diff options
author | Mark Whitley <markw@lineo.com> | 2000-07-10 23:07:54 +0000 |
---|---|---|
committer | Mark Whitley <markw@lineo.com> | 2000-07-10 23:07:54 +0000 |
commit | cc54d12ef716e9206344c18006b46386fe49ba9b (patch) | |
tree | c2f88581d8e81aef6de5010941380e1cb872f463 | |
parent | dd23b8bb431d27ea95534f528a1e1381e5841b8b (diff) | |
download | busybox-w32-cc54d12ef716e9206344c18006b46386fe49ba9b.tar.gz busybox-w32-cc54d12ef716e9206344c18006b46386fe49ba9b.tar.bz2 busybox-w32-cc54d12ef716e9206344c18006b46386fe49ba9b.zip |
Removed all trace of hand-tooled regexp routines. Bye bye, baby.
-rw-r--r-- | Makefile | 2 | ||||
-rw-r--r-- | regexp.c | 789 | ||||
-rw-r--r-- | regexp.h | 33 |
3 files changed, 1 insertions, 823 deletions
@@ -142,7 +142,7 @@ busybox: $(OBJECTS) | |||
142 | busybox.links: busybox.def.h | 142 | busybox.links: busybox.def.h |
143 | - ./busybox.mkll | sort >$@ | 143 | - ./busybox.mkll | sort >$@ |
144 | 144 | ||
145 | regexp.o nfsmount.o cmdedit.o: %.o: %.h | 145 | nfsmount.o cmdedit.o: %.o: %.h |
146 | $(OBJECTS): %.o: busybox.def.h internal.h %.c Makefile | 146 | $(OBJECTS): %.o: busybox.def.h internal.h %.c Makefile |
147 | 147 | ||
148 | test tests: | 148 | test tests: |
diff --git a/regexp.c b/regexp.c deleted file mode 100644 index 6fedb01ba..000000000 --- a/regexp.c +++ /dev/null | |||
@@ -1,789 +0,0 @@ | |||
1 | /* vi: set sw=4 ts=4: */ | ||
2 | /* regexp.c */ | ||
3 | |||
4 | #include "internal.h" | ||
5 | #include "regexp.h" | ||
6 | #include <setjmp.h> | ||
7 | #include <stdio.h> | ||
8 | #include <ctype.h> | ||
9 | |||
10 | |||
11 | #define NSUBEXP 10 | ||
12 | typedef struct regexp { | ||
13 | char *startp[NSUBEXP]; | ||
14 | char *endp[NSUBEXP]; | ||
15 | int minlen; /* length of shortest possible match */ | ||
16 | char first; /* first character, if known; else \0 */ | ||
17 | char bol; /* boolean: must start at beginning of line? */ | ||
18 | char program[1]; /* Unwarranted chumminess with compiler. */ | ||
19 | } regexp; | ||
20 | |||
21 | |||
22 | static regexp *regcomp(char* text); | ||
23 | static int regexec(struct regexp* re, char* str, int bol, int ignoreCase); | ||
24 | static void regsub(struct regexp* re, char* src, char* dst); | ||
25 | |||
26 | #if ( defined BB_GREP || defined BB_SED) | ||
27 | |||
28 | /* This also tries to find a needle in a haystack, but uses | ||
29 | * real regular expressions.... The fake regular expression | ||
30 | * version of find_match lives in utility.c. Using this version | ||
31 | * will add 3.9k to busybox... | ||
32 | * -Erik Andersen | ||
33 | */ | ||
34 | extern int find_match(char *haystack, char *needle, int ignoreCase) | ||
35 | { | ||
36 | int status; | ||
37 | struct regexp *re; | ||
38 | |||
39 | re = regcomp(needle); | ||
40 | status = regexec(re, haystack, FALSE, ignoreCase); | ||
41 | free(re); | ||
42 | return (status); | ||
43 | } | ||
44 | |||
45 | #if defined BB_SED | ||
46 | /* This performs substitutions after a regexp match has been found. | ||
47 | * The new string is returned. It is malloc'ed, and do must be freed. */ | ||
48 | extern int replace_match(char *haystack, char *needle, char *newNeedle, | ||
49 | int ignoreCase) | ||
50 | { | ||
51 | int status; | ||
52 | struct regexp *re; | ||
53 | char *s, buf[BUF_SIZE], *d = buf; | ||
54 | |||
55 | re = regcomp(needle); | ||
56 | status = regexec(re, haystack, FALSE, ignoreCase); | ||
57 | if (status == TRUE) { | ||
58 | s = haystack; | ||
59 | |||
60 | do { | ||
61 | /* copy stuff from before the match */ | ||
62 | while (s < re->startp[0]) | ||
63 | *d++ = *s++; | ||
64 | /* substitute for the matched part */ | ||
65 | regsub(re, newNeedle, d); | ||
66 | s = re->endp[0]; | ||
67 | d += strlen(d); | ||
68 | } while (regexec(re, s, FALSE, ignoreCase) == TRUE); | ||
69 | /* copy stuff from after the match */ | ||
70 | while ((*d++ = *s++)) { | ||
71 | } | ||
72 | d[0] = '\0'; | ||
73 | strcpy(haystack, buf); | ||
74 | } | ||
75 | free(re); | ||
76 | return (status); | ||
77 | } | ||
78 | #endif | ||
79 | |||
80 | |||
81 | /* code swiped from elvis-tiny 1.4 (a clone of vi) and adjusted to | ||
82 | * suit the needs of busybox by Erik Andersen. | ||
83 | * | ||
84 | * From the README: | ||
85 | * "Elvis is freely redistributable, in either source form or executable form. | ||
86 | * There are no restrictions on how you may use it". | ||
87 | * Elvis was written by Steve Kirkendall <kirkenda@cs.pdx.edu> | ||
88 | * | ||
89 | * | ||
90 | * This file contains the code that compiles regular expressions and executes | ||
91 | * them. It supports the same syntax and features as vi's regular expression | ||
92 | * code. Specifically, the meta characters are: | ||
93 | * ^ matches the beginning of a line | ||
94 | * $ matches the end of a line | ||
95 | * \< matches the beginning of a word | ||
96 | * \> matches the end of a word | ||
97 | * . matches any single character | ||
98 | * [] matches any character in a character class | ||
99 | * \( delimits the start of a subexpression | ||
100 | * \) delimits the end of a subexpression | ||
101 | * * repeats the preceding 0 or more times | ||
102 | * NOTE: You cannot follow a \) with a *. | ||
103 | * | ||
104 | * The physical structure of a compiled RE is as follows: | ||
105 | * - First, there is a one-byte value that says how many character classes | ||
106 | * are used in this regular expression | ||
107 | * - Next, each character class is stored as a bitmap that is 256 bits | ||
108 | * (32 bytes) long. | ||
109 | * - A mixture of literal characters and compiled meta characters follows. | ||
110 | * This begins with M_BEGIN(0) and ends with M_END(0). All meta chars | ||
111 | * are stored as a \n followed by a one-byte code, so they take up two | ||
112 | * bytes apiece. Literal characters take up one byte apiece. \n can't | ||
113 | * be used as a literal character. | ||
114 | * | ||
115 | */ | ||
116 | |||
117 | |||
118 | |||
119 | static char *previous; /* the previous regexp, used when null regexp is given */ | ||
120 | |||
121 | #if defined BB_SED | ||
122 | static char *previous1; /* a copy of the text from the previous substitution for regsub() */ | ||
123 | #endif | ||
124 | |||
125 | |||
126 | /* These are used to classify or recognize meta-characters */ | ||
127 | #define META '\0' | ||
128 | #define BASE_META(m) ((m) - 256) | ||
129 | #define INT_META(c) ((c) + 256) | ||
130 | #define IS_META(m) ((m) >= 256) | ||
131 | #define IS_CLASS(m) ((m) >= M_CLASS(0) && (m) <= M_CLASS(9)) | ||
132 | #define IS_START(m) ((m) >= M_START(0) && (m) <= M_START(9)) | ||
133 | #define IS_END(m) ((m) >= M_END(0) && (m) <= M_END(9)) | ||
134 | #define IS_CLOSURE(m) ((m) >= M_SPLAT && (m) <= M_QMARK) | ||
135 | #define ADD_META(s,m) (*(s)++ = META, *(s)++ = BASE_META(m)) | ||
136 | #define GET_META(s) (*(s) == META ? INT_META(*++(s)) : *s) | ||
137 | |||
138 | /* These are the internal codes used for each type of meta-character */ | ||
139 | #define M_BEGLINE 256 /* internal code for ^ */ | ||
140 | #define M_ENDLINE 257 /* internal code for $ */ | ||
141 | #define M_BEGWORD 258 /* internal code for \< */ | ||
142 | #define M_ENDWORD 259 /* internal code for \> */ | ||
143 | #define M_ANY 260 /* internal code for . */ | ||
144 | #define M_SPLAT 261 /* internal code for * */ | ||
145 | #define M_PLUS 262 /* internal code for \+ */ | ||
146 | #define M_QMARK 263 /* internal code for \? */ | ||
147 | #define M_CLASS(n) (264+(n)) /* internal code for [] */ | ||
148 | #define M_START(n) (274+(n)) /* internal code for \( */ | ||
149 | #define M_END(n) (284+(n)) /* internal code for \) */ | ||
150 | |||
151 | /* These are used during compilation */ | ||
152 | static int class_cnt; /* used to assign class IDs */ | ||
153 | static int start_cnt; /* used to assign start IDs */ | ||
154 | static int end_stk[NSUBEXP]; /* used to assign end IDs */ | ||
155 | static int end_sp; | ||
156 | static char *retext; /* points to the text being compiled */ | ||
157 | |||
158 | /* error-handling stuff */ | ||
159 | jmp_buf errorhandler; | ||
160 | |||
161 | #define FAIL(why) do {fprintf(stderr, why); longjmp(errorhandler, 1);} while (0) | ||
162 | |||
163 | |||
164 | |||
165 | |||
166 | /* This function builds a bitmap for a particular class */ | ||
167 | /* text -- start of the class */ | ||
168 | /* bmap -- the bitmap */ | ||
169 | static char *makeclass(char *text, char *bmap) | ||
170 | { | ||
171 | int i; | ||
172 | int complement = 0; | ||
173 | |||
174 | |||
175 | /* zero the bitmap */ | ||
176 | for (i = 0; bmap && i < 32; i++) { | ||
177 | bmap[i] = 0; | ||
178 | } | ||
179 | |||
180 | /* see if we're going to complement this class */ | ||
181 | if (*text == '^') { | ||
182 | text++; | ||
183 | complement = 1; | ||
184 | } | ||
185 | |||
186 | /* add in the characters */ | ||
187 | while (*text && *text != ']') { | ||
188 | /* is this a span of characters? */ | ||
189 | if (text[1] == '-' && text[2]) { | ||
190 | /* spans can't be backwards */ | ||
191 | if (text[0] > text[2]) { | ||
192 | FAIL("Backwards span in []"); | ||
193 | } | ||
194 | |||
195 | /* add each character in the span to the bitmap */ | ||
196 | for (i = text[0]; bmap && i <= text[2]; i++) { | ||
197 | bmap[i >> 3] |= (1 << (i & 7)); | ||
198 | } | ||
199 | |||
200 | /* move past this span */ | ||
201 | text += 3; | ||
202 | } else { | ||
203 | /* add this single character to the span */ | ||
204 | i = *text++; | ||
205 | if (bmap) { | ||
206 | bmap[i >> 3] |= (1 << (i & 7)); | ||
207 | } | ||
208 | } | ||
209 | } | ||
210 | |||
211 | /* make sure the closing ] is missing */ | ||
212 | if (*text++ != ']') { | ||
213 | FAIL("] missing"); | ||
214 | } | ||
215 | |||
216 | /* if we're supposed to complement this class, then do so */ | ||
217 | if (complement && bmap) { | ||
218 | for (i = 0; i < 32; i++) { | ||
219 | bmap[i] = ~bmap[i]; | ||
220 | } | ||
221 | } | ||
222 | |||
223 | return text; | ||
224 | } | ||
225 | |||
226 | |||
227 | |||
228 | |||
229 | /* This function gets the next character or meta character from a string. | ||
230 | * The pointer is incremented by 1, or by 2 for \-quoted characters. For [], | ||
231 | * a bitmap is generated via makeclass() (if re is given), and the | ||
232 | * character-class text is skipped. | ||
233 | */ | ||
234 | static int gettoken(sptr, re) | ||
235 | char **sptr; | ||
236 | regexp *re; | ||
237 | { | ||
238 | int c; | ||
239 | |||
240 | c = **sptr; | ||
241 | ++*sptr; | ||
242 | if (c == '\\') { | ||
243 | c = **sptr; | ||
244 | ++*sptr; | ||
245 | switch (c) { | ||
246 | case '<': | ||
247 | return M_BEGWORD; | ||
248 | |||
249 | case '>': | ||
250 | return M_ENDWORD; | ||
251 | |||
252 | case '(': | ||
253 | if (start_cnt >= NSUBEXP) { | ||
254 | FAIL("Too many \\(s"); | ||
255 | } | ||
256 | end_stk[end_sp++] = start_cnt; | ||
257 | return M_START(start_cnt++); | ||
258 | |||
259 | case ')': | ||
260 | if (end_sp <= 0) { | ||
261 | FAIL("Mismatched \\)"); | ||
262 | } | ||
263 | return M_END(end_stk[--end_sp]); | ||
264 | |||
265 | case '*': | ||
266 | return M_SPLAT; | ||
267 | |||
268 | case '.': | ||
269 | return M_ANY; | ||
270 | |||
271 | case '+': | ||
272 | return M_PLUS; | ||
273 | |||
274 | case '?': | ||
275 | return M_QMARK; | ||
276 | |||
277 | default: | ||
278 | return c; | ||
279 | } | ||
280 | } else { | ||
281 | switch (c) { | ||
282 | case '^': | ||
283 | if (*sptr == retext + 1) { | ||
284 | return M_BEGLINE; | ||
285 | } | ||
286 | return c; | ||
287 | |||
288 | case '$': | ||
289 | if (!**sptr) { | ||
290 | return M_ENDLINE; | ||
291 | } | ||
292 | return c; | ||
293 | |||
294 | case '.': | ||
295 | return M_ANY; | ||
296 | |||
297 | case '*': | ||
298 | return M_SPLAT; | ||
299 | |||
300 | case '[': | ||
301 | /* make sure we don't have too many classes */ | ||
302 | if (class_cnt >= 10) { | ||
303 | FAIL("Too many []s"); | ||
304 | } | ||
305 | |||
306 | /* process the character list for this class */ | ||
307 | if (re) { | ||
308 | /* generate the bitmap for this class */ | ||
309 | *sptr = makeclass(*sptr, re->program + 1 + 32 * class_cnt); | ||
310 | } else { | ||
311 | /* skip to end of the class */ | ||
312 | *sptr = makeclass(*sptr, (char *) 0); | ||
313 | } | ||
314 | return M_CLASS(class_cnt++); | ||
315 | |||
316 | default: | ||
317 | return c; | ||
318 | } | ||
319 | } | ||
320 | /*NOTREACHED*/} | ||
321 | |||
322 | |||
323 | |||
324 | |||
325 | /* This function calculates the number of bytes that will be needed for a | ||
326 | * compiled RE. Its argument is the uncompiled version. It is not clever | ||
327 | * about catching syntax errors; that is done in a later pass. | ||
328 | */ | ||
329 | static unsigned calcsize(text) | ||
330 | char *text; | ||
331 | { | ||
332 | unsigned size; | ||
333 | int token; | ||
334 | |||
335 | retext = text; | ||
336 | class_cnt = 0; | ||
337 | start_cnt = 1; | ||
338 | end_sp = 0; | ||
339 | size = 5; | ||
340 | while ((token = gettoken(&text, (regexp *) 0)) != 0) { | ||
341 | if (IS_CLASS(token)) { | ||
342 | size += 34; | ||
343 | } else if (IS_META(token)) { | ||
344 | size += 2; | ||
345 | } else { | ||
346 | size++; | ||
347 | } | ||
348 | } | ||
349 | |||
350 | return size; | ||
351 | } | ||
352 | |||
353 | |||
354 | |||
355 | /*---------------------------------------------------------------------------*/ | ||
356 | |||
357 | |||
358 | /* This function checks for a match between a character and a token which is | ||
359 | * known to represent a single character. It returns 0 if they match, or | ||
360 | * 1 if they don't. | ||
361 | */ | ||
362 | static int match1(regexp * re, char ch, int token, int ignoreCase) | ||
363 | { | ||
364 | if (!ch) { | ||
365 | /* the end of a line can't match any RE of width 1 */ | ||
366 | return 1; | ||
367 | } | ||
368 | if (token == M_ANY) { | ||
369 | return 0; | ||
370 | } else if (IS_CLASS(token)) { | ||
371 | if (re->program[1 + 32 * (token - M_CLASS(0)) + (ch >> 3)] & (1 << (ch & 7))) | ||
372 | return 0; | ||
373 | } | ||
374 | //fprintf(stderr, "match1: ch='%c' token='%c': ", ch, token); | ||
375 | if (ch == token || (ignoreCase == TRUE && tolower(ch) == tolower(token))) { | ||
376 | //fprintf(stderr, "match\n"); | ||
377 | return 0; | ||
378 | } | ||
379 | //fprintf(stderr, "no match\n"); | ||
380 | return 1; | ||
381 | } | ||
382 | |||
383 | |||
384 | |||
385 | /* This function checks characters up to and including the next closure, at | ||
386 | * which point it does a recursive call to check the rest of it. This function | ||
387 | * returns 0 if everything matches, or 1 if something doesn't match. | ||
388 | */ | ||
389 | /* re -- the regular expression */ | ||
390 | /* str -- the string */ | ||
391 | /* prog -- a portion of re->program, an compiled RE */ | ||
392 | /* here -- a portion of str, the string to compare it to */ | ||
393 | static int match(regexp * re, char *str, char *prog, char *here, | ||
394 | int ignoreCase) | ||
395 | { | ||
396 | int token; | ||
397 | int nmatched; | ||
398 | int closure; | ||
399 | |||
400 | for (token = GET_META(prog); !IS_CLOSURE(token); | ||
401 | prog++, token = GET_META(prog)) { | ||
402 | switch (token) { | ||
403 | /*case M_BEGLINE: can't happen; re->bol is used instead */ | ||
404 | case M_ENDLINE: | ||
405 | if (*here) | ||
406 | return 1; | ||
407 | break; | ||
408 | |||
409 | case M_BEGWORD: | ||
410 | if (here != str && | ||
411 | (here[-1] == '_' || | ||
412 | (isascii(here[-1]) && isalnum(here[-1])))) return 1; | ||
413 | break; | ||
414 | |||
415 | case M_ENDWORD: | ||
416 | if ((here[0] == '_' || isascii(here[0])) && isalnum(here[0])) | ||
417 | return 1; | ||
418 | break; | ||
419 | |||
420 | case M_START(0): | ||
421 | case M_START(1): | ||
422 | case M_START(2): | ||
423 | case M_START(3): | ||
424 | case M_START(4): | ||
425 | case M_START(5): | ||
426 | case M_START(6): | ||
427 | case M_START(7): | ||
428 | case M_START(8): | ||
429 | case M_START(9): | ||
430 | re->startp[token - M_START(0)] = (char *) here; | ||
431 | break; | ||
432 | |||
433 | case M_END(0): | ||
434 | case M_END(1): | ||
435 | case M_END(2): | ||
436 | case M_END(3): | ||
437 | case M_END(4): | ||
438 | case M_END(5): | ||
439 | case M_END(6): | ||
440 | case M_END(7): | ||
441 | case M_END(8): | ||
442 | case M_END(9): | ||
443 | re->endp[token - M_END(0)] = (char *) here; | ||
444 | if (token == M_END(0)) { | ||
445 | return 0; | ||
446 | } | ||
447 | break; | ||
448 | |||
449 | default: /* literal, M_CLASS(n), or M_ANY */ | ||
450 | if (match1(re, *here, token, ignoreCase) != 0) | ||
451 | return 1; | ||
452 | here++; | ||
453 | } | ||
454 | } | ||
455 | |||
456 | /* C L O S U R E */ | ||
457 | |||
458 | /* step 1: see what we have to match against, and move "prog" to point | ||
459 | * the the remainder of the compiled RE. | ||
460 | */ | ||
461 | closure = token; | ||
462 | prog++, token = GET_META(prog); | ||
463 | prog++; | ||
464 | |||
465 | /* step 2: see how many times we can match that token against the string */ | ||
466 | for (nmatched = 0; | ||
467 | (closure != M_QMARK || nmatched < 1) && *here | ||
468 | && match1(re, *here, token, ignoreCase) == 0; nmatched++, here++) { | ||
469 | } | ||
470 | |||
471 | /* step 3: try to match the remainder, and back off if it doesn't */ | ||
472 | while (nmatched >= 0 && match(re, str, prog, here, ignoreCase) != 0) { | ||
473 | nmatched--; | ||
474 | here--; | ||
475 | } | ||
476 | |||
477 | /* so how did it work out? */ | ||
478 | if (nmatched >= ((closure == M_PLUS) ? 1 : 0)) | ||
479 | return 0; | ||
480 | return 1; | ||
481 | } | ||
482 | |||
483 | |||
484 | /* This function compiles a regexp. */ | ||
485 | static regexp *regcomp(char *text) | ||
486 | { | ||
487 | int needfirst; | ||
488 | unsigned size; | ||
489 | int token; | ||
490 | int peek; | ||
491 | char *build; | ||
492 | regexp *re; | ||
493 | |||
494 | |||
495 | /* prepare for error handling */ | ||
496 | re = (regexp *) 0; | ||
497 | if (setjmp(errorhandler)) { | ||
498 | if (re) { | ||
499 | free(re); | ||
500 | } | ||
501 | return (regexp *) 0; | ||
502 | } | ||
503 | |||
504 | /* if an empty regexp string was given, use the previous one */ | ||
505 | if (*text == 0) { | ||
506 | if (!previous) { | ||
507 | FAIL("No previous RE"); | ||
508 | } | ||
509 | text = previous; | ||
510 | } else { /* non-empty regexp given, so remember it */ | ||
511 | |||
512 | if (previous) | ||
513 | free(previous); | ||
514 | previous = (char *) malloc((unsigned) (strlen(text) + 1)); | ||
515 | if (previous) | ||
516 | strcpy(previous, text); | ||
517 | } | ||
518 | |||
519 | /* allocate memory */ | ||
520 | class_cnt = 0; | ||
521 | start_cnt = 1; | ||
522 | end_sp = 0; | ||
523 | retext = text; | ||
524 | size = calcsize(text) + sizeof(regexp); | ||
525 | re = (regexp *) malloc((unsigned) size); | ||
526 | |||
527 | if (!re) { | ||
528 | FAIL("Not enough memory for this RE"); | ||
529 | } | ||
530 | |||
531 | /* compile it */ | ||
532 | build = &re->program[1 + 32 * class_cnt]; | ||
533 | re->program[0] = class_cnt; | ||
534 | for (token = 0; token < NSUBEXP; token++) { | ||
535 | re->startp[token] = re->endp[token] = (char *) 0; | ||
536 | } | ||
537 | re->first = 0; | ||
538 | re->bol = 0; | ||
539 | re->minlen = 0; | ||
540 | needfirst = 1; | ||
541 | class_cnt = 0; | ||
542 | start_cnt = 1; | ||
543 | end_sp = 0; | ||
544 | retext = text; | ||
545 | for (token = M_START(0), peek = gettoken(&text, re); | ||
546 | token; token = peek, peek = gettoken(&text, re)) { | ||
547 | |||
548 | /* special processing for the closure operator */ | ||
549 | if (IS_CLOSURE(peek)) { | ||
550 | |||
551 | /* detect misuse of closure operator */ | ||
552 | if (IS_START(token)) { | ||
553 | FAIL("* or \\+ or \\? follows nothing"); | ||
554 | } else if (IS_META(token) && token != M_ANY && !IS_CLASS(token)) { | ||
555 | FAIL("* or \\+ or \\? can only follow a normal character or . or []"); | ||
556 | } | ||
557 | |||
558 | /* it is okay -- make it prefix instead of postfix */ | ||
559 | ADD_META(build, peek); | ||
560 | |||
561 | /* take care of "needfirst" - is this the first char? */ | ||
562 | if (needfirst && peek == M_PLUS && !IS_META(token)) { | ||
563 | re->first = token; | ||
564 | } | ||
565 | needfirst = 0; | ||
566 | |||
567 | /* we used "peek" -- need to refill it */ | ||
568 | peek = gettoken(&text, re); | ||
569 | if (IS_CLOSURE(peek)) { | ||
570 | FAIL("* or \\+ or \\? doubled up"); | ||
571 | } | ||
572 | } else if (!IS_META(token)) { | ||
573 | /* normal char is NOT argument of closure */ | ||
574 | if (needfirst) { | ||
575 | re->first = token; | ||
576 | needfirst = 0; | ||
577 | } | ||
578 | re->minlen++; | ||
579 | } else if (token == M_ANY || IS_CLASS(token)) { | ||
580 | /* . or [] is NOT argument of closure */ | ||
581 | needfirst = 0; | ||
582 | re->minlen++; | ||
583 | } | ||
584 | |||
585 | /* the "token" character is not closure -- process it normally */ | ||
586 | if (token == M_BEGLINE) { | ||
587 | /* set the BOL flag instead of storing M_BEGLINE */ | ||
588 | re->bol = 1; | ||
589 | } else if (IS_META(token)) { | ||
590 | ADD_META(build, token); | ||
591 | } else { | ||
592 | *build++ = token; | ||
593 | } | ||
594 | } | ||
595 | |||
596 | /* end it with a \) which MUST MATCH the opening \( */ | ||
597 | ADD_META(build, M_END(0)); | ||
598 | if (end_sp > 0) { | ||
599 | FAIL("Not enough \\)s"); | ||
600 | } | ||
601 | |||
602 | return re; | ||
603 | } | ||
604 | |||
605 | |||
606 | |||
607 | |||
608 | /* This function searches through a string for text that matches an RE. */ | ||
609 | /* re -- the compiled regexp to search for */ | ||
610 | /* str -- the string to search through */ | ||
611 | /* bol -- does str start at the beginning of a line? (boolean) */ | ||
612 | /* ignoreCase -- ignoreCase or not */ | ||
613 | static int regexec(struct regexp *re, char *str, int bol, int ignoreCase) | ||
614 | { | ||
615 | char *prog; /* the entry point of re->program */ | ||
616 | int len; /* length of the string */ | ||
617 | char *here; | ||
618 | |||
619 | /* if must start at the beginning of a line, and this isn't, then fail */ | ||
620 | if (re->bol && bol == TRUE) { | ||
621 | return FALSE; | ||
622 | } | ||
623 | |||
624 | len = strlen(str); | ||
625 | prog = re->program + 1 + 32 * re->program[0]; | ||
626 | |||
627 | /* search for the RE in the string */ | ||
628 | if (re->bol) { | ||
629 | /* must occur at BOL */ | ||
630 | if ((re->first && match1(re, *(char *) str, re->first, ignoreCase)) /* wrong first letter? */ | ||
631 | ||len < re->minlen /* not long enough? */ | ||
632 | || match(re, (char *) str, prog, str, ignoreCase)) /* doesn't match? */ | ||
633 | return FALSE; /* THEN FAIL! */ | ||
634 | } else if (ignoreCase == FALSE) { | ||
635 | /* can occur anywhere in the line, noignorecase */ | ||
636 | for (here = (char *) str; (re->first && re->first != *here) | ||
637 | || match(re, (char *) str, prog, here, ignoreCase); | ||
638 | here++, len--) { | ||
639 | if (len < re->minlen) | ||
640 | return FALSE; | ||
641 | } | ||
642 | } else { | ||
643 | /* can occur anywhere in the line, ignorecase */ | ||
644 | for (here = (char *) str; | ||
645 | (re->first && match1(re, *here, (int) re->first, ignoreCase)) | ||
646 | || match(re, (char *) str, prog, here, ignoreCase); | ||
647 | here++, len--) { | ||
648 | if (len < re->minlen) | ||
649 | return FALSE; | ||
650 | } | ||
651 | } | ||
652 | |||
653 | /* if we didn't fail, then we must have succeeded */ | ||
654 | return TRUE; | ||
655 | } | ||
656 | |||
657 | |||
658 | |||
659 | |||
660 | #if defined BB_SED | ||
661 | /* This performs substitutions after a regexp match has been found. */ | ||
662 | static void regsub(regexp * re, char *src, char *dst) | ||
663 | { | ||
664 | char *cpy; | ||
665 | char *end; | ||
666 | char c; | ||
667 | char *start; | ||
668 | int mod; | ||
669 | |||
670 | mod = 0; | ||
671 | |||
672 | start = src; | ||
673 | while ((c = *src++) != '\0') { | ||
674 | /* recognize any meta characters */ | ||
675 | if (c == '&') { | ||
676 | cpy = re->startp[0]; | ||
677 | end = re->endp[0]; | ||
678 | } else if (c == '~') { | ||
679 | cpy = previous1; | ||
680 | if (cpy) | ||
681 | end = cpy + strlen(cpy); | ||
682 | } else if (c == '\\') { | ||
683 | c = *src++; | ||
684 | switch (c) { | ||
685 | case '0': | ||
686 | case '1': | ||
687 | case '2': | ||
688 | case '3': | ||
689 | case '4': | ||
690 | case '5': | ||
691 | case '6': | ||
692 | case '7': | ||
693 | case '8': | ||
694 | case '9': | ||
695 | /* \0 thru \9 mean "copy subexpression" */ | ||
696 | c -= '0'; | ||
697 | cpy = re->startp[(int) c]; | ||
698 | end = re->endp[(int) c]; | ||
699 | break; | ||
700 | case 'U': | ||
701 | case 'u': | ||
702 | case 'L': | ||
703 | case 'l': | ||
704 | /* \U and \L mean "convert to upper/lowercase" */ | ||
705 | mod = c; | ||
706 | continue; | ||
707 | |||
708 | case 'E': | ||
709 | case 'e': | ||
710 | /* \E ends the \U or \L */ | ||
711 | mod = 0; | ||
712 | continue; | ||
713 | case '&': | ||
714 | /* "\&" means "original text" */ | ||
715 | *dst++ = c; | ||
716 | continue; | ||
717 | |||
718 | case '~': | ||
719 | /* "\~" means "previous text, if any" */ | ||
720 | *dst++ = c; | ||
721 | continue; | ||
722 | default: | ||
723 | /* ordinary char preceded by backslash */ | ||
724 | *dst++ = c; | ||
725 | continue; | ||
726 | } | ||
727 | } else { | ||
728 | /* ordinary character, so just copy it */ | ||
729 | *dst++ = c; | ||
730 | continue; | ||
731 | } | ||
732 | |||
733 | /* Note: to reach this point in the code, we must have evaded | ||
734 | * all "continue" statements. To do that, we must have hit | ||
735 | * a metacharacter that involves copying. | ||
736 | */ | ||
737 | |||
738 | /* if there is nothing to copy, loop */ | ||
739 | if (!cpy) | ||
740 | continue; | ||
741 | |||
742 | /* copy over a portion of the original */ | ||
743 | while (cpy < end) { | ||
744 | switch (mod) { | ||
745 | case 'U': | ||
746 | case 'u': | ||
747 | /* convert to uppercase */ | ||
748 | if (isascii(*cpy) && islower(*cpy)) { | ||
749 | *dst++ = toupper(*cpy); | ||
750 | cpy++; | ||
751 | } else { | ||
752 | *dst++ = *cpy++; | ||
753 | } | ||
754 | break; | ||
755 | |||
756 | case 'L': | ||
757 | case 'l': | ||
758 | /* convert to lowercase */ | ||
759 | if (isascii(*cpy) && isupper(*cpy)) { | ||
760 | *dst++ = tolower(*cpy); | ||
761 | cpy++; | ||
762 | } else { | ||
763 | *dst++ = *cpy++; | ||
764 | } | ||
765 | break; | ||
766 | |||
767 | default: | ||
768 | /* copy without any conversion */ | ||
769 | *dst++ = *cpy++; | ||
770 | } | ||
771 | |||
772 | /* \u and \l end automatically after the first char */ | ||
773 | if (mod && (mod == 'u' || mod == 'l')) { | ||
774 | mod = 0; | ||
775 | } | ||
776 | } | ||
777 | } | ||
778 | *dst = '\0'; | ||
779 | |||
780 | /* remember what text we inserted this time */ | ||
781 | if (previous1) | ||
782 | free(previous1); | ||
783 | previous1 = (char *) malloc((unsigned) (strlen(start) + 1)); | ||
784 | if (previous1) | ||
785 | strcpy(previous1, start); | ||
786 | } | ||
787 | #endif | ||
788 | |||
789 | #endif /* BB_REGEXP */ | ||
diff --git a/regexp.h b/regexp.h deleted file mode 100644 index 9e0cb4022..000000000 --- a/regexp.h +++ /dev/null | |||
@@ -1,33 +0,0 @@ | |||
1 | /* vi: set sw=4 ts=4: */ | ||
2 | /* | ||
3 | * Busybox regexp header file | ||
4 | * | ||
5 | * | ||
6 | * This program 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 2 of the License, or | ||
9 | * (at your option) any later version. | ||
10 | * | ||
11 | * This program 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 GNU | ||
14 | * General Public License for more details. | ||
15 | * | ||
16 | * You should have received a copy of the GNU General Public License | ||
17 | * along with this program; if not, write to the Free Software | ||
18 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
19 | * | ||
20 | * Based in part on code from sash, Copyright (c) 1999 by David I. Bell | ||
21 | * Permission has been granted to redistribute this code under the GPL. | ||
22 | * | ||
23 | */ | ||
24 | #ifndef _REGEXP_H_ | ||
25 | #define _REGEXP_H_ | ||
26 | |||
27 | |||
28 | extern int find_match(char *haystack, char *needle, int ignoreCase); | ||
29 | extern int replace_match(char *haystack, char *needle, char *newNeedle, int ignoreCase); | ||
30 | |||
31 | #endif | ||
32 | |||
33 | |||