_forward); | |
1525 | defsubr (&Sskip_chars_backward); | |
17431c60 RS |
1526 | defsubr (&Sskip_syntax_forward); |
1527 | defsubr (&Sskip_syntax_backward); | |
ca1d1d23 JB |
1528 | defsubr (&Ssearch_forward); |
1529 | defsubr (&Ssearch_backward); | |
1530 | defsubr (&Sword_search_forward); | |
1531 | defsubr (&Sword_search_backward); | |
1532 | defsubr (&Sre_search_forward); | |
1533 | defsubr (&Sre_search_backward); | |
1534 | defsubr (&Sreplace_match); | |
1535 | defsubr (&Smatch_beginning); | |
1536 | defsubr (&Smatch_end); | |
1537 | defsubr (&Smatch_data); | |
1538 | defsubr (&Sstore_match_data); | |
1539 | defsubr (&Sregexp_quote); | |
1540 | } |