xref: /freebsd/contrib/llvm-project/llvm/lib/Support/regcomp.c (revision c8e7f78a3d28ff6e6223ed136ada8e1e2f34965e)
1 /*-
2  * This code is derived from OpenBSD's libc/regex, original license follows:
3  *
4  * Copyright (c) 1992, 1993, 1994 Henry Spencer.
5  * Copyright (c) 1992, 1993, 1994
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This code is derived from software contributed to Berkeley by
9  * Henry Spencer.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  * 3. Neither the name of the University nor the names of its contributors
20  *    may be used to endorse or promote products derived from this software
21  *    without specific prior written permission.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33  * SUCH DAMAGE.
34  *
35  *	@(#)regcomp.c	8.5 (Berkeley) 3/20/94
36  */
37 
38 #include <sys/types.h>
39 #include <stdint.h>
40 #include <stdio.h>
41 #include <string.h>
42 #include <ctype.h>
43 #include <limits.h>
44 #include <stdlib.h>
45 #include "regex_impl.h"
46 
47 #include "regutils.h"
48 #include "regex2.h"
49 
50 #include "llvm/Config/config.h"
51 #include "llvm/Support/Compiler.h"
52 
53 /* character-class table */
54 static struct cclass {
55 	const char *name;
56 	const char *chars;
57 	const char *multis;
58 } cclasses[] = {
59 	{ "alnum",	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\
60 0123456789",				""} ,
61 	{ "alpha",	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz",
62 					""} ,
63 	{ "blank",	" \t",		""} ,
64 	{ "cntrl",	"\007\b\t\n\v\f\r\1\2\3\4\5\6\16\17\20\21\22\23\24\
65 \25\26\27\30\31\32\33\34\35\36\37\177",	""} ,
66 	{ "digit",	"0123456789",	""} ,
67 	{ "graph",	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\
68 0123456789!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~",
69 					""} ,
70 	{ "lower",	"abcdefghijklmnopqrstuvwxyz",
71 					""} ,
72 	{ "print",	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\
73 0123456789!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~ ",
74 					""} ,
75 	{ "punct",	"!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~",
76 					""} ,
77 	{ "space",	"\t\n\v\f\r ",	""} ,
78 	{ "upper",	"ABCDEFGHIJKLMNOPQRSTUVWXYZ",
79 					""} ,
80 	{ "xdigit",	"0123456789ABCDEFabcdef",
81 					""} ,
82 	{ NULL,		0,		"" }
83 };
84 
85 /* character-name table */
86 static struct cname {
87 	const char *name;
88 	char code;
89 } cnames[] = {
90 	{ "NUL",			'\0' },
91 	{ "SOH",			'\001' },
92 	{ "STX",			'\002' },
93 	{ "ETX",			'\003' },
94 	{ "EOT",			'\004' },
95 	{ "ENQ",			'\005' },
96 	{ "ACK",			'\006' },
97 	{ "BEL",			'\007' },
98 	{ "alert",			'\007' },
99 	{ "BS",				'\010' },
100 	{ "backspace",			'\b' },
101 	{ "HT",				'\011' },
102 	{ "tab",			'\t' },
103 	{ "LF",				'\012' },
104 	{ "newline",			'\n' },
105 	{ "VT",				'\013' },
106 	{ "vertical-tab",		'\v' },
107 	{ "FF",				'\014' },
108 	{ "form-feed",			'\f' },
109 	{ "CR",				'\015' },
110 	{ "carriage-return",		'\r' },
111 	{ "SO",				'\016' },
112 	{ "SI",				'\017' },
113 	{ "DLE",			'\020' },
114 	{ "DC1",			'\021' },
115 	{ "DC2",			'\022' },
116 	{ "DC3",			'\023' },
117 	{ "DC4",			'\024' },
118 	{ "NAK",			'\025' },
119 	{ "SYN",			'\026' },
120 	{ "ETB",			'\027' },
121 	{ "CAN",			'\030' },
122 	{ "EM",				'\031' },
123 	{ "SUB",			'\032' },
124 	{ "ESC",			'\033' },
125 	{ "IS4",			'\034' },
126 	{ "FS",				'\034' },
127 	{ "IS3",			'\035' },
128 	{ "GS",				'\035' },
129 	{ "IS2",			'\036' },
130 	{ "RS",				'\036' },
131 	{ "IS1",			'\037' },
132 	{ "US",				'\037' },
133 	{ "space",			' ' },
134 	{ "exclamation-mark",		'!' },
135 	{ "quotation-mark",		'"' },
136 	{ "number-sign",		'#' },
137 	{ "dollar-sign",		'$' },
138 	{ "percent-sign",		'%' },
139 	{ "ampersand",			'&' },
140 	{ "apostrophe",			'\'' },
141 	{ "left-parenthesis",		'(' },
142 	{ "right-parenthesis",		')' },
143 	{ "asterisk",			'*' },
144 	{ "plus-sign",			'+' },
145 	{ "comma",			',' },
146 	{ "hyphen",			'-' },
147 	{ "hyphen-minus",		'-' },
148 	{ "period",			'.' },
149 	{ "full-stop",			'.' },
150 	{ "slash",			'/' },
151 	{ "solidus",			'/' },
152 	{ "zero",			'0' },
153 	{ "one",			'1' },
154 	{ "two",			'2' },
155 	{ "three",			'3' },
156 	{ "four",			'4' },
157 	{ "five",			'5' },
158 	{ "six",			'6' },
159 	{ "seven",			'7' },
160 	{ "eight",			'8' },
161 	{ "nine",			'9' },
162 	{ "colon",			':' },
163 	{ "semicolon",			';' },
164 	{ "less-than-sign",		'<' },
165 	{ "equals-sign",		'=' },
166 	{ "greater-than-sign",		'>' },
167 	{ "question-mark",		'?' },
168 	{ "commercial-at",		'@' },
169 	{ "left-square-bracket",	'[' },
170 	{ "backslash",			'\\' },
171 	{ "reverse-solidus",		'\\' },
172 	{ "right-square-bracket",	']' },
173 	{ "circumflex",			'^' },
174 	{ "circumflex-accent",		'^' },
175 	{ "underscore",			'_' },
176 	{ "low-line",			'_' },
177 	{ "grave-accent",		'`' },
178 	{ "left-brace",			'{' },
179 	{ "left-curly-bracket",		'{' },
180 	{ "vertical-line",		'|' },
181 	{ "right-brace",		'}' },
182 	{ "right-curly-bracket",	'}' },
183 	{ "tilde",			'~' },
184 	{ "DEL",			'\177' },
185 	{ NULL,				0 }
186 };
187 
188 /*
189  * parse structure, passed up and down to avoid global variables and
190  * other clumsinesses
191  */
192 struct parse {
193 	char *next;		/* next character in RE */
194 	char *end;		/* end of string (-> NUL normally) */
195 	int error;		/* has an error been seen? */
196 	sop *strip;		/* malloced strip */
197 	sopno ssize;		/* malloced strip size (allocated) */
198 	sopno slen;		/* malloced strip length (used) */
199 	int ncsalloc;		/* number of csets allocated */
200 	struct re_guts *g;
201 #	define	NPAREN	10	/* we need to remember () 1-9 for back refs */
202 	sopno pbegin[NPAREN];	/* -> ( ([0] unused) */
203 	sopno pend[NPAREN];	/* -> ) ([0] unused) */
204 };
205 
206 static void p_ere(struct parse *, int);
207 static void p_ere_exp(struct parse *);
208 static void p_str(struct parse *);
209 static void p_bre(struct parse *, int, int);
210 static int p_simp_re(struct parse *, int);
211 static int p_count(struct parse *);
212 static void p_bracket(struct parse *);
213 static void p_b_term(struct parse *, cset *);
214 static void p_b_cclass(struct parse *, cset *);
215 static void p_b_eclass(struct parse *, cset *);
216 static char p_b_symbol(struct parse *);
217 static char p_b_coll_elem(struct parse *, int);
218 static char othercase(int);
219 static void bothcases(struct parse *, int);
220 static void ordinary(struct parse *, int);
221 static void nonnewline(struct parse *);
222 static void repeat(struct parse *, sopno, int, int);
223 static int seterr(struct parse *, int);
224 static cset *allocset(struct parse *);
225 static void freeset(struct parse *, cset *);
226 static int freezeset(struct parse *, cset *);
227 static int firstch(struct parse *, cset *);
228 static int nch(struct parse *, cset *);
229 static void mcadd(struct parse *, cset *, const char *);
230 static void mcinvert(struct parse *, cset *);
231 static void mccase(struct parse *, cset *);
232 static int isinsets(struct re_guts *, int);
233 static int samesets(struct re_guts *, int, int);
234 static void categorize(struct parse *, struct re_guts *);
235 static sopno dupl(struct parse *, sopno, sopno);
236 static void doemit(struct parse *, sop, size_t);
237 static void doinsert(struct parse *, sop, size_t, sopno);
238 static void dofwd(struct parse *, sopno, sop);
239 static void enlarge(struct parse *, sopno);
240 static void stripsnug(struct parse *, struct re_guts *);
241 static void findmust(struct parse *, struct re_guts *);
242 static sopno pluscount(struct parse *, struct re_guts *);
243 
244 static char nuls[10];		/* place to point scanner in event of error */
245 
246 /*
247  * macros for use with parse structure
248  * BEWARE:  these know that the parse structure is named `p' !!!
249  */
250 #define	PEEK()	(*p->next)
251 #define	PEEK2()	(*(p->next+1))
252 #define	MORE()	(p->end - p->next > 0)
253 #define	MORE2()	(p->end - p->next > 1)
254 #define	SEE(c)	(MORE() && PEEK() == (c))
255 #define	SEETWO(a, b)	(MORE2() && PEEK() == (a) && PEEK2() == (b))
256 #define	EAT(c)	((SEE(c)) ? (NEXT(), 1) : 0)
257 #define	EATTWO(a, b)	((SEETWO(a, b)) ? (NEXT2(), 1) : 0)
258 #define	NEXT()	(p->next++)
259 #define	NEXT2()	(p->next += 2)
260 #define	NEXTn(n)	(p->next += (n))
261 #define	GETNEXT()	(*p->next++)
262 #define	SETERROR(e)	seterr(p, (e))
263 #define	REQUIRE(co, e)	(void)((co) || SETERROR(e))
264 #define	MUSTSEE(c, e)	(REQUIRE(MORE() && PEEK() == (c), e))
265 #define	MUSTEAT(c, e)	(REQUIRE(MORE() && GETNEXT() == (c), e))
266 #define	MUSTNOTSEE(c, e)	(REQUIRE(!MORE() || PEEK() != (c), e))
267 #define	EMIT(op, sopnd)	doemit(p, (sop)(op), (size_t)(sopnd))
268 #define	INSERT(op, pos)	doinsert(p, (sop)(op), HERE()-(pos)+1, pos)
269 #define	AHEAD(pos)		dofwd(p, pos, HERE()-(pos))
270 #define	ASTERN(sop, pos)	EMIT(sop, HERE()-pos)
271 #define	HERE()		(p->slen)
272 #define	THERE()		(p->slen - 1)
273 #define	THERETHERE()	(p->slen - 2)
274 #define	DROP(n)	(p->slen -= (n))
275 
276 #ifdef	_POSIX2_RE_DUP_MAX
277 #define	DUPMAX	_POSIX2_RE_DUP_MAX
278 #else
279 #define	DUPMAX	255
280 #endif
281 #define	INFINITY	(DUPMAX + 1)
282 
283 #ifndef NDEBUG
284 static int never = 0;		/* for use in asserts; shuts lint up */
285 #else
286 #define	never	0		/* some <assert.h>s have bugs too */
287 #endif
288 
289 /*
290  - llvm_regcomp - interface for parser and compilation
291  */
292 int				/* 0 success, otherwise REG_something */
293 llvm_regcomp(llvm_regex_t *preg, const char *pattern, int cflags)
294 {
295 	struct parse pa;
296 	struct re_guts *g;
297 	struct parse *p = &pa;
298 	int i;
299 	size_t len;
300 #ifdef REDEBUG
301 #	define	GOODFLAGS(f)	(f)
302 #else
303 #	define	GOODFLAGS(f)	((f)&~REG_DUMP)
304 #endif
305 
306 	cflags = GOODFLAGS(cflags);
307 	if ((cflags&REG_EXTENDED) && (cflags&REG_NOSPEC))
308 		return(REG_INVARG);
309 
310 	if (cflags&REG_PEND) {
311 		if (preg->re_endp < pattern)
312 			return(REG_INVARG);
313 		len = preg->re_endp - pattern;
314 	} else
315 		len = strlen((const char *)pattern);
316 
317 	/* do the mallocs early so failure handling is easy */
318 	g = (struct re_guts *)malloc(sizeof(struct re_guts) +
319 							(NC-1)*sizeof(cat_t));
320 	if (g == NULL)
321 		return(REG_ESPACE);
322 	p->ssize = len/(size_t)2*(size_t)3 + (size_t)1;	/* ugh */
323 	p->strip = (sop *)calloc(p->ssize, sizeof(sop));
324 	p->slen = 0;
325 	if (p->strip == NULL) {
326 		free((char *)g);
327 		return(REG_ESPACE);
328 	}
329 
330 	/* set things up */
331 	p->g = g;
332 	/* suppress warning from the following explicit cast. */
333 #ifdef __GNUC__
334 #pragma GCC diagnostic push
335 #pragma GCC diagnostic ignored "-Wcast-qual"
336 #endif /* __GNUC__ */
337 	p->next = (char *)pattern;	/* convenience; we do not modify it */
338 #ifdef __GNUC__
339 #pragma GCC diagnostic pop
340 #endif /* __GNUC__ */
341 	p->end = p->next + len;
342 	p->error = 0;
343 	p->ncsalloc = 0;
344 	for (i = 0; i < NPAREN; i++) {
345 		p->pbegin[i] = 0;
346 		p->pend[i] = 0;
347 	}
348 	g->csetsize = NC;
349 	g->sets = NULL;
350 	g->setbits = NULL;
351 	g->ncsets = 0;
352 	g->cflags = cflags;
353 	g->iflags = 0;
354 	g->nbol = 0;
355 	g->neol = 0;
356 	g->must = NULL;
357 	g->mlen = 0;
358 	g->nsub = 0;
359 	g->ncategories = 1;	/* category 0 is "everything else" */
360 	g->categories = &g->catspace[-(CHAR_MIN)];
361 	(void) memset((char *)g->catspace, 0, NC*sizeof(cat_t));
362 	g->backrefs = 0;
363 
364 	/* do it */
365 	EMIT(OEND, 0);
366 	g->firststate = THERE();
367 	if (cflags&REG_EXTENDED)
368 		p_ere(p, OUT);
369 	else if (cflags&REG_NOSPEC)
370 		p_str(p);
371 	else
372 		p_bre(p, OUT, OUT);
373 	EMIT(OEND, 0);
374 	g->laststate = THERE();
375 
376 	/* tidy up loose ends and fill things in */
377 	categorize(p, g);
378 	stripsnug(p, g);
379 	findmust(p, g);
380 	g->nplus = pluscount(p, g);
381 	g->magic = MAGIC2;
382 	preg->re_nsub = g->nsub;
383 	preg->re_g = g;
384 	preg->re_magic = MAGIC1;
385 #ifndef REDEBUG
386 	/* not debugging, so can't rely on the assert() in llvm_regexec() */
387 	if (g->iflags&REGEX_BAD)
388 		SETERROR(REG_ASSERT);
389 #endif
390 
391 	/* win or lose, we're done */
392 	if (p->error != 0)	/* lose */
393 		llvm_regfree(preg);
394 	return(p->error);
395 }
396 
397 /*
398  - p_ere - ERE parser top level, concatenation and alternation
399  */
400 static void
401 p_ere(struct parse *p, int stop)	/* character this ERE should end at */
402 {
403 	char c;
404 	sopno prevback = 0;
405 	sopno prevfwd = 0;
406 	sopno conc;
407 	int first = 1;		/* is this the first alternative? */
408 
409 	for (;;) {
410 		/* do a bunch of concatenated expressions */
411 		conc = HERE();
412 		while (MORE() && (c = PEEK()) != '|' && c != stop)
413 			p_ere_exp(p);
414 		REQUIRE(HERE() != conc, REG_EMPTY);	/* require nonempty */
415 
416 		if (!EAT('|'))
417 			break;		/* NOTE BREAK OUT */
418 
419 		if (first) {
420 			INSERT(OCH_, conc);	/* offset is wrong */
421 			prevfwd = conc;
422 			prevback = conc;
423 			first = 0;
424 		}
425 		ASTERN(OOR1, prevback);
426 		prevback = THERE();
427 		AHEAD(prevfwd);			/* fix previous offset */
428 		prevfwd = HERE();
429 		EMIT(OOR2, 0);			/* offset is very wrong */
430 	}
431 
432 	if (!first) {		/* tail-end fixups */
433 		AHEAD(prevfwd);
434 		ASTERN(O_CH, prevback);
435 	}
436 
437 	assert(!MORE() || SEE(stop));
438 }
439 
440 /*
441  - p_ere_exp - parse one subERE, an atom possibly followed by a repetition op
442  */
443 static void
444 p_ere_exp(struct parse *p)
445 {
446 	char c;
447 	sopno pos;
448 	int count;
449 	int count2;
450 	int backrefnum;
451 	sopno subno;
452 	int wascaret = 0;
453 
454 	assert(MORE());		/* caller should have ensured this */
455 	c = GETNEXT();
456 
457 	pos = HERE();
458 	switch (c) {
459 	case '(':
460 		REQUIRE(MORE(), REG_EPAREN);
461 		p->g->nsub++;
462 		subno = p->g->nsub;
463 		if (subno < NPAREN)
464 			p->pbegin[subno] = HERE();
465 		EMIT(OLPAREN, subno);
466 		if (!SEE(')'))
467 			p_ere(p, ')');
468 		if (subno < NPAREN) {
469 			p->pend[subno] = HERE();
470 			assert(p->pend[subno] != 0);
471 		}
472 		EMIT(ORPAREN, subno);
473 		MUSTEAT(')', REG_EPAREN);
474 		break;
475 #ifndef POSIX_MISTAKE
476 	case ')':		/* happens only if no current unmatched ( */
477 		/*
478 		 * You may ask, why the ifndef?  Because I didn't notice
479 		 * this until slightly too late for 1003.2, and none of the
480 		 * other 1003.2 regular-expression reviewers noticed it at
481 		 * all.  So an unmatched ) is legal POSIX, at least until
482 		 * we can get it fixed.
483 		 */
484 		SETERROR(REG_EPAREN);
485 		break;
486 #endif
487 	case '^':
488 		EMIT(OBOL, 0);
489 		p->g->iflags |= USEBOL;
490 		p->g->nbol++;
491 		wascaret = 1;
492 		break;
493 	case '$':
494 		EMIT(OEOL, 0);
495 		p->g->iflags |= USEEOL;
496 		p->g->neol++;
497 		break;
498 	case '|':
499 		SETERROR(REG_EMPTY);
500 		break;
501 	case '*':
502 	case '+':
503 	case '?':
504 		SETERROR(REG_BADRPT);
505 		break;
506 	case '.':
507 		if (p->g->cflags&REG_NEWLINE)
508 			nonnewline(p);
509 		else
510 			EMIT(OANY, 0);
511 		break;
512 	case '[':
513 		p_bracket(p);
514 		break;
515 	case '\\':
516 		REQUIRE(MORE(), REG_EESCAPE);
517 		c = GETNEXT();
518 		if (c >= '1' && c <= '9') {
519 			/* \[0-9] is taken to be a back-reference to a previously specified
520 			 * matching group. backrefnum will hold the number. The matching
521 			 * group must exist (i.e. if \4 is found there must have been at
522 			 * least 4 matching groups specified in the pattern previously).
523 			 */
524 			backrefnum = c - '0';
525 			if (p->pend[backrefnum] == 0) {
526 				SETERROR(REG_ESUBREG);
527 				break;
528 			}
529 
530 			/* Make sure everything checks out and emit the sequence
531 			 * that marks a back-reference to the parse structure.
532 			 */
533 			assert(backrefnum <= p->g->nsub);
534 			EMIT(OBACK_, backrefnum);
535 			assert(p->pbegin[backrefnum] != 0);
536 			assert(OP(p->strip[p->pbegin[backrefnum]]) == OLPAREN);
537 			assert(OP(p->strip[p->pend[backrefnum]]) == ORPAREN);
538 			(void) dupl(p, p->pbegin[backrefnum]+1, p->pend[backrefnum]);
539 			EMIT(O_BACK, backrefnum);
540 			p->g->backrefs = 1;
541 		} else {
542 			/* Other chars are simply themselves when escaped with a backslash.
543 			 */
544 			ordinary(p, c);
545 		}
546 		break;
547 	case '{':		/* okay as ordinary except if digit follows */
548 		REQUIRE(!MORE() || !isdigit((uch)PEEK()), REG_BADRPT);
549 		LLVM_FALLTHROUGH;
550 	default:
551 		ordinary(p, c);
552 		break;
553 	}
554 
555 	if (!MORE())
556 		return;
557 	c = PEEK();
558 	/* we call { a repetition if followed by a digit */
559 	if (!( c == '*' || c == '+' || c == '?' ||
560 				(c == '{' && MORE2() && isdigit((uch)PEEK2())) ))
561 		return;		/* no repetition, we're done */
562 	NEXT();
563 
564 	REQUIRE(!wascaret, REG_BADRPT);
565 	switch (c) {
566 	case '*':	/* implemented as +? */
567 		/* this case does not require the (y|) trick, noKLUDGE */
568 		INSERT(OPLUS_, pos);
569 		ASTERN(O_PLUS, pos);
570 		INSERT(OQUEST_, pos);
571 		ASTERN(O_QUEST, pos);
572 		break;
573 	case '+':
574 		INSERT(OPLUS_, pos);
575 		ASTERN(O_PLUS, pos);
576 		break;
577 	case '?':
578 		/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
579 		INSERT(OCH_, pos);		/* offset slightly wrong */
580 		ASTERN(OOR1, pos);		/* this one's right */
581 		AHEAD(pos);			/* fix the OCH_ */
582 		EMIT(OOR2, 0);			/* offset very wrong... */
583 		AHEAD(THERE());			/* ...so fix it */
584 		ASTERN(O_CH, THERETHERE());
585 		break;
586 	case '{':
587 		count = p_count(p);
588 		if (EAT(',')) {
589 			if (isdigit((uch)PEEK())) {
590 				count2 = p_count(p);
591 				REQUIRE(count <= count2, REG_BADBR);
592 			} else		/* single number with comma */
593 				count2 = INFINITY;
594 		} else		/* just a single number */
595 			count2 = count;
596 		repeat(p, pos, count, count2);
597 		if (!EAT('}')) {	/* error heuristics */
598 			while (MORE() && PEEK() != '}')
599 				NEXT();
600 			REQUIRE(MORE(), REG_EBRACE);
601 			SETERROR(REG_BADBR);
602 		}
603 		break;
604 	}
605 
606 	if (!MORE())
607 		return;
608 	c = PEEK();
609 	if (!( c == '*' || c == '+' || c == '?' ||
610 				(c == '{' && MORE2() && isdigit((uch)PEEK2())) ) )
611 		return;
612 	SETERROR(REG_BADRPT);
613 }
614 
615 /*
616  - p_str - string (no metacharacters) "parser"
617  */
618 static void
619 p_str(struct parse *p)
620 {
621 	REQUIRE(MORE(), REG_EMPTY);
622 	while (MORE())
623 		ordinary(p, GETNEXT());
624 }
625 
626 /*
627  - p_bre - BRE parser top level, anchoring and concatenation
628  * Giving end1 as OUT essentially eliminates the end1/end2 check.
629  *
630  * This implementation is a bit of a kludge, in that a trailing $ is first
631  * taken as an ordinary character and then revised to be an anchor.  The
632  * only undesirable side effect is that '$' gets included as a character
633  * category in such cases.  This is fairly harmless; not worth fixing.
634  * The amount of lookahead needed to avoid this kludge is excessive.
635  */
636 static void
637 p_bre(struct parse *p,
638     int end1,		/* first terminating character */
639     int end2)		/* second terminating character */
640 {
641 	sopno start = HERE();
642 	int first = 1;			/* first subexpression? */
643 	int wasdollar = 0;
644 
645 	if (EAT('^')) {
646 		EMIT(OBOL, 0);
647 		p->g->iflags |= USEBOL;
648 		p->g->nbol++;
649 	}
650 	while (MORE() && !SEETWO(end1, end2)) {
651 		wasdollar = p_simp_re(p, first);
652 		first = 0;
653 	}
654 	if (wasdollar) {	/* oops, that was a trailing anchor */
655 		DROP(1);
656 		EMIT(OEOL, 0);
657 		p->g->iflags |= USEEOL;
658 		p->g->neol++;
659 	}
660 
661 	REQUIRE(HERE() != start, REG_EMPTY);	/* require nonempty */
662 }
663 
664 /*
665  - p_simp_re - parse a simple RE, an atom possibly followed by a repetition
666  */
667 static int			/* was the simple RE an unbackslashed $? */
668 p_simp_re(struct parse *p,
669     int starordinary)		/* is a leading * an ordinary character? */
670 {
671 	int c;
672 	int count;
673 	int count2;
674 	sopno pos;
675 	int i;
676 	sopno subno;
677 #	define	BACKSL	(1<<CHAR_BIT)
678 
679         pos = HERE(); /* repetition op, if any, covers from here */
680 
681         assert(MORE()); /* caller should have ensured this */
682         c = GETNEXT();
683 	if (c == '\\') {
684 		REQUIRE(MORE(), REG_EESCAPE);
685 		c = BACKSL | GETNEXT();
686 	}
687 	switch (c) {
688 	case '.':
689 		if (p->g->cflags&REG_NEWLINE)
690 			nonnewline(p);
691 		else
692 			EMIT(OANY, 0);
693 		break;
694 	case '[':
695 		p_bracket(p);
696 		break;
697 	case BACKSL|'{':
698 		SETERROR(REG_BADRPT);
699 		break;
700 	case BACKSL|'(':
701 		p->g->nsub++;
702 		subno = p->g->nsub;
703 		if (subno < NPAREN)
704 			p->pbegin[subno] = HERE();
705 		EMIT(OLPAREN, subno);
706 		/* the MORE here is an error heuristic */
707 		if (MORE() && !SEETWO('\\', ')'))
708 			p_bre(p, '\\', ')');
709 		if (subno < NPAREN) {
710 			p->pend[subno] = HERE();
711 			assert(p->pend[subno] != 0);
712 		}
713 		EMIT(ORPAREN, subno);
714 		REQUIRE(EATTWO('\\', ')'), REG_EPAREN);
715 		break;
716 	case BACKSL|')':	/* should not get here -- must be user */
717 	case BACKSL|'}':
718 		SETERROR(REG_EPAREN);
719 		break;
720 	case BACKSL|'1':
721 	case BACKSL|'2':
722 	case BACKSL|'3':
723 	case BACKSL|'4':
724 	case BACKSL|'5':
725 	case BACKSL|'6':
726 	case BACKSL|'7':
727 	case BACKSL|'8':
728 	case BACKSL|'9':
729 		i = (c&~BACKSL) - '0';
730 		assert(i < NPAREN);
731 		if (p->pend[i] != 0) {
732 			assert(i <= p->g->nsub);
733 			EMIT(OBACK_, i);
734 			assert(p->pbegin[i] != 0);
735 			assert(OP(p->strip[p->pbegin[i]]) == OLPAREN);
736 			assert(OP(p->strip[p->pend[i]]) == ORPAREN);
737 			(void) dupl(p, p->pbegin[i]+1, p->pend[i]);
738 			EMIT(O_BACK, i);
739 		} else
740 			SETERROR(REG_ESUBREG);
741 		p->g->backrefs = 1;
742 		break;
743 	case '*':
744 		REQUIRE(starordinary, REG_BADRPT);
745 		LLVM_FALLTHROUGH;
746 	default:
747 		ordinary(p, (char)c);
748 		break;
749 	}
750 
751 	if (EAT('*')) {		/* implemented as +? */
752 		/* this case does not require the (y|) trick, noKLUDGE */
753 		INSERT(OPLUS_, pos);
754 		ASTERN(O_PLUS, pos);
755 		INSERT(OQUEST_, pos);
756 		ASTERN(O_QUEST, pos);
757 	} else if (EATTWO('\\', '{')) {
758 		count = p_count(p);
759 		if (EAT(',')) {
760 			if (MORE() && isdigit((uch)PEEK())) {
761 				count2 = p_count(p);
762 				REQUIRE(count <= count2, REG_BADBR);
763 			} else		/* single number with comma */
764 				count2 = INFINITY;
765 		} else		/* just a single number */
766 			count2 = count;
767 		repeat(p, pos, count, count2);
768 		if (!EATTWO('\\', '}')) {	/* error heuristics */
769 			while (MORE() && !SEETWO('\\', '}'))
770 				NEXT();
771 			REQUIRE(MORE(), REG_EBRACE);
772 			SETERROR(REG_BADBR);
773 		}
774 	} else if (c == '$')	/* $ (but not \$) ends it */
775 		return(1);
776 
777 	return(0);
778 }
779 
780 /*
781  - p_count - parse a repetition count
782  */
783 static int			/* the value */
784 p_count(struct parse *p)
785 {
786 	int count = 0;
787 	int ndigits = 0;
788 
789 	while (MORE() && isdigit((uch)PEEK()) && count <= DUPMAX) {
790 		count = count*10 + (GETNEXT() - '0');
791 		ndigits++;
792 	}
793 
794 	REQUIRE(ndigits > 0 && count <= DUPMAX, REG_BADBR);
795 	return(count);
796 }
797 
798 /*
799  - p_bracket - parse a bracketed character list
800  *
801  * Note a significant property of this code:  if the allocset() did SETERROR,
802  * no set operations are done.
803  */
804 static void
805 p_bracket(struct parse *p)
806 {
807 	cset *cs;
808 	int invert = 0;
809 
810 	/* Dept of Truly Sickening Special-Case Kludges */
811 	if (p->end - p->next > 5) {
812 		if (strncmp(p->next, "[:<:]]", 6) == 0) {
813 			EMIT(OBOW, 0);
814 			NEXTn(6);
815 			return;
816 		}
817 		if (strncmp(p->next, "[:>:]]", 6) == 0) {
818 			EMIT(OEOW, 0);
819 			NEXTn(6);
820 			return;
821 		}
822 	}
823 
824 	if ((cs = allocset(p)) == NULL) {
825 		/* allocset did set error status in p */
826 		return;
827 	}
828 
829 	if (EAT('^'))
830 		invert++;	/* make note to invert set at end */
831 	if (EAT(']'))
832 		CHadd(cs, ']');
833 	else if (EAT('-'))
834 		CHadd(cs, '-');
835 	while (MORE() && PEEK() != ']' && !SEETWO('-', ']'))
836 		p_b_term(p, cs);
837 	if (EAT('-'))
838 		CHadd(cs, '-');
839 	MUSTEAT(']', REG_EBRACK);
840 
841 	if (p->error != 0) {	/* don't mess things up further */
842 		freeset(p, cs);
843 		return;
844 	}
845 
846 	if (p->g->cflags&REG_ICASE) {
847 		int i;
848 		int ci;
849 
850 		for (i = p->g->csetsize - 1; i >= 0; i--)
851 			if (CHIN(cs, i) && isalpha(i)) {
852 				ci = othercase(i);
853 				if (ci != i)
854 					CHadd(cs, ci);
855 			}
856 		if (cs->multis != NULL)
857 			mccase(p, cs);
858 	}
859 	if (invert) {
860 		int i;
861 
862 		for (i = p->g->csetsize - 1; i >= 0; i--)
863 			if (CHIN(cs, i))
864 				CHsub(cs, i);
865 			else
866 				CHadd(cs, i);
867 		if (p->g->cflags&REG_NEWLINE)
868 			CHsub(cs, '\n');
869 		if (cs->multis != NULL)
870 			mcinvert(p, cs);
871 	}
872 
873 	assert(cs->multis == NULL);		/* xxx */
874 
875 	if (nch(p, cs) == 1) {		/* optimize singleton sets */
876 		ordinary(p, firstch(p, cs));
877 		freeset(p, cs);
878 	} else
879 		EMIT(OANYOF, freezeset(p, cs));
880 }
881 
882 /*
883  - p_b_term - parse one term of a bracketed character list
884  */
885 static void
886 p_b_term(struct parse *p, cset *cs)
887 {
888 	char c;
889 	char start, finish;
890 	int i;
891 
892 	/* classify what we've got */
893 	switch ((MORE()) ? PEEK() : '\0') {
894 	case '[':
895 		c = (MORE2()) ? PEEK2() : '\0';
896 		break;
897 	case '-':
898 		SETERROR(REG_ERANGE);
899 		return;			/* NOTE RETURN */
900 		break;
901 	default:
902 		c = '\0';
903 		break;
904 	}
905 
906 	switch (c) {
907 	case ':':		/* character class */
908 		NEXT2();
909 		REQUIRE(MORE(), REG_EBRACK);
910 		c = PEEK();
911 		REQUIRE(c != '-' && c != ']', REG_ECTYPE);
912 		p_b_cclass(p, cs);
913 		REQUIRE(MORE(), REG_EBRACK);
914 		REQUIRE(EATTWO(':', ']'), REG_ECTYPE);
915 		break;
916 	case '=':		/* equivalence class */
917 		NEXT2();
918 		REQUIRE(MORE(), REG_EBRACK);
919 		c = PEEK();
920 		REQUIRE(c != '-' && c != ']', REG_ECOLLATE);
921 		p_b_eclass(p, cs);
922 		REQUIRE(MORE(), REG_EBRACK);
923 		REQUIRE(EATTWO('=', ']'), REG_ECOLLATE);
924 		break;
925 	default:		/* symbol, ordinary character, or range */
926 /* xxx revision needed for multichar stuff */
927 		start = p_b_symbol(p);
928 		if (SEE('-') && MORE2() && PEEK2() != ']') {
929 			/* range */
930 			NEXT();
931 			if (EAT('-'))
932 				finish = '-';
933 			else
934 				finish = p_b_symbol(p);
935 		} else
936 			finish = start;
937 /* xxx what about signed chars here... */
938 		REQUIRE(start <= finish, REG_ERANGE);
939 		for (i = start; i <= finish; i++)
940 			CHadd(cs, i);
941 		break;
942 	}
943 }
944 
945 /*
946  - p_b_cclass - parse a character-class name and deal with it
947  */
948 static void
949 p_b_cclass(struct parse *p, cset *cs)
950 {
951 	char *sp = p->next;
952 	struct cclass *cp;
953 	size_t len;
954 	const char *u;
955 	char c;
956 
957 	while (MORE() && isalpha((uch)PEEK()))
958 		NEXT();
959 	len = p->next - sp;
960 	for (cp = cclasses; cp->name != NULL; cp++)
961 		if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
962 			break;
963 	if (cp->name == NULL) {
964 		/* oops, didn't find it */
965 		SETERROR(REG_ECTYPE);
966 		return;
967 	}
968 
969 	u = cp->chars;
970 	while ((c = *u++) != '\0')
971 		CHadd(cs, c);
972 	for (u = cp->multis; *u != '\0'; u += strlen(u) + 1)
973 		MCadd(p, cs, u);
974 }
975 
976 /*
977  - p_b_eclass - parse an equivalence-class name and deal with it
978  *
979  * This implementation is incomplete. xxx
980  */
981 static void
982 p_b_eclass(struct parse *p, cset *cs)
983 {
984 	char c;
985 
986 	c = p_b_coll_elem(p, '=');
987 	CHadd(cs, c);
988 }
989 
990 /*
991  - p_b_symbol - parse a character or [..]ed multicharacter collating symbol
992  */
993 static char			/* value of symbol */
994 p_b_symbol(struct parse *p)
995 {
996 	char value;
997 
998 	REQUIRE(MORE(), REG_EBRACK);
999 	if (!EATTWO('[', '.'))
1000 		return(GETNEXT());
1001 
1002 	/* collating symbol */
1003 	value = p_b_coll_elem(p, '.');
1004 	REQUIRE(EATTWO('.', ']'), REG_ECOLLATE);
1005 	return(value);
1006 }
1007 
1008 /*
1009  - p_b_coll_elem - parse a collating-element name and look it up
1010  */
1011 static char			/* value of collating element */
1012 p_b_coll_elem(struct parse *p,
1013     int endc)			/* name ended by endc,']' */
1014 {
1015 	char *sp = p->next;
1016 	struct cname *cp;
1017 	size_t len;
1018 
1019 	while (MORE() && !SEETWO(endc, ']'))
1020 		NEXT();
1021 	if (!MORE()) {
1022 		SETERROR(REG_EBRACK);
1023 		return(0);
1024 	}
1025 	len = p->next - sp;
1026 	for (cp = cnames; cp->name != NULL; cp++)
1027 		if (strncmp(cp->name, sp, len) == 0 && strlen(cp->name) == len)
1028 			return(cp->code);	/* known name */
1029 	if (len == 1)
1030 		return(*sp);	/* single character */
1031 	SETERROR(REG_ECOLLATE);			/* neither */
1032 	return(0);
1033 }
1034 
1035 /*
1036  - othercase - return the case counterpart of an alphabetic
1037  */
1038 static char			/* if no counterpart, return ch */
1039 othercase(int ch)
1040 {
1041 	ch = (uch)ch;
1042 	assert(isalpha(ch));
1043 	if (isupper(ch))
1044 		return ((uch)tolower(ch));
1045 	else if (islower(ch))
1046 		return ((uch)toupper(ch));
1047 	else			/* peculiar, but could happen */
1048 		return(ch);
1049 }
1050 
1051 /*
1052  - bothcases - emit a dualcase version of a two-case character
1053  *
1054  * Boy, is this implementation ever a kludge...
1055  */
1056 static void
1057 bothcases(struct parse *p, int ch)
1058 {
1059 	char *oldnext = p->next;
1060 	char *oldend = p->end;
1061 	char bracket[3];
1062 
1063 	ch = (uch)ch;
1064 	assert(othercase(ch) != ch);	/* p_bracket() would recurse */
1065 	p->next = bracket;
1066 	p->end = bracket+2;
1067 	bracket[0] = ch;
1068 	bracket[1] = ']';
1069 	bracket[2] = '\0';
1070 	p_bracket(p);
1071 	assert(p->next == bracket+2);
1072 	p->next = oldnext;
1073 	p->end = oldend;
1074 }
1075 
1076 /*
1077  - ordinary - emit an ordinary character
1078  */
1079 static void
1080 ordinary(struct parse *p, int ch)
1081 {
1082 	cat_t *cap = p->g->categories;
1083 
1084 	if ((p->g->cflags&REG_ICASE) && isalpha((uch)ch) && othercase(ch) != ch)
1085 		bothcases(p, ch);
1086 	else {
1087 		EMIT(OCHAR, (uch)ch);
1088 		if (cap[ch] == 0)
1089 			cap[ch] = p->g->ncategories++;
1090 	}
1091 }
1092 
1093 /*
1094  - nonnewline - emit REG_NEWLINE version of OANY
1095  *
1096  * Boy, is this implementation ever a kludge...
1097  */
1098 static void
1099 nonnewline(struct parse *p)
1100 {
1101 	char *oldnext = p->next;
1102 	char *oldend = p->end;
1103 	char bracket[4];
1104 
1105 	p->next = bracket;
1106 	p->end = bracket+3;
1107 	bracket[0] = '^';
1108 	bracket[1] = '\n';
1109 	bracket[2] = ']';
1110 	bracket[3] = '\0';
1111 	p_bracket(p);
1112 	assert(p->next == bracket+3);
1113 	p->next = oldnext;
1114 	p->end = oldend;
1115 }
1116 
1117 /*
1118  - repeat - generate code for a bounded repetition, recursively if needed
1119  */
1120 static void
1121 repeat(struct parse *p,
1122     sopno start,		/* operand from here to end of strip */
1123     int from,			/* repeated from this number */
1124     int to)			/* to this number of times (maybe INFINITY) */
1125 {
1126 	sopno finish = HERE();
1127 #	define	N	2
1128 #	define	INF	3
1129 #	define	REP(f, t)	((f)*8 + (t))
1130 #	define	MAP(n)	(((n) <= 1) ? (n) : ((n) == INFINITY) ? INF : N)
1131 	sopno copy;
1132 
1133 	if (p->error != 0)	/* head off possible runaway recursion */
1134 		return;
1135 
1136 	assert(from <= to);
1137 
1138 	switch (REP(MAP(from), MAP(to))) {
1139 	case REP(0, 0):			/* must be user doing this */
1140 		DROP(finish-start);	/* drop the operand */
1141 		break;
1142 	case REP(0, 1):			/* as x{1,1}? */
1143 	case REP(0, N):			/* as x{1,n}? */
1144 	case REP(0, INF):		/* as x{1,}? */
1145 		/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
1146 		INSERT(OCH_, start);		/* offset is wrong... */
1147 		repeat(p, start+1, 1, to);
1148 		ASTERN(OOR1, start);
1149 		AHEAD(start);			/* ... fix it */
1150 		EMIT(OOR2, 0);
1151 		AHEAD(THERE());
1152 		ASTERN(O_CH, THERETHERE());
1153 		break;
1154 	case REP(1, 1):			/* trivial case */
1155 		/* done */
1156 		break;
1157 	case REP(1, N):			/* as x?x{1,n-1} */
1158 		/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
1159 		INSERT(OCH_, start);
1160 		ASTERN(OOR1, start);
1161 		AHEAD(start);
1162 		EMIT(OOR2, 0);			/* offset very wrong... */
1163 		AHEAD(THERE());			/* ...so fix it */
1164 		ASTERN(O_CH, THERETHERE());
1165 		copy = dupl(p, start+1, finish+1);
1166 		assert(copy == finish+4);
1167 		repeat(p, copy, 1, to-1);
1168 		break;
1169 	case REP(1, INF):		/* as x+ */
1170 		INSERT(OPLUS_, start);
1171 		ASTERN(O_PLUS, start);
1172 		break;
1173 	case REP(N, N):			/* as xx{m-1,n-1} */
1174 		copy = dupl(p, start, finish);
1175 		repeat(p, copy, from-1, to-1);
1176 		break;
1177 	case REP(N, INF):		/* as xx{n-1,INF} */
1178 		copy = dupl(p, start, finish);
1179 		repeat(p, copy, from-1, to);
1180 		break;
1181 	default:			/* "can't happen" */
1182 		SETERROR(REG_ASSERT);	/* just in case */
1183 		break;
1184 	}
1185 }
1186 
1187 /*
1188  - seterr - set an error condition
1189  */
1190 static int			/* useless but makes type checking happy */
1191 seterr(struct parse *p, int e)
1192 {
1193 	if (p->error == 0)	/* keep earliest error condition */
1194 		p->error = e;
1195 	p->next = nuls;		/* try to bring things to a halt */
1196 	p->end = nuls;
1197 	return(0);		/* make the return value well-defined */
1198 }
1199 
1200 /*
1201  - allocset - allocate a set of characters for []
1202  */
1203 static cset *
1204 allocset(struct parse *p)
1205 {
1206 	int no = p->g->ncsets++;
1207 	size_t nc;
1208 	size_t nbytes;
1209 	cset *cs;
1210 	size_t css = (size_t)p->g->csetsize;
1211 	int i;
1212 
1213 	if (no >= p->ncsalloc) {	/* need another column of space */
1214 		void *ptr;
1215 
1216 		p->ncsalloc += CHAR_BIT;
1217 		nc = p->ncsalloc;
1218 		if (nc > SIZE_MAX / sizeof(cset))
1219 			goto nomem;
1220 		assert(nc % CHAR_BIT == 0);
1221 		nbytes = nc / CHAR_BIT * css;
1222 
1223 		ptr = (cset *)realloc((char *)p->g->sets, nc * sizeof(cset));
1224 		if (ptr == NULL)
1225 			goto nomem;
1226 		p->g->sets = ptr;
1227 
1228 		ptr = (uch *)realloc((char *)p->g->setbits, nbytes);
1229 		if (ptr == NULL)
1230 			goto nomem;
1231 		p->g->setbits = ptr;
1232 
1233 		for (i = 0; i < no; i++)
1234 			p->g->sets[i].ptr = p->g->setbits + css*(i/CHAR_BIT);
1235 
1236 		(void) memset((char *)p->g->setbits + (nbytes - css), 0, css);
1237 	}
1238 	/* XXX should not happen */
1239 	if (p->g->sets == NULL || p->g->setbits == NULL)
1240 		goto nomem;
1241 
1242 	cs = &p->g->sets[no];
1243 	cs->ptr = p->g->setbits + css*((no)/CHAR_BIT);
1244 	cs->mask = 1 << ((no) % CHAR_BIT);
1245 	cs->hash = 0;
1246 	cs->smultis = 0;
1247 	cs->multis = NULL;
1248 
1249 	return(cs);
1250 nomem:
1251 	free(p->g->sets);
1252 	p->g->sets = NULL;
1253 	free(p->g->setbits);
1254 	p->g->setbits = NULL;
1255 
1256 	SETERROR(REG_ESPACE);
1257 	/* caller's responsibility not to do set ops */
1258 	return(NULL);
1259 }
1260 
1261 /*
1262  - freeset - free a now-unused set
1263  */
1264 static void
1265 freeset(struct parse *p, cset *cs)
1266 {
1267 	size_t i;
1268 	cset *top = &p->g->sets[p->g->ncsets];
1269 	size_t css = (size_t)p->g->csetsize;
1270 
1271 	for (i = 0; i < css; i++)
1272 		CHsub(cs, i);
1273 	if (cs == top-1)	/* recover only the easy case */
1274 		p->g->ncsets--;
1275 }
1276 
1277 /*
1278  - freezeset - final processing on a set of characters
1279  *
1280  * The main task here is merging identical sets.  This is usually a waste
1281  * of time (although the hash code minimizes the overhead), but can win
1282  * big if REG_ICASE is being used.  REG_ICASE, by the way, is why the hash
1283  * is done using addition rather than xor -- all ASCII [aA] sets xor to
1284  * the same value!
1285  */
1286 static int			/* set number */
1287 freezeset(struct parse *p, cset *cs)
1288 {
1289 	uch h = cs->hash;
1290 	size_t i;
1291 	cset *top = &p->g->sets[p->g->ncsets];
1292 	cset *cs2;
1293 	size_t css = (size_t)p->g->csetsize;
1294 
1295 	/* look for an earlier one which is the same */
1296 	for (cs2 = &p->g->sets[0]; cs2 < top; cs2++)
1297 		if (cs2->hash == h && cs2 != cs) {
1298 			/* maybe */
1299 			for (i = 0; i < css; i++)
1300 				if (!!CHIN(cs2, i) != !!CHIN(cs, i))
1301 					break;		/* no */
1302 			if (i == css)
1303 				break;			/* yes */
1304 		}
1305 
1306 	if (cs2 < top) {	/* found one */
1307 		freeset(p, cs);
1308 		cs = cs2;
1309 	}
1310 
1311 	return((int)(cs - p->g->sets));
1312 }
1313 
1314 /*
1315  - firstch - return first character in a set (which must have at least one)
1316  */
1317 static int			/* character; there is no "none" value */
1318 firstch(struct parse *p, cset *cs)
1319 {
1320 	size_t i;
1321 	size_t css = (size_t)p->g->csetsize;
1322 
1323 	for (i = 0; i < css; i++)
1324 		if (CHIN(cs, i))
1325 			return((char)i);
1326 	assert(never);
1327 	return(0);		/* arbitrary */
1328 }
1329 
1330 /*
1331  - nch - number of characters in a set
1332  */
1333 static int
1334 nch(struct parse *p, cset *cs)
1335 {
1336 	size_t i;
1337 	size_t css = (size_t)p->g->csetsize;
1338 	int n = 0;
1339 
1340 	for (i = 0; i < css; i++)
1341 		if (CHIN(cs, i))
1342 			n++;
1343 	return(n);
1344 }
1345 
1346 /*
1347  - mcadd - add a collating element to a cset
1348  */
1349 static void
1350 mcadd( struct parse *p, cset *cs, const char *cp)
1351 {
1352 	size_t oldend = cs->smultis;
1353 	void *np;
1354 
1355 	cs->smultis += strlen(cp) + 1;
1356 	np = realloc(cs->multis, cs->smultis);
1357 	if (np == NULL) {
1358 		if (cs->multis)
1359 			free(cs->multis);
1360 		cs->multis = NULL;
1361 		SETERROR(REG_ESPACE);
1362 		return;
1363 	}
1364 	cs->multis = np;
1365 
1366 	llvm_strlcpy(cs->multis + oldend - 1, cp, cs->smultis - oldend + 1);
1367 }
1368 
1369 /*
1370  - mcinvert - invert the list of collating elements in a cset
1371  *
1372  * This would have to know the set of possibilities.  Implementation
1373  * is deferred.
1374  */
1375 /* ARGSUSED */
1376 static void
1377 mcinvert(struct parse *p, cset *cs)
1378 {
1379 	assert(cs->multis == NULL);	/* xxx */
1380 }
1381 
1382 /*
1383  - mccase - add case counterparts of the list of collating elements in a cset
1384  *
1385  * This would have to know the set of possibilities.  Implementation
1386  * is deferred.
1387  */
1388 /* ARGSUSED */
1389 static void
1390 mccase(struct parse *p, cset *cs)
1391 {
1392 	assert(cs->multis == NULL);	/* xxx */
1393 }
1394 
1395 /*
1396  - isinsets - is this character in any sets?
1397  */
1398 static int			/* predicate */
1399 isinsets(struct re_guts *g, int c)
1400 {
1401 	uch *col;
1402 	int i;
1403 	int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1404 	unsigned uc = (uch)c;
1405 
1406 	for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1407 		if (col[uc] != 0)
1408 			return(1);
1409 	return(0);
1410 }
1411 
1412 /*
1413  - samesets - are these two characters in exactly the same sets?
1414  */
1415 static int			/* predicate */
1416 samesets(struct re_guts *g, int c1, int c2)
1417 {
1418 	uch *col;
1419 	int i;
1420 	int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1421 	unsigned uc1 = (uch)c1;
1422 	unsigned uc2 = (uch)c2;
1423 
1424 	for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1425 		if (col[uc1] != col[uc2])
1426 			return(0);
1427 	return(1);
1428 }
1429 
1430 /*
1431  - categorize - sort out character categories
1432  */
1433 static void
1434 categorize(struct parse *p, struct re_guts *g)
1435 {
1436 	cat_t *cats = g->categories;
1437 	int c;
1438 	int c2;
1439 	cat_t cat;
1440 
1441 	/* avoid making error situations worse */
1442 	if (p->error != 0)
1443 		return;
1444 
1445 	for (c = CHAR_MIN; c <= CHAR_MAX; c++)
1446 		if (cats[c] == 0 && isinsets(g, c)) {
1447 			cat = g->ncategories++;
1448 			cats[c] = cat;
1449 			for (c2 = c+1; c2 <= CHAR_MAX; c2++)
1450 				if (cats[c2] == 0 && samesets(g, c, c2))
1451 					cats[c2] = cat;
1452 		}
1453 }
1454 
1455 /*
1456  - dupl - emit a duplicate of a bunch of sops
1457  */
1458 static sopno			/* start of duplicate */
1459 dupl(struct parse *p,
1460     sopno start,		/* from here */
1461     sopno finish)		/* to this less one */
1462 {
1463 	sopno ret = HERE();
1464 	sopno len = finish - start;
1465 
1466 	assert(finish >= start);
1467 	if (len == 0)
1468 		return(ret);
1469 	enlarge(p, p->ssize + len);	/* this many unexpected additions */
1470 	assert(p->ssize >= p->slen + len);
1471 	(void) memmove((char *)(p->strip + p->slen),
1472 		(char *)(p->strip + start), (size_t)len*sizeof(sop));
1473 	p->slen += len;
1474 	return(ret);
1475 }
1476 
1477 /*
1478  - doemit - emit a strip operator
1479  *
1480  * It might seem better to implement this as a macro with a function as
1481  * hard-case backup, but it's just too big and messy unless there are
1482  * some changes to the data structures.  Maybe later.
1483  */
1484 static void
1485 doemit(struct parse *p, sop op, size_t opnd)
1486 {
1487 	/* avoid making error situations worse */
1488 	if (p->error != 0)
1489 		return;
1490 
1491 	/* deal with oversize operands ("can't happen", more or less) */
1492 	assert(opnd < 1<<OPSHIFT);
1493 
1494 	/* deal with undersized strip */
1495 	if (p->slen >= p->ssize)
1496 		enlarge(p, (p->ssize+1) / 2 * 3);	/* +50% */
1497 	assert(p->slen < p->ssize);
1498 
1499 	/* finally, it's all reduced to the easy case */
1500 	p->strip[p->slen++] = SOP(op, opnd);
1501 }
1502 
1503 /*
1504  - doinsert - insert a sop into the strip
1505  */
1506 static void
1507 doinsert(struct parse *p, sop op, size_t opnd, sopno pos)
1508 {
1509 	sopno sn;
1510 	sop s;
1511 	int i;
1512 
1513 	/* avoid making error situations worse */
1514 	if (p->error != 0)
1515 		return;
1516 
1517 	sn = HERE();
1518 	EMIT(op, opnd);		/* do checks, ensure space */
1519 	assert(HERE() == sn+1);
1520 	s = p->strip[sn];
1521 
1522 	/* adjust paren pointers */
1523 	assert(pos > 0);
1524 	for (i = 1; i < NPAREN; i++) {
1525 		if (p->pbegin[i] >= pos) {
1526 			p->pbegin[i]++;
1527 		}
1528 		if (p->pend[i] >= pos) {
1529 			p->pend[i]++;
1530 		}
1531 	}
1532 
1533 	memmove((char *)&p->strip[pos+1], (char *)&p->strip[pos],
1534 						(HERE()-pos-1)*sizeof(sop));
1535 	p->strip[pos] = s;
1536 }
1537 
1538 /*
1539  - dofwd - complete a forward reference
1540  */
1541 static void
1542 dofwd(struct parse *p, sopno pos, sop value)
1543 {
1544 	/* avoid making error situations worse */
1545 	if (p->error != 0)
1546 		return;
1547 
1548 	assert(value < 1<<OPSHIFT);
1549 	p->strip[pos] = OP(p->strip[pos]) | value;
1550 }
1551 
1552 /*
1553  - enlarge - enlarge the strip
1554  */
1555 static void
1556 enlarge(struct parse *p, sopno size)
1557 {
1558 	sop *sp;
1559 
1560 	if (p->ssize >= size)
1561 		return;
1562 
1563 	if ((uintptr_t)size > SIZE_MAX / sizeof(sop)) {
1564 		SETERROR(REG_ESPACE);
1565 		return;
1566 	}
1567 
1568 	sp = (sop *)realloc(p->strip, size*sizeof(sop));
1569 	if (sp == NULL) {
1570 		SETERROR(REG_ESPACE);
1571 		return;
1572 	}
1573 	p->strip = sp;
1574 	p->ssize = size;
1575 }
1576 
1577 /*
1578  - stripsnug - compact the strip
1579  */
1580 static void
1581 stripsnug(struct parse *p, struct re_guts *g)
1582 {
1583 	g->nstates = p->slen;
1584 	if ((uintptr_t)p->slen > SIZE_MAX / sizeof(sop)) {
1585 		g->strip = p->strip;
1586 		SETERROR(REG_ESPACE);
1587 		return;
1588 	}
1589 
1590 	g->strip = (sop *)realloc((char *)p->strip, p->slen * sizeof(sop));
1591 	if (g->strip == NULL) {
1592 		SETERROR(REG_ESPACE);
1593 		g->strip = p->strip;
1594 	}
1595 }
1596 
1597 /*
1598  - findmust - fill in must and mlen with longest mandatory literal string
1599  *
1600  * This algorithm could do fancy things like analyzing the operands of |
1601  * for common subsequences.  Someday.  This code is simple and finds most
1602  * of the interesting cases.
1603  *
1604  * Note that must and mlen got initialized during setup.
1605  */
1606 static void
1607 findmust(struct parse *p, struct re_guts *g)
1608 {
1609 	sop *scan;
1610 	sop *start = 0; /* start initialized in the default case, after that */
1611 	sop *newstart = 0; /* newstart was initialized in the OCHAR case */
1612 	sopno newlen;
1613 	sop s;
1614 	char *cp;
1615 	sopno i;
1616 
1617 	/* avoid making error situations worse */
1618 	if (p->error != 0)
1619 		return;
1620 
1621 	/* find the longest OCHAR sequence in strip */
1622 	newlen = 0;
1623 	scan = g->strip + 1;
1624 	do {
1625 		s = *scan++;
1626 		switch (OP(s)) {
1627 		case OCHAR:		/* sequence member */
1628 			if (newlen == 0)		/* new sequence */
1629 				newstart = scan - 1;
1630 			newlen++;
1631 			break;
1632 		case OPLUS_:		/* things that don't break one */
1633 		case OLPAREN:
1634 		case ORPAREN:
1635 			break;
1636 		case OQUEST_:		/* things that must be skipped */
1637 		case OCH_:
1638 			scan--;
1639 			do {
1640 				scan += OPND(s);
1641 				s = *scan;
1642 				/* assert() interferes w debug printouts */
1643 				if (OP(s) != O_QUEST && OP(s) != O_CH &&
1644 							OP(s) != OOR2) {
1645 					g->iflags |= REGEX_BAD;
1646 					return;
1647 				}
1648 			} while (OP(s) != O_QUEST && OP(s) != O_CH);
1649 			LLVM_FALLTHROUGH;
1650 		default:		/* things that break a sequence */
1651 			if (newlen > g->mlen) {		/* ends one */
1652 				start = newstart;
1653 				g->mlen = newlen;
1654 			}
1655 			newlen = 0;
1656 			break;
1657 		}
1658 	} while (OP(s) != OEND);
1659 
1660 	if (g->mlen == 0)		/* there isn't one */
1661 		return;
1662 
1663 	/* turn it into a character string */
1664 	g->must = malloc((size_t)g->mlen + 1);
1665 	if (g->must == NULL) {		/* argh; just forget it */
1666 		g->mlen = 0;
1667 		return;
1668 	}
1669 	cp = g->must;
1670 	scan = start;
1671 	for (i = g->mlen; i > 0; i--) {
1672 		while (OP(s = *scan++) != OCHAR)
1673 			continue;
1674 		assert(cp < g->must + g->mlen);
1675 		*cp++ = (char)OPND(s);
1676 	}
1677 	assert(cp == g->must + g->mlen);
1678 	*cp++ = '\0';		/* just on general principles */
1679 }
1680 
1681 /*
1682  - pluscount - count + nesting
1683  */
1684 static sopno			/* nesting depth */
1685 pluscount(struct parse *p, struct re_guts *g)
1686 {
1687 	sop *scan;
1688 	sop s;
1689 	sopno plusnest = 0;
1690 	sopno maxnest = 0;
1691 
1692 	if (p->error != 0)
1693 		return(0);	/* there may not be an OEND */
1694 
1695 	scan = g->strip + 1;
1696 	do {
1697 		s = *scan++;
1698 		switch (OP(s)) {
1699 		case OPLUS_:
1700 			plusnest++;
1701 			break;
1702 		case O_PLUS:
1703 			if (plusnest > maxnest)
1704 				maxnest = plusnest;
1705 			plusnest--;
1706 			break;
1707 		}
1708 	} while (OP(s) != OEND);
1709 	if (plusnest != 0)
1710 		g->iflags |= REGEX_BAD;
1711 	return(maxnest);
1712 }
1713