lib/libc/gen/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  * 4. Neither the name of the University nor the names of its contributors
  17  *    may be used to endorse or promote products derived from this software
  18  *    without specific prior written permission.
  19  *
  20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  30  * SUCH DAMAGE.
  31  *
  32  * @(#)glob.c   8.3 (Berkeley) 10/13/93
  33  * $FreeBSD: src/lib/libc/gen/glob.c,v 1.27 2008/06/26 07:12:35 mtm Exp $
  34  * $DragonFly: src/lib/libc/gen/glob.c,v 1.6 2005/12/07 02:28:15 corecode Exp $
  35  */
  36
  37 /*
  38  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  39  *
  40  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  41  *
  42  * Optional extra services, controlled by flags not defined by POSIX:
  43  *
  44  * GLOB_QUOTE:
  45  *      Escaping convention: \ inhibits any special meaning the following
  46  *      character might have (except \ at end of string is retained).
  47  * GLOB_MAGCHAR:
  48  *      Set in gl_flags if pattern contained a globbing character.
  49  * GLOB_NOMAGIC:
  50  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  51  *      not contain any magic characters.  [Used in csh style globbing]
  52  * GLOB_ALTDIRFUNC:
  53  *      Use alternately specified directory access functions.
  54  * GLOB_TILDE:
  55  *      expand ~user/foo to the /home/dir/of/user/foo
  56  * GLOB_BRACE:
  57  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  58  * gl_matchc:
  59  *      Number of matches in the current invocation of glob.
  60  */
  61
  62 /*
  63  * Some notes on multibyte character support:
  64  * 1. Patterns with illegal byte sequences match nothing - even if
  65  *    GLOB_NOCHECK is specified.
  66  * 2. Illegal byte sequences in filenames are handled by treating them as
  67  *    single-byte characters with a value of the first byte of the sequence
  68  *    cast to wchar_t.
  69  * 3. State-dependent encodings are not currently supported.
  70  */
  71
  72 #include <sys/param.h>
  73 #include <sys/stat.h>
  74
  75 #include <ctype.h>
  76 #include <dirent.h>
  77 #include <errno.h>
  78 #include <glob.h>
  79 #include <limits.h>
  80 #include <pwd.h>
  81 #include <stdint.h>
  82 #include <stdio.h>
  83 #include <stdlib.h>
  84 #include <string.h>
  85 #include <unistd.h>
  86 #include <wchar.h>
  87
  88 #include "collate.h"
  89
  90 #define DOLLAR          '$'
  91 #define DOT             '.'
  92 #define EOS             '\0'
  93 #define LBRACKET        '['
  94 #define NOT             '!'
  95 #define QUESTION        '?'
  96 #define QUOTE           '\\'
  97 #define RANGE           '-'
  98 #define RBRACKET        ']'
  99 #define SEP             '/'
 100 #define STAR            '*'
 101 #define TILDE           '~'
 102 #define UNDERSCORE      '_'
 103 #define LBRACE          '{'
 104 #define RBRACE          '}'
 105 #define SLASH           '/'
 106 #define COMMA           ','
 107
 108 #ifndef DEBUG
 109
 110 #define M_QUOTE         0x8000000000ULL
 111 #define M_PROTECT       0x4000000000ULL
 112 #define M_MASK          0xffffffffffULL
 113 #define M_CHAR          0x00ffffffffULL
 114
 115 typedef uint_fast64_t Char;
 116
 117 #else
 118
 119 #define M_QUOTE         0x80
 120 #define M_PROTECT       0x40
 121 #define M_MASK          0xff
 122 #define M_CHAR          0x7f
 123
 124 typedef char Char;
 125
 126 #endif
 127
 128
 129 #define CHAR(c)         ((Char)((c)&M_CHAR))
 130 #define META(c)         ((Char)((c)|M_QUOTE))
 131 #define M_ALL           META('*')
 132 #define M_END           META(']')
 133 #define M_NOT           META('!')
 134 #define M_ONE           META('?')
 135 #define M_RNG           META('-')
 136 #define M_SET           META('[')
 137 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 138
 139
 140 static int       compare(const void *, const void *);
 141 static int       g_Ctoc(const Char *, char *, size_t);
 142 static int       g_lstat(const Char *, struct stat *, glob_t *);
 143 static DIR      *g_opendir(const Char *, glob_t *);
 144 static const Char *g_strchr(const Char *, wchar_t);
 145 #ifdef notdef
 146 static Char     *g_strcat(Char *, const Char *);
 147 #endif
 148 static int       g_stat(const Char *, struct stat *, glob_t *);
 149 static int       glob0(const Char *, glob_t *, size_t *);
 150 static int       glob1(const Char *, glob_t *, size_t *);
 151 static int       glob2(Char *, Char *, Char *, const Char *, glob_t *,
 152                        size_t *);
 153 static int       glob3(Char *, Char *, Char *, const Char *, const Char *,
 154                        glob_t *, size_t *);
 155 static int       globextend(const Char *, glob_t *, size_t *);
 156 static const Char *
 157                  globtilde(const Char *, Char *, size_t, glob_t *);
 158 static int       globexp1(const Char *, glob_t *, size_t *);
 159 static int       globexp2(const Char *, const Char *, glob_t *, int *, size_t *);
 160 static int       match(const Char *, const Char *, const Char *);
 161 #ifdef DEBUG
 162 static void      qprintf(const char *, const Char *);
 163 #endif
 164
 165 int
 166 glob(const char *pattern, int flags, int (*errfunc)(const char *, int),
 167      glob_t *pglob)
 168 {
 169         const char *patnext;
 170         size_t limit;
 171         Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
 172         mbstate_t mbs;
 173         wchar_t wc;
 174         size_t clen;
 175
 176         patnext = pattern;
 177         if (!(flags & GLOB_APPEND)) {
 178                 pglob->gl_pathc = 0;
 179                 pglob->gl_pathv = NULL;
 180                 if (!(flags & GLOB_DOOFFS))
 181                         pglob->gl_offs = 0;
 182         }
 183         if (flags & GLOB_LIMIT) {
 184                 limit = pglob->gl_matchc;
 185                 if (limit == 0)
 186                         limit = ARG_MAX;
 187         } else
 188                 limit = 0;
 189         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 190         pglob->gl_errfunc = errfunc;
 191         pglob->gl_matchc = 0;
 192
 193         bufnext = patbuf;
 194         bufend = bufnext + MAXPATHLEN - 1;
 195         if (flags & GLOB_NOESCAPE) {
 196                 memset(&mbs, 0, sizeof(mbs));
 197                 while (bufend - bufnext >= MB_CUR_MAX) {
 198                         clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
 199                         if (clen == (size_t)-1 || clen == (size_t)-2)
 200                                 return (GLOB_NOMATCH);
 201                         else if (clen == 0)
 202                                 break;
 203                         *bufnext++ = wc;
 204                         patnext += clen;
 205                 }
 206         } else {
 207                 /* Protect the quoted characters. */
 208                 memset(&mbs, 0, sizeof(mbs));
 209                 while (bufend - bufnext >= MB_CUR_MAX) {
 210                         if (*patnext == QUOTE) {
 211                                 if (*++patnext == EOS) {
 212                                         *bufnext++ = QUOTE | M_PROTECT;
 213                                         continue;
 214                                 }
 215                                 prot = M_PROTECT;
 216                         } else
 217                                 prot = 0;
 218                         clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
 219                         if (clen == (size_t)-1 || clen == (size_t)-2)
 220                                 return (GLOB_NOMATCH);
 221                         else if (clen == 0)
 222                                 break;
 223                         *bufnext++ = wc | prot;
 224                         patnext += clen;
 225                 }
 226         }
 227         *bufnext = EOS;
 228
 229         if (flags & GLOB_BRACE)
 230             return globexp1(patbuf, pglob, &limit);
 231         else
 232             return glob0(patbuf, pglob, &limit);
 233 }
 234
 235 /*
 236  * Expand recursively a glob {} pattern. When there is no more expansion
 237  * invoke the standard globbing routine to glob the rest of the magic
 238  * characters
 239  */
 240 static int
 241 globexp1(const Char *pattern, glob_t *pglob, size_t *limit)
 242 {
 243         const Char *ptr = pattern;
 244         int rv;
 245
 246         /* Protect a single {}, for find(1), like csh */
 247         if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
 248                 return glob0(pattern, pglob, limit);
 249
 250         while ((ptr = g_strchr(ptr, LBRACE)) != NULL)
 251                 if (!globexp2(ptr, pattern, pglob, &rv, limit))
 252                         return rv;
 253
 254         return glob0(pattern, pglob, limit);
 255 }
 256
 257
 258 /*
 259  * Recursive brace globbing helper. Tries to expand a single brace.
 260  * If it succeeds then it invokes globexp1 with the new pattern.
 261  * If it fails then it tries to glob the rest of the pattern and returns.
 262  */
 263 static int
 264 globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv,
 265          size_t *limit)
 266 {
 267         int     i;
 268         Char   *lm, *ls;
 269         const Char *pe, *pm, *pm1, *pl;
 270         Char    patbuf[MAXPATHLEN];
 271
 272         /* copy part up to the brace */
 273         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 274                 continue;
 275         *lm = EOS;
 276         ls = lm;
 277
 278         /* Find the balanced brace */
 279         for (i = 0, pe = ++ptr; *pe; pe++)
 280                 if (*pe == LBRACKET) {
 281                         /* Ignore everything between [] */
 282                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
 283                                 continue;
 284                         if (*pe == EOS) {
 285                                 /*
 286                                  * We could not find a matching RBRACKET.
 287                                  * Ignore and just look for RBRACE
 288                                  */
 289                                 pe = pm;
 290                         }
 291                 }
 292                 else if (*pe == LBRACE)
 293                         i++;
 294                 else if (*pe == RBRACE) {
 295                         if (i == 0)
 296                                 break;
 297                         i--;
 298                 }
 299
 300         /* Non matching braces; just glob the pattern */
 301         if (i != 0 || *pe == EOS) {
 302                 *rv = glob0(patbuf, pglob, limit);
 303                 return 0;
 304         }
 305
 306         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 307                 switch (*pm) {
 308                 case LBRACKET:
 309                         /* Ignore everything between [] */
 310                         for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++)
 311                                 continue;
 312                         if (*pm == EOS) {
 313                                 /*
 314                                  * We could not find a matching RBRACKET.
 315                                  * Ignore and just look for RBRACE
 316                                  */
 317                                 pm = pm1;
 318                         }
 319                         break;
 320
 321                 case LBRACE:
 322                         i++;
 323                         break;
 324
 325                 case RBRACE:
 326                         if (i) {
 327                             i--;
 328                             break;
 329                         }
 330                         /* FALLTHROUGH */
 331                 case COMMA:
 332                         if (i && *pm == COMMA)
 333                                 break;
 334                         else {
 335                                 /* Append the current string */
 336                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 337                                         continue;
 338                                 /*
 339                                  * Append the rest of the pattern after the
 340                                  * closing brace
 341                                  */
 342                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 343                                         continue;
 344
 345                                 /* Expand the current pattern */
 346 #ifdef DEBUG
 347                                 qprintf("globexp2:", patbuf);
 348 #endif
 349                                 *rv = globexp1(patbuf, pglob, limit);
 350
 351                                 /* move after the comma, to the next string */
 352                                 pl = pm + 1;
 353                         }
 354                         break;
 355
 356                 default:
 357                         break;
 358                 }
 359         *rv = 0;
 360         return 0;
 361 }
 362
 363
 364
 365 /*
 366  * expand tilde from the passwd file.
 367  */
 368 static const Char *
 369 globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
 370 {
 371         struct passwd *pwd;
 372         char *h;
 373         const Char *p;
 374         Char *b, *eb;
 375
 376         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 377                 return pattern;
 378
 379         /*
 380          * Copy up to the end of the string or /
 381          */
 382         eb = &patbuf[patbuf_len - 1];
 383         for (p = pattern + 1, h = (char *) patbuf;
 384             h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
 385                 continue;
 386
 387         *h = EOS;
 388
 389         if (((char *) patbuf)[0] == EOS) {
 390                 /*
 391                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 392                  * we're not running setuid or setgid) and then trying
 393                  * the password file
 394                  */
 395                 if (issetugid() != 0 ||
 396                     (h = getenv("HOME")) == NULL) {
 397                         if (((h = getlogin()) != NULL &&
 398                              (pwd = getpwnam(h)) != NULL) ||
 399                             (pwd = getpwuid(getuid())) != NULL)
 400                                 h = pwd->pw_dir;
 401                         else
 402                                 return pattern;
 403                 }
 404         }
 405         else {
 406                 /*
 407                  * Expand a ~user
 408                  */
 409                 if ((pwd = getpwnam((char*) patbuf)) == NULL)
 410                         return pattern;
 411                 else
 412                         h = pwd->pw_dir;
 413         }
 414
 415         /* Copy the home directory */
 416         for (b = patbuf; b < eb && *h; *b++ = *h++)
 417                 continue;
 418
 419         /* Append the rest of the pattern */
 420         while (b < eb && (*b++ = *p++) != EOS)
 421                 continue;
 422         *b = EOS;
 423
 424         return patbuf;
 425 }
 426
 427
 428 /*
 429  * The main glob() routine: compiles the pattern (optionally processing
 430  * quotes), calls glob1() to do the real pattern matching, and finally
 431  * sorts the list (unless unsorted operation is requested).  Returns 0
 432  * if things went well, nonzero if errors occurred.
 433  */
 434 static int
 435 glob0(const Char *pattern, glob_t *pglob, size_t *limit)
 436 {
 437         const Char *qpatnext;
 438         int c, err;
 439         size_t oldpathc;
 440         Char *bufnext, patbuf[MAXPATHLEN];
 441
 442         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
 443         oldpathc = pglob->gl_pathc;
 444         bufnext = patbuf;
 445
 446         /* We don't need to check for buffer overflow any more. */
 447         while ((c = *qpatnext++) != EOS) {
 448                 switch (c) {
 449                 case LBRACKET:
 450                         c = *qpatnext;
 451                         if (c == NOT)
 452                                 ++qpatnext;
 453                         if (*qpatnext == EOS ||
 454                             g_strchr(qpatnext+1, RBRACKET) == NULL) {
 455                                 *bufnext++ = LBRACKET;
 456                                 if (c == NOT)
 457                                         --qpatnext;
 458                                 break;
 459                         }
 460                         *bufnext++ = M_SET;
 461                         if (c == NOT)
 462                                 *bufnext++ = M_NOT;
 463                         c = *qpatnext++;
 464                         do {
 465                                 *bufnext++ = CHAR(c);
 466                                 if (*qpatnext == RANGE &&
 467                                     (c = qpatnext[1]) != RBRACKET) {
 468                                         *bufnext++ = M_RNG;
 469                                         *bufnext++ = CHAR(c);
 470                                         qpatnext += 2;
 471                                 }
 472                         } while ((c = *qpatnext++) != RBRACKET);
 473                         pglob->gl_flags |= GLOB_MAGCHAR;
 474                         *bufnext++ = M_END;
 475                         break;
 476                 case QUESTION:
 477                         pglob->gl_flags |= GLOB_MAGCHAR;
 478                         *bufnext++ = M_ONE;
 479                         break;
 480                 case STAR:
 481                         pglob->gl_flags |= GLOB_MAGCHAR;
 482                         /* collapse adjacent stars to one,
 483                          * to avoid exponential behavior
 484                          */
 485                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 486                             *bufnext++ = M_ALL;
 487                         break;
 488                 default:
 489                         *bufnext++ = CHAR(c);
 490                         break;
 491                 }
 492         }
 493         *bufnext = EOS;
 494 #ifdef DEBUG
 495         qprintf("glob0:", patbuf);
 496 #endif
 497
 498         if ((err = glob1(patbuf, pglob, limit)) != 0)
 499                 return(err);
 500
 501         /*
 502          * If there was no match we are going to append the pattern
 503          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
 504          * and the pattern did not contain any magic characters
 505          * GLOB_NOMAGIC is there just for compatibility with csh.
 506          */
 507         if (pglob->gl_pathc == oldpathc) {
 508                 if (((pglob->gl_flags & GLOB_NOCHECK) ||
 509                     ((pglob->gl_flags & GLOB_NOMAGIC) &&
 510                         !(pglob->gl_flags & GLOB_MAGCHAR))))
 511                         return(globextend(pattern, pglob, limit));
 512                 else
 513                         return(GLOB_NOMATCH);
 514         }
 515         if (!(pglob->gl_flags & GLOB_NOSORT))
 516                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 517                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 518         return(0);
 519 }
 520
 521 static int
 522 compare(const void *p, const void *q)
 523 {
 524         return(strcmp(*(const char * const *)p, *(const char * const *)q));
 525 }
 526
 527 static int
 528 glob1(const Char *pattern, glob_t *pglob, size_t *limit)
 529 {
 530         Char pathbuf[MAXPATHLEN];
 531
 532         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 533         if (*pattern == EOS)
 534                 return(0);
 535         return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 536             pattern, pglob, limit));
 537 }
 538
 539 /*
 540  * The functions glob2 and glob3 are mutually recursive; there is one level
 541  * of recursion for each segment in the pattern that contains one or more
 542  * meta characters.
 543  */
 544 static int
 545 glob2(Char *pathbuf, Char *pathend, Char *pathend_last, const Char *pattern,
 546       glob_t *pglob, size_t *limit)
 547 {
 548         struct stat sb;
 549         const Char *p;
 550         Char *q;
 551         int anymeta;
 552
 553         /*
 554          * Loop over pattern segments until end of pattern or until
 555          * segment with meta character found.
 556          */
 557         for (anymeta = 0;;) {
 558                 if (*pattern == EOS) {          /* End of pattern? */
 559                         *pathend = EOS;
 560                         if (g_lstat(pathbuf, &sb, pglob))
 561                                 return(0);
 562
 563                         if (((pglob->gl_flags & GLOB_MARK) &&
 564                             pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
 565                             || (S_ISLNK(sb.st_mode) &&
 566                             (g_stat(pathbuf, &sb, pglob) == 0) &&
 567                             S_ISDIR(sb.st_mode)))) {
 568                                 if (pathend + 1 > pathend_last)
 569                                         return (GLOB_ABORTED);
 570                                 *pathend++ = SEP;
 571                                 *pathend = EOS;
 572                         }
 573                         ++pglob->gl_matchc;
 574                         return(globextend(pathbuf, pglob, limit));
 575                 }
 576
 577                 /* Find end of next segment, copy tentatively to pathend. */
 578                 q = pathend;
 579                 p = pattern;
 580                 while (*p != EOS && *p != SEP) {
 581                         if (ismeta(*p))
 582                                 anymeta = 1;
 583                         if (q + 1 > pathend_last)
 584                                 return (GLOB_ABORTED);
 585                         *q++ = *p++;
 586                 }
 587
 588                 if (!anymeta) {         /* No expansion, do next segment. */
 589                         pathend = q;
 590                         pattern = p;
 591                         while (*pattern == SEP) {
 592                                 if (pathend + 1 > pathend_last)
 593                                         return (GLOB_ABORTED);
 594                                 *pathend++ = *pattern++;
 595                         }
 596                 } else                  /* Need expansion, recurse. */
 597                         return(glob3(pathbuf, pathend, pathend_last, pattern, p,
 598                             pglob, limit));
 599         }
 600         /* NOTREACHED */
 601 }
 602
 603 static int
 604 glob3(Char *pathbuf, Char *pathend, Char *pathend_last, const Char *pattern,
 605       const Char *restpattern, glob_t *pglob, size_t *limit)
 606 {
 607         struct dirent *dp;
 608         DIR *dirp;
 609         int err;
 610         char buf[MAXPATHLEN];
 611         struct dirent *(*readdirfunc)(DIR *);
 612
 613         if (pathend > pathend_last)
 614                 return (GLOB_ABORTED);
 615         *pathend = EOS;
 616         errno = 0;
 617
 618         if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
 619                 /* TODO: don't call for ENOENT or ENOTDIR? */
 620                 if (pglob->gl_errfunc) {
 621                         if (g_Ctoc(pathbuf, buf, sizeof(buf)))
 622                                 return (GLOB_ABORTED);
 623                         if (pglob->gl_errfunc(buf, errno) ||
 624                             pglob->gl_flags & GLOB_ERR)
 625                                 return (GLOB_ABORTED);
 626                 }
 627                 return(0);
 628         }
 629
 630         err = 0;
 631
 632         /* pglob->gl_readdir takes a void *, fix this manually. */
 633         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 634                 readdirfunc = (struct dirent *(*)(DIR *))pglob->gl_readdir;
 635         else
 636                 readdirfunc = readdir;
 637
 638         /* Search directory for matching names. */
 639         while ((dp = (*readdirfunc)(dirp)) != NULL) {
 640                 char *sc;
 641                 Char *dc;
 642                 wchar_t wc;
 643                 size_t clen;
 644                 mbstate_t mbs;
 645
 646                 /* Initial DOT must be matched literally. */
 647                 if (dp->d_name[0] == DOT && *pattern != DOT &&
 648                     !(pglob->gl_flags & GLOB_PERIOD))
 649                         continue;
 650                 memset(&mbs, 0, sizeof(mbs));
 651                 dc = pathend;
 652                 sc = dp->d_name;
 653                 while (dc < pathend_last) {
 654                         clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs);
 655                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 656                                 wc = *sc;
 657                                 clen = 1;
 658                                 memset(&mbs, 0, sizeof(mbs));
 659                         }
 660                         if ((*dc++ = wc) == EOS)
 661                                 break;
 662                         sc += clen;
 663                 }
 664                 if (!match(pathend, pattern, restpattern)) {
 665                         *pathend = EOS;
 666                         continue;
 667                 }
 668                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 669                     pglob, limit);
 670                 if (err)
 671                         break;
 672         }
 673
 674         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 675                 (*pglob->gl_closedir)(dirp);
 676         else
 677                 closedir(dirp);
 678         return(err);
 679 }
 680
 681
 682 /*
 683  * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
 684  * add the new item, and update gl_pathc.
 685  *
 686  * This assumes the BSD realloc, which only copies the block when its size
 687  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 688  * behavior.
 689  *
 690  * Return 0 if new item added, error code if memory couldn't be allocated.
 691  *
 692  * Invariant of the glob_t structure:
 693  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 694  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 695  */
 696 static int
 697 globextend(const Char *path, glob_t *pglob, size_t *limit)
 698 {
 699         char **pathv;
 700         size_t i, newsize, len;
 701         char *copy;
 702         const Char *p;
 703
 704         if (*limit && pglob->gl_pathc > *limit) {
 705                 errno = 0;
 706                 return (GLOB_NOSPACE);
 707         }
 708
 709         newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
 710         pathv = pglob->gl_pathv ?
 711                     realloc((char *)pglob->gl_pathv, newsize) :
 712                     malloc(newsize);
 713         if (pathv == NULL) {
 714                 if (pglob->gl_pathv) {
 715                         free(pglob->gl_pathv);
 716                         pglob->gl_pathv = NULL;
 717                 }
 718                 return(GLOB_NOSPACE);
 719         }
 720
 721         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 722                 /* first time around -- clear initial gl_offs items */
 723                 pathv += pglob->gl_offs;
 724                 for (i = pglob->gl_offs + 1; --i > 0; )
 725                         *--pathv = NULL;
 726         }
 727         pglob->gl_pathv = pathv;
 728
 729         for (p = path; *p++;)
 730                 continue;
 731         len = MB_CUR_MAX * (size_t)(p - path);  /* XXX overallocation */
 732         if ((copy = malloc(len)) != NULL) {
 733                 if (g_Ctoc(path, copy, len)) {
 734                         free(copy);
 735                         return (GLOB_NOSPACE);
 736                 }
 737                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 738         }
 739         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 740         return(copy == NULL ? GLOB_NOSPACE : 0);
 741 }
 742
 743 /*
 744  * pattern matching function for filenames.  Each occurrence of the *
 745  * pattern causes a recursion level.
 746  */
 747 static int
 748 match(const Char * name, const Char *pat, const Char *patend)
 749 {
 750         int ok, negate_range;
 751         Char c, k;
 752
 753         while (pat < patend) {
 754                 c = *pat++;
 755                 switch (c & M_MASK) {
 756                 case M_ALL:
 757                         if (pat == patend)
 758                                 return(1);
 759                         do
 760                             if (match(name, pat, patend))
 761                                     return(1);
 762                         while (*name++ != EOS);
 763                         return(0);
 764                 case M_ONE:
 765                         if (*name++ == EOS)
 766                                 return(0);
 767                         break;
 768                 case M_SET:
 769                         ok = 0;
 770                         if ((k = *name++) == EOS)
 771                                 return(0);
 772                         if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 773                                 ++pat;
 774                         while (((c = *pat++) & M_MASK) != M_END)
 775                                 if ((*pat & M_MASK) == M_RNG) {
 776                                         if (__collate_load_error ?
 777                                             CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1]) :
 778                                                __collate_range_cmp(CHAR(c), CHAR(k)) <= 0
 779                                             && __collate_range_cmp(CHAR(k), CHAR(pat[1])) <= 0
 780                                            )
 781                                                 ok = 1;
 782                                         pat += 2;
 783                                 } else if (c == k)
 784                                         ok = 1;
 785                         if (ok == negate_range)
 786                                 return(0);
 787                         break;
 788                 default:
 789                         if (*name++ != c)
 790                                 return(0);
 791                         break;
 792                 }
 793         }
 794         return(*name == EOS);
 795 }
 796
 797 /* Free allocated data belonging to a glob_t structure. */
 798 void
 799 globfree(glob_t *pglob)
 800 {
 801         size_t i;
 802         char **pp;
 803
 804         if (pglob->gl_pathv != NULL) {
 805                 pp = pglob->gl_pathv + pglob->gl_offs;
 806                 for (i = pglob->gl_pathc; i--; ++pp)
 807                         if (*pp)
 808                                 free(*pp);
 809                 free(pglob->gl_pathv);
 810                 pglob->gl_pathv = NULL;
 811         }
 812 }
 813
 814 static DIR *
 815 g_opendir(const Char *str, glob_t *pglob)
 816 {
 817         char buf[MAXPATHLEN];
 818
 819         if (*str == '\0')
 820                 strcpy(buf, ".");
 821         else if (g_Ctoc(str, buf, sizeof(buf)))
 822                 return (NULL);
 823
 824         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 825                 return((*pglob->gl_opendir)(buf));
 826
 827         return(opendir(buf));
 828 }
 829
 830 static int
 831 g_lstat(const Char *fn, struct stat *sb, glob_t *pglob)
 832 {
 833         char buf[MAXPATHLEN];
 834
 835         if (g_Ctoc(fn, buf, sizeof(buf))) {
 836                 errno = ENAMETOOLONG;
 837                 return (-1);
 838         }
 839         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 840                 return((*pglob->gl_lstat)(buf, sb));
 841         return(lstat(buf, sb));
 842 }
 843
 844 static int
 845 g_stat(const Char *fn, struct stat *sb, glob_t *pglob)
 846 {
 847         char buf[MAXPATHLEN];
 848
 849         if (g_Ctoc(fn, buf, sizeof(buf))) {
 850                 errno = ENAMETOOLONG;
 851                 return (-1);
 852         }
 853         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 854                 return((*pglob->gl_stat)(buf, sb));
 855         return(stat(buf, sb));
 856 }
 857
 858 static const Char *
 859 g_strchr(const Char *str, wchar_t ch)
 860 {
 861         for (;; ++str) {
 862                 if (*str == ch)
 863                         return(str);
 864                 if (*str == '\0')
 865                         return(NULL);
 866         }
 867 }
 868
 869 static int
 870 g_Ctoc(const Char *str, char *buf, size_t len)
 871 {
 872         mbstate_t mbs;
 873         size_t clen;
 874
 875         memset(&mbs, 0, sizeof(mbs));
 876         while (len >= MB_CUR_MAX) {
 877                 clen = wcrtomb(buf, *str, &mbs);
 878                 if (clen == (size_t)-1)
 879                         return (1);
 880                 if (*str == L'\0')
 881                         return (0);
 882                 str++;
 883                 buf += clen;
 884                 len -= clen;
 885         }
 886         return (1);
 887 }
 888
 889 #ifdef DEBUG
 890 static void
 891 qprintf(const char *str, const Char *s)
 892 {
 893         const Char *p;
 894
 895         printf("%s:\n", str);
 896         for (p = s; *p; p++)
 897                 printf("%c", CHAR(*p));
 898         printf("\n");
 899         for (p = s; *p; p++)
 900                 printf("%c", *p & M_PROTECT ? '"' : ' ');
 901         printf("\n");
 902         for (p = s; *p; p++)
 903                 printf("%c", ismeta(*p) ? '_' : ' ');
 904         printf("\n");
 905 }
 906 #endif