regex/regcomp.c

   1 /*-
   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.
   5  *
   6  * This code is derived from software contributed to Berkeley by
   7  * Henry Spencer of the University of Toronto.
   8  *
   9  * Redistribution and use in source and binary forms, with or without
  10  * modification, are permitted provided that the following conditions
  11  * are met:
  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  * 3. All advertising materials mentioning features or use of this software
  18  *    must display the following acknowledgement:
  19  *      This product includes software developed by the University of
  20  *      California, Berkeley and its contributors.
  21  * 4. Neither the name of the University nor the names of its contributors
  22  *    may be used to endorse or promote products derived from this software
  23  *    without specific prior written permission.
  24  *
  25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  35  * SUCH DAMAGE.
  36  *
  37  *      @(#)regcomp.c   8.4 (Berkeley) 3/19/94
  38  */
  39
  40 #if defined(LIBC_SCCS) && !defined(lint)
  41 static char sccsid[] = "@(#)regcomp.c   8.4 (Berkeley) 3/19/94";
  42 #endif /* LIBC_SCCS and not lint */
  43
  44 #include <sys/types.h>
  45 #include <stdio.h>
  46 #include <string.h>
  47 #include <ctype.h>
  48 #include <limits.h>
  49 #include <stdlib.h>
  50 #include <regex.h>
  51
  52 #include "utils.h"
  53 #include "regex2.h"
  54
  55 #include "cclass.h"
  56 #include "cname.h"
  57
  58 /*
  59  * parse structure, passed up and down to avoid global variables and
  60  * other clumsinesses
  61  */
  62 struct parse {
  63         char *next;             /* next character in RE */
  64         char *end;              /* end of string (-> NUL normally) */
  65         int error;              /* has an error been seen? */
  66         sop *strip;             /* malloced strip */
  67         sopno ssize;            /* malloced strip size (allocated) */
  68         sopno slen;             /* malloced strip length (used) */
  69         int ncsalloc;           /* number of csets allocated */
  70         struct re_guts *g;
  71 #       define  NPAREN  10      /* we need to remember () 1-9 for back refs */
  72         sopno pbegin[NPAREN];   /* -> ( ([0] unused) */
  73         sopno pend[NPAREN];     /* -> ) ([0] unused) */
  74 };
  75
  76 /* ========= begin header generated by ./mkh ========= */
  77 #ifdef __cplusplus
  78 extern "C" {
  79 #endif
  80
  81 /* === regcomp.c === */
  82 static void p_ere __P((struct parse *p, int stop));
  83 static void p_ere_exp __P((struct parse *p));
  84 static void p_str __P((struct parse *p));
  85 static void p_bre __P((struct parse *p, int end1, int end2));
  86 static int p_simp_re __P((struct parse *p, int starordinary));
  87 static int p_count __P((struct parse *p));
  88 static void p_bracket __P((struct parse *p));
  89 static void p_b_term __P((struct parse *p, cset *cs));
  90 static void p_b_cclass __P((struct parse *p, cset *cs));
  91 static void p_b_eclass __P((struct parse *p, cset *cs));
  92 static char p_b_symbol __P((struct parse *p));
  93 static char p_b_coll_elem __P((struct parse *p, int endc));
  94 static char othercase __P((int ch));
  95 static void bothcases __P((struct parse *p, int ch));
  96 static void ordinary __P((struct parse *p, int ch));
  97 static void nonnewline __P((struct parse *p));
  98 static void repeat __P((struct parse *p, sopno start, int from, int to));
  99 static int seterr __P((struct parse *p, int e));
 100 static cset *allocset __P((struct parse *p));
 101 static void freeset __P((struct parse *p, cset *cs));
 102 static int freezeset __P((struct parse *p, cset *cs));
 103 static int firstch __P((struct parse *p, cset *cs));
 104 static int nch __P((struct parse *p, cset *cs));
 105 static void mcadd __P((struct parse *p, cset *cs, char *cp));
 106 static void mcsub __P((cset *cs, char *cp));
 107 static int mcin __P((cset *cs, char *cp));
 108 static char *mcfind __P((cset *cs, char *cp));
 109 static void mcinvert __P((struct parse *p, cset *cs));
 110 static void mccase __P((struct parse *p, cset *cs));
 111 static int isinsets __P((struct re_guts *g, int c));
 112 static int samesets __P((struct re_guts *g, int c1, int c2));
 113 static void categorize __P((struct parse *p, struct re_guts *g));
 114 static sopno dupl __P((struct parse *p, sopno start, sopno finish));
 115 static void doemit __P((struct parse *p, sop op, size_t opnd));
 116 static void doinsert __P((struct parse *p, sop op, size_t opnd, sopno pos));
 117 static void dofwd __P((struct parse *p, sopno pos, sop value));
 118 static void enlarge __P((struct parse *p, sopno size));
 119 static void stripsnug __P((struct parse *p, struct re_guts *g));
 120 static void findmust __P((struct parse *p, struct re_guts *g));
 121 static sopno pluscount __P((struct parse *p, struct re_guts *g));
 122
 123 #ifdef __cplusplus
 124 }
 125 #endif
 126 /* ========= end header generated by ./mkh ========= */
 127
 128 static char nuls[10];           /* place to point scanner in event of error */
 129
 130 /*
 131  * macros for use with parse structure
 132  * BEWARE:  these know that the parse structure is named `p' !!!
 133  */
 134 #define PEEK()  (*p->next)
 135 #define PEEK2() (*(p->next+1))
 136 #define MORE()  (p->next < p->end)
 137 #define MORE2() (p->next+1 < p->end)
 138 #define SEE(c)  (MORE() && PEEK() == (c))
 139 #define SEETWO(a, b)    (MORE() && MORE2() && PEEK() == (a) && PEEK2() == (b))
 140 #define EAT(c)  ((SEE(c)) ? (NEXT(), 1) : 0)
 141 #define EATTWO(a, b)    ((SEETWO(a, b)) ? (NEXT2(), 1) : 0)
 142 #define NEXT()  (p->next++)
 143 #define NEXT2() (p->next += 2)
 144 #define NEXTn(n)        (p->next += (n))
 145 #define GETNEXT()       (*p->next++)
 146 #define SETERROR(e)     seterr(p, (e))
 147 #define REQUIRE(co, e)  ((co) || SETERROR(e))
 148 #define MUSTSEE(c, e)   (REQUIRE(MORE() && PEEK() == (c), e))
 149 #define MUSTEAT(c, e)   (REQUIRE(MORE() && GETNEXT() == (c), e))
 150 #define MUSTNOTSEE(c, e)        (REQUIRE(!MORE() || PEEK() != (c), e))
 151 #define EMIT(op, sopnd) doemit(p, (sop)(op), (size_t)(sopnd))
 152 #define INSERT(op, pos) doinsert(p, (sop)(op), HERE()-(pos)+1, pos)
 153 #define AHEAD(pos)              dofwd(p, pos, HERE()-(pos))
 154 #define ASTERN(sop, pos)        EMIT(sop, HERE()-pos)
 155 #define HERE()          (p->slen)
 156 #define THERE()         (p->slen - 1)
 157 #define THERETHERE()    (p->slen - 2)
 158 #define DROP(n) (p->slen -= (n))
 159
 160 #ifndef NDEBUG
 161 static int never = 0;           /* for use in asserts; shuts lint up */
 162 #else
 163 #define never   0               /* some <assert.h>s have bugs too */
 164 #endif
 165
 166 /*
 167  - regcomp - interface for parser and compilation
 168  = extern int regcomp(regex_t *, const char *, int);
 169  = #define      REG_BASIC       0000
 170  = #define      REG_EXTENDED    0001
 171  = #define      REG_ICASE       0002
 172  = #define      REG_NOSUB       0004
 173  = #define      REG_NEWLINE     0010
 174  = #define      REG_NOSPEC      0020
 175  = #define      REG_PEND        0040
 176  = #define      REG_DUMP        0200
 177  */
 178 int                             /* 0 success, otherwise REG_something */
 179 regcomp(preg, pattern, cflags)
 180 regex_t *preg;
 181 const char *pattern;
 182 int cflags;
 183 {
 184         struct parse pa;
 185         register struct re_guts *g;
 186         register struct parse *p = &pa;
 187         register int i;
 188         register size_t len;
 189 #ifdef REDEBUG
 190 #       define  GOODFLAGS(f)    (f)
 191 #else
 192 #       define  GOODFLAGS(f)    ((f)&~REG_DUMP)
 193 #endif
 194
 195         cflags = GOODFLAGS(cflags);
 196         if ((cflags&REG_EXTENDED) && (cflags&REG_NOSPEC))
 197                 return(REG_INVARG);
 198
 199         if (cflags&REG_PEND) {
 200                 if (preg->re_endp < pattern)
 201                         return(REG_INVARG);
 202                 len = preg->re_endp - pattern;
 203         } else
 204                 len = strlen((char *)pattern);
 205
 206         /* do the mallocs early so failure handling is easy */
 207         g = (struct re_guts *)malloc(sizeof(struct re_guts) +
 208                                                         (NC-1)*sizeof(cat_t));
 209         if (g == NULL)
 210                 return(REG_ESPACE);
 211         p->ssize = len/(size_t)2*(size_t)3 + (size_t)1; /* ugh */
 212         p->strip = (sop *)malloc(p->ssize * sizeof(sop));
 213         p->slen = 0;
 214         if (p->strip == NULL) {
 215                 free((char *)g);
 216                 return(REG_ESPACE);
 217         }
 218
 219         /* set things up */
 220         p->g = g;
 221         p->next = (char *)pattern;      /* convenience; we do not modify it */
 222         p->end = p->next + len;
 223         p->error = 0;
 224         p->ncsalloc = 0;
 225         for (i = 0; i < NPAREN; i++) {
 226                 p->pbegin[i] = 0;
 227                 p->pend[i] = 0;
 228         }
 229         g->csetsize = NC;
 230         g->sets = NULL;
 231         g->setbits = NULL;
 232         g->ncsets = 0;
 233         g->cflags = cflags;
 234         g->iflags = 0;
 235         g->nbol = 0;
 236         g->neol = 0;
 237         g->must = NULL;
 238         g->mlen = 0;
 239         g->nsub = 0;
 240         g->ncategories = 1;     /* category 0 is "everything else" */
 241         g->categories = &g->catspace[-(CHAR_MIN)];
 242         (void) memset((char *)g->catspace, 0, NC*sizeof(cat_t));
 243         g->backrefs = 0;
 244
 245         /* do it */
 246         EMIT(OEND, 0);
 247         g->firststate = THERE();
 248         if (cflags&REG_EXTENDED)
 249                 p_ere(p, OUT);
 250         else if (cflags&REG_NOSPEC)
 251                 p_str(p);
 252         else
 253                 p_bre(p, OUT, OUT);
 254         EMIT(OEND, 0);
 255         g->laststate = THERE();
 256
 257         /* tidy up loose ends and fill things in */
 258         categorize(p, g);
 259         stripsnug(p, g);
 260         findmust(p, g);
 261         g->nplus = pluscount(p, g);
 262         g->magic = MAGIC2;
 263         preg->re_nsub = g->nsub;
 264         preg->re_g = g;
 265         preg->re_magic = MAGIC1;
 266 #ifndef REDEBUG
 267         /* not debugging, so can't rely on the assert() in regexec() */
 268         if (g->iflags&BAD)
 269                 SETERROR(REG_ASSERT);
 270 #endif
 271
 272         /* win or lose, we're done */
 273         if (p->error != 0)      /* lose */
 274                 regfree(preg);
 275         return(p->error);
 276 }
 277
 278 /*
 279  - p_ere - ERE parser top level, concatenation and alternation
 280  == static void p_ere(register struct parse *p, int stop);
 281  */
 282 static void
 283 p_ere(p, stop)
 284 register struct parse *p;
 285 int stop;                       /* character this ERE should end at */
 286 {
 287         register char c;
 288         register sopno prevback;
 289         register sopno prevfwd;
 290         register sopno conc;
 291         register int first = 1;         /* is this the first alternative? */
 292
 293         for (;;) {
 294                 /* do a bunch of concatenated expressions */
 295                 conc = HERE();
 296                 while (MORE() && (c = PEEK()) != '|' && c != stop)
 297                         p_ere_exp(p);
 298                 REQUIRE(HERE() != conc, REG_EMPTY);     /* require nonempty */
 299
 300                 if (!EAT('|'))
 301                         break;          /* NOTE BREAK OUT */
 302
 303                 if (first) {
 304                         INSERT(OCH_, conc);     /* offset is wrong */
 305                         prevfwd = conc;
 306                         prevback = conc;
 307                         first = 0;
 308                 }
 309                 ASTERN(OOR1, prevback);
 310                 prevback = THERE();
 311                 AHEAD(prevfwd);                 /* fix previous offset */
 312                 prevfwd = HERE();
 313                 EMIT(OOR2, 0);                  /* offset is very wrong */
 314         }
 315
 316         if (!first) {           /* tail-end fixups */
 317                 AHEAD(prevfwd);
 318                 ASTERN(O_CH, prevback);
 319         }
 320
 321         assert(!MORE() || SEE(stop));
 322 }
 323
 324 /*
 325  - p_ere_exp - parse one subERE, an atom possibly followed by a repetition op
 326  == static void p_ere_exp(register struct parse *p);
 327  */
 328 static void
 329 p_ere_exp(p)
 330 register struct parse *p;
 331 {
 332         register char c;
 333         register sopno pos;
 334         register int count;
 335         register int count2;
 336         register sopno subno;
 337         int wascaret = 0;
 338
 339         assert(MORE());         /* caller should have ensured this */
 340         c = GETNEXT();
 341
 342         pos = HERE();
 343         switch (c) {
 344         case '(':
 345                 REQUIRE(MORE(), REG_EPAREN);
 346                 p->g->nsub++;
 347                 subno = p->g->nsub;
 348                 if (subno < NPAREN)
 349                         p->pbegin[subno] = HERE();
 350                 EMIT(OLPAREN, subno);
 351                 if (!SEE(')'))
 352                         p_ere(p, ')');
 353                 if (subno < NPAREN) {
 354                         p->pend[subno] = HERE();
 355                         assert(p->pend[subno] != 0);
 356                 }
 357                 EMIT(ORPAREN, subno);
 358                 MUSTEAT(')', REG_EPAREN);
 359                 break;
 360 #ifndef POSIX_MISTAKE
 361         case ')':               /* happens only if no current unmatched ( */
 362                 /*
 363                  * You may ask, why the ifndef?  Because I didn't notice
 364                  * this until slightly too late for 1003.2, and none of the
 365                  * other 1003.2 regular-expression reviewers noticed it at
 366                  * all.  So an unmatched ) is legal POSIX, at least until
 367                  * we can get it fixed.
 368                  */
 369                 SETERROR(REG_EPAREN);
 370                 break;
 371 #endif
 372         case '^':
 373                 EMIT(OBOL, 0);
 374                 p->g->iflags |= USEBOL;
 375                 p->g->nbol++;
 376                 wascaret = 1;
 377                 break;
 378         case '$':
 379                 EMIT(OEOL, 0);
 380                 p->g->iflags |= USEEOL;
 381                 p->g->neol++;
 382                 break;
 383         case '|':
 384                 SETERROR(REG_EMPTY);
 385                 break;
 386         case '*':
 387         case '+':
 388         case '?':
 389                 SETERROR(REG_BADRPT);
 390                 break;
 391         case '.':
 392                 if (p->g->cflags&REG_NEWLINE)
 393                         nonnewline(p);
 394                 else
 395                         EMIT(OANY, 0);
 396                 break;
 397         case '[':
 398                 p_bracket(p);
 399                 break;
 400         case '\\':
 401                 REQUIRE(MORE(), REG_EESCAPE);
 402                 c = GETNEXT();
 403                 ordinary(p, c);
 404                 break;
 405         case '{':               /* okay as ordinary except if digit follows */
 406                 REQUIRE(!MORE() || !isdigit(PEEK()), REG_BADRPT);
 407                 /* FALLTHROUGH */
 408         default:
 409                 ordinary(p, c);
 410                 break;
 411         }
 412
 413         if (!MORE())
 414                 return;
 415         c = PEEK();
 416         /* we call { a repetition if followed by a digit */
 417         if (!( c == '*' || c == '+' || c == '?' ||
 418                                 (c == '{' && MORE2() && isdigit(PEEK2())) ))
 419                 return;         /* no repetition, we're done */
 420         NEXT();
 421
 422         REQUIRE(!wascaret, REG_BADRPT);
 423         switch (c) {
 424         case '*':       /* implemented as +? */
 425                 /* this case does not require the (y|) trick, noKLUDGE */
 426                 INSERT(OPLUS_, pos);
 427                 ASTERN(O_PLUS, pos);
 428                 INSERT(OQUEST_, pos);
 429                 ASTERN(O_QUEST, pos);
 430                 break;
 431         case '+':
 432                 INSERT(OPLUS_, pos);
 433                 ASTERN(O_PLUS, pos);
 434                 break;
 435         case '?':
 436                 /* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
 437                 INSERT(OCH_, pos);              /* offset slightly wrong */
 438                 ASTERN(OOR1, pos);              /* this one's right */
 439                 AHEAD(pos);                     /* fix the OCH_ */
 440                 EMIT(OOR2, 0);                  /* offset very wrong... */
 441                 AHEAD(THERE());                 /* ...so fix it */
 442                 ASTERN(O_CH, THERETHERE());
 443                 break;
 444         case '{':
 445                 count = p_count(p);
 446                 if (EAT(',')) {
 447                         if (isdigit(PEEK())) {
 448                                 count2 = p_count(p);
 449                                 REQUIRE(count <= count2, REG_BADBR);
 450                         } else          /* single number with comma */
 451                                 count2 = INFINITY;
 452                 } else          /* just a single number */
 453                         count2 = count;
 454                 repeat(p, pos, count, count2);
 455                 if (!EAT('}')) {        /* error heuristics */
 456                         while (MORE() && PEEK() != '}')
 457                                 NEXT();
 458                         REQUIRE(MORE(), REG_EBRACE);
 459                         SETERROR(REG_BADBR);
 460                 }
 461                 break;
 462         }
 463
 464         if (!MORE())
 465                 return;
 466         c = PEEK();
 467         if (!( c == '*' || c == '+' || c == '?' ||
 468                                 (c == '{' && MORE2() && isdigit(PEEK2())) ) )
 469                 return;
 470         SETERROR(REG_BADRPT);
 471 }
 472
 473 /*
 474  - p_str - string (no metacharacters) "parser"
 475  == static void p_str(register struct parse *p);
 476  */
 477 static void
 478 p_str(p)
 479 register struct parse *p;
 480 {
 481         REQUIRE(MORE(), REG_EMPTY);
 482         while (MORE())
 483                 ordinary(p, GETNEXT());
 484 }
 485
 486 /*
 487  - p_bre - BRE parser top level, anchoring and concatenation
 488  == static void p_bre(register struct parse *p, register int end1, \
 489  ==     register int end2);
 490  * Giving end1 as OUT essentially eliminates the end1/end2 check.
 491  *
 492  * This implementation is a bit of a kludge, in that a trailing $ is first
 493  * taken as an ordinary character and then revised to be an anchor.  The
 494  * only undesirable side effect is that '$' gets included as a character
 495  * category in such cases.  This is fairly harmless; not worth fixing.
 496  * The amount of lookahead needed to avoid this kludge is excessive.
 497  */
 498 static void
 499 p_bre(p, end1, end2)
 500 register struct parse *p;
 501 register int end1;              /* first terminating character */
 502 register int end2;              /* second terminating character */
 503 {
 504         register sopno start = HERE();
 505         register int first = 1;                 /* first subexpression? */
 506         register int wasdollar = 0;
 507
 508         if (EAT('^')) {
 509                 EMIT(OBOL, 0);
 510                 p->g->iflags |= USEBOL;
 511                 p->g->nbol++;
 512         }
 513         while (MORE() && !SEETWO(end1, end2)) {
 514                 wasdollar = p_simp_re(p, first);
 515                 first = 0;
 516         }
 517         if (wasdollar) {        /* oops, that was a trailing anchor */
 518                 DROP(1);
 519                 EMIT(OEOL, 0);
 520                 p->g->iflags |= USEEOL;
 521                 p->g->neol++;
 522         }
 523
 524         REQUIRE(HERE() != start, REG_EMPTY);    /* require nonempty */
 525 }
 526
 527 /*
 528  - p_simp_re - parse a simple RE, an atom possibly followed by a repetition
 529  == static int p_simp_re(register struct parse *p, int starordinary);
 530  */
 531 static int                      /* was the simple RE an unbackslashed $? */
 532 p_simp_re(p, starordinary)
 533 register struct parse *p;
 534 int starordinary;               /* is a leading * an ordinary character? */
 535 {
 536         register int c;
 537         register int count;
 538         register int count2;
 539         register sopno pos;
 540         register int i;
 541         register sopno subno;
 542 #       define  BACKSL  (1<<CHAR_BIT)
 543
 544         pos = HERE();           /* repetion op, if any, covers from here */
 545
 546         assert(MORE());         /* caller should have ensured this */
 547         c = GETNEXT();
 548         if (c == '\\') {
 549                 REQUIRE(MORE(), REG_EESCAPE);
 550                 c = BACKSL | (unsigned char)GETNEXT();
 551         }
 552         switch (c) {
 553         case '.':
 554                 if (p->g->cflags&REG_NEWLINE)
 555                         nonnewline(p);
 556                 else
 557                         EMIT(OANY, 0);
 558                 break;
 559         case '[':
 560                 p_bracket(p);
 561                 break;
 562         case BACKSL|'{':
 563                 SETERROR(REG_BADRPT);
 564                 break;
 565         case BACKSL|'(':
 566                 p->g->nsub++;
 567                 subno = p->g->nsub;
 568                 if (subno < NPAREN)
 569                         p->pbegin[subno] = HERE();
 570                 EMIT(OLPAREN, subno);
 571                 /* the MORE here is an error heuristic */
 572                 if (MORE() && !SEETWO('\\', ')'))
 573                         p_bre(p, '\\', ')');
 574                 if (subno < NPAREN) {
 575                         p->pend[subno] = HERE();
 576                         assert(p->pend[subno] != 0);
 577                 }
 578                 EMIT(ORPAREN, subno);
 579                 REQUIRE(EATTWO('\\', ')'), REG_EPAREN);
 580                 break;
 581         case BACKSL|')':        /* should not get here -- must be user */
 582         case BACKSL|'}':
 583                 SETERROR(REG_EPAREN);
 584                 break;
 585         case BACKSL|'1':
 586         case BACKSL|'2':
 587         case BACKSL|'3':
 588         case BACKSL|'4':
 589         case BACKSL|'5':
 590         case BACKSL|'6':
 591         case BACKSL|'7':
 592         case BACKSL|'8':
 593         case BACKSL|'9':
 594                 i = (c&~BACKSL) - '0';
 595                 assert(i < NPAREN);
 596                 if (p->pend[i] != 0) {
 597                         assert(i <= p->g->nsub);
 598                         EMIT(OBACK_, i);
 599                         assert(p->pbegin[i] != 0);
 600                         assert(OP(p->strip[p->pbegin[i]]) == OLPAREN);
 601                         assert(OP(p->strip[p->pend[i]]) == ORPAREN);
 602                         (void) dupl(p, p->pbegin[i]+1, p->pend[i]);
 603                         EMIT(O_BACK, i);
 604                 } else
 605                         SETERROR(REG_ESUBREG);
 606                 p->g->backrefs = 1;
 607                 break;
 608         case '*':
 609                 REQUIRE(starordinary, REG_BADRPT);
 610                 /* FALLTHROUGH */
 611         default:
 612                 ordinary(p, c &~ BACKSL);
 613                 break;
 614         }
 615
 616         if (EAT('*')) {         /* implemented as +? */
 617                 /* this case does not require the (y|) trick, noKLUDGE */
 618                 INSERT(OPLUS_, pos);
 619                 ASTERN(O_PLUS, pos);
 620                 INSERT(OQUEST_, pos);
 621                 ASTERN(O_QUEST, pos);
 622         } else if (EATTWO('\\', '{')) {
 623                 count = p_count(p);
 624                 if (EAT(',')) {
 625                         if (MORE() && isdigit(PEEK())) {
 626                                 count2 = p_count(p);
 627                                 REQUIRE(count <= count2, REG_BADBR);
 628                         } else          /* single number with comma */
 629                                 count2 = INFINITY;
 630                 } else          /* just a single number */
 631                         count2 = count;
 632                 repeat(p, pos, count, count2);
 633                 if (!EATTWO('\\', '}')) {       /* error heuristics */
 634                         while (MORE() && !SEETWO('\\', '}'))
 635                                 NEXT();
 636                         REQUIRE(MORE(), REG_EBRACE);
 637                         SETERROR(REG_BADBR);
 638                 }
 639         } else if (c == (unsigned char)'$')     /* $ (but not \$) ends it */
 640                 return(1);
 641
 642         return(0);
 643 }
 644
 645 /*
 646  - p_count - parse a repetition count
 647  == static int p_count(register struct parse *p);
 648  */
 649 static int                      /* the value */
 650 p_count(p)
 651 register struct parse *p;
 652 {
 653         register int count = 0;
 654         register int ndigits = 0;
 655
 656         while (MORE() && isdigit(PEEK()) && count <= DUPMAX) {
 657                 count = count*10 + (GETNEXT() - '0');
 658                 ndigits++;
 659         }
 660
 661         REQUIRE(ndigits > 0 && count <= DUPMAX, REG_BADBR);
 662         return(count);
 663 }
 664
 665 /*
 666  - p_bracket - parse a bracketed character list
 667  == static void p_bracket(register struct parse *p);
 668  *
 669  * Note a significant property of this code:  if the allocset() did SETERROR,
 670  * no set operations are done.
 671  */
 672 static void
 673 p_bracket(p)
 674 register struct parse *p;
 675 {
 676         register char c;
 677         register cset *cs = allocset(p);
 678         register int invert = 0;
 679
 680         /* Dept of Truly Sickening Special-Case Kludges */
 681         if (p->next + 5 < p->end && strncmp(p->next, "[:<:]]", 6) == 0) {
 682                 EMIT(OBOW, 0);
 683                 NEXTn(6);
 684                 return;
 685         }
 686         if (p->next + 5 < p->end && strncmp(p->next, "[:>:]]", 6) == 0) {
 687                 EMIT(OEOW, 0);
 688                 NEXTn(6);
 689                 return;
 690         }
 691
 692         if (EAT('^'))
 693                 invert++;       /* make note to invert set at end */
 694         if (EAT(']'))
 695                 CHadd(cs, ']');
 696         else if (EAT('-'))
 697                 CHadd(cs, '-');
 698         while (MORE() && PEEK() != ']' && !SEETWO('-', ']'))
 699                 p_b_term(p, cs);
 700         if (EAT('-'))
 701                 CHadd(cs, '-');
 702         MUSTEAT(']', REG_EBRACK);
 703
 704         if (p->error != 0)      /* don't mess things up further */
 705                 return;
 706
 707         if (p->g->cflags&REG_ICASE) {
 708                 register int i;
 709                 register int ci;
 710
 711                 for (i = p->g->csetsize - 1; i >= 0; i--)
 712                         if (CHIN(cs, i) && isalpha(i)) {
 713                                 ci = othercase(i);
 714                                 if (ci != i)
 715                                         CHadd(cs, ci);
 716                         }
 717                 if (cs->multis != NULL)
 718                         mccase(p, cs);
 719         }
 720         if (invert) {
 721                 register int i;
 722
 723                 for (i = p->g->csetsize - 1; i >= 0; i--)
 724                         if (CHIN(cs, i))
 725                                 CHsub(cs, i);
 726                         else
 727                                 CHadd(cs, i);
 728                 if (p->g->cflags&REG_NEWLINE)
 729                         CHsub(cs, '\n');
 730                 if (cs->multis != NULL)
 731                         mcinvert(p, cs);
 732         }
 733
 734         assert(cs->multis == NULL);             /* xxx */
 735
 736         if (nch(p, cs) == 1) {          /* optimize singleton sets */
 737                 ordinary(p, firstch(p, cs));
 738                 freeset(p, cs);
 739         } else
 740                 EMIT(OANYOF, freezeset(p, cs));
 741 }
 742
 743 /*
 744  - p_b_term - parse one term of a bracketed character list
 745  == static void p_b_term(register struct parse *p, register cset *cs);
 746  */
 747 static void
 748 p_b_term(p, cs)
 749 register struct parse *p;
 750 register cset *cs;
 751 {
 752         register char c;
 753         register char start, finish;
 754         register int i;
 755
 756         /* classify what we've got */
 757         switch ((MORE()) ? PEEK() : '\0') {
 758         case '[':
 759                 c = (MORE2()) ? PEEK2() : '\0';
 760                 break;
 761         case '-':
 762                 SETERROR(REG_ERANGE);
 763                 return;                 /* NOTE RETURN */
 764                 break;
 765         default:
 766                 c = '\0';
 767                 break;
 768         }
 769
 770         switch (c) {
 771         case ':':               /* character class */
 772                 NEXT2();
 773                 REQUIRE(MORE(), REG_EBRACK);
 774                 c = PEEK();
 775                 REQUIRE(c != '-' && c != ']', REG_ECTYPE);
 776                 p_b_cclass(p, cs);
 777                 REQUIRE(MORE(), REG_EBRACK);
 778                 REQUIRE(EATTWO(':', ']'), REG_ECTYPE);
 779                 break;
 780         case '=':               /* equivalence class */
 781                 NEXT2();
 782                 REQUIRE(MORE(), REG_EBRACK);
 783                 c = PEEK();
 784                 REQUIRE(c != '-' && c != ']', REG_ECOLLATE);
 785                 p_b_eclass(p, cs);
 786                 REQUIRE(MORE(), REG_EBRACK);
 787                 REQUIRE(EATTWO('=', ']'), REG_ECOLLATE);
 788                 break;
 789         default:                /* symbol, ordinary character, or range */
 790 /* xxx revision needed for multichar stuff */
 791                 start = p_b_symbol(p);
 792                 if (SEE('-') && MORE2() && PEEK2() != ']') {
 793                         /* range */
 794                         NEXT();
 795                         if (EAT('-'))
 796                                 finish = '-';
 797                         else
 798                                 finish = p_b_symbol(p);
 799                 } else
 800                         finish = start;
 801 /* xxx what about signed chars here... */
 802                 REQUIRE(start <= finish, REG_ERANGE);
 803                 for (i = start; i <= finish; i++)
 804                         CHadd(cs, i);
 805                 break;
 806         }
 807 }
 808
 809 /*
 810  - p_b_cclass - parse a character-class name and deal with it
 811  == static void p_b_cclass(register struct parse *p, register cset *cs);
 812  */
 813 static void
 814 p_b_cclass(p, cs)
 815 register struct parse *p;
 816 register cset *cs;
 817 {
 818         register char *sp = p->next;
 819         register struct cclass *cp;
 820         register size_t len;
 821         register char *u;
 822         register char c;
 823
 824         while (MORE() && isalpha(PEEK()))
 825                 NEXT();
 826         len = p->next - sp;
 827         for (cp = cclasses; cp->name != NULL; cp++)
 828                 if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
 829                         break;
 830         if (cp->name == NULL) {
 831                 /* oops, didn't find it */
 832                 SETERROR(REG_ECTYPE);
 833                 return;
 834         }
 835
 836         u = cp->chars;
 837         while ((c = *u++) != '\0')
 838                 CHadd(cs, c);
 839         for (u = cp->multis; *u != '\0'; u += strlen(u) + 1)
 840                 MCadd(p, cs, u);
 841 }
 842
 843 /*
 844  - p_b_eclass - parse an equivalence-class name and deal with it
 845  == static void p_b_eclass(register struct parse *p, register cset *cs);
 846  *
 847  * This implementation is incomplete. xxx
 848  */
 849 static void
 850 p_b_eclass(p, cs)
 851 register struct parse *p;
 852 register cset *cs;
 853 {
 854         register char c;
 855
 856         c = p_b_coll_elem(p, '=');
 857         CHadd(cs, c);
 858 }
 859
 860 /*
 861  - p_b_symbol - parse a character or [..]ed multicharacter collating symbol
 862  == static char p_b_symbol(register struct parse *p);
 863  */
 864 static char                     /* value of symbol */
 865 p_b_symbol(p)
 866 register struct parse *p;
 867 {
 868         register char value;
 869
 870         REQUIRE(MORE(), REG_EBRACK);
 871         if (!EATTWO('[', '.'))
 872                 return(GETNEXT());
 873
 874         /* collating symbol */
 875         value = p_b_coll_elem(p, '.');
 876         REQUIRE(EATTWO('.', ']'), REG_ECOLLATE);
 877         return(value);
 878 }
 879
 880 /*
 881  - p_b_coll_elem - parse a collating-element name and look it up
 882  == static char p_b_coll_elem(register struct parse *p, int endc);
 883  */
 884 static char                     /* value of collating element */
 885 p_b_coll_elem(p, endc)
 886 register struct parse *p;
 887 int endc;                       /* name ended by endc,']' */
 888 {
 889         register char *sp = p->next;
 890         register struct cname *cp;
 891         register int len;
 892         register char c;
 893
 894         while (MORE() && !SEETWO(endc, ']'))
 895                 NEXT();
 896         if (!MORE()) {
 897                 SETERROR(REG_EBRACK);
 898                 return(0);
 899         }
 900         len = p->next - sp;
 901         for (cp = cnames; cp->name != NULL; cp++)
 902                 if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
 903                         return(cp->code);       /* known name */
 904         if (len == 1)
 905                 return(*sp);    /* single character */
 906         SETERROR(REG_ECOLLATE);                 /* neither */
 907         return(0);
 908 }
 909
 910 /*
 911  - othercase - return the case counterpart of an alphabetic
 912  == static char othercase(int ch);
 913  */
 914 static char                     /* if no counterpart, return ch */
 915 othercase(ch)
 916 int ch;
 917 {
 918         assert(isalpha(ch));
 919         if (isupper(ch))
 920                 return(tolower(ch));
 921         else if (islower(ch))
 922                 return(toupper(ch));
 923         else                    /* peculiar, but could happen */
 924                 return(ch);
 925 }
 926
 927 /*
 928  - bothcases - emit a dualcase version of a two-case character
 929  == static void bothcases(register struct parse *p, int ch);
 930  *
 931  * Boy, is this implementation ever a kludge...
 932  */
 933 static void
 934 bothcases(p, ch)
 935 register struct parse *p;
 936 int ch;
 937 {
 938         register char *oldnext = p->next;
 939         register char *oldend = p->end;
 940         char bracket[3];
 941
 942         assert(othercase(ch) != ch);    /* p_bracket() would recurse */
 943         p->next = bracket;
 944         p->end = bracket+2;
 945         bracket[0] = ch;
 946         bracket[1] = ']';
 947         bracket[2] = '\0';
 948         p_bracket(p);
 949         assert(p->next == bracket+2);
 950         p->next = oldnext;
 951         p->end = oldend;
 952 }
 953
 954 /*
 955  - ordinary - emit an ordinary character
 956  == static void ordinary(register struct parse *p, register int ch);
 957  */
 958 static void
 959 ordinary(p, ch)
 960 register struct parse *p;
 961 register int ch;
 962 {
 963         register cat_t *cap = p->g->categories;
 964
 965         if ((p->g->cflags&REG_ICASE) && isalpha(ch) && othercase(ch) != ch)
 966                 bothcases(p, ch);
 967         else {
 968                 EMIT(OCHAR, (unsigned char)ch);
 969                 if (cap[ch] == 0)
 970                         cap[ch] = p->g->ncategories++;
 971         }
 972 }
 973
 974 /*
 975  - nonnewline - emit REG_NEWLINE version of OANY
 976  == static void nonnewline(register struct parse *p);
 977  *
 978  * Boy, is this implementation ever a kludge...
 979  */
 980 static void
 981 nonnewline(p)
 982 register struct parse *p;
 983 {
 984         register char *oldnext = p->next;
 985         register char *oldend = p->end;
 986         char bracket[4];
 987
 988         p->next = bracket;
 989         p->end = bracket+3;
 990         bracket[0] = '^';
 991         bracket[1] = '\n';
 992         bracket[2] = ']';
 993         bracket[3] = '\0';
 994         p_bracket(p);
 995         assert(p->next == bracket+3);
 996         p->next = oldnext;
 997         p->end = oldend;
 998 }
 999
1000 /*
1001  - repeat - generate code for a bounded repetition, recursively if needed
1002  == static void repeat(register struct parse *p, sopno start, int from, int to);
1003  */
1004 static void
1005 repeat(p, start, from, to)
1006 register struct parse *p;
1007 sopno start;                    /* operand from here to end of strip */
1008 int from;                       /* repeated from this number */
1009 int to;                         /* to this number of times (maybe INFINITY) */
1010 {
1011         register sopno finish = HERE();
1012 #       define  N       2
1013 #       define  INF     3
1014 #       define  REP(f, t)       ((f)*8 + (t))
1015 #       define  MAP(n)  (((n) <= 1) ? (n) : ((n) == INFINITY) ? INF : N)
1016         register sopno copy;
1017
1018         if (p->error != 0)      /* head off possible runaway recursion */
1019                 return;
1020
1021         assert(from <= to);
1022
1023         switch (REP(MAP(from), MAP(to))) {
1024         case REP(0, 0):                 /* must be user doing this */
1025                 DROP(finish-start);     /* drop the operand */
1026                 break;
1027         case REP(0, 1):                 /* as x{1,1}? */
1028         case REP(0, N):                 /* as x{1,n}? */
1029         case REP(0, INF):               /* as x{1,}? */
1030                 /* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
1031                 INSERT(OCH_, start);            /* offset is wrong... */
1032                 repeat(p, start+1, 1, to);
1033                 ASTERN(OOR1, start);
1034                 AHEAD(start);                   /* ... fix it */
1035                 EMIT(OOR2, 0);
1036                 AHEAD(THERE());
1037                 ASTERN(O_CH, THERETHERE());
1038                 break;
1039         case REP(1, 1):                 /* trivial case */
1040                 /* done */
1041                 break;
1042         case REP(1, N):                 /* as x?x{1,n-1} */
1043                 /* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
1044                 INSERT(OCH_, start);
1045                 ASTERN(OOR1, start);
1046                 AHEAD(start);
1047                 EMIT(OOR2, 0);                  /* offset very wrong... */
1048                 AHEAD(THERE());                 /* ...so fix it */
1049                 ASTERN(O_CH, THERETHERE());
1050                 copy = dupl(p, start+1, finish+1);
1051                 assert(copy == finish+4);
1052                 repeat(p, copy, 1, to-1);
1053                 break;
1054         case REP(1, INF):               /* as x+ */
1055                 INSERT(OPLUS_, start);
1056                 ASTERN(O_PLUS, start);
1057                 break;
1058         case REP(N, N):                 /* as xx{m-1,n-1} */
1059                 copy = dupl(p, start, finish);
1060                 repeat(p, copy, from-1, to-1);
1061                 break;
1062         case REP(N, INF):               /* as xx{n-1,INF} */
1063                 copy = dupl(p, start, finish);
1064                 repeat(p, copy, from-1, to);
1065                 break;
1066         default:                        /* "can't happen" */
1067                 SETERROR(REG_ASSERT);   /* just in case */
1068                 break;
1069         }
1070 }
1071
1072 /*
1073  - seterr - set an error condition
1074  == static int seterr(register struct parse *p, int e);
1075  */
1076 static int                      /* useless but makes type checking happy */
1077 seterr(p, e)
1078 register struct parse *p;
1079 int e;
1080 {
1081         if (p->error == 0)      /* keep earliest error condition */
1082                 p->error = e;
1083         p->next = nuls;         /* try to bring things to a halt */
1084         p->end = nuls;
1085         return(0);              /* make the return value well-defined */
1086 }
1087
1088 /*
1089  - allocset - allocate a set of characters for []
1090  == static cset *allocset(register struct parse *p);
1091  */
1092 static cset *
1093 allocset(p)
1094 register struct parse *p;
1095 {
1096         register int no = p->g->ncsets++;
1097         register size_t nc;
1098         register size_t nbytes;
1099         register cset *cs;
1100         register size_t css = (size_t)p->g->csetsize;
1101         register int i;
1102
1103         if (no >= p->ncsalloc) {        /* need another column of space */
1104                 p->ncsalloc += CHAR_BIT;
1105                 nc = p->ncsalloc;
1106                 assert(nc % CHAR_BIT == 0);
1107                 nbytes = nc / CHAR_BIT * css;
1108                 if (p->g->sets == NULL)
1109                         p->g->sets = (cset *)malloc(nc * sizeof(cset));
1110                 else
1111                         p->g->sets = (cset *)realloc((char *)p->g->sets,
1112                                                         nc * sizeof(cset));
1113                 if (p->g->setbits == NULL)
1114                         p->g->setbits = (uch *)malloc(nbytes);
1115                 else {
1116                         p->g->setbits = (uch *)realloc((char *)p->g->setbits,
1117                                                                 nbytes);
1118                         /* xxx this isn't right if setbits is now NULL */
1119                         for (i = 0; i < no; i++)
1120                                 p->g->sets[i].ptr = p->g->setbits + css*(i/CHAR_BIT);
1121                 }
1122                 if (p->g->sets != NULL && p->g->setbits != NULL)
1123                         (void) memset((char *)p->g->setbits + (nbytes - css),
1124                                                                 0, css);
1125                 else {
1126                         no = 0;
1127                         SETERROR(REG_ESPACE);
1128                         /* caller's responsibility not to do set ops */
1129                 }
1130         }
1131
1132         assert(p->g->sets != NULL);     /* xxx */
1133         cs = &p->g->sets[no];
1134         cs->ptr = p->g->setbits + css*((no)/CHAR_BIT);
1135         cs->mask = 1 << ((no) % CHAR_BIT);
1136         cs->hash = 0;
1137         cs->smultis = 0;
1138         cs->multis = NULL;
1139
1140         return(cs);
1141 }
1142
1143 /*
1144  - freeset - free a now-unused set
1145  == static void freeset(register struct parse *p, register cset *cs);
1146  */
1147 static void
1148 freeset(p, cs)
1149 register struct parse *p;
1150 register cset *cs;
1151 {
1152         register int i;
1153         register cset *top = &p->g->sets[p->g->ncsets];
1154         register size_t css = (size_t)p->g->csetsize;
1155
1156         for (i = 0; i < css; i++)
1157                 CHsub(cs, i);
1158         if (cs == top-1)        /* recover only the easy case */
1159                 p->g->ncsets--;
1160 }
1161
1162 /*
1163  - freezeset - final processing on a set of characters
1164  == static int freezeset(register struct parse *p, register cset *cs);
1165  *
1166  * The main task here is merging identical sets.  This is usually a waste
1167  * of time (although the hash code minimizes the overhead), but can win
1168  * big if REG_ICASE is being used.  REG_ICASE, by the way, is why the hash
1169  * is done using addition rather than xor -- all ASCII [aA] sets xor to
1170  * the same value!
1171  */
1172 static int                      /* set number */
1173 freezeset(p, cs)
1174 register struct parse *p;
1175 register cset *cs;
1176 {
1177         register uch h = cs->hash;
1178         register int i;
1179         register cset *top = &p->g->sets[p->g->ncsets];
1180         register cset *cs2;
1181         register size_t css = (size_t)p->g->csetsize;
1182
1183         /* look for an earlier one which is the same */
1184         for (cs2 = &p->g->sets[0]; cs2 < top; cs2++)
1185                 if (cs2->hash == h && cs2 != cs) {
1186                         /* maybe */
1187                         for (i = 0; i < css; i++)
1188                                 if (!!CHIN(cs2, i) != !!CHIN(cs, i))
1189                                         break;          /* no */
1190                         if (i == css)
1191                                 break;                  /* yes */
1192                 }
1193
1194         if (cs2 < top) {        /* found one */
1195                 freeset(p, cs);
1196                 cs = cs2;
1197         }
1198
1199         return((int)(cs - p->g->sets));
1200 }
1201
1202 /*
1203  - firstch - return first character in a set (which must have at least one)
1204  == static int firstch(register struct parse *p, register cset *cs);
1205  */
1206 static int                      /* character; there is no "none" value */
1207 firstch(p, cs)
1208 register struct parse *p;
1209 register cset *cs;
1210 {
1211         register int i;
1212         register size_t css = (size_t)p->g->csetsize;
1213
1214         for (i = 0; i < css; i++)
1215                 if (CHIN(cs, i))
1216                         return((char)i);
1217         assert(never);
1218         return(0);              /* arbitrary */
1219 }
1220
1221 /*
1222  - nch - number of characters in a set
1223  == static int nch(register struct parse *p, register cset *cs);
1224  */
1225 static int
1226 nch(p, cs)
1227 register struct parse *p;
1228 register cset *cs;
1229 {
1230         register int i;
1231         register size_t css = (size_t)p->g->csetsize;
1232         register int n = 0;
1233
1234         for (i = 0; i < css; i++)
1235                 if (CHIN(cs, i))
1236                         n++;
1237         return(n);
1238 }
1239
1240 /*
1241  - mcadd - add a collating element to a cset
1242  == static void mcadd(register struct parse *p, register cset *cs, \
1243  ==     register char *cp);
1244  */
1245 static void
1246 mcadd(p, cs, cp)
1247 register struct parse *p;
1248 register cset *cs;
1249 register char *cp;
1250 {
1251         register size_t oldend = cs->smultis;
1252
1253         cs->smultis += strlen(cp) + 1;
1254         if (cs->multis == NULL)
1255                 cs->multis = malloc(cs->smultis);
1256         else
1257                 cs->multis = realloc(cs->multis, cs->smultis);
1258         if (cs->multis == NULL) {
1259                 SETERROR(REG_ESPACE);
1260                 return;
1261         }
1262
1263         (void) strcpy(cs->multis + oldend - 1, cp);
1264         cs->multis[cs->smultis - 1] = '\0';
1265 }
1266
1267 /*
1268  - mcsub - subtract a collating element from a cset
1269  == static void mcsub(register cset *cs, register char *cp);
1270  */
1271 static void
1272 mcsub(cs, cp)
1273 register cset *cs;
1274 register char *cp;
1275 {
1276         register char *fp = mcfind(cs, cp);
1277         register size_t len = strlen(fp);
1278
1279         assert(fp != NULL);
1280         (void) memmove(fp, fp + len + 1,
1281                                 cs->smultis - (fp + len + 1 - cs->multis));
1282         cs->smultis -= len;
1283
1284         if (cs->smultis == 0) {
1285                 free(cs->multis);
1286                 cs->multis = NULL;
1287                 return;
1288         }
1289
1290         cs->multis = realloc(cs->multis, cs->smultis);
1291         assert(cs->multis != NULL);
1292 }
1293
1294 /*
1295  - mcin - is a collating element in a cset?
1296  == static int mcin(register cset *cs, register char *cp);
1297  */
1298 static int
1299 mcin(cs, cp)
1300 register cset *cs;
1301 register char *cp;
1302 {
1303         return(mcfind(cs, cp) != NULL);
1304 }
1305
1306 /*
1307  - mcfind - find a collating element in a cset
1308  == static char *mcfind(register cset *cs, register char *cp);
1309  */
1310 static char *
1311 mcfind(cs, cp)
1312 register cset *cs;
1313 register char *cp;
1314 {
1315         register char *p;
1316
1317         if (cs->multis == NULL)
1318                 return(NULL);
1319         for (p = cs->multis; *p != '\0'; p += strlen(p) + 1)
1320                 if (strcmp(cp, p) == 0)
1321                         return(p);
1322         return(NULL);
1323 }
1324
1325 /*
1326  - mcinvert - invert the list of collating elements in a cset
1327  == static void mcinvert(register struct parse *p, register cset *cs);
1328  *
1329  * This would have to know the set of possibilities.  Implementation
1330  * is deferred.
1331  */
1332 static void
1333 mcinvert(p, cs)
1334 register struct parse *p;
1335 register cset *cs;
1336 {
1337         assert(cs->multis == NULL);     /* xxx */
1338 }
1339
1340 /*
1341  - mccase - add case counterparts of the list of collating elements in a cset
1342  == static void mccase(register struct parse *p, register cset *cs);
1343  *
1344  * This would have to know the set of possibilities.  Implementation
1345  * is deferred.
1346  */
1347 static void
1348 mccase(p, cs)
1349 register struct parse *p;
1350 register cset *cs;
1351 {
1352         assert(cs->multis == NULL);     /* xxx */
1353 }
1354
1355 /*
1356  - isinsets - is this character in any sets?
1357  == static int isinsets(register struct re_guts *g, int c);
1358  */
1359 static int                      /* predicate */
1360 isinsets(g, c)
1361 register struct re_guts *g;
1362 int c;
1363 {
1364         register uch *col;
1365         register int i;
1366         register int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1367         register unsigned uc = (unsigned char)c;
1368
1369         for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1370                 if (col[uc] != 0)
1371                         return(1);
1372         return(0);
1373 }
1374
1375 /*
1376  - samesets - are these two characters in exactly the same sets?
1377  == static int samesets(register struct re_guts *g, int c1, int c2);
1378  */
1379 static int                      /* predicate */
1380 samesets(g, c1, c2)
1381 register struct re_guts *g;
1382 int c1;
1383 int c2;
1384 {
1385         register uch *col;
1386         register int i;
1387         register int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1388         register unsigned uc1 = (unsigned char)c1;
1389         register unsigned uc2 = (unsigned char)c2;
1390
1391         for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1392                 if (col[uc1] != col[uc2])
1393                         return(0);
1394         return(1);
1395 }
1396
1397 /*
1398  - categorize - sort out character categories
1399  == static void categorize(struct parse *p, register struct re_guts *g);
1400  */
1401 static void
1402 categorize(p, g)
1403 struct parse *p;
1404 register struct re_guts *g;
1405 {
1406         register cat_t *cats = g->categories;
1407         register int c;
1408         register int c2;
1409         register cat_t cat;
1410
1411         /* avoid making error situations worse */
1412         if (p->error != 0)
1413                 return;
1414
1415         for (c = CHAR_MIN; c <= CHAR_MAX; c++)
1416                 if (cats[c] == 0 && isinsets(g, c)) {
1417                         cat = g->ncategories++;
1418                         cats[c] = cat;
1419                         for (c2 = c+1; c2 <= CHAR_MAX; c2++)
1420                                 if (cats[c2] == 0 && samesets(g, c, c2))
1421                                         cats[c2] = cat;
1422                 }
1423 }
1424
1425 /*
1426  - dupl - emit a duplicate of a bunch of sops
1427  == static sopno dupl(register struct parse *p, sopno start, sopno finish);
1428  */
1429 static sopno                    /* start of duplicate */
1430 dupl(p, start, finish)
1431 register struct parse *p;
1432 sopno start;                    /* from here */
1433 sopno finish;                   /* to this less one */
1434 {
1435         register sopno ret = HERE();
1436         register sopno len = finish - start;
1437
1438         assert(finish >= start);
1439         if (len == 0)
1440                 return(ret);
1441         enlarge(p, p->ssize + len);     /* this many unexpected additions */
1442         assert(p->ssize >= p->slen + len);
1443         (void) memcpy((char *)(p->strip + p->slen),
1444                 (char *)(p->strip + start), (size_t)len*sizeof(sop));
1445         p->slen += len;
1446         return(ret);
1447 }
1448
1449 /*
1450  - doemit - emit a strip operator
1451  == static void doemit(register struct parse *p, sop op, size_t opnd);
1452  *
1453  * It might seem better to implement this as a macro with a function as
1454  * hard-case backup, but it's just too big and messy unless there are
1455  * some changes to the data structures.  Maybe later.
1456  */
1457 static void
1458 doemit(p, op, opnd)
1459 register struct parse *p;
1460 sop op;
1461 size_t opnd;
1462 {
1463         /* avoid making error situations worse */
1464         if (p->error != 0)
1465                 return;
1466
1467         /* deal with oversize operands ("can't happen", more or less) */
1468         assert(opnd < 1<<OPSHIFT);
1469
1470         /* deal with undersized strip */
1471         if (p->slen >= p->ssize)
1472                 enlarge(p, (p->ssize+1) / 2 * 3);       /* +50% */
1473         assert(p->slen < p->ssize);
1474
1475         /* finally, it's all reduced to the easy case */
1476         p->strip[p->slen++] = SOP(op, opnd);
1477 }
1478
1479 /*
1480  - doinsert - insert a sop into the strip
1481  == static void doinsert(register struct parse *p, sop op, size_t opnd, sopno pos);
1482  */
1483 static void
1484 doinsert(p, op, opnd, pos)
1485 register struct parse *p;
1486 sop op;
1487 size_t opnd;
1488 sopno pos;
1489 {
1490         register sopno sn;
1491         register sop s;
1492         register int i;
1493
1494         /* avoid making error situations worse */
1495         if (p->error != 0)
1496                 return;
1497
1498         sn = HERE();
1499         EMIT(op, opnd);         /* do checks, ensure space */
1500         assert(HERE() == sn+1);
1501         s = p->strip[sn];
1502
1503         /* adjust paren pointers */
1504         assert(pos > 0);
1505         for (i = 1; i < NPAREN; i++) {
1506                 if (p->pbegin[i] >= pos) {
1507                         p->pbegin[i]++;
1508                 }
1509                 if (p->pend[i] >= pos) {
1510                         p->pend[i]++;
1511                 }
1512         }
1513
1514         memmove((char *)&p->strip[pos+1], (char *)&p->strip[pos],
1515                                                 (HERE()-pos-1)*sizeof(sop));
1516         p->strip[pos] = s;
1517 }
1518
1519 /*
1520  - dofwd - complete a forward reference
1521  == static void dofwd(register struct parse *p, sopno pos, sop value);
1522  */
1523 static void
1524 dofwd(p, pos, value)
1525 register struct parse *p;
1526 register sopno pos;
1527 sop value;
1528 {
1529         /* avoid making error situations worse */
1530         if (p->error != 0)
1531                 return;
1532
1533         assert(value < 1<<OPSHIFT);
1534         p->strip[pos] = OP(p->strip[pos]) | value;
1535 }
1536
1537 /*
1538  - enlarge - enlarge the strip
1539  == static void enlarge(register struct parse *p, sopno size);
1540  */
1541 static void
1542 enlarge(p, size)
1543 register struct parse *p;
1544 register sopno size;
1545 {
1546         register sop *sp;
1547
1548         if (p->ssize >= size)
1549                 return;
1550
1551         sp = (sop *)realloc(p->strip, size*sizeof(sop));
1552         if (sp == NULL) {
1553                 SETERROR(REG_ESPACE);
1554                 return;
1555         }
1556         p->strip = sp;
1557         p->ssize = size;
1558 }
1559
1560 /*
1561  - stripsnug - compact the strip
1562  == static void stripsnug(register struct parse *p, register struct re_guts *g);
1563  */
1564 static void
1565 stripsnug(p, g)
1566 register struct parse *p;
1567 register struct re_guts *g;
1568 {
1569         g->nstates = p->slen;
1570         g->strip = (sop *)realloc((char *)p->strip, p->slen * sizeof(sop));
1571         if (g->strip == NULL) {
1572                 SETERROR(REG_ESPACE);
1573                 g->strip = p->strip;
1574         }
1575 }
1576
1577 /*
1578  - findmust - fill in must and mlen with longest mandatory literal string
1579  == static void findmust(register struct parse *p, register struct re_guts *g);
1580  *
1581  * This algorithm could do fancy things like analyzing the operands of |
1582  * for common subsequences.  Someday.  This code is simple and finds most
1583  * of the interesting cases.
1584  *
1585  * Note that must and mlen got initialized during setup.
1586  */
1587 static void
1588 findmust(p, g)
1589 struct parse *p;
1590 register struct re_guts *g;
1591 {
1592         register sop *scan;
1593         sop *start;
1594         register sop *newstart;
1595         register sopno newlen;
1596         register sop s;
1597         register char *cp;
1598         register sopno i;
1599
1600         /* avoid making error situations worse */
1601         if (p->error != 0)
1602                 return;
1603
1604         /* find the longest OCHAR sequence in strip */
1605         newlen = 0;
1606         scan = g->strip + 1;
1607         do {
1608                 s = *scan++;
1609                 switch (OP(s)) {
1610                 case OCHAR:             /* sequence member */
1611                         if (newlen == 0)                /* new sequence */
1612                                 newstart = scan - 1;
1613                         newlen++;
1614                         break;
1615                 case OPLUS_:            /* things that don't break one */
1616                 case OLPAREN:
1617                 case ORPAREN:
1618                         break;
1619                 case OQUEST_:           /* things that must be skipped */
1620                 case OCH_:
1621                         scan--;
1622                         do {
1623                                 scan += OPND(s);
1624                                 s = *scan;
1625                                 /* assert() interferes w debug printouts */
1626                                 if (OP(s) != O_QUEST && OP(s) != O_CH &&
1627                                                         OP(s) != OOR2) {
1628                                         g->iflags |= BAD;
1629                                         return;
1630                                 }
1631                         } while (OP(s) != O_QUEST && OP(s) != O_CH);
1632                         /* fallthrough */
1633                 default:                /* things that break a sequence */
1634                         if (newlen > g->mlen) {         /* ends one */
1635                                 start = newstart;
1636                                 g->mlen = newlen;
1637                         }
1638                         newlen = 0;
1639                         break;
1640                 }
1641         } while (OP(s) != OEND);
1642
1643         if (g->mlen == 0)               /* there isn't one */
1644                 return;
1645
1646         /* turn it into a character string */
1647         g->must = malloc((size_t)g->mlen + 1);
1648         if (g->must == NULL) {          /* argh; just forget it */
1649                 g->mlen = 0;
1650                 return;
1651         }
1652         cp = g->must;
1653         scan = start;
1654         for (i = g->mlen; i > 0; i--) {
1655                 while (OP(s = *scan++) != OCHAR)
1656                         continue;
1657                 assert(cp < g->must + g->mlen);
1658                 *cp++ = (char)OPND(s);
1659         }
1660         assert(cp == g->must + g->mlen);
1661         *cp++ = '\0';           /* just on general principles */
1662 }
1663
1664 /*
1665  - pluscount - count + nesting
1666  == static sopno pluscount(register struct parse *p, register struct re_guts *g);
1667  */
1668 static sopno                    /* nesting depth */
1669 pluscount(p, g)
1670 struct parse *p;
1671 register struct re_guts *g;
1672 {
1673         register sop *scan;
1674         register sop s;
1675         register sopno plusnest = 0;
1676         register sopno maxnest = 0;
1677
1678         if (p->error != 0)
1679                 return(0);      /* there may not be an OEND */
1680
1681         scan = g->strip + 1;
1682         do {
1683                 s = *scan++;
1684                 switch (OP(s)) {
1685                 case OPLUS_:
1686                         plusnest++;
1687                         break;
1688                 case O_PLUS:
1689                         if (plusnest > maxnest)
1690                                 maxnest = plusnest;
1691                         plusnest--;
1692                         break;
1693                 }
1694         } while (OP(s) != OEND);
1695         if (plusnest != 0)
1696                 g->iflags |= BAD;
1697         return(maxnest);
1698 }