gen/FreeBSD/glob.c

   1 /*
   2  * Copyright (c) 1989, 1993
   3  *      The Regents of the University of California.  All rights reserved.
   4  *
   5  * This code is derived from software contributed to Berkeley by
   6  * Guido van Rossum.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 3. All advertising materials mentioning features or use of this software
  17  *    must display the following acknowledgement:
  18  *      This product includes software developed by the University of
  19  *      California, Berkeley and its contributors.
  20  * 4. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  */
  36
  37 #if defined(LIBC_SCCS) && !defined(lint)
  38 static char sccsid[] = "@(#)glob.c      8.3 (Berkeley) 10/13/93";
  39 #endif /* LIBC_SCCS and not lint */
  40 #include <sys/cdefs.h>
  41 __FBSDID("$FreeBSD: src/lib/libc/gen/glob.c,v 1.22 2004/07/29 03:48:52 tjr Exp $");
  42
  43 /*
  44  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  45  *
  46  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  47  *
  48  * Optional extra services, controlled by flags not defined by POSIX:
  49  *
  50  * GLOB_QUOTE:
  51  *      Escaping convention: \ inhibits any special meaning the following
  52  *      character might have (except \ at end of string is retained).
  53  * GLOB_MAGCHAR:
  54  *      Set in gl_flags if pattern contained a globbing character.
  55  * GLOB_NOMAGIC:
  56  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  57  *      not contain any magic characters.  [Used in csh style globbing]
  58  * GLOB_ALTDIRFUNC:
  59  *      Use alternately specified directory access functions.
  60  * GLOB_TILDE:
  61  *      expand ~user/foo to the /home/dir/of/user/foo
  62  * GLOB_BRACE:
  63  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  64  * gl_matchc:
  65  *      Number of matches in the current invocation of glob.
  66  */
  67
  68 /*
  69  * Some notes on multibyte character support:
  70  * 1. Patterns with illegal byte sequences match nothing - even if
  71  *    GLOB_NOCHECK is specified.
  72  * 2. Illegal byte sequences in filenames are handled by treating them as
  73  *    single-byte characters with a value of the first byte of the sequence
  74  *    cast to wchar_t.
  75  * 3. State-dependent encodings are not currently supported.
  76  */
  77
  78 #include <sys/param.h>
  79 #include <sys/stat.h>
  80
  81 #include <ctype.h>
  82 #include <dirent.h>
  83 #include <errno.h>
  84 #include <glob.h>
  85 #include <limits.h>
  86 #include <pwd.h>
  87 #include <stdint.h>
  88 #include <stdio.h>
  89 #include <stdlib.h>
  90 #include <string.h>
  91 #include <unistd.h>
  92 #include <wchar.h>
  93
  94 #include "collate.h"
  95
  96 #define DOLLAR          '$'
  97 #define DOT             '.'
  98 #define EOS             '\0'
  99 #define LBRACKET        '['
 100 #define NOT             '!'
 101 #define QUESTION        '?'
 102 #define QUOTE           '\\'
 103 #define RANGE           '-'
 104 #define RBRACKET        ']'
 105 #define SEP             '/'
 106 #define STAR            '*'
 107 #define TILDE           '~'
 108 #define UNDERSCORE      '_'
 109 #define LBRACE          '{'
 110 #define RBRACE          '}'
 111 #define SLASH           '/'
 112 #define COMMA           ','
 113
 114 #ifndef DEBUG
 115
 116 #define M_QUOTE         0x8000000000ULL
 117 #define M_PROTECT       0x4000000000ULL
 118 #define M_MASK          0xffffffffffULL
 119 #define M_CHAR          0x00ffffffffULL
 120
 121 typedef uint_fast64_t Char;
 122
 123 #else
 124
 125 #define M_QUOTE         0x80
 126 #define M_PROTECT       0x40
 127 #define M_MASK          0xff
 128 #define M_CHAR          0x7f
 129
 130 typedef char Char;
 131
 132 #endif
 133
 134
 135 #define CHAR(c)         ((Char)((c)&M_CHAR))
 136 #define META(c)         ((Char)((c)|M_QUOTE))
 137 #define M_ALL           META('*')
 138 #define M_END           META(']')
 139 #define M_NOT           META('!')
 140 #define M_ONE           META('?')
 141 #define M_RNG           META('-')
 142 #define M_SET           META('[')
 143 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 144
 145
 146 static int       compare(const void *, const void *);
 147 static int       g_Ctoc(const Char *, char *, u_int);
 148 static int       g_lstat(Char *, struct stat *, glob_t *);
 149 static DIR      *g_opendir(Char *, glob_t *);
 150 static Char     *g_strchr(Char *, wchar_t);
 151 #ifdef notdef
 152 static Char     *g_strcat(Char *, const Char *);
 153 #endif
 154 static int       g_stat(Char *, struct stat *, glob_t *);
 155 static int       glob0(const Char *, glob_t *, int *);
 156 static int       glob1(Char *, glob_t *, int *);
 157 static int       glob2(Char *, Char *, Char *, Char *, glob_t *, int *);
 158 static int       glob3(Char *, Char *, Char *, Char *, Char *, glob_t *, int *);
 159 static int       globextend(const Char *, glob_t *, int *);
 160 static const Char *
 161                  globtilde(const Char *, Char *, size_t, glob_t *);
 162 static int       globexp1(const Char *, glob_t *, int *);
 163 static int       globexp2(const Char *, const Char *, glob_t *, int *, int *);
 164 static int       match(Char *, Char *, Char *);
 165 #ifdef DEBUG
 166 static void      qprintf(const char *, Char *);
 167 #endif
 168
 169 int
 170 glob(pattern, flags, errfunc, pglob)
 171         const char *pattern;
 172         int flags, (*errfunc)(const char *, int);
 173         glob_t *pglob;
 174 {
 175         const u_char *patnext;
 176         int limit;
 177         Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
 178         mbstate_t mbs;
 179         wchar_t wc;
 180         size_t clen;
 181
 182         patnext = (u_char *) pattern;
 183         if (!(flags & GLOB_APPEND)) {
 184                 pglob->gl_pathc = 0;
 185                 pglob->gl_pathv = NULL;
 186                 if (!(flags & GLOB_DOOFFS))
 187                         pglob->gl_offs = 0;
 188         }
 189         if (flags & GLOB_LIMIT) {
 190                 limit = pglob->gl_matchc;
 191                 if (limit == 0)
 192                         limit = ARG_MAX;
 193         } else
 194                 limit = 0;
 195         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 196         pglob->gl_errfunc = errfunc;
 197         pglob->gl_matchc = 0;
 198
 199         bufnext = patbuf;
 200         bufend = bufnext + MAXPATHLEN - 1;
 201         if (flags & GLOB_NOESCAPE) {
 202                 memset(&mbs, 0, sizeof(mbs));
 203                 while (bufend - bufnext >= MB_CUR_MAX) {
 204                         clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
 205                         if (clen == (size_t)-1 || clen == (size_t)-2)
 206                                 return (GLOB_NOMATCH);
 207                         else if (clen == 0)
 208                                 break;
 209                         *bufnext++ = wc;
 210                         patnext += clen;
 211                 }
 212         } else {
 213                 /* Protect the quoted characters. */
 214                 memset(&mbs, 0, sizeof(mbs));
 215                 while (bufend - bufnext >= MB_CUR_MAX) {
 216                         if (*patnext == QUOTE) {
 217                                 if (*++patnext == EOS) {
 218                                         *bufnext++ = QUOTE | M_PROTECT;
 219                                         continue;
 220                                 }
 221                                 prot = M_PROTECT;
 222                         } else
 223                                 prot = 0;
 224                         clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
 225                         if (clen == (size_t)-1 || clen == (size_t)-2)
 226                                 return (GLOB_NOMATCH);
 227                         else if (clen == 0)
 228                                 break;
 229                         *bufnext++ = wc | prot;
 230                         patnext += clen;
 231                 }
 232         }
 233         *bufnext = EOS;
 234
 235         if (flags & GLOB_BRACE)
 236             return globexp1(patbuf, pglob, &limit);
 237         else
 238             return glob0(patbuf, pglob, &limit);
 239 }
 240
 241 /*
 242  * Expand recursively a glob {} pattern. When there is no more expansion
 243  * invoke the standard globbing routine to glob the rest of the magic
 244  * characters
 245  */
 246 static int
 247 globexp1(pattern, pglob, limit)
 248         const Char *pattern;
 249         glob_t *pglob;
 250         int *limit;
 251 {
 252         const Char* ptr = pattern;
 253         int rv;
 254
 255         /* Protect a single {}, for find(1), like csh */
 256         if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
 257                 return glob0(pattern, pglob, limit);
 258
 259         while ((ptr = (const Char *) g_strchr((Char *) ptr, LBRACE)) != NULL)
 260                 if (!globexp2(ptr, pattern, pglob, &rv, limit))
 261                         return rv;
 262
 263         return glob0(pattern, pglob, limit);
 264 }
 265
 266
 267 /*
 268  * Recursive brace globbing helper. Tries to expand a single brace.
 269  * If it succeeds then it invokes globexp1 with the new pattern.
 270  * If it fails then it tries to glob the rest of the pattern and returns.
 271  */
 272 static int
 273 globexp2(ptr, pattern, pglob, rv, limit)
 274         const Char *ptr, *pattern;
 275         glob_t *pglob;
 276         int *rv, *limit;
 277 {
 278         int     i;
 279         Char   *lm, *ls;
 280         const Char *pe, *pm, *pl;
 281         Char    patbuf[MAXPATHLEN];
 282
 283         /* copy part up to the brace */
 284         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 285                 continue;
 286         *lm = EOS;
 287         ls = lm;
 288
 289         /* Find the balanced brace */
 290         for (i = 0, pe = ++ptr; *pe; pe++)
 291                 if (*pe == LBRACKET) {
 292                         /* Ignore everything between [] */
 293                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
 294                                 continue;
 295                         if (*pe == EOS) {
 296                                 /*
 297                                  * We could not find a matching RBRACKET.
 298                                  * Ignore and just look for RBRACE
 299                                  */
 300                                 pe = pm;
 301                         }
 302                 }
 303                 else if (*pe == LBRACE)
 304                         i++;
 305                 else if (*pe == RBRACE) {
 306                         if (i == 0)
 307                                 break;
 308                         i--;
 309                 }
 310
 311         /* Non matching braces; just glob the pattern */
 312         if (i != 0 || *pe == EOS) {
 313                 *rv = glob0(patbuf, pglob, limit);
 314                 return 0;
 315         }
 316
 317         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 318                 switch (*pm) {
 319                 case LBRACKET:
 320                         /* Ignore everything between [] */
 321                         for (pl = pm++; *pm != RBRACKET && *pm != EOS; pm++)
 322                                 continue;
 323                         if (*pm == EOS) {
 324                                 /*
 325                                  * We could not find a matching RBRACKET.
 326                                  * Ignore and just look for RBRACE
 327                                  */
 328                                 pm = pl;
 329                         }
 330                         break;
 331
 332                 case LBRACE:
 333                         i++;
 334                         break;
 335
 336                 case RBRACE:
 337                         if (i) {
 338                             i--;
 339                             break;
 340                         }
 341                         /* FALLTHROUGH */
 342                 case COMMA:
 343                         if (i && *pm == COMMA)
 344                                 break;
 345                         else {
 346                                 /* Append the current string */
 347                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 348                                         continue;
 349                                 /*
 350                                  * Append the rest of the pattern after the
 351                                  * closing brace
 352                                  */
 353                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 354                                         continue;
 355
 356                                 /* Expand the current pattern */
 357 #ifdef DEBUG
 358                                 qprintf("globexp2:", patbuf);
 359 #endif
 360                                 *rv = globexp1(patbuf, pglob, limit);
 361
 362                                 /* move after the comma, to the next string */
 363                                 pl = pm + 1;
 364                         }
 365                         break;
 366
 367                 default:
 368                         break;
 369                 }
 370         *rv = 0;
 371         return 0;
 372 }
 373
 374
 375
 376 /*
 377  * expand tilde from the passwd file.
 378  */
 379 static const Char *
 380 globtilde(pattern, patbuf, patbuf_len, pglob)
 381         const Char *pattern;
 382         Char *patbuf;
 383         size_t patbuf_len;
 384         glob_t *pglob;
 385 {
 386         struct passwd *pwd;
 387         char *h;
 388         const Char *p;
 389         Char *b, *eb;
 390
 391         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 392                 return pattern;
 393
 394         /*
 395          * Copy up to the end of the string or /
 396          */
 397         eb = &patbuf[patbuf_len - 1];
 398         for (p = pattern + 1, h = (char *) patbuf;
 399             h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
 400                 continue;
 401
 402         *h = EOS;
 403
 404         if (((char *) patbuf)[0] == EOS) {
 405                 /*
 406                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 407                  * we're not running setuid or setgid) and then trying
 408                  * the password file
 409                  */
 410                 if (issetugid() != 0 ||
 411                     (h = getenv("HOME")) == NULL) {
 412                         if (((h = getlogin()) != NULL &&
 413                              (pwd = getpwnam(h)) != NULL) ||
 414                             (pwd = getpwuid(getuid())) != NULL)
 415                                 h = pwd->pw_dir;
 416                         else
 417                                 return pattern;
 418                 }
 419         }
 420         else {
 421                 /*
 422                  * Expand a ~user
 423                  */
 424                 if ((pwd = getpwnam((char*) patbuf)) == NULL)
 425                         return pattern;
 426                 else
 427                         h = pwd->pw_dir;
 428         }
 429
 430         /* Copy the home directory */
 431         for (b = patbuf; b < eb && *h; *b++ = *h++)
 432                 continue;
 433
 434         /* Append the rest of the pattern */
 435         while (b < eb && (*b++ = *p++) != EOS)
 436                 continue;
 437         *b = EOS;
 438
 439         return patbuf;
 440 }
 441
 442
 443 /*
 444  * The main glob() routine: compiles the pattern (optionally processing
 445  * quotes), calls glob1() to do the real pattern matching, and finally
 446  * sorts the list (unless unsorted operation is requested).  Returns 0
 447  * if things went well, nonzero if errors occurred.
 448  */
 449 static int
 450 glob0(pattern, pglob, limit)
 451         const Char *pattern;
 452         glob_t *pglob;
 453         int *limit;
 454 {
 455         const Char *qpatnext;
 456         int c, err, oldpathc;
 457         Char *bufnext, patbuf[MAXPATHLEN];
 458
 459         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
 460         oldpathc = pglob->gl_pathc;
 461         bufnext = patbuf;
 462
 463         /* We don't need to check for buffer overflow any more. */
 464         while ((c = *qpatnext++) != EOS) {
 465                 switch (c) {
 466                 case LBRACKET:
 467                         c = *qpatnext;
 468                         if (c == NOT)
 469                                 ++qpatnext;
 470                         if (*qpatnext == EOS ||
 471                             g_strchr((Char *) qpatnext+1, RBRACKET) == NULL) {
 472                                 *bufnext++ = LBRACKET;
 473                                 if (c == NOT)
 474                                         --qpatnext;
 475                                 break;
 476                         }
 477                         *bufnext++ = M_SET;
 478                         if (c == NOT)
 479                                 *bufnext++ = M_NOT;
 480                         c = *qpatnext++;
 481                         do {
 482                                 *bufnext++ = CHAR(c);
 483                                 if (*qpatnext == RANGE &&
 484                                     (c = qpatnext[1]) != RBRACKET) {
 485                                         *bufnext++ = M_RNG;
 486                                         *bufnext++ = CHAR(c);
 487                                         qpatnext += 2;
 488                                 }
 489                         } while ((c = *qpatnext++) != RBRACKET);
 490                         pglob->gl_flags |= GLOB_MAGCHAR;
 491                         *bufnext++ = M_END;
 492                         break;
 493                 case QUESTION:
 494                         pglob->gl_flags |= GLOB_MAGCHAR;
 495                         *bufnext++ = M_ONE;
 496                         break;
 497                 case STAR:
 498                         pglob->gl_flags |= GLOB_MAGCHAR;
 499                         /* collapse adjacent stars to one,
 500                          * to avoid exponential behavior
 501                          */
 502                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 503                             *bufnext++ = M_ALL;
 504                         break;
 505                 default:
 506                         *bufnext++ = CHAR(c);
 507                         break;
 508                 }
 509         }
 510         *bufnext = EOS;
 511 #ifdef DEBUG
 512         qprintf("glob0:", patbuf);
 513 #endif
 514
 515         if ((err = glob1(patbuf, pglob, limit)) != 0)
 516                 return(err);
 517
 518         /*
 519          * If there was no match we are going to append the pattern
 520          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
 521          * and the pattern did not contain any magic characters
 522          * GLOB_NOMAGIC is there just for compatibility with csh.
 523          */
 524         if (pglob->gl_pathc == oldpathc) {
 525                 if (((pglob->gl_flags & GLOB_NOCHECK) ||
 526                     ((pglob->gl_flags & GLOB_NOMAGIC) &&
 527                         !(pglob->gl_flags & GLOB_MAGCHAR))))
 528                         return(globextend(pattern, pglob, limit));
 529                 else
 530                         return(GLOB_NOMATCH);
 531         }
 532         if (!(pglob->gl_flags & GLOB_NOSORT))
 533                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 534                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 535         return(0);
 536 }
 537
 538 static int
 539 compare(p, q)
 540         const void *p, *q;
 541 {
 542         return(strcmp(*(char **)p, *(char **)q));
 543 }
 544
 545 static int
 546 glob1(pattern, pglob, limit)
 547         Char *pattern;
 548         glob_t *pglob;
 549         int *limit;
 550 {
 551         Char pathbuf[MAXPATHLEN];
 552
 553         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 554         if (*pattern == EOS)
 555                 return(0);
 556         return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 557             pattern, pglob, limit));
 558 }
 559
 560 /*
 561  * The functions glob2 and glob3 are mutually recursive; there is one level
 562  * of recursion for each segment in the pattern that contains one or more
 563  * meta characters.
 564  */
 565 static int
 566 glob2(pathbuf, pathend, pathend_last, pattern, pglob, limit)
 567         Char *pathbuf, *pathend, *pathend_last, *pattern;
 568         glob_t *pglob;
 569         int *limit;
 570 {
 571         struct stat sb;
 572         Char *p, *q;
 573         int anymeta;
 574
 575         /*
 576          * Loop over pattern segments until end of pattern or until
 577          * segment with meta character found.
 578          */
 579         for (anymeta = 0;;) {
 580                 if (*pattern == EOS) {          /* End of pattern? */
 581                         *pathend = EOS;
 582                         if (g_lstat(pathbuf, &sb, pglob))
 583                                 return(0);
 584
 585                         if (((pglob->gl_flags & GLOB_MARK) &&
 586                             pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
 587                             || (S_ISLNK(sb.st_mode) &&
 588                             (g_stat(pathbuf, &sb, pglob) == 0) &&
 589                             S_ISDIR(sb.st_mode)))) {
 590                                 if (pathend + 1 > pathend_last)
 591                                         return (GLOB_ABORTED);
 592                                 *pathend++ = SEP;
 593                                 *pathend = EOS;
 594                         }
 595                         ++pglob->gl_matchc;
 596                         return(globextend(pathbuf, pglob, limit));
 597                 }
 598
 599                 /* Find end of next segment, copy tentatively to pathend. */
 600                 q = pathend;
 601                 p = pattern;
 602                 while (*p != EOS && *p != SEP) {
 603                         if (ismeta(*p))
 604                                 anymeta = 1;
 605                         if (q + 1 > pathend_last)
 606                                 return (GLOB_ABORTED);
 607                         *q++ = *p++;
 608                 }
 609
 610                 if (!anymeta) {         /* No expansion, do next segment. */
 611                         pathend = q;
 612                         pattern = p;
 613                         while (*pattern == SEP) {
 614                                 if (pathend + 1 > pathend_last)
 615                                         return (GLOB_ABORTED);
 616                                 *pathend++ = *pattern++;
 617                         }
 618                 } else                  /* Need expansion, recurse. */
 619                         return(glob3(pathbuf, pathend, pathend_last, pattern, p,
 620                             pglob, limit));
 621         }
 622         /* NOTREACHED */
 623 }
 624
 625 static int
 626 glob3(pathbuf, pathend, pathend_last, pattern, restpattern, pglob, limit)
 627         Char *pathbuf, *pathend, *pathend_last, *pattern, *restpattern;
 628         glob_t *pglob;
 629         int *limit;
 630 {
 631         struct dirent *dp;
 632         DIR *dirp;
 633         int err;
 634         char buf[MAXPATHLEN];
 635
 636         /*
 637          * The readdirfunc declaration can't be prototyped, because it is
 638          * assigned, below, to two functions which are prototyped in glob.h
 639          * and dirent.h as taking pointers to differently typed opaque
 640          * structures.
 641          */
 642         struct dirent *(*readdirfunc)();
 643
 644         if (pathend > pathend_last)
 645                 return (GLOB_ABORTED);
 646         *pathend = EOS;
 647         errno = 0;
 648
 649         if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
 650                 /* TODO: don't call for ENOENT or ENOTDIR? */
 651                 if (pglob->gl_errfunc) {
 652                         if (g_Ctoc(pathbuf, buf, sizeof(buf)))
 653                                 return (GLOB_ABORTED);
 654                         if (pglob->gl_errfunc(buf, errno) ||
 655                             pglob->gl_flags & GLOB_ERR)
 656                                 return (GLOB_ABORTED);
 657                 }
 658                 return(0);
 659         }
 660
 661         err = 0;
 662
 663         /* Search directory for matching names. */
 664         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 665                 readdirfunc = pglob->gl_readdir;
 666         else
 667                 readdirfunc = readdir;
 668         while ((dp = (*readdirfunc)(dirp))) {
 669                 u_char *sc;
 670                 Char *dc;
 671                 wchar_t wc;
 672                 size_t clen;
 673                 mbstate_t mbs;
 674
 675                 /* Initial DOT must be matched literally. */
 676                 if (dp->d_name[0] == DOT && *pattern != DOT)
 677                         continue;
 678                 memset(&mbs, 0, sizeof(mbs));
 679                 dc = pathend;
 680                 sc = (u_char *) dp->d_name;
 681                 while (dc < pathend_last) {
 682                         clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs);
 683                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 684                                 wc = *sc;
 685                                 clen = 1;
 686                                 memset(&mbs, 0, sizeof(mbs));
 687                         }
 688                         if ((*dc++ = wc) == EOS)
 689                                 break;
 690                         sc += clen;
 691                 }
 692                 if (!match(pathend, pattern, restpattern)) {
 693                         *pathend = EOS;
 694                         continue;
 695                 }
 696                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 697                     pglob, limit);
 698                 if (err)
 699                         break;
 700         }
 701
 702         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 703                 (*pglob->gl_closedir)(dirp);
 704         else
 705                 closedir(dirp);
 706         return(err);
 707 }
 708
 709
 710 /*
 711  * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
 712  * add the new item, and update gl_pathc.
 713  *
 714  * This assumes the BSD realloc, which only copies the block when its size
 715  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 716  * behavior.
 717  *
 718  * Return 0 if new item added, error code if memory couldn't be allocated.
 719  *
 720  * Invariant of the glob_t structure:
 721  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 722  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 723  */
 724 static int
 725 globextend(path, pglob, limit)
 726         const Char *path;
 727         glob_t *pglob;
 728         int *limit;
 729 {
 730         char **pathv;
 731         int i;
 732         u_int newsize, len;
 733         char *copy;
 734         const Char *p;
 735
 736         if (*limit && pglob->gl_pathc > *limit) {
 737                 errno = 0;
 738                 return (GLOB_NOSPACE);
 739         }
 740
 741         newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
 742         pathv = pglob->gl_pathv ?
 743                     realloc((char *)pglob->gl_pathv, newsize) :
 744                     malloc(newsize);
 745         if (pathv == NULL) {
 746                 if (pglob->gl_pathv) {
 747                         free(pglob->gl_pathv);
 748                         pglob->gl_pathv = NULL;
 749                 }
 750                 return(GLOB_NOSPACE);
 751         }
 752
 753         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 754                 /* first time around -- clear initial gl_offs items */
 755                 pathv += pglob->gl_offs;
 756                 for (i = pglob->gl_offs; --i >= 0; )
 757                         *--pathv = NULL;
 758         }
 759         pglob->gl_pathv = pathv;
 760
 761         for (p = path; *p++;)
 762                 continue;
 763         len = MB_CUR_MAX * (size_t)(p - path);  /* XXX overallocation */
 764         if ((copy = malloc(len)) != NULL) {
 765                 if (g_Ctoc(path, copy, len)) {
 766                         free(copy);
 767                         return (GLOB_NOSPACE);
 768                 }
 769                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 770         }
 771         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 772         return(copy == NULL ? GLOB_NOSPACE : 0);
 773 }
 774
 775 /*
 776  * pattern matching function for filenames.  Each occurrence of the *
 777  * pattern causes a recursion level.
 778  */
 779 static int
 780 match(name, pat, patend)
 781         Char *name, *pat, *patend;
 782 {
 783         int ok, negate_range;
 784         Char c, k;
 785
 786         while (pat < patend) {
 787                 c = *pat++;
 788                 switch (c & M_MASK) {
 789                 case M_ALL:
 790                         if (pat == patend)
 791                                 return(1);
 792                         do
 793                             if (match(name, pat, patend))
 794                                     return(1);
 795                         while (*name++ != EOS);
 796                         return(0);
 797                 case M_ONE:
 798                         if (*name++ == EOS)
 799                                 return(0);
 800                         break;
 801                 case M_SET:
 802                         ok = 0;
 803                         if ((k = *name++) == EOS)
 804                                 return(0);
 805                         if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 806                                 ++pat;
 807                         while (((c = *pat++) & M_MASK) != M_END)
 808                                 if ((*pat & M_MASK) == M_RNG) {
 809                                         if (__collate_load_error ?
 810                                             CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1]) :
 811                                                __collate_range_cmp(CHAR(c), CHAR(k)) <= 0
 812                                             && __collate_range_cmp(CHAR(k), CHAR(pat[1])) <= 0
 813                                            )
 814                                                 ok = 1;
 815                                         pat += 2;
 816                                 } else if (c == k)
 817                                         ok = 1;
 818                         if (ok == negate_range)
 819                                 return(0);
 820                         break;
 821                 default:
 822                         if (*name++ != c)
 823                                 return(0);
 824                         break;
 825                 }
 826         }
 827         return(*name == EOS);
 828 }
 829
 830 /* Free allocated data belonging to a glob_t structure. */
 831 void
 832 globfree(pglob)
 833         glob_t *pglob;
 834 {
 835         int i;
 836         char **pp;
 837
 838         if (pglob->gl_pathv != NULL) {
 839                 pp = pglob->gl_pathv + pglob->gl_offs;
 840                 for (i = pglob->gl_pathc; i--; ++pp)
 841                         if (*pp)
 842                                 free(*pp);
 843                 free(pglob->gl_pathv);
 844                 pglob->gl_pathv = NULL;
 845         }
 846 }
 847
 848 static DIR *
 849 g_opendir(str, pglob)
 850         Char *str;
 851         glob_t *pglob;
 852 {
 853         char buf[MAXPATHLEN];
 854
 855         if (!*str)
 856                 strcpy(buf, ".");
 857         else {
 858                 if (g_Ctoc(str, buf, sizeof(buf)))
 859                         return (NULL);
 860         }
 861
 862         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 863                 return((*pglob->gl_opendir)(buf));
 864
 865         return(opendir(buf));
 866 }
 867
 868 static int
 869 g_lstat(fn, sb, pglob)
 870         Char *fn;
 871         struct stat *sb;
 872         glob_t *pglob;
 873 {
 874         char buf[MAXPATHLEN];
 875
 876         if (g_Ctoc(fn, buf, sizeof(buf))) {
 877                 errno = ENAMETOOLONG;
 878                 return (-1);
 879         }
 880         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 881                 return((*pglob->gl_lstat)(buf, sb));
 882         return(lstat(buf, sb));
 883 }
 884
 885 static int
 886 g_stat(fn, sb, pglob)
 887         Char *fn;
 888         struct stat *sb;
 889         glob_t *pglob;
 890 {
 891         char buf[MAXPATHLEN];
 892
 893         if (g_Ctoc(fn, buf, sizeof(buf))) {
 894                 errno = ENAMETOOLONG;
 895                 return (-1);
 896         }
 897         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 898                 return((*pglob->gl_stat)(buf, sb));
 899         return(stat(buf, sb));
 900 }
 901
 902 static Char *
 903 g_strchr(str, ch)
 904         Char *str;
 905         wchar_t ch;
 906 {
 907         do {
 908                 if (*str == ch)
 909                         return (str);
 910         } while (*str++);
 911         return (NULL);
 912 }
 913
 914 static int
 915 g_Ctoc(str, buf, len)
 916         const Char *str;
 917         char *buf;
 918         u_int len;
 919 {
 920         mbstate_t mbs;
 921         size_t clen;
 922
 923         memset(&mbs, 0, sizeof(mbs));
 924         while (len >= MB_CUR_MAX) {
 925                 clen = wcrtomb(buf, *str, &mbs);
 926                 if (clen == (size_t)-1)
 927                         return (1);
 928                 if (*str == L'\0')
 929                         return (0);
 930                 str++;
 931                 buf += clen;
 932                 len -= clen;
 933         }
 934         return (1);
 935 }
 936
 937 #ifdef DEBUG
 938 static void
 939 qprintf(str, s)
 940         const char *str;
 941         Char *s;
 942 {
 943         Char *p;
 944
 945         (void)printf("%s:\n", str);
 946         for (p = s; *p; p++)
 947                 (void)printf("%c", CHAR(*p));
 948         (void)printf("\n");
 949         for (p = s; *p; p++)
 950                 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
 951         (void)printf("\n");
 952         for (p = s; *p; p++)
 953                 (void)printf("%c", ismeta(*p) ? '_' : ' ');
 954         (void)printf("\n");
 955 }
 956 #endif