2 * Copyright (c) 1992, 1993, 1994 Henry Spencer.
3 * Copyright (c) 1992, 1993, 1994
4 * The Regents of the University of California. All rights reserved.
6 * This code is derived from software contributed to Berkeley by
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * @(#)engine.c 8.5 (Berkeley) 3/20/94
36 #include <sys/cdefs.h>
37 __FBSDID("$FreeBSD$");
40 * The matching engine and friends. This file is #included by regexec.c
41 * after suitable #defines of a variety of macros used herein, so that
42 * different state representations can be used without duplicating masses
47 #define matcher smatcher
50 #define dissect sdissect
51 #define backref sbackref
58 #define matcher lmatcher
61 #define dissect ldissect
62 #define backref lbackref
69 #define matcher mmatcher
72 #define dissect mdissect
73 #define backref mbackref
80 /* another structure passed up and down to avoid zillions of parameters */
84 regmatch_t *pmatch; /* [nsub+1] (0 element unused) */
85 const char *offp; /* offsets work from here */
86 const char *beginp; /* start of string -- virtual NUL precedes */
87 const char *endp; /* end of string -- virtual NUL here */
88 const char *coldp; /* can be no match starting before here */
89 const char **lastpos; /* [nplus+1] */
91 states st; /* current states */
92 states fresh; /* states for a fresh start */
93 states tmp; /* temporary */
94 states empty; /* empty set of states */
95 mbstate_t mbs; /* multibyte conversion state */
98 /* ========= begin header generated by ./mkh ========= */
103 /* === engine.c === */
104 static int matcher(struct re_guts *g, const char *string, size_t nmatch, regmatch_t pmatch[], int eflags);
105 static const char *dissect(struct match *m, const char *start, const char *stop, sopno startst, sopno stopst);
106 static const char *backref(struct match *m, const char *start, const char *stop, sopno startst, sopno stopst, sopno lev, int);
107 static const char *fast(struct match *m, const char *start, const char *stop, sopno startst, sopno stopst);
108 static const char *slow(struct match *m, const char *start, const char *stop, sopno startst, sopno stopst);
109 static states step(struct re_guts *g, sopno start, sopno stop, states bef, wint_t ch, states aft);
110 #define MAX_RECURSION 100
113 #define BOLEOL (BOL-2)
114 #define NOTHING (BOL-3)
117 #define BADCHAR (BOL-6)
118 #define NONCHAR(c) ((c) <= OUT)
120 static void print(struct match *m, const char *caption, states st, int ch, FILE *d);
123 static void at(struct match *m, const char *title, const char *start, const char *stop, sopno startst, sopno stopst);
126 static const char *pchar(int ch);
132 /* ========= end header generated by ./mkh ========= */
135 #define SP(t, s, c) print(m, t, s, c, stdout)
136 #define AT(t, p1, p2, s1, s2) at(m, t, p1, p2, s1, s2)
137 #define NOTE(str) { if (m->eflags®_TRACE) printf("=%s\n", (str)); }
139 #define SP(t, s, c) /* nothing */
140 #define AT(t, p1, p2, s1, s2) /* nothing */
141 #define NOTE(s) /* nothing */
145 - matcher - the actual matching engine
146 == static int matcher(struct re_guts *g, const char *string, \
147 == size_t nmatch, regmatch_t pmatch[], int eflags);
149 static int /* 0 success, REG_NOMATCH failure */
150 matcher(struct re_guts *g,
159 struct match *m = &mv;
161 const sopno gf = g->firststate+1; /* +1 for OEND */
162 const sopno gl = g->laststate;
165 /* Boyer-Moore algorithms variables */
168 const char *mustfirst;
169 const char *mustlast;
173 /* simplify the situation where possible */
174 if (g->cflags®_NOSUB)
176 if (eflags®_STARTEND) {
177 start = string + pmatch[0].rm_so;
178 stop = string + pmatch[0].rm_eo;
181 stop = start + strlen(start);
186 /* prescreening; this does wonders for this rather slow code */
187 if (g->must != NULL) {
188 if (g->charjump != NULL && g->matchjump != NULL) {
190 mustlast = g->must + g->mlen - 1;
191 charjump = g->charjump;
192 matchjump = g->matchjump;
194 for (dp = start+g->mlen-1; dp < stop;) {
195 /* Fast skip non-matches */
196 while (dp < stop && charjump[(int)*dp])
197 dp += charjump[(int)*dp];
203 /* We depend on not being used for
204 * for strings of length 1
206 while (*--dp == *--pp && pp != mustfirst);
211 /* Jump to next possible match */
212 mj = matchjump[pp - mustfirst];
213 cj = charjump[(int)*dp];
214 dp += (cj < mj ? mj : cj);
220 for (dp = start; dp < stop; dp++)
221 if (*dp == g->must[0] &&
222 stop - dp >= g->mlen &&
223 memcmp(dp, g->must, (size_t)g->mlen) == 0)
225 if (dp == stop) /* we didn't find g->must */
230 /* match struct setup */
246 /* Adjust start according to moffset, to speed things up */
248 start = ((dp - g->moffset) < start) ? start : dp - g->moffset;
250 /* this loop does only one repetition except for backrefs */
252 endp = fast(m, start, stop, gf, gl);
253 if (endp == NULL) { /* a miss */
254 if (m->pmatch != NULL)
255 free((char *)m->pmatch);
256 if (m->lastpos != NULL)
257 free((char *)m->lastpos);
261 if (nmatch == 0 && !g->backrefs)
262 break; /* no further info needed */
265 assert(m->coldp != NULL);
267 NOTE("finding start");
268 endp = slow(m, m->coldp, stop, gf, gl);
271 assert(m->coldp < m->endp);
272 m->coldp += XMBRTOWC(NULL, m->coldp,
273 m->endp - m->coldp, &m->mbs, 0);
275 if (nmatch == 1 && !g->backrefs)
276 break; /* no further info needed */
278 /* oh my, he wants the subexpressions... */
279 if (m->pmatch == NULL)
280 m->pmatch = (regmatch_t *)malloc((m->g->nsub + 1) *
282 if (m->pmatch == NULL) {
286 for (i = 1; i <= m->g->nsub; i++)
287 m->pmatch[i].rm_so = m->pmatch[i].rm_eo = -1;
288 if (!g->backrefs && !(m->eflags®_BACKR)) {
290 dp = dissect(m, m->coldp, endp, gf, gl);
292 if (g->nplus > 0 && m->lastpos == NULL)
293 m->lastpos = malloc((g->nplus+1) *
294 sizeof(const char *));
295 if (g->nplus > 0 && m->lastpos == NULL) {
300 NOTE("backref dissect");
301 dp = backref(m, m->coldp, endp, gf, gl, (sopno)0, 0);
306 /* uh-oh... we couldn't find a subexpression-level match */
307 assert(g->backrefs); /* must be back references doing it */
308 assert(g->nplus == 0 || m->lastpos != NULL);
310 if (dp != NULL || endp <= m->coldp)
313 endp = slow(m, m->coldp, endp-1, gf, gl);
316 /* try it on a shorter possibility */
318 for (i = 1; i <= m->g->nsub; i++) {
319 assert(m->pmatch[i].rm_so == -1);
320 assert(m->pmatch[i].rm_eo == -1);
323 NOTE("backoff dissect");
324 dp = backref(m, m->coldp, endp, gf, gl, (sopno)0, 0);
326 assert(dp == NULL || dp == endp);
327 if (dp != NULL) /* found a shorter one */
330 /* despite initial appearances, there is no match here */
332 /* recycle starting later */
333 start = m->coldp + XMBRTOWC(NULL, m->coldp,
334 stop - m->coldp, &m->mbs, 0);
335 assert(start <= stop);
338 /* fill in the details if requested */
340 pmatch[0].rm_so = m->coldp - m->offp;
341 pmatch[0].rm_eo = endp - m->offp;
344 assert(m->pmatch != NULL);
345 for (i = 1; i < nmatch; i++)
347 pmatch[i] = m->pmatch[i];
349 pmatch[i].rm_so = -1;
350 pmatch[i].rm_eo = -1;
354 if (m->pmatch != NULL)
355 free((char *)m->pmatch);
356 if (m->lastpos != NULL)
357 free((char *)m->lastpos);
363 - dissect - figure out what matched what, no back references
364 == static const char *dissect(struct match *m, const char *start, \
365 == const char *stop, sopno startst, sopno stopst);
367 static const char * /* == stop (success) always */
368 dissect(struct match *m,
375 sopno ss; /* start sop of current subRE */
376 sopno es; /* end sop of current subRE */
377 const char *sp; /* start of string matched by it */
378 const char *stp; /* string matched by it cannot pass here */
379 const char *rest; /* start of rest of string */
380 const char *tail; /* string unmatched by rest of RE */
381 sopno ssub; /* start sop of subsubRE */
382 sopno esub; /* end sop of subsubRE */
383 const char *ssp; /* start of string matched by subsubRE */
384 const char *sep; /* end of string matched by subsubRE */
385 const char *oldssp; /* previous ssp */
388 AT("diss", start, stop, startst, stopst);
390 for (ss = startst; ss < stopst; ss = es) {
391 /* identify end of subRE */
393 switch (OP(m->g->strip[es])) {
396 es += OPND(m->g->strip[es]);
399 while (OP(m->g->strip[es]) != O_CH)
400 es += OPND(m->g->strip[es]);
405 /* figure out what it matched */
406 switch (OP(m->g->strip[ss])) {
411 sp += XMBRTOWC(NULL, sp, stop - start, &m->mbs, 0);
420 sp += XMBRTOWC(NULL, sp, stop - start, &m->mbs, 0);
426 /* cases where length of match is hard to find */
430 /* how long could this one be? */
431 rest = slow(m, sp, stp, ss, es);
432 assert(rest != NULL); /* it did match */
433 /* could the rest match the rest? */
434 tail = slow(m, rest, stop, es, stopst);
437 /* no -- try a shorter match for this one */
439 assert(stp >= sp); /* it did work */
443 /* did innards match? */
444 if (slow(m, sp, rest, ssub, esub) != NULL) {
445 dp = dissect(m, sp, rest, ssub, esub);
454 /* how long could this one be? */
455 rest = slow(m, sp, stp, ss, es);
456 assert(rest != NULL); /* it did match */
457 /* could the rest match the rest? */
458 tail = slow(m, rest, stop, es, stopst);
461 /* no -- try a shorter match for this one */
463 assert(stp >= sp); /* it did work */
469 for (;;) { /* find last match of innards */
470 sep = slow(m, ssp, rest, ssub, esub);
471 if (sep == NULL || sep == ssp)
472 break; /* failed or matched null */
473 oldssp = ssp; /* on to next try */
477 /* last successful match */
481 assert(sep == rest); /* must exhaust substring */
482 assert(slow(m, ssp, sep, ssub, esub) == rest);
483 dp = dissect(m, ssp, sep, ssub, esub);
490 /* how long could this one be? */
491 rest = slow(m, sp, stp, ss, es);
492 assert(rest != NULL); /* it did match */
493 /* could the rest match the rest? */
494 tail = slow(m, rest, stop, es, stopst);
497 /* no -- try a shorter match for this one */
499 assert(stp >= sp); /* it did work */
502 esub = ss + OPND(m->g->strip[ss]) - 1;
503 assert(OP(m->g->strip[esub]) == OOR1);
504 for (;;) { /* find first matching branch */
505 if (slow(m, sp, rest, ssub, esub) == rest)
506 break; /* it matched all of it */
507 /* that one missed, try next one */
508 assert(OP(m->g->strip[esub]) == OOR1);
510 assert(OP(m->g->strip[esub]) == OOR2);
512 esub += OPND(m->g->strip[esub]);
513 if (OP(m->g->strip[esub]) == OOR2)
516 assert(OP(m->g->strip[esub]) == O_CH);
518 dp = dissect(m, sp, rest, ssub, esub);
530 i = OPND(m->g->strip[ss]);
531 assert(0 < i && i <= m->g->nsub);
532 m->pmatch[i].rm_so = sp - m->offp;
535 i = OPND(m->g->strip[ss]);
536 assert(0 < i && i <= m->g->nsub);
537 m->pmatch[i].rm_eo = sp - m->offp;
550 - backref - figure out what matched what, figuring in back references
551 == static const char *backref(struct match *m, const char *start, \
552 == const char *stop, sopno startst, sopno stopst, sopno lev);
554 static const char * /* == stop (success) or NULL (failure) */
555 backref(struct match *m,
560 sopno lev, /* PLUS nesting level */
564 sopno ss; /* start sop of current subRE */
565 const char *sp; /* start of string matched by it */
566 sopno ssub; /* start sop of subsubRE */
567 sopno esub; /* end sop of subsubRE */
568 const char *ssp; /* start of string matched by subsubRE */
577 AT("back", start, stop, startst, stopst);
580 /* get as far as we can with easy stuff */
582 for (ss = startst; !hard && ss < stopst; ss++)
583 switch (OP(s = m->g->strip[ss])) {
587 sp += XMBRTOWC(&wc, sp, stop - sp, &m->mbs, BADCHAR);
594 sp += XMBRTOWC(&wc, sp, stop - sp, &m->mbs, BADCHAR);
601 cs = &m->g->sets[OPND(s)];
602 sp += XMBRTOWC(&wc, sp, stop - sp, &m->mbs, BADCHAR);
603 if (wc == BADCHAR || !CHIN(cs, wc))
607 if ( (sp == m->beginp && !(m->eflags®_NOTBOL)) ||
608 (sp < m->endp && *(sp-1) == '\n' &&
609 (m->g->cflags®_NEWLINE)) )
615 if ( (sp == m->endp && !(m->eflags®_NOTEOL)) ||
616 (sp < m->endp && *sp == '\n' &&
617 (m->g->cflags®_NEWLINE)) )
623 if (( (sp == m->beginp && !(m->eflags®_NOTBOL)) ||
624 (sp < m->endp && *(sp-1) == '\n' &&
625 (m->g->cflags®_NEWLINE)) ||
627 !ISWORD(*(sp-1))) ) &&
628 (sp < m->endp && ISWORD(*sp)) )
634 if (( (sp == m->endp && !(m->eflags®_NOTEOL)) ||
635 (sp < m->endp && *sp == '\n' &&
636 (m->g->cflags®_NEWLINE)) ||
637 (sp < m->endp && !ISWORD(*sp)) ) &&
638 (sp > m->beginp && ISWORD(*(sp-1))) )
645 case OOR1: /* matches null but needs to skip */
649 assert(OP(s) == OOR2);
651 } while (OP(s = m->g->strip[ss]) != O_CH);
652 /* note that the ss++ gets us past the O_CH */
654 default: /* have to make a choice */
658 if (!hard) { /* that was it! */
663 ss--; /* adjust for the for's final increment */
666 AT("hard", sp, stop, ss, stopst);
669 case OBACK_: /* the vilest depths */
671 assert(0 < i && i <= m->g->nsub);
672 if (m->pmatch[i].rm_eo == -1)
674 assert(m->pmatch[i].rm_so != -1);
675 len = m->pmatch[i].rm_eo - m->pmatch[i].rm_so;
676 if (len == 0 && rec++ > MAX_RECURSION)
678 assert(stop - m->beginp >= len);
680 return(NULL); /* not enough left to match */
681 ssp = m->offp + m->pmatch[i].rm_so;
682 if (memcmp(sp, ssp, len) != 0)
684 while (m->g->strip[ss] != SOP(O_BACK, i))
686 return(backref(m, sp+len, stop, ss+1, stopst, lev, rec));
688 case OQUEST_: /* to null or not */
689 dp = backref(m, sp, stop, ss+1, stopst, lev, rec);
691 return(dp); /* not */
692 return(backref(m, sp, stop, ss+OPND(s)+1, stopst, lev, rec));
695 assert(m->lastpos != NULL);
696 assert(lev+1 <= m->g->nplus);
697 m->lastpos[lev+1] = sp;
698 return(backref(m, sp, stop, ss+1, stopst, lev+1, rec));
701 if (sp == m->lastpos[lev]) /* last pass matched null */
702 return(backref(m, sp, stop, ss+1, stopst, lev-1, rec));
703 /* try another pass */
704 m->lastpos[lev] = sp;
705 dp = backref(m, sp, stop, ss-OPND(s)+1, stopst, lev, rec);
707 return(backref(m, sp, stop, ss+1, stopst, lev-1, rec));
711 case OCH_: /* find the right one, if any */
713 esub = ss + OPND(s) - 1;
714 assert(OP(m->g->strip[esub]) == OOR1);
715 for (;;) { /* find first matching branch */
716 dp = backref(m, sp, stop, ssub, esub, lev, rec);
719 /* that one missed, try next one */
720 if (OP(m->g->strip[esub]) == O_CH)
721 return(NULL); /* there is none */
723 assert(OP(m->g->strip[esub]) == OOR2);
725 esub += OPND(m->g->strip[esub]);
726 if (OP(m->g->strip[esub]) == OOR2)
729 assert(OP(m->g->strip[esub]) == O_CH);
732 case OLPAREN: /* must undo assignment if rest fails */
734 assert(0 < i && i <= m->g->nsub);
735 offsave = m->pmatch[i].rm_so;
736 m->pmatch[i].rm_so = sp - m->offp;
737 dp = backref(m, sp, stop, ss+1, stopst, lev, rec);
740 m->pmatch[i].rm_so = offsave;
743 case ORPAREN: /* must undo assignment if rest fails */
745 assert(0 < i && i <= m->g->nsub);
746 offsave = m->pmatch[i].rm_eo;
747 m->pmatch[i].rm_eo = sp - m->offp;
748 dp = backref(m, sp, stop, ss+1, stopst, lev, rec);
751 m->pmatch[i].rm_eo = offsave;
762 return "shut up gcc";
766 - fast - step through the string at top speed
767 == static const char *fast(struct match *m, const char *start, \
768 == const char *stop, sopno startst, sopno stopst);
770 static const char * /* where tentative match ended, or NULL */
771 fast( struct match *m,
778 states fresh = m->fresh;
780 const char *p = start;
782 wint_t lastc; /* previous c */
785 const char *coldp; /* last p after which no match was underway */
790 st = step(m->g, startst, stopst, st, NOTHING, st);
794 if (start == m->beginp)
798 * XXX Wrong if the previous character was multi-byte.
799 * Newline never is (in encodings supported by FreeBSD),
800 * so this only breaks the ISWORD tests below.
802 c = (uch)*(start - 1);
811 clen = XMBRTOWC(&c, p, m->endp - p, &m->mbs, BADCHAR);
815 /* is there an EOL and/or BOL between lastc and c? */
818 if ( (lastc == '\n' && m->g->cflags®_NEWLINE) ||
819 (lastc == OUT && !(m->eflags®_NOTBOL)) ) {
823 if ( (c == '\n' && m->g->cflags®_NEWLINE) ||
824 (c == OUT && !(m->eflags®_NOTEOL)) ) {
825 flagch = (flagch == BOL) ? BOLEOL : EOL;
830 st = step(m->g, startst, stopst, st, flagch, st);
834 /* how about a word boundary? */
835 if ( (flagch == BOL || (lastc != OUT && !ISWORD(lastc))) &&
836 (c != OUT && ISWORD(c)) ) {
839 if ( (lastc != OUT && ISWORD(lastc)) &&
840 (flagch == EOL || (c != OUT && !ISWORD(c))) ) {
843 if (flagch == BOW || flagch == EOW) {
844 st = step(m->g, startst, stopst, st, flagch, st);
849 if (ISSET(st, stopst) || p == stop || clen > stop - p)
850 break; /* NOTE BREAK OUT */
852 /* no, we must deal with this character */
856 st = step(m->g, startst, stopst, tmp, c, st);
858 assert(EQ(step(m->g, startst, stopst, st, NOTHING, st), st));
862 assert(coldp != NULL);
864 if (ISSET(st, stopst))
865 return(p+XMBRTOWC(NULL, p, stop - p, &m->mbs, 0));
871 - slow - step through the string more deliberately
872 == static const char *slow(struct match *m, const char *start, \
873 == const char *stop, sopno startst, sopno stopst);
875 static const char * /* where it ended */
876 slow( struct match *m,
883 states empty = m->empty;
885 const char *p = start;
887 wint_t lastc; /* previous c */
890 const char *matchp; /* last p at which a match ended */
893 AT("slow", start, stop, startst, stopst);
896 SP("sstart", st, *p);
897 st = step(m->g, startst, stopst, st, NOTHING, st);
899 if (start == m->beginp)
903 * XXX Wrong if the previous character was multi-byte.
904 * Newline never is (in encodings supported by FreeBSD),
905 * so this only breaks the ISWORD tests below.
907 c = (uch)*(start - 1);
916 clen = XMBRTOWC(&c, p, m->endp - p, &m->mbs, BADCHAR);
918 /* is there an EOL and/or BOL between lastc and c? */
921 if ( (lastc == '\n' && m->g->cflags®_NEWLINE) ||
922 (lastc == OUT && !(m->eflags®_NOTBOL)) ) {
926 if ( (c == '\n' && m->g->cflags®_NEWLINE) ||
927 (c == OUT && !(m->eflags®_NOTEOL)) ) {
928 flagch = (flagch == BOL) ? BOLEOL : EOL;
933 st = step(m->g, startst, stopst, st, flagch, st);
934 SP("sboleol", st, c);
937 /* how about a word boundary? */
938 if ( (flagch == BOL || (lastc != OUT && !ISWORD(lastc))) &&
939 (c != OUT && ISWORD(c)) ) {
942 if ( (lastc != OUT && ISWORD(lastc)) &&
943 (flagch == EOL || (c != OUT && !ISWORD(c))) ) {
946 if (flagch == BOW || flagch == EOW) {
947 st = step(m->g, startst, stopst, st, flagch, st);
948 SP("sboweow", st, c);
952 if (ISSET(st, stopst))
954 if (EQ(st, empty) || p == stop || clen > stop - p)
955 break; /* NOTE BREAK OUT */
957 /* no, we must deal with this character */
961 st = step(m->g, startst, stopst, tmp, c, st);
963 assert(EQ(step(m->g, startst, stopst, st, NOTHING, st), st));
972 - step - map set of states reachable before char to set reachable after
973 == static states step(struct re_guts *g, sopno start, sopno stop, \
974 == states bef, int ch, states aft);
975 == #define BOL (OUT-1)
976 == #define EOL (BOL-1)
977 == #define BOLEOL (BOL-2)
978 == #define NOTHING (BOL-3)
979 == #define BOW (BOL-4)
980 == #define EOW (BOL-5)
981 == #define BADCHAR (BOL-6)
982 == #define NONCHAR(c) ((c) <= OUT)
985 step(struct re_guts *g,
986 sopno start, /* start state within strip */
987 sopno stop, /* state after stop state within strip */
988 states bef, /* states reachable before */
989 wint_t ch, /* character or NONCHAR code */
990 states aft) /* states already known reachable after */
995 onestate here; /* note, macros know this name */
999 for (pc = start, INIT(here, pc); pc != stop; pc++, INC(here)) {
1003 assert(pc == stop-1);
1006 /* only characters can match */
1007 assert(!NONCHAR(ch) || ch != OPND(s));
1012 if (ch == BOL || ch == BOLEOL)
1016 if (ch == EOL || ch == BOLEOL)
1032 cs = &g->sets[OPND(s)];
1033 if (!NONCHAR(ch) && CHIN(cs, ch))
1036 case OBACK_: /* ignored here */
1040 case OPLUS_: /* forward, this is just an empty */
1043 case O_PLUS: /* both forward and back */
1045 i = ISSETBACK(aft, OPND(s));
1046 BACK(aft, aft, OPND(s));
1047 if (!i && ISSETBACK(aft, OPND(s))) {
1048 /* oho, must reconsider loop body */
1053 case OQUEST_: /* two branches, both forward */
1055 FWD(aft, aft, OPND(s));
1057 case O_QUEST: /* just an empty */
1060 case OLPAREN: /* not significant here */
1064 case OCH_: /* mark the first two branches */
1066 assert(OP(g->strip[pc+OPND(s)]) == OOR2);
1067 FWD(aft, aft, OPND(s));
1069 case OOR1: /* done a branch, find the O_CH */
1070 if (ISSTATEIN(aft, here)) {
1072 OP(s = g->strip[pc+look]) != O_CH;
1074 assert(OP(s) == OOR2);
1075 FWD(aft, aft, look);
1078 case OOR2: /* propagate OCH_'s marking */
1080 if (OP(g->strip[pc+OPND(s)]) != O_CH) {
1081 assert(OP(g->strip[pc+OPND(s)]) == OOR2);
1082 FWD(aft, aft, OPND(s));
1085 case O_CH: /* just empty */
1088 default: /* ooooops... */
1099 - print - print a set of states
1101 == static void print(struct match *m, const char *caption, states st, \
1102 == int ch, FILE *d);
1106 print(struct match *m,
1107 const char *caption,
1112 struct re_guts *g = m->g;
1116 if (!(m->eflags®_TRACE))
1119 fprintf(d, "%s", caption);
1121 fprintf(d, " %s", pchar(ch));
1122 for (i = 0; i < g->nstates; i++)
1124 fprintf(d, "%s%d", (first) ? "\t" : ", ", i);
1131 - at - print current situation
1133 == static void at(struct match *m, const char *title, const char *start, \
1134 == const char *stop, sopno startst, sopno stopst);
1138 at( struct match *m,
1145 if (!(m->eflags®_TRACE))
1148 printf("%s %s-", title, pchar(*start));
1149 printf("%s ", pchar(*stop));
1150 printf("%ld-%ld\n", (long)startst, (long)stopst);
1154 #define PCHARDONE /* never again */
1156 - pchar - make a character printable
1158 == static const char *pchar(int ch);
1161 * Is this identical to regchar() over in debug.c? Well, yes. But a
1162 * duplicate here avoids having a debugging-capable regexec.o tied to
1163 * a matching debug.o, and this is convenient. It all disappears in
1164 * the non-debug compilation anyway, so it doesn't matter much.
1166 static const char * /* -> representation */
1169 static char pbuf[10];
1171 if (isprint((uch)ch) || ch == ' ')
1172 sprintf(pbuf, "%c", ch);
1174 sprintf(pbuf, "\\%o", ch);