lib/libc/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.
   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   8.5 (Berkeley) 3/20/94
  38  */
  39
  40 #if defined(LIBC_SCCS) && !defined(lint)
  41 static char sccsid[] = "@(#)regcomp.c   8.5 (Berkeley) 3/20/94";
  42 #endif /* LIBC_SCCS and not lint */
  43
  44 #include <sys/types.h>
  45 #include <stdio.h>
  46 #include <string.h>
  47 #ifdef __minix_vmd
  48 #include <bsd/asciictype.h>
  49 #else
  50 #include <ctype.h>
  51 #endif
  52 #include <limits.h>
  53 #include <stdlib.h>
  54 #include <regex.h>
  55
  56 #include "utils.h"
  57 #include "regex2.h"
  58
  59 #include "cclass.h"
  60 #include "cname.h"
  61
  62 /*
  63  * parse structure, passed up and down to avoid global variables and
  64  * other clumsinesses
  65  */
  66 struct parse {
  67         char *next;             /* next character in RE */
  68         char *end;              /* end of string (-> NUL normally) */
  69         int error;              /* has an error been seen? */
  70         sop *strip;             /* malloced strip */
  71         sopno ssize;            /* malloced strip size (allocated) */
  72         sopno slen;             /* malloced strip length (used) */
  73         int ncsalloc;           /* number of csets allocated */
  74         struct re_guts *g;
  75 #       define  NPAREN  10      /* we need to remember () 1-9 for back refs */
  76         sopno pbegin[NPAREN];   /* -> ( ([0] unused) */
  77         sopno pend[NPAREN];     /* -> ) ([0] unused) */
  78 };
  79
  80 /* ========= begin header generated by ./mkh ========= */
  81 #ifdef __cplusplus
  82 extern "C" {
  83 #endif
  84
  85 /* === regcomp.c === */
  86 static void p_ere(struct parse *p, int stop);
  87 static void p_ere_exp(struct parse *p);
  88 static void p_str(struct parse *p);
  89 static void p_bre(struct parse *p, int end1, int end2);
  90 static int p_simp_re(struct parse *p, int starordinary);
  91 static int p_count(struct parse *p);
  92 static void p_bracket(struct parse *p);
  93 static void p_b_term(struct parse *p, cset *cs);
  94 static void p_b_cclass(struct parse *p, cset *cs);
  95 static void p_b_eclass(struct parse *p, cset *cs);
  96 static char p_b_symbol(struct parse *p);
  97 static char p_b_coll_elem(struct parse *p, int endc);
  98 static char othercase(int ch);
  99 static void bothcases(struct parse *p, int ch);
 100 static void ordinary(struct parse *p, int ch);
 101 static void nonnewline(struct parse *p);
 102 static void repeat(struct parse *p, sopno start, int from, int to);
 103 static int seterr(struct parse *p, int e);
 104 static cset *allocset(struct parse *p);
 105 static void freeset(struct parse *p, cset *cs);
 106 static int freezeset(struct parse *p, cset *cs);
 107 static int firstch(struct parse *p, cset *cs);
 108 static int nch(struct parse *p, cset *cs);
 109 static void mcadd(struct parse *p, cset *cs, char *cp);
 110 static void mcsub(cset *cs, char *cp);
 111 static int mcin(cset *cs, char *cp);
 112 static char *mcfind(cset *cs, char *cp);
 113 static void mcinvert(struct parse *p, cset *cs);
 114 static void mccase(struct parse *p, cset *cs);
 115 static int isinsets(struct re_guts *g, int c);
 116 static int samesets(struct re_guts *g, int c1, int c2);
 117 static void categorize(struct parse *p, struct re_guts *g);
 118 static sopno dupl(struct parse *p, sopno start, sopno finish);
 119 static void doemit(struct parse *p, sop op, size_t opnd);
 120 static void doinsert(struct parse *p, sop op, size_t opnd, sopno pos);
 121 static void dofwd(struct parse *p, sopno pos, sop value);
 122 static void enlarge(struct parse *p, sopno size);
 123 static void stripsnug(struct parse *p, struct re_guts *g);
 124 static void findmust(struct parse *p, struct re_guts *g);
 125 static sopno pluscount(struct parse *p, struct re_guts *g);
 126
 127 #ifdef __cplusplus
 128 }
 129 #endif
 130 /* ========= end header generated by ./mkh ========= */
 131
 132 static char nuls[10];           /* place to point scanner in event of error */
 133
 134 /*
 135  * macros for use with parse structure
 136  * BEWARE:  these know that the parse structure is named `p' !!!
 137  */
 138 #define PEEK()  (*p->next)
 139 #define PEEK2() (*(p->next+1))
 140 #define MORE()  (p->next < p->end)
 141 #define MORE2() (p->next+1 < p->end)
 142 #define SEE(c)  (MORE() && PEEK() == (c))
 143 #define SEETWO(a, b)    (MORE() && MORE2() && PEEK() == (a) && PEEK2() == (b))
 144 #define EAT(c)  ((SEE(c)) ? (NEXT(), 1) : 0)
 145 #define EATTWO(a, b)    ((SEETWO(a, b)) ? (NEXT2(), 1) : 0)
 146 #define NEXT()  (p->next++)
 147 #define NEXT2() (p->next += 2)
 148 #define NEXTn(n)        (p->next += (n))
 149 #define GETNEXT()       (*p->next++)
 150 #define SETERROR(e)     seterr(p, (e))
 151 #define REQUIRE(co, e)  ((co) || SETERROR(e))
 152 #define MUSTSEE(c, e)   (REQUIRE(MORE() && PEEK() == (c), e))
 153 #define MUSTEAT(c, e)   (REQUIRE(MORE() && GETNEXT() == (c), e))
 154 #define MUSTNOTSEE(c, e)        (REQUIRE(!MORE() || PEEK() != (c), e))
 155 #define EMIT(op, sopnd) doemit(p, (sop)(op), (size_t)(sopnd))
 156 #define INSERT(op, pos) doinsert(p, (sop)(op), HERE()-(pos)+1, pos)
 157 #define AHEAD(pos)              dofwd(p, pos, HERE()-(pos))
 158 #define ASTERN(sop, pos)        EMIT(sop, HERE()-pos)
 159 #define HERE()          (p->slen)
 160 #define THERE()         (p->slen - 1)
 161 #define THERETHERE()    (p->slen - 2)
 162 #define DROP(n) (p->slen -= (n))
 163
 164 #ifndef NDEBUG
 165 static int never = 0;           /* for use in asserts; shuts lint up */
 166 #else
 167 #define never   0               /* some <assert.h>s have bugs too */
 168 #endif
 169
 170 /*
 171  - regcomp - interface for parser and compilation
 172  = extern int regcomp(regex_t *, const char *, int);
 173  = #define      REG_BASIC       0000
 174  = #define      REG_EXTENDED    0001
 175  = #define      REG_ICASE       0002
 176  = #define      REG_NOSUB       0004
 177  = #define      REG_NEWLINE     0010
 178  = #define      REG_NOSPEC      0020
 179  = #define      REG_PEND        0040
 180  = #define      REG_DUMP        0200
 181  */
 182 int                             /* 0 success, otherwise REG_something */
 183 regcomp(preg, pattern, cflags)
 184 regex_t *preg;
 185 const char *pattern;
 186 int cflags;
 187 {
 188         struct parse pa;
 189         register struct re_guts *g;
 190         register struct parse *p = &pa;
 191         register int i;
 192         register size_t len;
 193 #ifdef REDEBUG
 194 #       define  GOODFLAGS(f)    (f)
 195 #else
 196 #       define  GOODFLAGS(f)    ((f)&~REG_DUMP)
 197 #endif
 198
 199         cflags = GOODFLAGS(cflags);
 200         if ((cflags&REG_EXTENDED) && (cflags&REG_NOSPEC))
 201                 return(REG_INVARG);
 202
 203         if (cflags&REG_PEND) {
 204                 if (preg->re_endp < pattern)
 205                         return(REG_INVARG);
 206                 len = preg->re_endp - pattern;
 207         } else
 208                 len = strlen((char *)pattern);
 209
 210         /* do the mallocs early so failure handling is easy */
 211         g = (struct re_guts *)malloc(sizeof(struct re_guts) +
 212                                                         (NC-1)*sizeof(cat_t));
 213         if (g == NULL)
 214                 return(REG_ESPACE);
 215         p->ssize = len/(size_t)2*(size_t)3 + (size_t)1; /* ugh */
 216         p->strip = (sop *)malloc(p->ssize * sizeof(sop));
 217         p->slen = 0;
 218         if (p->strip == NULL) {
 219                 free((char *)g);
 220                 return(REG_ESPACE);
 221         }
 222
 223         /* set things up */
 224         p->g = g;
 225         p->next = (char *)pattern;      /* convenience; we do not modify it */
 226         p->end = p->next + len;
 227         p->error = 0;
 228         p->ncsalloc = 0;
 229         for (i = 0; i < NPAREN; i++) {
 230                 p->pbegin[i] = 0;
 231                 p->pend[i] = 0;
 232         }
 233         g->csetsize = NC;
 234         g->sets = NULL;
 235         g->setbits = NULL;
 236         g->ncsets = 0;
 237         g->cflags = cflags;
 238         g->iflags = 0;
 239         g->nbol = 0;
 240         g->neol = 0;
 241         g->must = NULL;
 242         g->mlen = 0;
 243         g->nsub = 0;
 244         g->ncategories = 1;     /* category 0 is "everything else" */
 245         g->categories = &g->catspace[-(CHAR_MIN)];
 246         (void) memset((char *)g->catspace, 0, NC*sizeof(cat_t));
 247         g->backrefs = 0;
 248
 249         /* do it */
 250         EMIT(OEND, 0);
 251         g->firststate = THERE();
 252         if (cflags&REG_EXTENDED)
 253                 p_ere(p, OUT);
 254         else if (cflags&REG_NOSPEC)
 255                 p_str(p);
 256         else
 257                 p_bre(p, OUT, OUT);
 258         EMIT(OEND, 0);
 259         g->laststate = THERE();
 260
 261         /* tidy up loose ends and fill things in */
 262         categorize(p, g);
 263         stripsnug(p, g);
 264         findmust(p, g);
 265         g->nplus = pluscount(p, g);
 266         g->magic = MAGIC2;
 267         preg->re_nsub = g->nsub;
 268         preg->re_g = g;
 269         preg->re_magic = MAGIC1;
 270 #ifndef REDEBUG
 271         /* not debugging, so can't rely on the assert() in regexec() */
 272         if (g->iflags&BAD)
 273                 SETERROR(REG_ASSERT);
 274 #endif
 275
 276         /* win or lose, we're done */
 277         if (p->error != 0)      /* lose */
 278                 regfree(preg);
 279         return(p->error);
 280 }
 281
 282 /*
 283  - p_ere - ERE parser top level, concatenation and alternation
 284  == static void p_ere(register struct parse *p, int stop);
 285  */
 286 static void
 287 p_ere(p, stop)
 288 register struct parse *p;
 289 int stop;                       /* character this ERE should end at */
 290 {
 291         register char c;
 292         register sopno prevback;
 293         register sopno prevfwd;
 294         register sopno conc;
 295         register int first = 1;         /* is this the first alternative? */
 296
 297         for (;;) {
 298                 /* do a bunch of concatenated expressions */
 299                 conc = HERE();
 300                 while (MORE() && (c = PEEK()) != '|' && c != stop)
 301                         p_ere_exp(p);
 302                 REQUIRE(HERE() != conc, REG_EMPTY);     /* require nonempty */
 303
 304                 if (!EAT('|'))
 305                         break;          /* NOTE BREAK OUT */
 306
 307                 if (first) {
 308                         INSERT(OCH_, conc);     /* offset is wrong */
 309                         prevfwd = conc;
 310                         prevback = conc;
 311                         first = 0;
 312                 }
 313                 ASTERN(OOR1, prevback);
 314                 prevback = THERE();
 315                 AHEAD(prevfwd);                 /* fix previous offset */
 316                 prevfwd = HERE();
 317                 EMIT(OOR2, 0);                  /* offset is very wrong */
 318         }
 319
 320         if (!first) {           /* tail-end fixups */
 321                 AHEAD(prevfwd);
 322                 ASTERN(O_CH, prevback);
 323         }
 324
 325         assert(!MORE() || SEE(stop));
 326 }
 327
 328 /*
 329  - p_ere_exp - parse one subERE, an atom possibly followed by a repetition op
 330  == static void p_ere_exp(register struct parse *p);
 331  */
 332 static void
 333 p_ere_exp(p)
 334 register struct parse *p;
 335 {
 336         register char c;
 337         register sopno pos;
 338         register int count;
 339         register int count2;
 340         register sopno subno;
 341         int wascaret = 0;
 342
 343         assert(MORE());         /* caller should have ensured this */
 344         c = GETNEXT();
 345
 346         pos = HERE();
 347         switch (c) {
 348         case '(':
 349                 REQUIRE(MORE(), REG_EPAREN);
 350                 p->g->nsub++;
 351                 subno = p->g->nsub;
 352                 if (subno < NPAREN)
 353                         p->pbegin[subno] = HERE();
 354                 EMIT(OLPAREN, subno);
 355                 if (!SEE(')'))
 356                         p_ere(p, ')');
 357                 if (subno < NPAREN) {
 358                         p->pend[subno] = HERE();
 359                         assert(p->pend[subno] != 0);
 360                 }
 361                 EMIT(ORPAREN, subno);
 362                 MUSTEAT(')', REG_EPAREN);
 363                 break;
 364 #ifndef POSIX_MISTAKE
 365         case ')':               /* happens only if no current unmatched ( */
 366                 /*
 367                  * You may ask, why the ifndef?  Because I didn't notice
 368                  * this until slightly too late for 1003.2, and none of the
 369                  * other 1003.2 regular-expression reviewers noticed it at
 370                  * all.  So an unmatched ) is legal POSIX, at least until
 371                  * we can get it fixed.
 372                  */
 373                 SETERROR(REG_EPAREN);
 374                 break;
 375 #endif
 376         case '^':
 377                 EMIT(OBOL, 0);
 378                 p->g->iflags |= USEBOL;
 379                 p->g->nbol++;
 380                 wascaret = 1;
 381                 break;
 382         case '$':
 383                 EMIT(OEOL, 0);
 384                 p->g->iflags |= USEEOL;
 385                 p->g->neol++;
 386                 break;
 387         case '|':
 388                 SETERROR(REG_EMPTY);
 389                 break;
 390         case '*':
 391         case '+':
 392         case '?':
 393                 SETERROR(REG_BADRPT);
 394                 break;
 395         case '.':
 396                 if (p->g->cflags&REG_NEWLINE)
 397                         nonnewline(p);
 398                 else
 399                         EMIT(OANY, 0);
 400                 break;
 401         case '[':
 402                 p_bracket(p);
 403                 break;
 404         case '\\':
 405                 REQUIRE(MORE(), REG_EESCAPE);
 406                 c = GETNEXT();
 407                 ordinary(p, c);
 408                 break;
 409         case '{':               /* okay as ordinary except if digit follows */
 410                 REQUIRE(!MORE() || !isdigit(PEEK()), REG_BADRPT);
 411                 /* FALLTHROUGH */
 412         default:
 413                 ordinary(p, c);
 414                 break;
 415         }
 416
 417         if (!MORE())
 418                 return;
 419         c = PEEK();
 420         /* we call { a repetition if followed by a digit */
 421         if (!( c == '*' || c == '+' || c == '?' ||
 422                                 (c == '{' && MORE2() && isdigit(PEEK2())) ))
 423                 return;         /* no repetition, we're done */
 424         NEXT();
 425
 426         REQUIRE(!wascaret, REG_BADRPT);
 427         switch (c) {
 428         case '*':       /* implemented as +? */
 429                 /* this case does not require the (y|) trick, noKLUDGE */
 430                 INSERT(OPLUS_, pos);
 431                 ASTERN(O_PLUS, pos);
 432                 INSERT(OQUEST_, pos);
 433                 ASTERN(O_QUEST, pos);
 434                 break;
 435         case '+':
 436                 INSERT(OPLUS_, pos);
 437                 ASTERN(O_PLUS, pos);
 438                 break;
 439         case '?':
 440                 /* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
 441                 INSERT(OCH_, pos);              /* offset slightly wrong */
 442                 ASTERN(OOR1, pos);              /* this one's right */
 443                 AHEAD(pos);                     /* fix the OCH_ */
 444                 EMIT(OOR2, 0);                  /* offset very wrong... */
 445                 AHEAD(THERE());                 /* ...so fix it */
 446                 ASTERN(O_CH, THERETHERE());
 447                 break;
 448         case '{':
 449                 count = p_count(p);
 450                 if (EAT(',')) {
 451                         if (isdigit(PEEK())) {
 452                                 count2 = p_count(p);
 453                                 REQUIRE(count <= count2, REG_BADBR);
 454                         } else          /* single number with comma */
 455                                 count2 = INFINITY;
 456                 } else          /* just a single number */
 457                         count2 = count;
 458                 repeat(p, pos, count, count2);
 459                 if (!EAT('}')) {        /* error heuristics */
 460                         while (MORE() && PEEK() != '}')
 461                                 NEXT();
 462                         REQUIRE(MORE(), REG_EBRACE);
 463                         SETERROR(REG_BADBR);
 464                 }
 465                 break;
 466         }
 467
 468         if (!MORE())
 469                 return;
 470         c = PEEK();
 471         if (!( c == '*' || c == '+' || c == '?' ||
 472                                 (c == '{' && MORE2() && isdigit(PEEK2())) ) )
 473                 return;
 474         SETERROR(REG_BADRPT);
 475 }
 476
 477 /*
 478  - p_str - string (no metacharacters) "parser"
 479  == static void p_str(register struct parse *p);
 480  */
 481 static void
 482 p_str(p)
 483 register struct parse *p;
 484 {
 485         REQUIRE(MORE(), REG_EMPTY);
 486         while (MORE())
 487                 ordinary(p, GETNEXT());
 488 }
 489
 490 /*
 491  - p_bre - BRE parser top level, anchoring and concatenation
 492  == static void p_bre(register struct parse *p, register int end1, \
 493  ==     register int end2);
 494  * Giving end1 as OUT essentially eliminates the end1/end2 check.
 495  *
 496  * This implementation is a bit of a kludge, in that a trailing $ is first
 497  * taken as an ordinary character and then revised to be an anchor.  The
 498  * only undesirable side effect is that '$' gets included as a character
 499  * category in such cases.  This is fairly harmless; not worth fixing.
 500  * The amount of lookahead needed to avoid this kludge is excessive.
 501  */
 502 static void
 503 p_bre(p, end1, end2)
 504 register struct parse *p;
 505 register int end1;              /* first terminating character */
 506 register int end2;              /* second terminating character */
 507 {
 508         register sopno start = HERE();
 509         register int first = 1;                 /* first subexpression? */
 510         register int wasdollar = 0;
 511
 512         if (EAT('^')) {
 513                 EMIT(OBOL, 0);
 514                 p->g->iflags |= USEBOL;
 515                 p->g->nbol++;
 516         }
 517         while (MORE() && !SEETWO(end1, end2)) {
 518                 wasdollar = p_simp_re(p, first);
 519                 first = 0;
 520         }
 521         if (wasdollar) {        /* oops, that was a trailing anchor */
 522                 DROP(1);
 523                 EMIT(OEOL, 0);
 524                 p->g->iflags |= USEEOL;
 525                 p->g->neol++;
 526         }
 527
 528         REQUIRE(HERE() != start, REG_EMPTY);    /* require nonempty */
 529 }
 530
 531 /*
 532  - p_simp_re - parse a simple RE, an atom possibly followed by a repetition
 533  == static int p_simp_re(register struct parse *p, int starordinary);
 534  */
 535 static int                      /* was the simple RE an unbackslashed $? */
 536 p_simp_re(p, starordinary)
 537 register struct parse *p;
 538 int starordinary;               /* is a leading * an ordinary character? */
 539 {
 540         register int c;
 541         register int count;
 542         register int count2;
 543         register sopno pos;
 544         register int i;
 545         register sopno subno;
 546 #       define  BACKSL  (1<<CHAR_BIT)
 547
 548         pos = HERE();           /* repetion op, if any, covers from here */
 549
 550         assert(MORE());         /* caller should have ensured this */
 551         c = GETNEXT();
 552         if (c == '\\') {
 553                 REQUIRE(MORE(), REG_EESCAPE);
 554                 c = BACKSL | (unsigned char)GETNEXT();
 555         }
 556         switch (c) {
 557         case '.':
 558                 if (p->g->cflags&REG_NEWLINE)
 559                         nonnewline(p);
 560                 else
 561                         EMIT(OANY, 0);
 562                 break;
 563         case '[':
 564                 p_bracket(p);
 565                 break;
 566         case BACKSL|'{':
 567                 SETERROR(REG_BADRPT);
 568                 break;
 569         case BACKSL|'(':
 570                 p->g->nsub++;
 571                 subno = p->g->nsub;
 572                 if (subno < NPAREN)
 573                         p->pbegin[subno] = HERE();
 574                 EMIT(OLPAREN, subno);
 575                 /* the MORE here is an error heuristic */
 576                 if (MORE() && !SEETWO('\\', ')'))
 577                         p_bre(p, '\\', ')');
 578                 if (subno < NPAREN) {
 579                         p->pend[subno] = HERE();
 580                         assert(p->pend[subno] != 0);
 581                 }
 582                 EMIT(ORPAREN, subno);
 583                 REQUIRE(EATTWO('\\', ')'), REG_EPAREN);
 584                 break;
 585         case BACKSL|')':        /* should not get here -- must be user */
 586         case BACKSL|'}':
 587                 SETERROR(REG_EPAREN);
 588                 break;
 589         case BACKSL|'1':
 590         case BACKSL|'2':
 591         case BACKSL|'3':
 592         case BACKSL|'4':
 593         case BACKSL|'5':
 594         case BACKSL|'6':
 595         case BACKSL|'7':
 596         case BACKSL|'8':
 597         case BACKSL|'9':
 598                 i = (c&~BACKSL) - '0';
 599                 assert(i < NPAREN);
 600                 if (p->pend[i] != 0) {
 601                         assert(i <= p->g->nsub);
 602                         EMIT(OBACK_, i);
 603                         assert(p->pbegin[i] != 0);
 604                         assert(OP(p->strip[p->pbegin[i]]) == OLPAREN);
 605                         assert(OP(p->strip[p->pend[i]]) == ORPAREN);
 606                         (void) dupl(p, p->pbegin[i]+1, p->pend[i]);
 607                         EMIT(O_BACK, i);
 608                 } else
 609                         SETERROR(REG_ESUBREG);
 610                 p->g->backrefs = 1;
 611                 break;
 612         case '*':
 613                 REQUIRE(starordinary, REG_BADRPT);
 614                 /* FALLTHROUGH */
 615         default:
 616                 ordinary(p, c &~ BACKSL);
 617                 break;
 618         }
 619
 620         if (EAT('*')) {         /* implemented as +? */
 621                 /* this case does not require the (y|) trick, noKLUDGE */
 622                 INSERT(OPLUS_, pos);
 623                 ASTERN(O_PLUS, pos);
 624                 INSERT(OQUEST_, pos);
 625                 ASTERN(O_QUEST, pos);
 626         } else if (EATTWO('\\', '{')) {
 627                 count = p_count(p);
 628                 if (EAT(',')) {
 629                         if (MORE() && isdigit(PEEK())) {
 630                                 count2 = p_count(p);
 631                                 REQUIRE(count <= count2, REG_BADBR);
 632                         } else          /* single number with comma */
 633                                 count2 = INFINITY;
 634                 } else          /* just a single number */
 635                         count2 = count;
 636                 repeat(p, pos, count, count2);
 637                 if (!EATTWO('\\', '}')) {       /* error heuristics */
 638                         while (MORE() && !SEETWO('\\', '}'))
 639                                 NEXT();
 640                         REQUIRE(MORE(), REG_EBRACE);
 641                         SETERROR(REG_BADBR);
 642                 }
 643         } else if (c == (unsigned char)'$')     /* $ (but not \$) ends it */
 644                 return(1);
 645
 646         return(0);
 647 }
 648
 649 /*
 650  - p_count - parse a repetition count
 651  == static int p_count(register struct parse *p);
 652  */
 653 static int                      /* the value */
 654 p_count(p)
 655 register struct parse *p;
 656 {
 657         register int count = 0;
 658         register int ndigits = 0;
 659
 660         while (MORE() && isdigit(PEEK()) && count <= DUPMAX) {
 661                 count = count*10 + (GETNEXT() - '0');
 662                 ndigits++;
 663         }
 664
 665         REQUIRE(ndigits > 0 && count <= DUPMAX, REG_BADBR);
 666         return(count);
 667 }
 668
 669 /*
 670  - p_bracket - parse a bracketed character list
 671  == static void p_bracket(register struct parse *p);
 672  *
 673  * Note a significant property of this code:  if the allocset() did SETERROR,
 674  * no set operations are done.
 675  */
 676 static void
 677 p_bracket(p)
 678 register struct parse *p;
 679 {
 680         register char c;
 681         register cset *cs = allocset(p);
 682         register int invert = 0;
 683
 684         /* Dept of Truly Sickening Special-Case Kludges */
 685         if (p->next + 5 < p->end && strncmp(p->next, "[:<:]]", 6) == 0) {
 686                 EMIT(OBOW, 0);
 687                 NEXTn(6);
 688                 return;
 689         }
 690         if (p->next + 5 < p->end && strncmp(p->next, "[:>:]]", 6) == 0) {
 691                 EMIT(OEOW, 0);
 692                 NEXTn(6);
 693                 return;
 694         }
 695
 696         if (EAT('^'))
 697                 invert++;       /* make note to invert set at end */
 698         if (EAT(']'))
 699                 CHadd(cs, ']');
 700         else if (EAT('-'))
 701                 CHadd(cs, '-');
 702         while (MORE() && PEEK() != ']' && !SEETWO('-', ']'))
 703                 p_b_term(p, cs);
 704         if (EAT('-'))
 705                 CHadd(cs, '-');
 706         MUSTEAT(']', REG_EBRACK);
 707
 708         if (p->error != 0)      /* don't mess things up further */
 709                 return;
 710
 711         if (p->g->cflags&REG_ICASE) {
 712                 register int i;
 713                 register int ci;
 714
 715                 for (i = p->g->csetsize - 1; i >= 0; i--)
 716                         if (CHIN(cs, i) && isalpha(i)) {
 717                                 ci = othercase(i);
 718                                 if (ci != i)
 719                                         CHadd(cs, ci);
 720                         }
 721                 if (cs->multis != NULL)
 722                         mccase(p, cs);
 723         }
 724         if (invert) {
 725                 register int i;
 726
 727                 for (i = p->g->csetsize - 1; i >= 0; i--)
 728                         if (CHIN(cs, i))
 729                                 CHsub(cs, i);
 730                         else
 731                                 CHadd(cs, i);
 732                 if (p->g->cflags&REG_NEWLINE)
 733                         CHsub(cs, '\n');
 734                 if (cs->multis != NULL)
 735                         mcinvert(p, cs);
 736         }
 737
 738         assert(cs->multis == NULL);             /* xxx */
 739
 740         if (nch(p, cs) == 1) {          /* optimize singleton sets */
 741                 ordinary(p, firstch(p, cs));
 742                 freeset(p, cs);
 743         } else
 744                 EMIT(OANYOF, freezeset(p, cs));
 745 }
 746
 747 /*
 748  - p_b_term - parse one term of a bracketed character list
 749  == static void p_b_term(register struct parse *p, register cset *cs);
 750  */
 751 static void
 752 p_b_term(p, cs)
 753 register struct parse *p;
 754 register cset *cs;
 755 {
 756         register char c;
 757         register char start, finish;
 758         register int i;
 759
 760         /* classify what we've got */
 761         switch ((MORE()) ? PEEK() : '\0') {
 762         case '[':
 763                 c = (MORE2()) ? PEEK2() : '\0';
 764                 break;
 765         case '-':
 766                 SETERROR(REG_ERANGE);
 767                 return;                 /* NOTE RETURN */
 768                 break;
 769         default:
 770                 c = '\0';
 771                 break;
 772         }
 773
 774         switch (c) {
 775         case ':':               /* character class */
 776                 NEXT2();
 777                 REQUIRE(MORE(), REG_EBRACK);
 778                 c = PEEK();
 779                 REQUIRE(c != '-' && c != ']', REG_ECTYPE);
 780                 p_b_cclass(p, cs);
 781                 REQUIRE(MORE(), REG_EBRACK);
 782                 REQUIRE(EATTWO(':', ']'), REG_ECTYPE);
 783                 break;
 784         case '=':               /* equivalence class */
 785                 NEXT2();
 786                 REQUIRE(MORE(), REG_EBRACK);
 787                 c = PEEK();
 788                 REQUIRE(c != '-' && c != ']', REG_ECOLLATE);
 789                 p_b_eclass(p, cs);
 790                 REQUIRE(MORE(), REG_EBRACK);
 791                 REQUIRE(EATTWO('=', ']'), REG_ECOLLATE);
 792                 break;
 793         default:                /* symbol, ordinary character, or range */
 794 /* xxx revision needed for multichar stuff */
 795                 start = p_b_symbol(p);
 796                 if (SEE('-') && MORE2() && PEEK2() != ']') {
 797                         /* range */
 798                         NEXT();
 799                         if (EAT('-'))
 800                                 finish = '-';
 801                         else
 802                                 finish = p_b_symbol(p);
 803                 } else
 804                         finish = start;
 805 /* xxx what about signed chars here... */
 806                 REQUIRE(start <= finish, REG_ERANGE);
 807                 for (i = start; i <= finish; i++)
 808                         CHadd(cs, i);
 809                 break;
 810         }
 811 }
 812
 813 /*
 814  - p_b_cclass - parse a character-class name and deal with it
 815  == static void p_b_cclass(register struct parse *p, register cset *cs);
 816  */
 817 static void
 818 p_b_cclass(p, cs)
 819 register struct parse *p;
 820 register cset *cs;
 821 {
 822         register char *sp = p->next;
 823         register struct cclass *cp;
 824         register size_t len;
 825         register char *u;
 826         register char c;
 827
 828         while (MORE() && isalpha(PEEK()))
 829                 NEXT();
 830         len = p->next - sp;
 831         for (cp = cclasses; cp->name != NULL; cp++)
 832                 if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
 833                         break;
 834         if (cp->name == NULL) {
 835                 /* oops, didn't find it */
 836                 SETERROR(REG_ECTYPE);
 837                 return;
 838         }
 839
 840         u = cp->chars;
 841         while ((c = *u++) != '\0')
 842                 CHadd(cs, c);
 843         for (u = cp->multis; *u != '\0'; u += strlen(u) + 1)
 844                 MCadd(p, cs, u);
 845 }
 846
 847 /*
 848  - p_b_eclass - parse an equivalence-class name and deal with it
 849  == static void p_b_eclass(register struct parse *p, register cset *cs);
 850  *
 851  * This implementation is incomplete. xxx
 852  */
 853 static void
 854 p_b_eclass(p, cs)
 855 register struct parse *p;
 856 register cset *cs;
 857 {
 858         register char c;
 859
 860         c = p_b_coll_elem(p, '=');
 861         CHadd(cs, c);
 862 }
 863
 864 /*
 865  - p_b_symbol - parse a character or [..]ed multicharacter collating symbol
 866  == static char p_b_symbol(register struct parse *p);
 867  */
 868 static char                     /* value of symbol */
 869 p_b_symbol(p)
 870 register struct parse *p;
 871 {
 872         register char value;
 873
 874         REQUIRE(MORE(), REG_EBRACK);
 875         if (!EATTWO('[', '.'))
 876                 return(GETNEXT());
 877
 878         /* collating symbol */
 879         value = p_b_coll_elem(p, '.');
 880         REQUIRE(EATTWO('.', ']'), REG_ECOLLATE);
 881         return(value);
 882 }
 883
 884 /*
 885  - p_b_coll_elem - parse a collating-element name and look it up
 886  == static char p_b_coll_elem(register struct parse *p, int endc);
 887  */
 888 static char                     /* value of collating element */
 889 p_b_coll_elem(p, endc)
 890 register struct parse *p;
 891 int endc;                       /* name ended by endc,']' */
 892 {
 893         register char *sp = p->next;
 894         register struct cname *cp;
 895         register int len;
 896         register char c;
 897
 898         while (MORE() && !SEETWO(endc, ']'))
 899                 NEXT();
 900         if (!MORE()) {
 901                 SETERROR(REG_EBRACK);
 902                 return(0);
 903         }
 904         len = p->next - sp;
 905         for (cp = cnames; cp->name != NULL; cp++)
 906                 if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
 907                         return(cp->code);       /* known name */
 908         if (len == 1)
 909                 return(*sp);    /* single character */
 910         SETERROR(REG_ECOLLATE);                 /* neither */
 911         return(0);
 912 }
 913
 914 /*
 915  - othercase - return the case counterpart of an alphabetic
 916  == static char othercase(int ch);
 917  */
 918 static char                     /* if no counterpart, return ch */
 919 othercase(ch)
 920 int ch;
 921 {
 922         assert(isalpha(ch));
 923         if (isupper(ch))
 924                 return(tolower(ch));
 925         else if (islower(ch))
 926                 return(toupper(ch));
 927         else                    /* peculiar, but could happen */
 928                 return(ch);
 929 }
 930
 931 /*
 932  - bothcases - emit a dualcase version of a two-case character
 933  == static void bothcases(register struct parse *p, int ch);
 934  *
 935  * Boy, is this implementation ever a kludge...
 936  */
 937 static void
 938 bothcases(p, ch)
 939 register struct parse *p;
 940 int ch;
 941 {
 942         register char *oldnext = p->next;
 943         register char *oldend = p->end;
 944         char bracket[3];
 945
 946         assert(othercase(ch) != ch);    /* p_bracket() would recurse */
 947         p->next = bracket;
 948         p->end = bracket+2;
 949         bracket[0] = ch;
 950         bracket[1] = ']';
 951         bracket[2] = '\0';
 952         p_bracket(p);
 953         assert(p->next == bracket+2);
 954         p->next = oldnext;
 955         p->end = oldend;
 956 }
 957
 958 /*
 959  - ordinary - emit an ordinary character
 960  == static void ordinary(register struct parse *p, register int ch);
 961  */
 962 static void
 963 ordinary(p, ch)
 964 register struct parse *p;
 965 register int ch;
 966 {
 967         register cat_t *cap = p->g->categories;
 968
 969         if ((p->g->cflags&REG_ICASE) && isalpha(ch) && othercase(ch) != ch)
 970                 bothcases(p, ch);
 971         else {
 972                 EMIT(OCHAR, (unsigned char)ch);
 973                 if (cap[ch] == 0)
 974                         cap[ch] = p->g->ncategories++;
 975         }
 976 }
 977
 978 /*
 979  - nonnewline - emit REG_NEWLINE version of OANY
 980  == static void nonnewline(register struct parse *p);
 981  *
 982  * Boy, is this implementation ever a kludge...
 983  */
 984 static void
 985 nonnewline(p)
 986 register struct parse *p;
 987 {
 988         register char *oldnext = p->next;
 989         register char *oldend = p->end;
 990         char bracket[4];
 991
 992         p->next = bracket;
 993         p->end = bracket+3;
 994         bracket[0] = '^';
 995         bracket[1] = '\n';
 996         bracket[2] = ']';
 997         bracket[3] = '\0';
 998         p_bracket(p);
 999         assert(p->next == bracket+3);
1000         p->next = oldnext;
1001         p->end = oldend;
1002 }
1003
1004 /*
1005  - repeat - generate code for a bounded repetition, recursively if needed
1006  == static void repeat(register struct parse *p, sopno start, int from, int to);
1007  */
1008 static void
1009 repeat(p, start, from, to)
1010 register struct parse *p;
1011 sopno start;                    /* operand from here to end of strip */
1012 int from;                       /* repeated from this number */
1013 int to;                         /* to this number of times (maybe INFINITY) */
1014 {
1015         register sopno finish = HERE();
1016 #       define  N       2
1017 #       define  INF     3
1018 #       define  REP(f, t)       ((f)*8 + (t))
1019 #       define  MAP(n)  (((n) <= 1) ? (n) : ((n) == INFINITY) ? INF : N)
1020         register sopno copy;
1021
1022         if (p->error != 0)      /* head off possible runaway recursion */
1023                 return;
1024
1025         assert(from <= to);
1026
1027         switch (REP(MAP(from), MAP(to))) {
1028         case REP(0, 0):                 /* must be user doing this */
1029                 DROP(finish-start);     /* drop the operand */
1030                 break;
1031         case REP(0, 1):                 /* as x{1,1}? */
1032         case REP(0, N):                 /* as x{1,n}? */
1033         case REP(0, INF):               /* as x{1,}? */
1034                 /* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
1035                 INSERT(OCH_, start);            /* offset is wrong... */
1036                 repeat(p, start+1, 1, to);
1037                 ASTERN(OOR1, start);
1038                 AHEAD(start);                   /* ... fix it */
1039                 EMIT(OOR2, 0);
1040                 AHEAD(THERE());
1041                 ASTERN(O_CH, THERETHERE());
1042                 break;
1043         case REP(1, 1):                 /* trivial case */
1044                 /* done */
1045                 break;
1046         case REP(1, N):                 /* as x?x{1,n-1} */
1047                 /* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
1048                 INSERT(OCH_, start);
1049                 ASTERN(OOR1, start);
1050                 AHEAD(start);
1051                 EMIT(OOR2, 0);                  /* offset very wrong... */
1052                 AHEAD(THERE());                 /* ...so fix it */
1053                 ASTERN(O_CH, THERETHERE());
1054                 copy = dupl(p, start+1, finish+1);
1055                 assert(copy == finish+4);
1056                 repeat(p, copy, 1, to-1);
1057                 break;
1058         case REP(1, INF):               /* as x+ */
1059                 INSERT(OPLUS_, start);
1060                 ASTERN(O_PLUS, start);
1061                 break;
1062         case REP(N, N):                 /* as xx{m-1,n-1} */
1063                 copy = dupl(p, start, finish);
1064                 repeat(p, copy, from-1, to-1);
1065                 break;
1066         case REP(N, INF):               /* as xx{n-1,INF} */
1067                 copy = dupl(p, start, finish);
1068                 repeat(p, copy, from-1, to);
1069                 break;
1070         default:                        /* "can't happen" */
1071                 SETERROR(REG_ASSERT);   /* just in case */
1072                 break;
1073         }
1074 }
1075
1076 /*
1077  - seterr - set an error condition
1078  == static int seterr(register struct parse *p, int e);
1079  */
1080 static int                      /* useless but makes type checking happy */
1081 seterr(p, e)
1082 register struct parse *p;
1083 int e;
1084 {
1085         if (p->error == 0)      /* keep earliest error condition */
1086                 p->error = e;
1087         p->next = nuls;         /* try to bring things to a halt */
1088         p->end = nuls;
1089         return(0);              /* make the return value well-defined */
1090 }
1091
1092 /*
1093  - allocset - allocate a set of characters for []
1094  == static cset *allocset(register struct parse *p);
1095  */
1096 static cset *
1097 allocset(p)
1098 register struct parse *p;
1099 {
1100         register int no = p->g->ncsets++;
1101         register size_t nc;
1102         register size_t nbytes;
1103         register cset *cs;
1104         register size_t css = (size_t)p->g->csetsize;
1105         register int i;
1106
1107         if (no >= p->ncsalloc) {        /* need another column of space */
1108                 p->ncsalloc += CHAR_BIT;
1109                 nc = p->ncsalloc;
1110                 assert(nc % CHAR_BIT == 0);
1111                 nbytes = nc / CHAR_BIT * css;
1112                 if (p->g->sets == NULL)
1113                         p->g->sets = (cset *)malloc(nc * sizeof(cset));
1114                 else
1115                         p->g->sets = (cset *)realloc((char *)p->g->sets,
1116                                                         nc * sizeof(cset));
1117                 if (p->g->setbits == NULL)
1118                         p->g->setbits = (uch *)malloc(nbytes);
1119                 else {
1120                         p->g->setbits = (uch *)realloc((char *)p->g->setbits,
1121                                                                 nbytes);
1122                         /* xxx this isn't right if setbits is now NULL */
1123                         for (i = 0; i < no; i++)
1124                                 p->g->sets[i].ptr = p->g->setbits + css*(i/CHAR_BIT);
1125                 }
1126                 if (p->g->sets != NULL && p->g->setbits != NULL)
1127                         (void) memset((char *)p->g->setbits + (nbytes - css),
1128                                                                 0, css);
1129                 else {
1130                         no = 0;
1131                         SETERROR(REG_ESPACE);
1132                         /* caller's responsibility not to do set ops */
1133                 }
1134         }
1135
1136         assert(p->g->sets != NULL);     /* xxx */
1137         cs = &p->g->sets[no];
1138         cs->ptr = p->g->setbits + css*((no)/CHAR_BIT);
1139         cs->mask = 1 << ((no) % CHAR_BIT);
1140         cs->hash = 0;
1141         cs->smultis = 0;
1142         cs->multis = NULL;
1143
1144         return(cs);
1145 }
1146
1147 /*
1148  - freeset - free a now-unused set
1149  == static void freeset(register struct parse *p, register cset *cs);
1150  */
1151 static void
1152 freeset(p, cs)
1153 register struct parse *p;
1154 register cset *cs;
1155 {
1156         register int i;
1157         register cset *top = &p->g->sets[p->g->ncsets];
1158         register size_t css = (size_t)p->g->csetsize;
1159
1160         for (i = 0; i < css; i++)
1161                 CHsub(cs, i);
1162         if (cs == top-1)        /* recover only the easy case */
1163                 p->g->ncsets--;
1164 }
1165
1166 /*
1167  - freezeset - final processing on a set of characters
1168  == static int freezeset(register struct parse *p, register cset *cs);
1169  *
1170  * The main task here is merging identical sets.  This is usually a waste
1171  * of time (although the hash code minimizes the overhead), but can win
1172  * big if REG_ICASE is being used.  REG_ICASE, by the way, is why the hash
1173  * is done using addition rather than xor -- all ASCII [aA] sets xor to
1174  * the same value!
1175  */
1176 static int                      /* set number */
1177 freezeset(p, cs)
1178 register struct parse *p;
1179 register cset *cs;
1180 {
1181         register uch h = cs->hash;
1182         register int i;
1183         register cset *top = &p->g->sets[p->g->ncsets];
1184         register cset *cs2;
1185         register size_t css = (size_t)p->g->csetsize;
1186
1187         /* look for an earlier one which is the same */
1188         for (cs2 = &p->g->sets[0]; cs2 < top; cs2++)
1189                 if (cs2->hash == h && cs2 != cs) {
1190                         /* maybe */
1191                         for (i = 0; i < css; i++)
1192                                 if (!!CHIN(cs2, i) != !!CHIN(cs, i))
1193                                         break;          /* no */
1194                         if (i == css)
1195                                 break;                  /* yes */
1196                 }
1197
1198         if (cs2 < top) {        /* found one */
1199                 freeset(p, cs);
1200                 cs = cs2;
1201         }
1202
1203         return((int)(cs - p->g->sets));
1204 }
1205
1206 /*
1207  - firstch - return first character in a set (which must have at least one)
1208  == static int firstch(register struct parse *p, register cset *cs);
1209  */
1210 static int                      /* character; there is no "none" value */
1211 firstch(p, cs)
1212 register struct parse *p;
1213 register cset *cs;
1214 {
1215         register int i;
1216         register size_t css = (size_t)p->g->csetsize;
1217
1218         for (i = 0; i < css; i++)
1219                 if (CHIN(cs, i))
1220                         return((char)i);
1221         assert(never);
1222         return(0);              /* arbitrary */
1223 }
1224
1225 /*
1226  - nch - number of characters in a set
1227  == static int nch(register struct parse *p, register cset *cs);
1228  */
1229 static int
1230 nch(p, cs)
1231 register struct parse *p;
1232 register cset *cs;
1233 {
1234         register int i;
1235         register size_t css = (size_t)p->g->csetsize;
1236         register int n = 0;
1237
1238         for (i = 0; i < css; i++)
1239                 if (CHIN(cs, i))
1240                         n++;
1241         return(n);
1242 }
1243
1244 /*
1245  - mcadd - add a collating element to a cset
1246  == static void mcadd(register struct parse *p, register cset *cs, \
1247  ==     register char *cp);
1248  */
1249 static void
1250 mcadd(p, cs, cp)
1251 register struct parse *p;
1252 register cset *cs;
1253 register char *cp;
1254 {
1255         register size_t oldend = cs->smultis;
1256
1257         cs->smultis += strlen(cp) + 1;
1258         if (cs->multis == NULL)
1259                 cs->multis = malloc(cs->smultis);
1260         else
1261                 cs->multis = realloc(cs->multis, cs->smultis);
1262         if (cs->multis == NULL) {
1263                 SETERROR(REG_ESPACE);
1264                 return;
1265         }
1266
1267         (void) strcpy(cs->multis + oldend - 1, cp);
1268         cs->multis[cs->smultis - 1] = '\0';
1269 }
1270
1271 /*
1272  - mcsub - subtract a collating element from a cset
1273  == static void mcsub(register cset *cs, register char *cp);
1274  */
1275 static void
1276 mcsub(cs, cp)
1277 register cset *cs;
1278 register char *cp;
1279 {
1280         register char *fp = mcfind(cs, cp);
1281         register size_t len = strlen(fp);
1282
1283         assert(fp != NULL);
1284         (void) memmove(fp, fp + len + 1,
1285                                 cs->smultis - (fp + len + 1 - cs->multis));
1286         cs->smultis -= len;
1287
1288         if (cs->smultis == 0) {
1289                 free(cs->multis);
1290                 cs->multis = NULL;
1291                 return;
1292         }
1293
1294         cs->multis = realloc(cs->multis, cs->smultis);
1295         assert(cs->multis != NULL);
1296 }
1297
1298 /*
1299  - mcin - is a collating element in a cset?
1300  == static int mcin(register cset *cs, register char *cp);
1301  */
1302 static int
1303 mcin(cs, cp)
1304 register cset *cs;
1305 register char *cp;
1306 {
1307         return(mcfind(cs, cp) != NULL);
1308 }
1309
1310 /*
1311  - mcfind - find a collating element in a cset
1312  == static char *mcfind(register cset *cs, register char *cp);
1313  */
1314 static char *
1315 mcfind(cs, cp)
1316 register cset *cs;
1317 register char *cp;
1318 {
1319         register char *p;
1320
1321         if (cs->multis == NULL)
1322                 return(NULL);
1323         for (p = cs->multis; *p != '\0'; p += strlen(p) + 1)
1324                 if (strcmp(cp, p) == 0)
1325                         return(p);
1326         return(NULL);
1327 }
1328
1329 /*
1330  - mcinvert - invert the list of collating elements in a cset
1331  == static void mcinvert(register struct parse *p, register cset *cs);
1332  *
1333  * This would have to know the set of possibilities.  Implementation
1334  * is deferred.
1335  */
1336 static void
1337 mcinvert(p, cs)
1338 register struct parse *p;
1339 register cset *cs;
1340 {
1341         assert(cs->multis == NULL);     /* xxx */
1342 }
1343
1344 /*
1345  - mccase - add case counterparts of the list of collating elements in a cset
1346  == static void mccase(register struct parse *p, register cset *cs);
1347  *
1348  * This would have to know the set of possibilities.  Implementation
1349  * is deferred.
1350  */
1351 static void
1352 mccase(p, cs)
1353 register struct parse *p;
1354 register cset *cs;
1355 {
1356         assert(cs->multis == NULL);     /* xxx */
1357 }
1358
1359 /*
1360  - isinsets - is this character in any sets?
1361  == static int isinsets(register struct re_guts *g, int c);
1362  */
1363 static int                      /* predicate */
1364 isinsets(g, c)
1365 register struct re_guts *g;
1366 int c;
1367 {
1368         register uch *col;
1369         register int i;
1370         register int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1371         register unsigned uc = (unsigned char)c;
1372
1373         for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1374                 if (col[uc] != 0)
1375                         return(1);
1376         return(0);
1377 }
1378
1379 /*
1380  - samesets - are these two characters in exactly the same sets?
1381  == static int samesets(register struct re_guts *g, int c1, int c2);
1382  */
1383 static int                      /* predicate */
1384 samesets(g, c1, c2)
1385 register struct re_guts *g;
1386 int c1;
1387 int c2;
1388 {
1389         register uch *col;
1390         register int i;
1391         register int ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
1392         register unsigned uc1 = (unsigned char)c1;
1393         register unsigned uc2 = (unsigned char)c2;
1394
1395         for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
1396                 if (col[uc1] != col[uc2])
1397                         return(0);
1398         return(1);
1399 }
1400
1401 /*
1402  - categorize - sort out character categories
1403  == static void categorize(struct parse *p, register struct re_guts *g);
1404  */
1405 static void
1406 categorize(p, g)
1407 struct parse *p;
1408 register struct re_guts *g;
1409 {
1410         register cat_t *cats = g->categories;
1411         register int c;
1412         register int c2;
1413         register cat_t cat;
1414
1415         /* avoid making error situations worse */
1416         if (p->error != 0)
1417                 return;
1418
1419         for (c = CHAR_MIN; c <= CHAR_MAX; c++)
1420                 if (cats[c] == 0 && isinsets(g, c)) {
1421                         cat = g->ncategories++;
1422                         cats[c] = cat;
1423                         for (c2 = c+1; c2 <= CHAR_MAX; c2++)
1424                                 if (cats[c2] == 0 && samesets(g, c, c2))
1425                                         cats[c2] = cat;
1426                 }
1427 }
1428
1429 /*
1430  - dupl - emit a duplicate of a bunch of sops
1431  == static sopno dupl(register struct parse *p, sopno start, sopno finish);
1432  */
1433 static sopno                    /* start of duplicate */
1434 dupl(p, start, finish)
1435 register struct parse *p;
1436 sopno start;                    /* from here */
1437 sopno finish;                   /* to this less one */
1438 {
1439         register sopno ret = HERE();
1440         register sopno len = finish - start;
1441
1442         assert(finish >= start);
1443         if (len == 0)
1444                 return(ret);
1445         enlarge(p, p->ssize + len);     /* this many unexpected additions */
1446         assert(p->ssize >= p->slen + len);
1447         (void) memcpy((char *)(p->strip + p->slen),
1448                 (char *)(p->strip + start), (size_t)len*sizeof(sop));
1449         p->slen += len;
1450         return(ret);
1451 }
1452
1453 /*
1454  - doemit - emit a strip operator
1455  == static void doemit(register struct parse *p, sop op, size_t opnd);
1456  *
1457  * It might seem better to implement this as a macro with a function as
1458  * hard-case backup, but it's just too big and messy unless there are
1459  * some changes to the data structures.  Maybe later.
1460  */
1461 static void
1462 doemit(p, op, opnd)
1463 register struct parse *p;
1464 sop op;
1465 size_t opnd;
1466 {
1467         /* avoid making error situations worse */
1468         if (p->error != 0)
1469                 return;
1470
1471         /* deal with oversize operands ("can't happen", more or less) */
1472         assert(opnd < 1<<OPSHIFT);
1473
1474         /* deal with undersized strip */
1475         if (p->slen >= p->ssize)
1476                 enlarge(p, (p->ssize+1) / 2 * 3);       /* +50% */
1477         assert(p->slen < p->ssize);
1478
1479         /* finally, it's all reduced to the easy case */
1480         p->strip[p->slen++] = SOP(op, opnd);
1481 }
1482
1483 /*
1484  - doinsert - insert a sop into the strip
1485  == static void doinsert(register struct parse *p, sop op, size_t opnd, sopno pos);
1486  */
1487 static void
1488 doinsert(p, op, opnd, pos)
1489 register struct parse *p;
1490 sop op;
1491 size_t opnd;
1492 sopno pos;
1493 {
1494         register sopno sn;
1495         register sop s;
1496         register int i;
1497
1498         /* avoid making error situations worse */
1499         if (p->error != 0)
1500                 return;
1501
1502         sn = HERE();
1503         EMIT(op, opnd);         /* do checks, ensure space */
1504         assert(HERE() == sn+1);
1505         s = p->strip[sn];
1506
1507         /* adjust paren pointers */
1508         assert(pos > 0);
1509         for (i = 1; i < NPAREN; i++) {
1510                 if (p->pbegin[i] >= pos) {
1511                         p->pbegin[i]++;
1512                 }
1513                 if (p->pend[i] >= pos) {
1514                         p->pend[i]++;
1515                 }
1516         }
1517
1518         memmove((char *)&p->strip[pos+1], (char *)&p->strip[pos],
1519                                                 (HERE()-pos-1)*sizeof(sop));
1520         p->strip[pos] = s;
1521 }
1522
1523 /*
1524  - dofwd - complete a forward reference
1525  == static void dofwd(register struct parse *p, sopno pos, sop value);
1526  */
1527 static void
1528 dofwd(p, pos, value)
1529 register struct parse *p;
1530 register sopno pos;
1531 sop value;
1532 {
1533         /* avoid making error situations worse */
1534         if (p->error != 0)
1535                 return;
1536
1537         assert(value < 1<<OPSHIFT);
1538         p->strip[pos] = OP(p->strip[pos]) | value;
1539 }
1540
1541 /*
1542  - enlarge - enlarge the strip
1543  == static void enlarge(register struct parse *p, sopno size);
1544  */
1545 static void
1546 enlarge(p, size)
1547 register struct parse *p;
1548 register sopno size;
1549 {
1550         register sop *sp;
1551
1552         if (p->ssize >= size)
1553                 return;
1554
1555         sp = (sop *)realloc(p->strip, size*sizeof(sop));
1556         if (sp == NULL) {
1557                 SETERROR(REG_ESPACE);
1558                 return;
1559         }
1560         p->strip = sp;
1561         p->ssize = size;
1562 }
1563
1564 /*
1565  - stripsnug - compact the strip
1566  == static void stripsnug(register struct parse *p, register struct re_guts *g);
1567  */
1568 static void
1569 stripsnug(p, g)
1570 register struct parse *p;
1571 register struct re_guts *g;
1572 {
1573         g->nstates = p->slen;
1574         g->strip = (sop *)realloc((char *)p->strip, p->slen * sizeof(sop));
1575         if (g->strip == NULL) {
1576                 SETERROR(REG_ESPACE);
1577                 g->strip = p->strip;
1578         }
1579 }
1580
1581 /*
1582  - findmust - fill in must and mlen with longest mandatory literal string
1583  == static void findmust(register struct parse *p, register struct re_guts *g);
1584  *
1585  * This algorithm could do fancy things like analyzing the operands of |
1586  * for common subsequences.  Someday.  This code is simple and finds most
1587  * of the interesting cases.
1588  *
1589  * Note that must and mlen got initialized during setup.
1590  */
1591 static void
1592 findmust(p, g)
1593 struct parse *p;
1594 register struct re_guts *g;
1595 {
1596         register sop *scan;
1597         sop *start;
1598         register sop *newstart;
1599         register sopno newlen;
1600         register sop s;
1601         register char *cp;
1602         register sopno i;
1603
1604         /* avoid making error situations worse */
1605         if (p->error != 0)
1606                 return;
1607
1608         /* find the longest OCHAR sequence in strip */
1609         newlen = 0;
1610         scan = g->strip + 1;
1611         do {
1612                 s = *scan++;
1613                 switch (OP(s)) {
1614                 case OCHAR:             /* sequence member */
1615                         if (newlen == 0)                /* new sequence */
1616                                 newstart = scan - 1;
1617                         newlen++;
1618                         break;
1619                 case OPLUS_:            /* things that don't break one */
1620                 case OLPAREN:
1621                 case ORPAREN:
1622                         break;
1623                 case OQUEST_:           /* things that must be skipped */
1624                 case OCH_:
1625                         scan--;
1626                         do {
1627                                 scan += OPND(s);
1628                                 s = *scan;
1629                                 /* assert() interferes w debug printouts */
1630                                 if (OP(s) != O_QUEST && OP(s) != O_CH &&
1631                                                         OP(s) != OOR2) {
1632                                         g->iflags |= BAD;
1633                                         return;
1634                                 }
1635                         } while (OP(s) != O_QUEST && OP(s) != O_CH);
1636                         /* fallthrough */
1637                 default:                /* things that break a sequence */
1638                         if (newlen > g->mlen) {         /* ends one */
1639                                 start = newstart;
1640                                 g->mlen = newlen;
1641                         }
1642                         newlen = 0;
1643                         break;
1644                 }
1645         } while (OP(s) != OEND);
1646
1647         if (g->mlen == 0)               /* there isn't one */
1648                 return;
1649
1650         /* turn it into a character string */
1651         g->must = malloc((size_t)g->mlen + 1);
1652         if (g->must == NULL) {          /* argh; just forget it */
1653                 g->mlen = 0;
1654                 return;
1655         }
1656         cp = g->must;
1657         scan = start;
1658         for (i = g->mlen; i > 0; i--) {
1659                 while (OP(s = *scan++) != OCHAR)
1660                         continue;
1661                 assert(cp < g->must + g->mlen);
1662                 *cp++ = (char)OPND(s);
1663         }
1664         assert(cp == g->must + g->mlen);
1665         *cp++ = '\0';           /* just on general principles */
1666 }
1667
1668 /*
1669  - pluscount - count + nesting
1670  == static sopno pluscount(register struct parse *p, register struct re_guts *g);
1671  */
1672 static sopno                    /* nesting depth */
1673 pluscount(p, g)
1674 struct parse *p;
1675 register struct re_guts *g;
1676 {
1677         register sop *scan;
1678         register sop s;
1679         register sopno plusnest = 0;
1680         register sopno maxnest = 0;
1681
1682         if (p->error != 0)
1683                 return(0);      /* there may not be an OEND */
1684
1685         scan = g->strip + 1;
1686         do {
1687                 s = *scan++;
1688                 switch (OP(s)) {
1689                 case OPLUS_:
1690                         plusnest++;
1691                         break;
1692                 case O_PLUS:
1693                         if (plusnest > maxnest)
1694                                 maxnest = plusnest;
1695                         plusnest--;
1696                         break;
1697                 }
1698         } while (OP(s) != OEND);
1699         if (plusnest != 0)
1700                 g->iflags |= BAD;
1701         return(maxnest);
1702 }
1703
1704 /*
1705  * $PchId: regcomp.c,v 1.2 1996/03/12 19:10:15 philip Exp $
1706  */