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  * 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
  33 #if defined(LIBC_SCCS) && !defined(lint)
  34 static char sccsid[] = "@(#)glob.c      8.3 (Berkeley) 10/13/93";
  35 #endif /* LIBC_SCCS and not lint */
  36 #include <sys/cdefs.h>
  37 __FBSDID("$FreeBSD: src/lib/libc/gen/glob.c,v 1.28 2010/05/12 17:44:00 gordon Exp $");
  38
  39 /*
  40  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  41  *
  42  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  43  *
  44  * Optional extra services, controlled by flags not defined by POSIX:
  45  *
  46  * GLOB_QUOTE:
  47  *      Escaping convention: \ inhibits any special meaning the following
  48  *      character might have (except \ at end of string is retained).
  49  * GLOB_MAGCHAR:
  50  *      Set in gl_flags if pattern contained a globbing character.
  51  * GLOB_NOMAGIC:
  52  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  53  *      not contain any magic characters.  [Used in csh style globbing]
  54  * GLOB_ALTDIRFUNC:
  55  *      Use alternately specified directory access functions.
  56  * GLOB_TILDE:
  57  *      expand ~user/foo to the /home/dir/of/user/foo
  58  * GLOB_BRACE:
  59  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  60  * gl_matchc:
  61  *      Number of matches in the current invocation of glob.
  62  */
  63
  64 /*
  65  * Some notes on multibyte character support:
  66  * 1. Patterns with illegal byte sequences match nothing - even if
  67  *    GLOB_NOCHECK is specified.
  68  * 2. Illegal byte sequences in filenames are handled by treating them as
  69  *    single-byte characters with a value of the first byte of the sequence
  70  *    cast to wchar_t.
  71  * 3. State-dependent encodings are not currently supported.
  72  */
  73
  74 #include <sys/param.h>
  75 #include <sys/stat.h>
  76
  77 #include <ctype.h>
  78 #include <dirent.h>
  79 #include <errno.h>
  80 #include <glob.h>
  81 #include <limits.h>
  82 #include <pwd.h>
  83 #include <stdint.h>
  84 #include <stdio.h>
  85 #include <stdlib.h>
  86 #include <string.h>
  87 #include <unistd.h>
  88 #include <wchar.h>
  89
  90 #include "collate.h"
  91
  92 #define DOLLAR          '$'
  93 #define DOT             '.'
  94 #define EOS             '\0'
  95 #define LBRACKET        '['
  96 #define NOT             '!'
  97 #define QUESTION        '?'
  98 #define QUOTE           '\\'
  99 #define RANGE           '-'
 100 #define RBRACKET        ']'
 101 #define SEP             '/'
 102 #define STAR            '*'
 103 #define TILDE           '~'
 104 #define UNDERSCORE      '_'
 105 #define LBRACE          '{'
 106 #define RBRACE          '}'
 107 #define SLASH           '/'
 108 #define COMMA           ','
 109
 110 #ifndef DEBUG
 111
 112 #define M_QUOTE         0x8000000000ULL
 113 #define M_PROTECT       0x4000000000ULL
 114 #define M_MASK          0xffffffffffULL
 115 #define M_CHAR          0x00ffffffffULL
 116
 117 typedef uint_fast64_t Char;
 118
 119 #else
 120
 121 #define M_QUOTE         0x80
 122 #define M_PROTECT       0x40
 123 #define M_MASK          0xff
 124 #define M_CHAR          0x7f
 125
 126 typedef char Char;
 127
 128 #endif
 129
 130
 131 #define CHAR(c)         ((Char)((c)&M_CHAR))
 132 #define META(c)         ((Char)((c)|M_QUOTE))
 133 #define M_ALL           META('*')
 134 #define M_END           META(']')
 135 #define M_NOT           META('!')
 136 #define M_ONE           META('?')
 137 #define M_RNG           META('-')
 138 #define M_SET           META('[')
 139 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 140
 141
 142 static int       compare(const void *, const void *);
 143 static int       g_Ctoc(const Char *, char *, size_t);
 144 static int       g_lstat(Char *, struct stat *, glob_t *);
 145 static DIR      *g_opendir(Char *, glob_t *);
 146 static const Char *g_strchr(const Char *, wchar_t);
 147 #ifdef notdef
 148 static Char     *g_strcat(Char *, const Char *);
 149 #endif
 150 static int       g_stat(Char *, struct stat *, glob_t *);
 151 static int       glob0(const Char *, glob_t *, size_t *);
 152 static int       glob1(Char *, glob_t *, size_t *);
 153 static int       glob2(Char *, Char *, Char *, Char *, glob_t *, size_t *);
 154 static int       glob3(Char *, Char *, Char *, Char *, Char *, 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(Char *, Char *, Char *);
 161 #ifdef DEBUG
 162 static void      qprintf(const char *, Char *);
 163 #endif
 164
 165 int
 166 glob(const char *pattern, int flags, int (*errfunc)(const char *, int), glob_t *pglob)
 167 {
 168         const char *patnext;
 169         size_t limit;
 170         Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
 171         mbstate_t mbs;
 172         wchar_t wc;
 173         size_t clen;
 174
 175         patnext = pattern;
 176         if (!(flags & GLOB_APPEND)) {
 177                 pglob->gl_pathc = 0;
 178                 pglob->gl_pathv = NULL;
 179                 if (!(flags & GLOB_DOOFFS))
 180                         pglob->gl_offs = 0;
 181         }
 182         if (flags & GLOB_LIMIT) {
 183                 limit = pglob->gl_matchc;
 184                 if (limit == 0)
 185                         limit = ARG_MAX;
 186         } else
 187                 limit = 0;
 188         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 189         pglob->gl_errfunc = errfunc;
 190         pglob->gl_matchc = 0;
 191
 192         bufnext = patbuf;
 193         bufend = bufnext + MAXPATHLEN - 1;
 194         if (flags & GLOB_NOESCAPE) {
 195                 memset(&mbs, 0, sizeof(mbs));
 196                 while (bufend - bufnext >= MB_CUR_MAX) {
 197                         clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
 198                         if (clen == (size_t)-1 || clen == (size_t)-2)
 199                                 return (GLOB_NOMATCH);
 200                         else if (clen == 0)
 201                                 break;
 202                         *bufnext++ = wc;
 203                         patnext += clen;
 204                 }
 205         } else {
 206                 /* Protect the quoted characters. */
 207                 memset(&mbs, 0, sizeof(mbs));
 208                 while (bufend - bufnext >= MB_CUR_MAX) {
 209                         if (*patnext == QUOTE) {
 210                                 if (*++patnext == EOS) {
 211                                         *bufnext++ = QUOTE | M_PROTECT;
 212                                         continue;
 213                                 }
 214                                 prot = M_PROTECT;
 215                         } else
 216                                 prot = 0;
 217                         clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
 218                         if (clen == (size_t)-1 || clen == (size_t)-2)
 219                                 return (GLOB_NOMATCH);
 220                         else if (clen == 0)
 221                                 break;
 222                         *bufnext++ = wc | prot;
 223                         patnext += clen;
 224                 }
 225         }
 226         *bufnext = EOS;
 227
 228         if (flags & GLOB_BRACE)
 229             return globexp1(patbuf, pglob, &limit);
 230         else
 231             return glob0(patbuf, pglob, &limit);
 232 }
 233
 234 /*
 235  * Expand recursively a glob {} pattern. When there is no more expansion
 236  * invoke the standard globbing routine to glob the rest of the magic
 237  * characters
 238  */
 239 static int
 240 globexp1(const Char *pattern, glob_t *pglob, size_t *limit)
 241 {
 242         const Char* ptr = pattern;
 243         int rv;
 244
 245         /* Protect a single {}, for find(1), like csh */
 246         if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
 247                 return glob0(pattern, pglob, limit);
 248
 249         while ((ptr = g_strchr(ptr, LBRACE)) != NULL)
 250                 if (!globexp2(ptr, pattern, pglob, &rv, limit))
 251                         return rv;
 252
 253         return glob0(pattern, pglob, limit);
 254 }
 255
 256
 257 /*
 258  * Recursive brace globbing helper. Tries to expand a single brace.
 259  * If it succeeds then it invokes globexp1 with the new pattern.
 260  * If it fails then it tries to glob the rest of the pattern and returns.
 261  */
 262 static int
 263 globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv, size_t *limit)
 264 {
 265         int     i;
 266         Char   *lm, *ls;
 267         const Char *pe, *pm, *pm1, *pl;
 268         Char    patbuf[MAXPATHLEN];
 269
 270         /* copy part up to the brace */
 271         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 272                 continue;
 273         *lm = EOS;
 274         ls = lm;
 275
 276         /* Find the balanced brace */
 277         for (i = 0, pe = ++ptr; *pe; pe++)
 278                 if (*pe == LBRACKET) {
 279                         /* Ignore everything between [] */
 280                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
 281                                 continue;
 282                         if (*pe == EOS) {
 283                                 /*
 284                                  * We could not find a matching RBRACKET.
 285                                  * Ignore and just look for RBRACE
 286                                  */
 287                                 pe = pm;
 288                         }
 289                 }
 290                 else if (*pe == LBRACE)
 291                         i++;
 292                 else if (*pe == RBRACE) {
 293                         if (i == 0)
 294                                 break;
 295                         i--;
 296                 }
 297
 298         /* Non matching braces; just glob the pattern */
 299         if (i != 0 || *pe == EOS) {
 300                 *rv = glob0(patbuf, pglob, limit);
 301                 return 0;
 302         }
 303
 304         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 305                 switch (*pm) {
 306                 case LBRACKET:
 307                         /* Ignore everything between [] */
 308                         for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++)
 309                                 continue;
 310                         if (*pm == EOS) {
 311                                 /*
 312                                  * We could not find a matching RBRACKET.
 313                                  * Ignore and just look for RBRACE
 314                                  */
 315                                 pm = pm1;
 316                         }
 317                         break;
 318
 319                 case LBRACE:
 320                         i++;
 321                         break;
 322
 323                 case RBRACE:
 324                         if (i) {
 325                             i--;
 326                             break;
 327                         }
 328                         /* FALLTHROUGH */
 329                 case COMMA:
 330                         if (i && *pm == COMMA)
 331                                 break;
 332                         else {
 333                                 /* Append the current string */
 334                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 335                                         continue;
 336                                 /*
 337                                  * Append the rest of the pattern after the
 338                                  * closing brace
 339                                  */
 340                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 341                                         continue;
 342
 343                                 /* Expand the current pattern */
 344 #ifdef DEBUG
 345                                 qprintf("globexp2:", patbuf);
 346 #endif
 347                                 *rv = globexp1(patbuf, pglob, limit);
 348
 349                                 /* move after the comma, to the next string */
 350                                 pl = pm + 1;
 351                         }
 352                         break;
 353
 354                 default:
 355                         break;
 356                 }
 357         *rv = 0;
 358         return 0;
 359 }
 360
 361
 362
 363 /*
 364  * expand tilde from the passwd file.
 365  */
 366 static const Char *
 367 globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
 368 {
 369         struct passwd *pwd;
 370         char *h;
 371         const Char *p;
 372         Char *b, *eb;
 373
 374         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 375                 return pattern;
 376
 377         /*
 378          * Copy up to the end of the string or /
 379          */
 380         eb = &patbuf[patbuf_len - 1];
 381         for (p = pattern + 1, h = (char *) patbuf;
 382             h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
 383                 continue;
 384
 385         *h = EOS;
 386
 387         if (((char *) patbuf)[0] == EOS) {
 388                 /*
 389                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 390                  * we're not running setuid or setgid) and then trying
 391                  * the password file
 392                  */
 393                 if (issetugid() != 0 ||
 394                     (h = getenv("HOME")) == NULL) {
 395                         if (((h = getlogin()) != NULL &&
 396                              (pwd = getpwnam(h)) != NULL) ||
 397                             (pwd = getpwuid(getuid())) != NULL)
 398                                 h = pwd->pw_dir;
 399                         else
 400                                 return pattern;
 401                 }
 402         }
 403         else {
 404                 /*
 405                  * Expand a ~user
 406                  */
 407                 if ((pwd = getpwnam((char*) patbuf)) == NULL)
 408                         return pattern;
 409                 else
 410                         h = pwd->pw_dir;
 411         }
 412
 413         /* Copy the home directory */
 414         for (b = patbuf; b < eb && *h; *b++ = *h++)
 415                 continue;
 416
 417         /* Append the rest of the pattern */
 418         while (b < eb && (*b++ = *p++) != EOS)
 419                 continue;
 420         *b = EOS;
 421
 422         return patbuf;
 423 }
 424
 425
 426 /*
 427  * The main glob() routine: compiles the pattern (optionally processing
 428  * quotes), calls glob1() to do the real pattern matching, and finally
 429  * sorts the list (unless unsorted operation is requested).  Returns 0
 430  * if things went well, nonzero if errors occurred.
 431  */
 432 static int
 433 glob0(const Char *pattern, glob_t *pglob, size_t *limit)
 434 {
 435         const Char *qpatnext;
 436         int err;
 437         size_t oldpathc;
 438         Char *bufnext, c, patbuf[MAXPATHLEN];
 439
 440         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
 441         oldpathc = pglob->gl_pathc;
 442         bufnext = patbuf;
 443
 444         /* We don't need to check for buffer overflow any more. */
 445         while ((c = *qpatnext++) != EOS) {
 446                 switch (c) {
 447                 case LBRACKET:
 448                         c = *qpatnext;
 449                         if (c == NOT)
 450                                 ++qpatnext;
 451                         if (*qpatnext == EOS ||
 452                             g_strchr(qpatnext+1, RBRACKET) == NULL) {
 453                                 *bufnext++ = LBRACKET;
 454                                 if (c == NOT)
 455                                         --qpatnext;
 456                                 break;
 457                         }
 458                         *bufnext++ = M_SET;
 459                         if (c == NOT)
 460                                 *bufnext++ = M_NOT;
 461                         c = *qpatnext++;
 462                         do {
 463                                 *bufnext++ = CHAR(c);
 464                                 if (*qpatnext == RANGE &&
 465                                     (c = qpatnext[1]) != RBRACKET) {
 466                                         *bufnext++ = M_RNG;
 467                                         *bufnext++ = CHAR(c);
 468                                         qpatnext += 2;
 469                                 }
 470                         } while ((c = *qpatnext++) != RBRACKET);
 471                         pglob->gl_flags |= GLOB_MAGCHAR;
 472                         *bufnext++ = M_END;
 473                         break;
 474                 case QUESTION:
 475                         pglob->gl_flags |= GLOB_MAGCHAR;
 476                         *bufnext++ = M_ONE;
 477                         break;
 478                 case STAR:
 479                         pglob->gl_flags |= GLOB_MAGCHAR;
 480                         /* collapse adjacent stars to one,
 481                          * to avoid exponential behavior
 482                          */
 483                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 484                             *bufnext++ = M_ALL;
 485                         break;
 486                 default:
 487                         *bufnext++ = CHAR(c);
 488                         break;
 489                 }
 490         }
 491         *bufnext = EOS;
 492 #ifdef DEBUG
 493         qprintf("glob0:", patbuf);
 494 #endif
 495
 496         if ((err = glob1(patbuf, pglob, limit)) != 0)
 497                 return(err);
 498
 499         /*
 500          * If there was no match we are going to append the pattern
 501          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
 502          * and the pattern did not contain any magic characters
 503          * GLOB_NOMAGIC is there just for compatibility with csh.
 504          */
 505         if (pglob->gl_pathc == oldpathc) {
 506                 if (((pglob->gl_flags & GLOB_NOCHECK) ||
 507                     ((pglob->gl_flags & GLOB_NOMAGIC) &&
 508                         !(pglob->gl_flags & GLOB_MAGCHAR))))
 509                         return(globextend(pattern, pglob, limit));
 510                 else
 511                         return(GLOB_NOMATCH);
 512         }
 513         if (!(pglob->gl_flags & GLOB_NOSORT))
 514                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 515                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 516         return(0);
 517 }
 518
 519 static int
 520 compare(const void *p, const void *q)
 521 {
 522         return(strcmp(*(char **)p, *(char **)q));
 523 }
 524
 525 static int
 526 glob1(Char *pattern, glob_t *pglob, size_t *limit)
 527 {
 528         Char pathbuf[MAXPATHLEN];
 529
 530         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 531         if (*pattern == EOS)
 532                 return(0);
 533         return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 534             pattern, pglob, limit));
 535 }
 536
 537 /*
 538  * The functions glob2 and glob3 are mutually recursive; there is one level
 539  * of recursion for each segment in the pattern that contains one or more
 540  * meta characters.
 541  */
 542 static int
 543 glob2(Char *pathbuf, Char *pathend, Char *pathend_last, Char *pattern,
 544       glob_t *pglob, size_t *limit)
 545 {
 546         struct stat sb;
 547         Char *p, *q;
 548         int anymeta;
 549
 550         /*
 551          * Loop over pattern segments until end of pattern or until
 552          * segment with meta character found.
 553          */
 554         for (anymeta = 0;;) {
 555                 if (*pattern == EOS) {          /* End of pattern? */
 556                         *pathend = EOS;
 557                         if (g_lstat(pathbuf, &sb, pglob))
 558                                 return(0);
 559
 560                         if (((pglob->gl_flags & GLOB_MARK) &&
 561                             pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
 562                             || (S_ISLNK(sb.st_mode) &&
 563                             (g_stat(pathbuf, &sb, pglob) == 0) &&
 564                             S_ISDIR(sb.st_mode)))) {
 565                                 if (pathend + 1 > pathend_last)
 566                                         return (GLOB_ABORTED);
 567                                 *pathend++ = SEP;
 568                                 *pathend = EOS;
 569                         }
 570                         ++pglob->gl_matchc;
 571                         return(globextend(pathbuf, pglob, limit));
 572                 }
 573
 574                 /* Find end of next segment, copy tentatively to pathend. */
 575                 q = pathend;
 576                 p = pattern;
 577                 while (*p != EOS && *p != SEP) {
 578                         if (ismeta(*p))
 579                                 anymeta = 1;
 580                         if (q + 1 > pathend_last)
 581                                 return (GLOB_ABORTED);
 582                         *q++ = *p++;
 583                 }
 584
 585                 if (!anymeta) {         /* No expansion, do next segment. */
 586                         pathend = q;
 587                         pattern = p;
 588                         while (*pattern == SEP) {
 589                                 if (pathend + 1 > pathend_last)
 590                                         return (GLOB_ABORTED);
 591                                 *pathend++ = *pattern++;
 592                         }
 593                 } else                  /* Need expansion, recurse. */
 594                         return(glob3(pathbuf, pathend, pathend_last, pattern, p,
 595                             pglob, limit));
 596         }
 597         /* NOTREACHED */
 598 }
 599
 600 static int
 601 glob3(Char *pathbuf, Char *pathend, Char *pathend_last,
 602       Char *pattern, Char *restpattern,
 603       glob_t *pglob, size_t *limit)
 604 {
 605         struct dirent *dp;
 606         DIR *dirp;
 607         int err;
 608         char buf[MAXPATHLEN];
 609
 610         /*
 611          * The readdirfunc declaration can't be prototyped, because it is
 612          * assigned, below, to two functions which are prototyped in glob.h
 613          * and dirent.h as taking pointers to differently typed opaque
 614          * structures.
 615          */
 616         struct dirent *(*readdirfunc)();
 617
 618         if (pathend > pathend_last)
 619                 return (GLOB_ABORTED);
 620         *pathend = EOS;
 621         errno = 0;
 622
 623         if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
 624                 /* TODO: don't call for ENOENT or ENOTDIR? */
 625                 if (pglob->gl_errfunc) {
 626                         if (g_Ctoc(pathbuf, buf, sizeof(buf)))
 627                                 return (GLOB_ABORTED);
 628                         if (pglob->gl_errfunc(buf, errno) ||
 629                             pglob->gl_flags & GLOB_ERR)
 630                                 return (GLOB_ABORTED);
 631                 }
 632                 return(0);
 633         }
 634
 635         err = 0;
 636
 637         /* Search directory for matching names. */
 638         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 639                 readdirfunc = pglob->gl_readdir;
 640         else
 641                 readdirfunc = readdir;
 642         while ((dp = (*readdirfunc)(dirp))) {
 643                 char *sc;
 644                 Char *dc;
 645                 wchar_t wc;
 646                 size_t clen;
 647                 mbstate_t mbs;
 648
 649                 /* Initial DOT must be matched literally. */
 650                 if (dp->d_name[0] == DOT && *pattern != DOT)
 651                         continue;
 652                 memset(&mbs, 0, sizeof(mbs));
 653                 dc = pathend;
 654                 sc = dp->d_name;
 655                 while (dc < pathend_last) {
 656                         clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs);
 657                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 658                                 wc = *sc;
 659                                 clen = 1;
 660                                 memset(&mbs, 0, sizeof(mbs));
 661                         }
 662                         if ((*dc++ = wc) == EOS)
 663                                 break;
 664                         sc += clen;
 665                 }
 666                 if (!match(pathend, pattern, restpattern)) {
 667                         *pathend = EOS;
 668                         continue;
 669                 }
 670                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 671                     pglob, limit);
 672                 if (err)
 673                         break;
 674         }
 675
 676         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 677                 (*pglob->gl_closedir)(dirp);
 678         else
 679                 closedir(dirp);
 680         return(err);
 681 }
 682
 683
 684 /*
 685  * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
 686  * add the new item, and update gl_pathc.
 687  *
 688  * This assumes the BSD realloc, which only copies the block when its size
 689  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 690  * behavior.
 691  *
 692  * Return 0 if new item added, error code if memory couldn't be allocated.
 693  *
 694  * Invariant of the glob_t structure:
 695  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 696  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 697  */
 698 static int
 699 globextend(const Char *path, glob_t *pglob, size_t *limit)
 700 {
 701         char **pathv;
 702         size_t i, newsize, len;
 703         char *copy;
 704         const Char *p;
 705
 706         if (*limit && pglob->gl_pathc > *limit) {
 707                 errno = 0;
 708                 return (GLOB_NOSPACE);
 709         }
 710
 711         newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
 712         pathv = pglob->gl_pathv ?
 713                     realloc((char *)pglob->gl_pathv, newsize) :
 714                     malloc(newsize);
 715         if (pathv == NULL) {
 716                 if (pglob->gl_pathv) {
 717                         free(pglob->gl_pathv);
 718                         pglob->gl_pathv = NULL;
 719                 }
 720                 return(GLOB_NOSPACE);
 721         }
 722
 723         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 724                 /* first time around -- clear initial gl_offs items */
 725                 pathv += pglob->gl_offs;
 726                 for (i = pglob->gl_offs + 1; --i > 0; )
 727                         *--pathv = NULL;
 728         }
 729         pglob->gl_pathv = pathv;
 730
 731         for (p = path; *p++;)
 732                 continue;
 733         len = MB_CUR_MAX * (size_t)(p - path);  /* XXX overallocation */
 734         if ((copy = malloc(len)) != NULL) {
 735                 if (g_Ctoc(path, copy, len)) {
 736                         free(copy);
 737                         return (GLOB_NOSPACE);
 738                 }
 739                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 740         }
 741         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 742         return(copy == NULL ? GLOB_NOSPACE : 0);
 743 }
 744
 745 /*
 746  * pattern matching function for filenames.  Each occurrence of the *
 747  * pattern causes a recursion level.
 748  */
 749 static int
 750 match(Char *name, Char *pat, Char *patend)
 751 {
 752         int ok, negate_range;
 753         Char c, k;
 754
 755         while (pat < patend) {
 756                 c = *pat++;
 757                 switch (c & M_MASK) {
 758                 case M_ALL:
 759                         if (pat == patend)
 760                                 return(1);
 761                         do
 762                             if (match(name, pat, patend))
 763                                     return(1);
 764                         while (*name++ != EOS);
 765                         return(0);
 766                 case M_ONE:
 767                         if (*name++ == EOS)
 768                                 return(0);
 769                         break;
 770                 case M_SET:
 771                         ok = 0;
 772                         if ((k = *name++) == EOS)
 773                                 return(0);
 774                         if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 775                                 ++pat;
 776                         while (((c = *pat++) & M_MASK) != M_END)
 777                                 if ((*pat & M_MASK) == M_RNG) {
 778                                         if (__collate_load_error ?
 779                                             CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1]) :
 780                                                __collate_range_cmp(CHAR(c), CHAR(k)) <= 0
 781                                             && __collate_range_cmp(CHAR(k), CHAR(pat[1])) <= 0
 782                                            )
 783                                                 ok = 1;
 784                                         pat += 2;
 785                                 } else if (c == k)
 786                                         ok = 1;
 787                         if (ok == negate_range)
 788                                 return(0);
 789                         break;
 790                 default:
 791                         if (*name++ != c)
 792                                 return(0);
 793                         break;
 794                 }
 795         }
 796         return(*name == EOS);
 797 }
 798
 799 /* Free allocated data belonging to a glob_t structure. */
 800 void
 801 globfree(glob_t *pglob)
 802 {
 803         size_t i;
 804         char **pp;
 805
 806         if (pglob->gl_pathv != NULL) {
 807                 pp = pglob->gl_pathv + pglob->gl_offs;
 808                 for (i = pglob->gl_pathc; i--; ++pp)
 809                         if (*pp)
 810                                 free(*pp);
 811                 free(pglob->gl_pathv);
 812                 pglob->gl_pathv = NULL;
 813         }
 814 }
 815
 816 static DIR *
 817 g_opendir(Char *str, glob_t *pglob)
 818 {
 819         char buf[MAXPATHLEN];
 820
 821         if (!*str)
 822                 strcpy(buf, ".");
 823         else {
 824                 if (g_Ctoc(str, buf, sizeof(buf)))
 825                         return (NULL);
 826         }
 827
 828         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 829                 return((*pglob->gl_opendir)(buf));
 830
 831         return(opendir(buf));
 832 }
 833
 834 static int
 835 g_lstat(Char *fn, struct stat *sb, glob_t *pglob)
 836 {
 837         char buf[MAXPATHLEN];
 838
 839         if (g_Ctoc(fn, buf, sizeof(buf))) {
 840                 errno = ENAMETOOLONG;
 841                 return (-1);
 842         }
 843         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 844                 return((*pglob->gl_lstat)(buf, sb));
 845         return(lstat(buf, sb));
 846 }
 847
 848 static int
 849 g_stat(Char *fn, struct stat *sb, glob_t *pglob)
 850 {
 851         char buf[MAXPATHLEN];
 852
 853         if (g_Ctoc(fn, buf, sizeof(buf))) {
 854                 errno = ENAMETOOLONG;
 855                 return (-1);
 856         }
 857         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 858                 return((*pglob->gl_stat)(buf, sb));
 859         return(stat(buf, sb));
 860 }
 861
 862 static const Char *
 863 g_strchr(const Char *str, wchar_t ch)
 864 {
 865
 866         do {
 867                 if (*str == ch)
 868                         return (str);
 869         } while (*str++);
 870         return (NULL);
 871 }
 872
 873 static int
 874 g_Ctoc(const Char *str, char *buf, size_t len)
 875 {
 876         mbstate_t mbs;
 877         size_t clen;
 878
 879         memset(&mbs, 0, sizeof(mbs));
 880         while (len >= MB_CUR_MAX) {
 881                 clen = wcrtomb(buf, *str, &mbs);
 882                 if (clen == (size_t)-1)
 883                         return (1);
 884                 if (*str == L'\0')
 885                         return (0);
 886                 str++;
 887                 buf += clen;
 888                 len -= clen;
 889         }
 890         return (1);
 891 }
 892
 893 #ifdef DEBUG
 894 static void
 895 qprintf(const char *str, Char *s)
 896 {
 897         Char *p;
 898
 899         (void)printf("%s:\n", str);
 900         for (p = s; *p; p++)
 901                 (void)printf("%c", CHAR(*p));
 902         (void)printf("\n");
 903         for (p = s; *p; p++)
 904                 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
 905         (void)printf("\n");
 906         for (p = s; *p; p++)
 907                 (void)printf("%c", ismeta(*p) ? '_' : ' ');
 908         (void)printf("\n");
 909 }
 910 #endif