xref: /openbsd-src/lib/libc/regex/regcomp.c (revision aa5e9e10509ffd51558f081f01cd78bfa3c4f2a5)
1 /*	$OpenBSD: regcomp.c,v 1.22 2013/04/17 17:39:29 tedu Exp $ */
2 /*-
3  * Copyright (c) 1992, 1993, 1994 Henry Spencer.
4  * Copyright (c) 1992, 1993, 1994
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Henry Spencer.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  *
34  *	@(#)regcomp.c	8.5 (Berkeley) 3/20/94
35  */
36 
37 #include <sys/types.h>
38 #include <stdio.h>
39 #include <string.h>
40 #include <ctype.h>
41 #include <limits.h>
42 #include <stdlib.h>
43 #include <regex.h>
44 
45 #include "utils.h"
46 #include "regex2.h"
47 
48 #include "cclass.h"
49 #include "cname.h"
50 
51 /*
52  * parse structure, passed up and down to avoid global variables and
53  * other clumsinesses
54  */
55 struct parse {
56 	char *next;		/* next character in RE */
57 	char *end;		/* end of string (-> NUL normally) */
58 	int error;		/* has an error been seen? */
59 	sop *strip;		/* malloced strip */
60 	sopno ssize;		/* malloced strip size (allocated) */
61 	sopno slen;		/* malloced strip length (used) */
62 	int ncsalloc;		/* number of csets allocated */
63 	struct re_guts *g;
64 #	define	NPAREN	10	/* we need to remember () 1-9 for back refs */
65 	sopno pbegin[NPAREN];	/* -> ( ([0] unused) */
66 	sopno pend[NPAREN];	/* -> ) ([0] unused) */
67 };
68 
69 static void p_ere(struct parse *, int);
70 static void p_ere_exp(struct parse *);
71 static void p_str(struct parse *);
72 static void p_bre(struct parse *, int, int);
73 static int p_simp_re(struct parse *, int);
74 static int p_count(struct parse *);
75 static void p_bracket(struct parse *);
76 static void p_b_term(struct parse *, cset *);
77 static void p_b_cclass(struct parse *, cset *);
78 static void p_b_eclass(struct parse *, cset *);
79 static char p_b_symbol(struct parse *);
80 static char p_b_coll_elem(struct parse *, int);
81 static char othercase(int);
82 static void bothcases(struct parse *, int);
83 static void ordinary(struct parse *, int);
84 static void nonnewline(struct parse *);
85 static void repeat(struct parse *, sopno, int, int);
86 static int seterr(struct parse *, int);
87 static cset *allocset(struct parse *);
88 static void freeset(struct parse *, cset *);
89 static int freezeset(struct parse *, cset *);
90 static int firstch(struct parse *, cset *);
91 static int nch(struct parse *, cset *);
92 static void mcadd(struct parse *, cset *, char *);
93 static void mcinvert(struct parse *, cset *);
94 static void mccase(struct parse *, cset *);
95 static int isinsets(struct re_guts *, int);
96 static int samesets(struct re_guts *, int, int);
97 static void categorize(struct parse *, struct re_guts *);
98 static sopno dupl(struct parse *, sopno, sopno);
99 static void doemit(struct parse *, sop, size_t);
100 static void doinsert(struct parse *, sop, size_t, sopno);
101 static void dofwd(struct parse *, sopno, sop);
102 static int enlarge(struct parse *, sopno);
103 static void stripsnug(struct parse *, struct re_guts *);
104 static void findmust(struct parse *, struct re_guts *);
105 static sopno pluscount(struct parse *, struct re_guts *);
106 
107 static char nuls[10];		/* place to point scanner in event of error */
108 
109 /*
110  * macros for use with parse structure
111  * BEWARE:  these know that the parse structure is named `p' !!!
112  */
113 #define	PEEK()	(*p->next)
114 #define	PEEK2()	(*(p->next+1))
115 #define	MORE()	(p->next < p->end)
116 #define	MORE2()	(p->next+1 < p->end)
117 #define	SEE(c)	(MORE() && PEEK() == (c))
118 #define	SEETWO(a, b)	(MORE() && MORE2() && PEEK() == (a) && PEEK2() == (b))
119 #define	EAT(c)	((SEE(c)) ? (NEXT(), 1) : 0)
120 #define	EATTWO(a, b)	((SEETWO(a, b)) ? (NEXT2(), 1) : 0)
121 #define	NEXT()	(p->next++)
122 #define	NEXT2()	(p->next += 2)
123 #define	NEXTn(n)	(p->next += (n))
124 #define	GETNEXT()	(*p->next++)
125 #define	SETERROR(e)	seterr(p, (e))
126 #define	REQUIRE(co, e)	((co) || SETERROR(e))
127 #define	MUSTSEE(c, e)	(REQUIRE(MORE() && PEEK() == (c), e))
128 #define	MUSTEAT(c, e)	(REQUIRE(MORE() && GETNEXT() == (c), e))
129 #define	MUSTNOTSEE(c, e)	(REQUIRE(!MORE() || PEEK() != (c), e))
130 #define	EMIT(op, sopnd)	doemit(p, (sop)(op), (size_t)(sopnd))
131 #define	INSERT(op, pos)	doinsert(p, (sop)(op), HERE()-(pos)+1, pos)
132 #define	AHEAD(pos)		dofwd(p, pos, HERE()-(pos))
133 #define	ASTERN(sop, pos)	EMIT(sop, HERE()-pos)
134 #define	HERE()		(p->slen)
135 #define	THERE()		(p->slen - 1)
136 #define	THERETHERE()	(p->slen - 2)
137 #define	DROP(n)	(p->slen -= (n))
138 
139 #ifndef NDEBUG
140 static int never = 0;		/* for use in asserts; shuts lint up */
141 #else
142 #define	never	0		/* some <assert.h>s have bugs too */
143 #endif
144 
145 /*
146  - regcomp - interface for parser and compilation
147  */
148 int				/* 0 success, otherwise REG_something */
149 regcomp(regex_t *preg, const char *pattern, int cflags)
150 {
151 	struct parse pa;
152 	struct re_guts *g;
153 	struct parse *p = &pa;
154 	int i;
155 	size_t len;
156 #ifdef REDEBUG
157 #	define	GOODFLAGS(f)	(f)
158 #else
159 #	define	GOODFLAGS(f)	((f)&~REG_DUMP)
160 #endif
161 
162 	cflags = GOODFLAGS(cflags);
163 	if ((cflags&REG_EXTENDED) && (cflags&REG_NOSPEC))
164 		return(REG_INVARG);
165 
166 	if (cflags&REG_PEND) {
167 		if (preg->re_endp < pattern)
168 			return(REG_INVARG);
169 		len = preg->re_endp - pattern;
170 	} else
171 		len = strlen((char *)pattern);
172 
173 	/* do the mallocs early so failure handling is easy */
174 	g = (struct re_guts *)malloc(sizeof(struct re_guts));
175 	if (g == NULL)
176 		return(REG_ESPACE);
177 	p->ssize = len/(size_t)2*(size_t)3 + (size_t)1;	/* ugh */
178 	p->strip = (sop *)calloc(p->ssize, sizeof(sop));
179 	p->slen = 0;
180 	if (p->strip == NULL) {
181 		free((char *)g);
182 		return(REG_ESPACE);
183 	}
184 
185 	/* set things up */
186 	p->g = g;
187 	p->next = (char *)pattern;	/* convenience; we do not modify it */
188 	p->end = p->next + len;
189 	p->error = 0;
190 	p->ncsalloc = 0;
191 	for (i = 0; i < NPAREN; i++) {
192 		p->pbegin[i] = 0;
193 		p->pend[i] = 0;
194 	}
195 	g->csetsize = NC;
196 	g->sets = NULL;
197 	g->setbits = NULL;
198 	g->ncsets = 0;
199 	g->cflags = cflags;
200 	g->iflags = 0;
201 	g->nbol = 0;
202 	g->neol = 0;
203 	g->must = NULL;
204 	g->mlen = 0;
205 	g->nsub = 0;
206 	g->ncategories = 1;	/* category 0 is "everything else" */
207 	g->categories = &g->catspace[-(CHAR_MIN)];
208 	memset(g->catspace, 0, sizeof(g->catspace));
209 	g->backrefs = 0;
210 
211 	/* do it */
212 	EMIT(OEND, 0);
213 	g->firststate = THERE();
214 	if (cflags&REG_EXTENDED)
215 		p_ere(p, OUT);
216 	else if (cflags&REG_NOSPEC)
217 		p_str(p);
218 	else
219 		p_bre(p, OUT, OUT);
220 	EMIT(OEND, 0);
221 	g->laststate = THERE();
222 
223 	/* tidy up loose ends and fill things in */
224 	categorize(p, g);
225 	stripsnug(p, g);
226 	findmust(p, g);
227 	g->nplus = pluscount(p, g);
228 	g->magic = MAGIC2;
229 	preg->re_nsub = g->nsub;
230 	preg->re_g = g;
231 	preg->re_magic = MAGIC1;
232 #ifndef REDEBUG
233 	/* not debugging, so can't rely on the assert() in regexec() */
234 	if (g->iflags&BAD)
235 		SETERROR(REG_ASSERT);
236 #endif
237 
238 	/* win or lose, we're done */
239 	if (p->error != 0)	/* lose */
240 		regfree(preg);
241 	return(p->error);
242 }
243 
244 /*
245  - p_ere - ERE parser top level, concatenation and alternation
246  */
247 static void
248 p_ere(struct parse *p, int stop)	/* character this ERE should end at */
249 {
250 	char c;
251 	sopno prevback;
252 	sopno prevfwd;
253 	sopno conc;
254 	int first = 1;		/* is this the first alternative? */
255 
256 	for (;;) {
257 		/* do a bunch of concatenated expressions */
258 		conc = HERE();
259 		while (MORE() && (c = PEEK()) != '|' && c != stop)
260 			p_ere_exp(p);
261 		REQUIRE(HERE() != conc, REG_EMPTY);	/* require nonempty */
262 
263 		if (!EAT('|'))
264 			break;		/* NOTE BREAK OUT */
265 
266 		if (first) {
267 			INSERT(OCH_, conc);	/* offset is wrong */
268 			prevfwd = conc;
269 			prevback = conc;
270 			first = 0;
271 		}
272 		ASTERN(OOR1, prevback);
273 		prevback = THERE();
274 		AHEAD(prevfwd);			/* fix previous offset */
275 		prevfwd = HERE();
276 		EMIT(OOR2, 0);			/* offset is very wrong */
277 	}
278 
279 	if (!first) {		/* tail-end fixups */
280 		AHEAD(prevfwd);
281 		ASTERN(O_CH, prevback);
282 	}
283 
284 	assert(!MORE() || SEE(stop));
285 }
286 
287 /*
288  - p_ere_exp - parse one subERE, an atom possibly followed by a repetition op
289  */
290 static void
291 p_ere_exp(struct parse *p)
292 {
293 	char c;
294 	sopno pos;
295 	int count;
296 	int count2;
297 	sopno subno;
298 	int wascaret = 0;
299 
300 	assert(MORE());		/* caller should have ensured this */
301 	c = GETNEXT();
302 
303 	pos = HERE();
304 	switch (c) {
305 	case '(':
306 		REQUIRE(MORE(), REG_EPAREN);
307 		p->g->nsub++;
308 		subno = p->g->nsub;
309 		if (subno < NPAREN)
310 			p->pbegin[subno] = HERE();
311 		EMIT(OLPAREN, subno);
312 		if (!SEE(')'))
313 			p_ere(p, ')');
314 		if (subno < NPAREN) {
315 			p->pend[subno] = HERE();
316 			assert(p->pend[subno] != 0);
317 		}
318 		EMIT(ORPAREN, subno);
319 		MUSTEAT(')', REG_EPAREN);
320 		break;
321 	case '^':
322 		EMIT(OBOL, 0);
323 		p->g->iflags |= USEBOL;
324 		p->g->nbol++;
325 		wascaret = 1;
326 		break;
327 	case '$':
328 		EMIT(OEOL, 0);
329 		p->g->iflags |= USEEOL;
330 		p->g->neol++;
331 		break;
332 	case '|':
333 		SETERROR(REG_EMPTY);
334 		break;
335 	case '*':
336 	case '+':
337 	case '?':
338 		SETERROR(REG_BADRPT);
339 		break;
340 	case '.':
341 		if (p->g->cflags&REG_NEWLINE)
342 			nonnewline(p);
343 		else
344 			EMIT(OANY, 0);
345 		break;
346 	case '[':
347 		p_bracket(p);
348 		break;
349 	case '\\':
350 		REQUIRE(MORE(), REG_EESCAPE);
351 		c = GETNEXT();
352 		ordinary(p, c);
353 		break;
354 	case '{':		/* okay as ordinary except if digit follows */
355 		REQUIRE(!MORE() || !isdigit((uch)PEEK()), REG_BADRPT);
356 		/* FALLTHROUGH */
357 	default:
358 		ordinary(p, c);
359 		break;
360 	}
361 
362 	if (!MORE())
363 		return;
364 	c = PEEK();
365 	/* we call { a repetition if followed by a digit */
366 	if (!( c == '*' || c == '+' || c == '?' ||
367 				(c == '{' && MORE2() && isdigit((uch)PEEK2())) ))
368 		return;		/* no repetition, we're done */
369 	NEXT();
370 
371 	REQUIRE(!wascaret, REG_BADRPT);
372 	switch (c) {
373 	case '*':	/* implemented as +? */
374 		/* this case does not require the (y|) trick, noKLUDGE */
375 		INSERT(OPLUS_, pos);
376 		ASTERN(O_PLUS, pos);
377 		INSERT(OQUEST_, pos);
378 		ASTERN(O_QUEST, pos);
379 		break;
380 	case '+':
381 		INSERT(OPLUS_, pos);
382 		ASTERN(O_PLUS, pos);
383 		break;
384 	case '?':
385 		/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
386 		INSERT(OCH_, pos);		/* offset slightly wrong */
387 		ASTERN(OOR1, pos);		/* this one's right */
388 		AHEAD(pos);			/* fix the OCH_ */
389 		EMIT(OOR2, 0);			/* offset very wrong... */
390 		AHEAD(THERE());			/* ...so fix it */
391 		ASTERN(O_CH, THERETHERE());
392 		break;
393 	case '{':
394 		count = p_count(p);
395 		if (EAT(',')) {
396 			if (isdigit((uch)PEEK())) {
397 				count2 = p_count(p);
398 				REQUIRE(count <= count2, REG_BADBR);
399 			} else		/* single number with comma */
400 				count2 = INFINITY;
401 		} else		/* just a single number */
402 			count2 = count;
403 		repeat(p, pos, count, count2);
404 		if (!EAT('}')) {	/* error heuristics */
405 			while (MORE() && PEEK() != '}')
406 				NEXT();
407 			REQUIRE(MORE(), REG_EBRACE);
408 			SETERROR(REG_BADBR);
409 		}
410 		break;
411 	}
412 
413 	if (!MORE())
414 		return;
415 	c = PEEK();
416 	if (!( c == '*' || c == '+' || c == '?' ||
417 				(c == '{' && MORE2() && isdigit((uch)PEEK2())) ) )
418 		return;
419 	SETERROR(REG_BADRPT);
420 }
421 
422 /*
423  - p_str - string (no metacharacters) "parser"
424  */
425 static void
426 p_str(struct parse *p)
427 {
428 	REQUIRE(MORE(), REG_EMPTY);
429 	while (MORE())
430 		ordinary(p, GETNEXT());
431 }
432 
433 /*
434  - p_bre - BRE parser top level, anchoring and concatenation
435  * Giving end1 as OUT essentially eliminates the end1/end2 check.
436  *
437  * This implementation is a bit of a kludge, in that a trailing $ is first
438  * taken as an ordinary character and then revised to be an anchor.  The
439  * only undesirable side effect is that '$' gets included as a character
440  * category in such cases.  This is fairly harmless; not worth fixing.
441  * The amount of lookahead needed to avoid this kludge is excessive.
442  */
443 static void
444 p_bre(struct parse *p,
445     int end1,		/* first terminating character */
446     int end2)		/* second terminating character */
447 {
448 	sopno start = HERE();
449 	int first = 1;			/* first subexpression? */
450 	int wasdollar = 0;
451 
452 	if (EAT('^')) {
453 		EMIT(OBOL, 0);
454 		p->g->iflags |= USEBOL;
455 		p->g->nbol++;
456 	}
457 	while (MORE() && !SEETWO(end1, end2)) {
458 		wasdollar = p_simp_re(p, first);
459 		first = 0;
460 	}
461 	if (wasdollar) {	/* oops, that was a trailing anchor */
462 		DROP(1);
463 		EMIT(OEOL, 0);
464 		p->g->iflags |= USEEOL;
465 		p->g->neol++;
466 	}
467 
468 	REQUIRE(HERE() != start, REG_EMPTY);	/* require nonempty */
469 }
470 
471 /*
472  - p_simp_re - parse a simple RE, an atom possibly followed by a repetition
473  */
474 static int			/* was the simple RE an unbackslashed $? */
475 p_simp_re(struct parse *p,
476     int starordinary)		/* is a leading * an ordinary character? */
477 {
478 	int c;
479 	int count;
480 	int count2;
481 	sopno pos;
482 	int i;
483 	sopno subno;
484 #	define	BACKSL	(1<<CHAR_BIT)
485 
486 	pos = HERE();		/* repetion op, if any, covers from here */
487 
488 	assert(MORE());		/* caller should have ensured this */
489 	c = GETNEXT();
490 	if (c == '\\') {
491 		REQUIRE(MORE(), REG_EESCAPE);
492 		c = BACKSL | GETNEXT();
493 	}
494 	switch (c) {
495 	case '.':
496 		if (p->g->cflags&REG_NEWLINE)
497 			nonnewline(p);
498 		else
499 			EMIT(OANY, 0);
500 		break;
501 	case '[':
502 		p_bracket(p);
503 		break;
504 	case BACKSL|'{':
505 		SETERROR(REG_BADRPT);
506 		break;
507 	case BACKSL|'(':
508 		p->g->nsub++;
509 		subno = p->g->nsub;
510 		if (subno < NPAREN)
511 			p->pbegin[subno] = HERE();
512 		EMIT(OLPAREN, subno);
513 		/* the MORE here is an error heuristic */
514 		if (MORE() && !SEETWO('\\', ')'))
515 			p_bre(p, '\\', ')');
516 		if (subno < NPAREN) {
517 			p->pend[subno] = HERE();
518 			assert(p->pend[subno] != 0);
519 		}
520 		EMIT(ORPAREN, subno);
521 		REQUIRE(EATTWO('\\', ')'), REG_EPAREN);
522 		break;
523 	case BACKSL|')':	/* should not get here -- must be user */
524 	case BACKSL|'}':
525 		SETERROR(REG_EPAREN);
526 		break;
527 	case BACKSL|'1':
528 	case BACKSL|'2':
529 	case BACKSL|'3':
530 	case BACKSL|'4':
531 	case BACKSL|'5':
532 	case BACKSL|'6':
533 	case BACKSL|'7':
534 	case BACKSL|'8':
535 	case BACKSL|'9':
536 		i = (c&~BACKSL) - '0';
537 		assert(i < NPAREN);
538 		if (p->pend[i] != 0) {
539 			assert(i <= p->g->nsub);
540 			EMIT(OBACK_, i);
541 			assert(p->pbegin[i] != 0);
542 			assert(OP(p->strip[p->pbegin[i]]) == OLPAREN);
543 			assert(OP(p->strip[p->pend[i]]) == ORPAREN);
544 			(void) dupl(p, p->pbegin[i]+1, p->pend[i]);
545 			EMIT(O_BACK, i);
546 		} else
547 			SETERROR(REG_ESUBREG);
548 		p->g->backrefs = 1;
549 		break;
550 	case '*':
551 		REQUIRE(starordinary, REG_BADRPT);
552 		/* FALLTHROUGH */
553 	default:
554 		ordinary(p, (char)c);
555 		break;
556 	}
557 
558 	if (EAT('*')) {		/* implemented as +? */
559 		/* this case does not require the (y|) trick, noKLUDGE */
560 		INSERT(OPLUS_, pos);
561 		ASTERN(O_PLUS, pos);
562 		INSERT(OQUEST_, pos);
563 		ASTERN(O_QUEST, pos);
564 	} else if (EATTWO('\\', '{')) {
565 		count = p_count(p);
566 		if (EAT(',')) {
567 			if (MORE() && isdigit((uch)PEEK())) {
568 				count2 = p_count(p);
569 				REQUIRE(count <= count2, REG_BADBR);
570 			} else		/* single number with comma */
571 				count2 = INFINITY;
572 		} else		/* just a single number */
573 			count2 = count;
574 		repeat(p, pos, count, count2);
575 		if (!EATTWO('\\', '}')) {	/* error heuristics */
576 			while (MORE() && !SEETWO('\\', '}'))
577 				NEXT();
578 			REQUIRE(MORE(), REG_EBRACE);
579 			SETERROR(REG_BADBR);
580 		}
581 	} else if (c == '$')	/* $ (but not \$) ends it */
582 		return(1);
583 
584 	return(0);
585 }
586 
587 /*
588  - p_count - parse a repetition count
589  */
590 static int			/* the value */
591 p_count(struct parse *p)
592 {
593 	int count = 0;
594 	int ndigits = 0;
595 
596 	while (MORE() && isdigit((uch)PEEK()) && count <= DUPMAX) {
597 		count = count*10 + (GETNEXT() - '0');
598 		ndigits++;
599 	}
600 
601 	REQUIRE(ndigits > 0 && count <= DUPMAX, REG_BADBR);
602 	return(count);
603 }
604 
605 /*
606  - p_bracket - parse a bracketed character list
607  *
608  * Note a significant property of this code:  if the allocset() did SETERROR,
609  * no set operations are done.
610  */
611 static void
612 p_bracket(struct parse *p)
613 {
614 	cset *cs;
615 	int invert = 0;
616 
617 	/* Dept of Truly Sickening Special-Case Kludges */
618 	if (p->next + 5 < p->end && strncmp(p->next, "[:<:]]", 6) == 0) {
619 		EMIT(OBOW, 0);
620 		NEXTn(6);
621 		return;
622 	}
623 	if (p->next + 5 < p->end && strncmp(p->next, "[:>:]]", 6) == 0) {
624 		EMIT(OEOW, 0);
625 		NEXTn(6);
626 		return;
627 	}
628 
629 	if ((cs = allocset(p)) == NULL) {
630 		/* allocset did set error status in p */
631 		return;
632 	}
633 
634 	if (EAT('^'))
635 		invert++;	/* make note to invert set at end */
636 	if (EAT(']'))
637 		CHadd(cs, ']');
638 	else if (EAT('-'))
639 		CHadd(cs, '-');
640 	while (MORE() && PEEK() != ']' && !SEETWO('-', ']'))
641 		p_b_term(p, cs);
642 	if (EAT('-'))
643 		CHadd(cs, '-');
644 	MUSTEAT(']', REG_EBRACK);
645 
646 	if (p->error != 0) {	/* don't mess things up further */
647 		freeset(p, cs);
648 		return;
649 	}
650 
651 	if (p->g->cflags&REG_ICASE) {
652 		int i;
653 		int ci;
654 
655 		for (i = p->g->csetsize - 1; i >= 0; i--)
656 			if (CHIN(cs, i) && isalpha(i)) {
657 				ci = othercase(i);
658 				if (ci != i)
659 					CHadd(cs, ci);
660 			}
661 		if (cs->multis != NULL)
662 			mccase(p, cs);
663 	}
664 	if (invert) {
665 		int i;
666 
667 		for (i = p->g->csetsize - 1; i >= 0; i--)
668 			if (CHIN(cs, i))
669 				CHsub(cs, i);
670 			else
671 				CHadd(cs, i);
672 		if (p->g->cflags&REG_NEWLINE)
673 			CHsub(cs, '\n');
674 		if (cs->multis != NULL)
675 			mcinvert(p, cs);
676 	}
677 
678 	assert(cs->multis == NULL);		/* xxx */
679 
680 	if (nch(p, cs) == 1) {		/* optimize singleton sets */
681 		ordinary(p, firstch(p, cs));
682 		freeset(p, cs);
683 	} else
684 		EMIT(OANYOF, freezeset(p, cs));
685 }
686 
687 /*
688  - p_b_term - parse one term of a bracketed character list
689  */
690 static void
691 p_b_term(struct parse *p, cset *cs)
692 {
693 	char c;
694 	char start, finish;
695 	int i;
696 
697 	/* classify what we've got */
698 	switch ((MORE()) ? PEEK() : '\0') {
699 	case '[':
700 		c = (MORE2()) ? PEEK2() : '\0';
701 		break;
702 	case '-':
703 		SETERROR(REG_ERANGE);
704 		return;			/* NOTE RETURN */
705 		break;
706 	default:
707 		c = '\0';
708 		break;
709 	}
710 
711 	switch (c) {
712 	case ':':		/* character class */
713 		NEXT2();
714 		REQUIRE(MORE(), REG_EBRACK);
715 		c = PEEK();
716 		REQUIRE(c != '-' && c != ']', REG_ECTYPE);
717 		p_b_cclass(p, cs);
718 		REQUIRE(MORE(), REG_EBRACK);
719 		REQUIRE(EATTWO(':', ']'), REG_ECTYPE);
720 		break;
721 	case '=':		/* equivalence class */
722 		NEXT2();
723 		REQUIRE(MORE(), REG_EBRACK);
724 		c = PEEK();
725 		REQUIRE(c != '-' && c != ']', REG_ECOLLATE);
726 		p_b_eclass(p, cs);
727 		REQUIRE(MORE(), REG_EBRACK);
728 		REQUIRE(EATTWO('=', ']'), REG_ECOLLATE);
729 		break;
730 	default:		/* symbol, ordinary character, or range */
731 /* xxx revision needed for multichar stuff */
732 		start = p_b_symbol(p);
733 		if (SEE('-') && MORE2() && PEEK2() != ']') {
734 			/* range */
735 			NEXT();
736 			if (EAT('-'))
737 				finish = '-';
738 			else
739 				finish = p_b_symbol(p);
740 		} else
741 			finish = start;
742 /* xxx what about signed chars here... */
743 		REQUIRE(start <= finish, REG_ERANGE);
744 		for (i = start; i <= finish; i++)
745 			CHadd(cs, i);
746 		break;
747 	}
748 }
749 
750 /*
751  - p_b_cclass - parse a character-class name and deal with it
752  */
753 static void
754 p_b_cclass(struct parse *p, cset *cs)
755 {
756 	char *sp = p->next;
757 	struct cclass *cp;
758 	size_t len;
759 	char *u;
760 	char c;
761 
762 	while (MORE() && isalpha(PEEK()))
763 		NEXT();
764 	len = p->next - sp;
765 	for (cp = cclasses; cp->name != NULL; cp++)
766 		if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
767 			break;
768 	if (cp->name == NULL) {
769 		/* oops, didn't find it */
770 		SETERROR(REG_ECTYPE);
771 		return;
772 	}
773 
774 	u = cp->chars;
775 	while ((c = *u++) != '\0')
776 		CHadd(cs, c);
777 	for (u = cp->multis; *u != '\0'; u += strlen(u) + 1)
778 		MCadd(p, cs, u);
779 }
780 
781 /*
782  - p_b_eclass - parse an equivalence-class name and deal with it
783  *
784  * This implementation is incomplete. xxx
785  */
786 static void
787 p_b_eclass(struct parse *p, cset *cs)
788 {
789 	char c;
790 
791 	c = p_b_coll_elem(p, '=');
792 	CHadd(cs, c);
793 }
794 
795 /*
796  - p_b_symbol - parse a character or [..]ed multicharacter collating symbol
797  */
798 static char			/* value of symbol */
799 p_b_symbol(struct parse *p)
800 {
801 	char value;
802 
803 	REQUIRE(MORE(), REG_EBRACK);
804 	if (!EATTWO('[', '.'))
805 		return(GETNEXT());
806 
807 	/* collating symbol */
808 	value = p_b_coll_elem(p, '.');
809 	REQUIRE(EATTWO('.', ']'), REG_ECOLLATE);
810 	return(value);
811 }
812 
813 /*
814  - p_b_coll_elem - parse a collating-element name and look it up
815  */
816 static char			/* value of collating element */
817 p_b_coll_elem(struct parse *p,
818     int endc)			/* name ended by endc,']' */
819 {
820 	char *sp = p->next;
821 	struct cname *cp;
822 	int len;
823 
824 	while (MORE() && !SEETWO(endc, ']'))
825 		NEXT();
826 	if (!MORE()) {
827 		SETERROR(REG_EBRACK);
828 		return(0);
829 	}
830 	len = p->next - sp;
831 	for (cp = cnames; cp->name != NULL; cp++)
832 		if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
833 			return(cp->code);	/* known name */
834 	if (len == 1)
835 		return(*sp);	/* single character */
836 	SETERROR(REG_ECOLLATE);			/* neither */
837 	return(0);
838 }
839 
840 /*
841  - othercase - return the case counterpart of an alphabetic
842  */
843 static char			/* if no counterpart, return ch */
844 othercase(int ch)
845 {
846 	ch = (uch)ch;
847 	assert(isalpha(ch));
848 	if (isupper(ch))
849 		return ((uch)tolower(ch));
850 	else if (islower(ch))
851 		return ((uch)toupper(ch));
852 	else			/* peculiar, but could happen */
853 		return(ch);
854 }
855 
856 /*
857  - bothcases - emit a dualcase version of a two-case character
858  *
859  * Boy, is this implementation ever a kludge...
860  */
861 static void
862 bothcases(struct parse *p, int ch)
863 {
864 	char *oldnext = p->next;
865 	char *oldend = p->end;
866 	char bracket[3];
867 
868 	ch = (uch)ch;
869 	assert(othercase(ch) != ch);	/* p_bracket() would recurse */
870 	p->next = bracket;
871 	p->end = bracket+2;
872 	bracket[0] = ch;
873 	bracket[1] = ']';
874 	bracket[2] = '\0';
875 	p_bracket(p);
876 	assert(p->next == bracket+2);
877 	p->next = oldnext;
878 	p->end = oldend;
879 }
880 
881 /*
882  - ordinary - emit an ordinary character
883  */
884 static void
885 ordinary(struct parse *p, int ch)
886 {
887 	cat_t *cap = p->g->categories;
888 
889 	if ((p->g->cflags&REG_ICASE) && isalpha((uch)ch) && othercase(ch) != ch)
890 		bothcases(p, ch);
891 	else {
892 		EMIT(OCHAR, (uch)ch);
893 		if (cap[ch] == 0)
894 			cap[ch] = p->g->ncategories++;
895 	}
896 }
897 
898 /*
899  - nonnewline - emit REG_NEWLINE version of OANY
900  *
901  * Boy, is this implementation ever a kludge...
902  */
903 static void
904 nonnewline(struct parse *p)
905 {
906 	char *oldnext = p->next;
907 	char *oldend = p->end;
908 	char bracket[4];
909 
910 	p->next = bracket;
911 	p->end = bracket+3;
912 	bracket[0] = '^';
913 	bracket[1] = '\n';
914 	bracket[2] = ']';
915 	bracket[3] = '\0';
916 	p_bracket(p);
917 	assert(p->next == bracket+3);
918 	p->next = oldnext;
919 	p->end = oldend;
920 }
921 
922 /*
923  - repeat - generate code for a bounded repetition, recursively if needed
924  */
925 static void
926 repeat(struct parse *p,
927     sopno start,		/* operand from here to end of strip */
928     int from,			/* repeated from this number */
929     int to)			/* to this number of times (maybe INFINITY) */
930 {
931 	sopno finish = HERE();
932 #	define	N	2
933 #	define	INF	3
934 #	define	REP(f, t)	((f)*8 + (t))
935 #	define	MAP(n)	(((n) <= 1) ? (n) : ((n) == INFINITY) ? INF : N)
936 	sopno copy;
937 
938 	if (p->error != 0)	/* head off possible runaway recursion */
939 		return;
940 
941 	assert(from <= to);
942 
943 	switch (REP(MAP(from), MAP(to))) {
944 	case REP(0, 0):			/* must be user doing this */
945 		DROP(finish-start);	/* drop the operand */
946 		break;
947 	case REP(0, 1):			/* as x{1,1}? */
948 	case REP(0, N):			/* as x{1,n}? */
949 	case REP(0, INF):		/* as x{1,}? */
950 		/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
951 		INSERT(OCH_, start);		/* offset is wrong... */
952 		repeat(p, start+1, 1, to);
953 		ASTERN(OOR1, start);
954 		AHEAD(start);			/* ... fix it */
955 		EMIT(OOR2, 0);
956 		AHEAD(THERE());
957 		ASTERN(O_CH, THERETHERE());
958 		break;
959 	case REP(1, 1):			/* trivial case */
960 		/* done */
961 		break;
962 	case REP(1, N):			/* as x?x{1,n-1} */
963 		/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
964 		INSERT(OCH_, start);
965 		ASTERN(OOR1, start);
966 		AHEAD(start);
967 		EMIT(OOR2, 0);			/* offset very wrong... */
968 		AHEAD(THERE());			/* ...so fix it */
969 		ASTERN(O_CH, THERETHERE());
970 		copy = dupl(p, start+1, finish+1);
971 		assert(copy == finish+4);
972 		repeat(p, copy, 1, to-1);
973 		break;
974 	case REP(1, INF):		/* as x+ */
975 		INSERT(OPLUS_, start);
976 		ASTERN(O_PLUS, start);
977 		break;
978 	case REP(N, N):			/* as xx{m-1,n-1} */
979 		copy = dupl(p, start, finish);
980 		repeat(p, copy, from-1, to-1);
981 		break;
982 	case REP(N, INF):		/* as xx{n-1,INF} */
983 		copy = dupl(p, start, finish);
984 		repeat(p, copy, from-1, to);
985 		break;
986 	default:			/* "can't happen" */
987 		SETERROR(REG_ASSERT);	/* just in case */
988 		break;
989 	}
990 }
991 
992 /*
993  - seterr - set an error condition
994  */
995 static int			/* useless but makes type checking happy */
996 seterr(struct parse *p, int e)
997 {
998 	if (p->error == 0)	/* keep earliest error condition */
999 		p->error = e;
1000 	p->next = nuls;		/* try to bring things to a halt */
1001 	p->end = nuls;
1002 	return(0);		/* make the return value well-defined */
1003 }
1004 
1005 /*
1006  - allocset - allocate a set of characters for []
1007  */
1008 static cset *
1009 allocset(struct parse *p)
1010 {
1011 	int no = p->g->ncsets++;
1012 	size_t nc;
1013 	size_t nbytes;
1014 	cset *cs;
1015 	size_t css = (size_t)p->g->csetsize;
1016 	int i;
1017 
1018 	if (no >= p->ncsalloc) {	/* need another column of space */
1019 		void *ptr;
1020 
1021 		p->ncsalloc += CHAR_BIT;
1022 		nc = p->ncsalloc;
1023 		assert(nc % CHAR_BIT == 0);
1024 		nbytes = nc / CHAR_BIT * css;
1025 
1026 		ptr = (cset *)realloc((char *)p->g->sets, nc * sizeof(cset));
1027 		if (ptr == NULL)
1028 			goto nomem;
1029 		p->g->sets = ptr;
1030 
1031 		ptr = (uch *)realloc((char *)p->g->setbits, nbytes);
1032 		if (ptr == NULL)
1033 			goto nomem;
1034 		p->g->setbits = ptr;
1035 
1036 		for (i = 0; i < no; i++)
1037 			p->g->sets[i].ptr = p->g->setbits + css*(i/CHAR_BIT);
1038 
1039 		(void) memset((char *)p->g->setbits + (nbytes - css), 0, css);
1040 	}
1041 	/* XXX should not happen */
1042 	if (p->g->sets == NULL || p->g->setbits == NULL)
1043 		goto nomem;
1044 
1045 	cs = &p->g->sets[no];
1046 	cs->ptr = p->g->setbits + css*((no)/CHAR_BIT);
1047 	cs->mask = 1 << ((no) % CHAR_BIT);
1048 	cs->hash = 0;
1049 	cs->smultis = 0;
1050 	cs->multis = NULL;
1051 
1052 	return(cs);
1053 nomem:
1054 	free(p->g->sets);
1055 	p->g->sets = NULL;
1056 	free(p->g->setbits);
1057 	p->g->setbits = NULL;
1058 
1059 	SETERROR(REG_ESPACE);
1060 	/* caller's responsibility not to do set ops */
1061 	return(NULL);
1062 }
1063 
1064 /*
1065  - freeset - free a now-unused set
1066  */
1067 static void
1068 freeset(struct parse *p, cset *cs)
1069 {
1070 	int i;
1071 	cset *top = &p->g->sets[p->g->ncsets];
1072 	size_t css = (size_t)p->g->csetsize;
1073 
1074 	for (i = 0; i < css; i++)
1075 		CHsub(cs, i);
1076 	if (cs == top-1)	/* recover only the easy case */
1077 		p->g->ncsets--;
1078 }
1079 
1080 /*
1081  - freezeset - final processing on a set of characters
1082  *
1083  * The main task here is merging identical sets.  This is usually a waste
1084  * of time (although the hash code minimizes the overhead), but can win
1085  * big if REG_ICASE is being used.  REG_ICASE, by the way, is why the hash
1086  * is done using addition rather than xor -- all ASCII [aA] sets xor to
1087  * the same value!
1088  */
1089 static int			/* set number */
1090 freezeset(struct parse *p, cset *cs)
1091 {
1092 	uch h = cs->hash;
1093 	int i;
1094 	cset *top = &p->g->sets[p->g->ncsets];
1095 	cset *cs2;
1096 	size_t css = (size_t)p->g->csetsize;
1097 
1098 	/* look for an earlier one which is the same */
1099 	for (cs2 = &p->g->sets[0]; cs2 < top; cs2++)
1100 		if (cs2->hash == h && cs2 != cs) {
1101 			/* maybe */
1102 			for (i = 0; i < css; i++)
1103 				if (!!CHIN(cs2, i) != !!CHIN(cs, i))
1104 					break;		/* no */
1105 			if (i == css)
1106 				break;			/* yes */
1107 		}
1108 
1109 	if (cs2 < top) {	/* found one */
1110 		freeset(p, cs);
1111 		cs = cs2;
1112 	}
1113 
1114 	return((int)(cs - p->g->sets));
1115 }
1116 
1117 /*
1118  - firstch - return first character in a set (which must have at least one)
1119  */
1120 static int			/* character; there is no "none" value */
1121 firstch(struct parse *p, cset *cs)
1122 {
1123 	int i;
1124 	size_t css = (size_t)p->g->csetsize;
1125 
1126 	for (i = 0; i < css; i++)
1127 		if (CHIN(cs, i))
1128 			return((char)i);
1129 	assert(never);
1130 	return(0);		/* arbitrary */
1131 }
1132 
1133 /*
1134  - nch - number of characters in a set
1135  */
1136 static int
1137 nch(struct parse *p, cset *cs)
1138 {
1139 	int i;
1140 	size_t css = (size_t)p->g->csetsize;
1141 	int n = 0;
1142 
1143 	for (i = 0; i < css; i++)
1144 		if (CHIN(cs, i))
1145 			n++;
1146 	return(n);
1147 }
1148 
1149 /*
1150  - mcadd - add a collating element to a cset
1151  */
1152 static void
1153 mcadd( struct parse *p, cset *cs, char *cp)
1154 {
1155 	size_t oldend = cs->smultis;
1156 	void *np;
1157 
1158 	cs->smultis += strlen(cp) + 1;
1159 	np = realloc(cs->multis, cs->smultis);
1160 	if (np == NULL) {
1161 		if (cs->multis)
1162 			free(cs->multis);
1163 		cs->multis = NULL;
1164 		SETERROR(REG_ESPACE);
1165 		return;
1166 	}
1167 	cs->multis = np;
1168 
1169 	strlcpy(cs->multis + oldend - 1, cp, cs->smultis - oldend + 1);
1170 }
1171 
1172 /*
1173  - mcinvert - invert the list of collating elements in a cset
1174  *
1175  * This would have to know the set of possibilities.  Implementation
1176  * is deferred.
1177  */
1178 /* ARGSUSED */
1179 static void
1180 mcinvert(struct parse *p, cset *cs)
1181 {
1182 	assert(cs->multis == NULL);	/* xxx */
1183 }
1184 
1185 /*
1186  - mccase - add case counterparts of the list of collating elements in a cset
1187  *
1188  * This would have to know the set of possibilities.  Implementation
1189  * is deferred.
1190  */
1191 /* ARGSUSED */
1192 static void
1193 mccase(struct parse *p, cset *cs)
1194 {
1195 	assert(cs->multis == NULL);	/* xxx */
1196 }
1197 
1198 /*
1199  - isinsets - is this character in any sets?
1200  */
1201 static int			/* predicate */
1202 isinsets(struct re_guts *g, int c)
1203 {
1204 	uch *col;
1205 	int i;
1206 	int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1207 	unsigned uc = (uch)c;
1208 
1209 	for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1210 		if (col[uc] != 0)
1211 			return(1);
1212 	return(0);
1213 }
1214 
1215 /*
1216  - samesets - are these two characters in exactly the same sets?
1217  */
1218 static int			/* predicate */
1219 samesets(struct re_guts *g, int c1, int c2)
1220 {
1221 	uch *col;
1222 	int i;
1223 	int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1224 	unsigned uc1 = (uch)c1;
1225 	unsigned uc2 = (uch)c2;
1226 
1227 	for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1228 		if (col[uc1] != col[uc2])
1229 			return(0);
1230 	return(1);
1231 }
1232 
1233 /*
1234  - categorize - sort out character categories
1235  */
1236 static void
1237 categorize(struct parse *p, struct re_guts *g)
1238 {
1239 	cat_t *cats = g->categories;
1240 	int c;
1241 	int c2;
1242 	cat_t cat;
1243 
1244 	/* avoid making error situations worse */
1245 	if (p->error != 0)
1246 		return;
1247 
1248 	for (c = CHAR_MIN; c <= CHAR_MAX; c++)
1249 		if (cats[c] == 0 && isinsets(g, c)) {
1250 			cat = g->ncategories++;
1251 			cats[c] = cat;
1252 			for (c2 = c+1; c2 <= CHAR_MAX; c2++)
1253 				if (cats[c2] == 0 && samesets(g, c, c2))
1254 					cats[c2] = cat;
1255 		}
1256 }
1257 
1258 /*
1259  - dupl - emit a duplicate of a bunch of sops
1260  */
1261 static sopno			/* start of duplicate */
1262 dupl(struct parse *p,
1263     sopno start,		/* from here */
1264     sopno finish)		/* to this less one */
1265 {
1266 	sopno ret = HERE();
1267 	sopno len = finish - start;
1268 
1269 	assert(finish >= start);
1270 	if (len == 0)
1271 		return(ret);
1272 	if (!enlarge(p, p->ssize + len)) /* this many unexpected additions */
1273 		return(ret);
1274 	(void) memcpy((char *)(p->strip + p->slen),
1275 		(char *)(p->strip + start), (size_t)len*sizeof(sop));
1276 	p->slen += len;
1277 	return(ret);
1278 }
1279 
1280 /*
1281  - doemit - emit a strip operator
1282  *
1283  * It might seem better to implement this as a macro with a function as
1284  * hard-case backup, but it's just too big and messy unless there are
1285  * some changes to the data structures.  Maybe later.
1286  */
1287 static void
1288 doemit(struct parse *p, sop op, size_t opnd)
1289 {
1290 	/* avoid making error situations worse */
1291 	if (p->error != 0)
1292 		return;
1293 
1294 	/* deal with oversize operands ("can't happen", more or less) */
1295 	assert(opnd < 1<<OPSHIFT);
1296 
1297 	/* deal with undersized strip */
1298 	if (p->slen >= p->ssize)
1299 		if (!enlarge(p, (p->ssize+1) / 2 * 3))	/* +50% */
1300 			return;
1301 
1302 	/* finally, it's all reduced to the easy case */
1303 	p->strip[p->slen++] = SOP(op, opnd);
1304 }
1305 
1306 /*
1307  - doinsert - insert a sop into the strip
1308  */
1309 static void
1310 doinsert(struct parse *p, sop op, size_t opnd, sopno pos)
1311 {
1312 	sopno sn;
1313 	sop s;
1314 	int i;
1315 
1316 	/* avoid making error situations worse */
1317 	if (p->error != 0)
1318 		return;
1319 
1320 	sn = HERE();
1321 	EMIT(op, opnd);		/* do checks, ensure space */
1322 	assert(HERE() == sn+1);
1323 	s = p->strip[sn];
1324 
1325 	/* adjust paren pointers */
1326 	assert(pos > 0);
1327 	for (i = 1; i < NPAREN; i++) {
1328 		if (p->pbegin[i] >= pos) {
1329 			p->pbegin[i]++;
1330 		}
1331 		if (p->pend[i] >= pos) {
1332 			p->pend[i]++;
1333 		}
1334 	}
1335 
1336 	memmove((char *)&p->strip[pos+1], (char *)&p->strip[pos],
1337 						(HERE()-pos-1)*sizeof(sop));
1338 	p->strip[pos] = s;
1339 }
1340 
1341 /*
1342  - dofwd - complete a forward reference
1343  */
1344 static void
1345 dofwd(struct parse *p, sopno pos, sop value)
1346 {
1347 	/* avoid making error situations worse */
1348 	if (p->error != 0)
1349 		return;
1350 
1351 	assert(value < 1<<OPSHIFT);
1352 	p->strip[pos] = OP(p->strip[pos]) | value;
1353 }
1354 
1355 /*
1356  - enlarge - enlarge the strip
1357  */
1358 static int
1359 enlarge(struct parse *p, sopno size)
1360 {
1361 	sop *sp;
1362 
1363 	if (p->ssize >= size)
1364 		return 1;
1365 
1366 	sp = (sop *)realloc(p->strip, size*sizeof(sop));
1367 	if (sp == NULL) {
1368 		SETERROR(REG_ESPACE);
1369 		return 0;
1370 	}
1371 	p->strip = sp;
1372 	p->ssize = size;
1373 	return 1;
1374 }
1375 
1376 /*
1377  - stripsnug - compact the strip
1378  */
1379 static void
1380 stripsnug(struct parse *p, struct re_guts *g)
1381 {
1382 	g->nstates = p->slen;
1383 	g->strip = (sop *)realloc((char *)p->strip, p->slen * sizeof(sop));
1384 	if (g->strip == NULL) {
1385 		SETERROR(REG_ESPACE);
1386 		g->strip = p->strip;
1387 	}
1388 }
1389 
1390 /*
1391  - findmust - fill in must and mlen with longest mandatory literal string
1392  *
1393  * This algorithm could do fancy things like analyzing the operands of |
1394  * for common subsequences.  Someday.  This code is simple and finds most
1395  * of the interesting cases.
1396  *
1397  * Note that must and mlen got initialized during setup.
1398  */
1399 static void
1400 findmust(struct parse *p, struct re_guts *g)
1401 {
1402 	sop *scan;
1403 	sop *start;    /* start initialized in the default case, after that */
1404 	sop *newstart; /* newstart was initialized in the OCHAR case */
1405 	sopno newlen;
1406 	sop s;
1407 	char *cp;
1408 	sopno i;
1409 
1410 	/* avoid making error situations worse */
1411 	if (p->error != 0)
1412 		return;
1413 
1414 	/* find the longest OCHAR sequence in strip */
1415 	newlen = 0;
1416 	scan = g->strip + 1;
1417 	do {
1418 		s = *scan++;
1419 		switch (OP(s)) {
1420 		case OCHAR:		/* sequence member */
1421 			if (newlen == 0)		/* new sequence */
1422 				newstart = scan - 1;
1423 			newlen++;
1424 			break;
1425 		case OPLUS_:		/* things that don't break one */
1426 		case OLPAREN:
1427 		case ORPAREN:
1428 			break;
1429 		case OQUEST_:		/* things that must be skipped */
1430 		case OCH_:
1431 			scan--;
1432 			do {
1433 				scan += OPND(s);
1434 				s = *scan;
1435 				/* assert() interferes w debug printouts */
1436 				if (OP(s) != O_QUEST && OP(s) != O_CH &&
1437 							OP(s) != OOR2) {
1438 					g->iflags |= BAD;
1439 					return;
1440 				}
1441 			} while (OP(s) != O_QUEST && OP(s) != O_CH);
1442 			/* fallthrough */
1443 		default:		/* things that break a sequence */
1444 			if (newlen > g->mlen) {		/* ends one */
1445 				start = newstart;
1446 				g->mlen = newlen;
1447 			}
1448 			newlen = 0;
1449 			break;
1450 		}
1451 	} while (OP(s) != OEND);
1452 
1453 	if (g->mlen == 0)		/* there isn't one */
1454 		return;
1455 
1456 	/* turn it into a character string */
1457 	g->must = malloc((size_t)g->mlen + 1);
1458 	if (g->must == NULL) {		/* argh; just forget it */
1459 		g->mlen = 0;
1460 		return;
1461 	}
1462 	cp = g->must;
1463 	scan = start;
1464 	for (i = g->mlen; i > 0; i--) {
1465 		while (OP(s = *scan++) != OCHAR)
1466 			continue;
1467 		assert(cp < g->must + g->mlen);
1468 		*cp++ = (char)OPND(s);
1469 	}
1470 	assert(cp == g->must + g->mlen);
1471 	*cp++ = '\0';		/* just on general principles */
1472 }
1473 
1474 /*
1475  - pluscount - count + nesting
1476  */
1477 static sopno			/* nesting depth */
1478 pluscount(struct parse *p, struct re_guts *g)
1479 {
1480 	sop *scan;
1481 	sop s;
1482 	sopno plusnest = 0;
1483 	sopno maxnest = 0;
1484 
1485 	if (p->error != 0)
1486 		return(0);	/* there may not be an OEND */
1487 
1488 	scan = g->strip + 1;
1489 	do {
1490 		s = *scan++;
1491 		switch (OP(s)) {
1492 		case OPLUS_:
1493 			plusnest++;
1494 			break;
1495 		case O_PLUS:
1496 			if (plusnest > maxnest)
1497 				maxnest = plusnest;
1498 			plusnest--;
1499 			break;
1500 		}
1501 	} while (OP(s) != OEND);
1502 	if (plusnest != 0)
1503 		g->iflags |= BAD;
1504 	return(maxnest);
1505 }
1506