lib/Support/regcomp.c

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