mdw@git.distorted.org.uk Git - termux-packages/blob - packages/libandroid-glob/glob.c

   1 /*
   2  * Natanael Arndt, 2011: removed collate.h dependencies
   3  *  (my changes are trivial)
   4  *
   5  * Copyright (c) 1989, 1993
   6  *      The Regents of the University of California.  All rights reserved.
   7  *
   8  * This code is derived from software contributed to Berkeley by
   9  * Guido van Rossum.
  10  *
  11  * Redistribution and use in source and binary forms, with or without
  12  * modification, are permitted provided that the following conditions
  13  * are met:
  14  * 1. Redistributions of source code must retain the above copyright
  15  *    notice, this list of conditions and the following disclaimer.
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in the
  18  *    documentation and/or other materials provided with the distribution.
  19  * 4. Neither the name of the University nor the names of its contributors
  20  *    may be used to endorse or promote products derived from this software
  21  *    without specific prior written permission.
  22  *
  23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  33  * SUCH DAMAGE.
  34  */
  35
  36 #if defined(LIBC_SCCS) && !defined(lint)
  37 static char sccsid[] = "@(#)glob.c      8.3 (Berkeley) 10/13/93";
  38 #endif /* LIBC_SCCS and not lint */
  39 #include <sys/cdefs.h>
  40 __FBSDID("$FreeBSD$");
  41
  42 /*
  43  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  44  *
  45  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  46  *
  47  * Optional extra services, controlled by flags not defined by POSIX:
  48  *
  49  * GLOB_QUOTE:
  50  *      Escaping convention: \ inhibits any special meaning the following
  51  *      character might have (except \ at end of string is retained).
  52  * GLOB_MAGCHAR:
  53  *      Set in gl_flags if pattern contained a globbing character.
  54  * GLOB_NOMAGIC:
  55  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  56  *      not contain any magic characters.  [Used in csh style globbing]
  57  * GLOB_ALTDIRFUNC:
  58  *      Use alternately specified directory access functions.
  59  * GLOB_TILDE:
  60  *      expand ~user/foo to the /home/dir/of/user/foo
  61  * GLOB_BRACE:
  62  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  63  * gl_matchc:
  64  *      Number of matches in the current invocation of glob.
  65  */
  66
  67 /*
  68  * Some notes on multibyte character support:
  69  * 1. Patterns with illegal byte sequences match nothing - even if
  70  *    GLOB_NOCHECK is specified.
  71  * 2. Illegal byte sequences in filenames are handled by treating them as
  72  *    single-byte characters with a value of the first byte of the sequence
  73  *    cast to wchar_t.
  74  * 3. State-dependent encodings are not currently supported.
  75  */
  76
  77 #include <sys/param.h>
  78 #include <sys/stat.h>
  79
  80 #include <ctype.h>
  81 #include <dirent.h>
  82 #include <errno.h>
  83 #include <glob.h>
  84 #include <limits.h>
  85 #include <pwd.h>
  86 #include <stdint.h>
  87 #include <stdio.h>
  88 #include <stdlib.h>
  89 #include <string.h>
  90 #include <unistd.h>
  91 #include <wchar.h>
  92
  93 #define DOLLAR          '$'
  94 #define DOT             '.'
  95 #define EOS             '\0'
  96 #define LBRACKET        '['
  97 #define NOT             '!'
  98 #define QUESTION        '?'
  99 #define QUOTE           '\\'
 100 #define RANGE           '-'
 101 #define RBRACKET        ']'
 102 #define SEP             '/'
 103 #define STAR            '*'
 104 #define TILDE           '~'
 105 #define UNDERSCORE      '_'
 106 #define LBRACE          '{'
 107 #define RBRACE          '}'
 108 #define SLASH           '/'
 109 #define COMMA           ','
 110
 111 #ifndef DEBUG
 112
 113 #define M_QUOTE         0x8000000000ULL
 114 #define M_PROTECT       0x4000000000ULL
 115 #define M_MASK          0xffffffffffULL
 116 #define M_CHAR          0x00ffffffffULL
 117
 118 typedef uint_fast64_t Char;
 119
 120 #else
 121
 122 #define M_QUOTE         0x80
 123 #define M_PROTECT       0x40
 124 #define M_MASK          0xff
 125 #define M_CHAR          0x7f
 126
 127 typedef char Char;
 128
 129 #endif
 130
 131
 132 #define CHAR(c)         ((Char)((c)&M_CHAR))
 133 #define META(c)         ((Char)((c)|M_QUOTE))
 134 #define M_ALL           META('*')
 135 #define M_END           META(']')
 136 #define M_NOT           META('!')
 137 #define M_ONE           META('?')
 138 #define M_RNG           META('-')
 139 #define M_SET           META('[')
 140 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 141
 142
 143 static int       compare(const void *, const void *);
 144 static int       g_Ctoc(const Char *, char *, size_t);
 145 static int       g_lstat(Char *, struct stat *, glob_t *);
 146 static DIR      *g_opendir(Char *, glob_t *);
 147 static const Char *g_strchr(const Char *, wchar_t);
 148 #ifdef notdef
 149 static Char     *g_strcat(Char *, const Char *);
 150 #endif
 151 static int       g_stat(Char *, struct stat *, glob_t *);
 152 static int       glob0(const Char *, glob_t *, size_t *);
 153 static int       glob1(Char *, glob_t *, size_t *);
 154 static int       glob2(Char *, Char *, Char *, Char *, glob_t *, size_t *);
 155 static int       glob3(Char *, Char *, Char *, Char *, Char *, glob_t *, size_t *);
 156 static int       globextend(const Char *, glob_t *, size_t *);
 157 static const Char *
 158                  globtilde(const Char *, Char *, size_t, glob_t *);
 159 static int       globexp1(const Char *, glob_t *, size_t *);
 160 static int       globexp2(const Char *, const Char *, glob_t *, int *, size_t *);
 161 static int       match(Char *, Char *, Char *);
 162 #ifdef DEBUG
 163 static void      qprintf(const char *, Char *);
 164 #endif
 165
 166 int
 167 glob(const char *pattern, int flags, int (*errfunc)(const char *, int), 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, size_t *limit)
 265 {
 266         int     i;
 267         Char   *lm, *ls;
 268         const Char *pe, *pm, *pm1, *pl;
 269         Char    patbuf[MAXPATHLEN];
 270
 271         /* copy part up to the brace */
 272         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 273                 continue;
 274         *lm = EOS;
 275         ls = lm;
 276
 277         /* Find the balanced brace */
 278         for (i = 0, pe = ++ptr; *pe; pe++)
 279                 if (*pe == LBRACKET) {
 280                         /* Ignore everything between [] */
 281                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
 282                                 continue;
 283                         if (*pe == EOS) {
 284                                 /*
 285                                  * We could not find a matching RBRACKET.
 286                                  * Ignore and just look for RBRACE
 287                                  */
 288                                 pe = pm;
 289                         }
 290                 }
 291                 else if (*pe == LBRACE)
 292                         i++;
 293                 else if (*pe == RBRACE) {
 294                         if (i == 0)
 295                                 break;
 296                         i--;
 297                 }
 298
 299         /* Non matching braces; just glob the pattern */
 300         if (i != 0 || *pe == EOS) {
 301                 *rv = glob0(patbuf, pglob, limit);
 302                 return 0;
 303         }
 304
 305         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 306                 switch (*pm) {
 307                 case LBRACKET:
 308                         /* Ignore everything between [] */
 309                         for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++)
 310                                 continue;
 311                         if (*pm == EOS) {
 312                                 /*
 313                                  * We could not find a matching RBRACKET.
 314                                  * Ignore and just look for RBRACE
 315                                  */
 316                                 pm = pm1;
 317                         }
 318                         break;
 319
 320                 case LBRACE:
 321                         i++;
 322                         break;
 323
 324                 case RBRACE:
 325                         if (i) {
 326                             i--;
 327                             break;
 328                         }
 329                         /* FALLTHROUGH */
 330                 case COMMA:
 331                         if (i && *pm == COMMA)
 332                                 break;
 333                         else {
 334                                 /* Append the current string */
 335                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 336                                         continue;
 337                                 /*
 338                                  * Append the rest of the pattern after the
 339                                  * closing brace
 340                                  */
 341                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 342                                         continue;
 343
 344                                 /* Expand the current pattern */
 345 #ifdef DEBUG
 346                                 qprintf("globexp2:", patbuf);
 347 #endif
 348                                 *rv = globexp1(patbuf, pglob, limit);
 349
 350                                 /* move after the comma, to the next string */
 351                                 pl = pm + 1;
 352                         }
 353                         break;
 354
 355                 default:
 356                         break;
 357                 }
 358         *rv = 0;
 359         return 0;
 360 }
 361
 362
 363
 364 /*
 365  * expand tilde from the passwd file.
 366  */
 367 static const Char *
 368 globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
 369 {
 370         struct passwd *pwd;
 371         char *h;
 372         const Char *p;
 373         Char *b, *eb;
 374
 375         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 376                 return pattern;
 377
 378         /*
 379          * Copy up to the end of the string or /
 380          */
 381         eb = &patbuf[patbuf_len - 1];
 382         for (p = pattern + 1, h = (char *) patbuf;
 383             h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
 384                 continue;
 385
 386         *h = EOS;
 387
 388         if (((char *) patbuf)[0] == EOS) {
 389                 /*
 390                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 391                  * we're not running setuid or setgid) and then trying
 392                  * the password file
 393                  */
 394                 if (issetugid() != 0 ||
 395                     (h = getenv("HOME")) == NULL) {
 396                         if (((h = getlogin()) != NULL &&
 397                              (pwd = getpwnam(h)) != NULL) ||
 398                             (pwd = getpwuid(getuid())) != NULL)
 399                                 h = pwd->pw_dir;
 400                         else
 401                                 return pattern;
 402                 }
 403         }
 404         else {
 405                 /*
 406                  * Expand a ~user
 407                  */
 408                 if ((pwd = getpwnam((char*) patbuf)) == NULL)
 409                         return pattern;
 410                 else
 411                         h = pwd->pw_dir;
 412         }
 413
 414         /* Copy the home directory */
 415         for (b = patbuf; b < eb && *h; *b++ = *h++)
 416                 continue;
 417
 418         /* Append the rest of the pattern */
 419         while (b < eb && (*b++ = *p++) != EOS)
 420                 continue;
 421         *b = EOS;
 422
 423         return patbuf;
 424 }
 425
 426
 427 /*
 428  * The main glob() routine: compiles the pattern (optionally processing
 429  * quotes), calls glob1() to do the real pattern matching, and finally
 430  * sorts the list (unless unsorted operation is requested).  Returns 0
 431  * if things went well, nonzero if errors occurred.
 432  */
 433 static int
 434 glob0(const Char *pattern, glob_t *pglob, size_t *limit)
 435 {
 436         const Char *qpatnext;
 437         int err;
 438         size_t oldpathc;
 439         Char *bufnext, c, patbuf[MAXPATHLEN];
 440
 441         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
 442         oldpathc = pglob->gl_pathc;
 443         bufnext = patbuf;
 444
 445         /* We don't need to check for buffer overflow any more. */
 446         while ((c = *qpatnext++) != EOS) {
 447                 switch (c) {
 448                 case LBRACKET:
 449                         c = *qpatnext;
 450                         if (c == NOT)
 451                                 ++qpatnext;
 452                         if (*qpatnext == EOS ||
 453                             g_strchr(qpatnext+1, RBRACKET) == NULL) {
 454                                 *bufnext++ = LBRACKET;
 455                                 if (c == NOT)
 456                                         --qpatnext;
 457                                 break;
 458                         }
 459                         *bufnext++ = M_SET;
 460                         if (c == NOT)
 461                                 *bufnext++ = M_NOT;
 462                         c = *qpatnext++;
 463                         do {
 464                                 *bufnext++ = CHAR(c);
 465                                 if (*qpatnext == RANGE &&
 466                                     (c = qpatnext[1]) != RBRACKET) {
 467                                         *bufnext++ = M_RNG;
 468                                         *bufnext++ = CHAR(c);
 469                                         qpatnext += 2;
 470                                 }
 471                         } while ((c = *qpatnext++) != RBRACKET);
 472                         pglob->gl_flags |= GLOB_MAGCHAR;
 473                         *bufnext++ = M_END;
 474                         break;
 475                 case QUESTION:
 476                         pglob->gl_flags |= GLOB_MAGCHAR;
 477                         *bufnext++ = M_ONE;
 478                         break;
 479                 case STAR:
 480                         pglob->gl_flags |= GLOB_MAGCHAR;
 481                         /* collapse adjacent stars to one,
 482                          * to avoid exponential behavior
 483                          */
 484                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 485                             *bufnext++ = M_ALL;
 486                         break;
 487                 default:
 488                         *bufnext++ = CHAR(c);
 489                         break;
 490                 }
 491         }
 492         *bufnext = EOS;
 493 #ifdef DEBUG
 494         qprintf("glob0:", patbuf);
 495 #endif
 496
 497         if ((err = glob1(patbuf, pglob, limit)) != 0)
 498                 return(err);
 499
 500         /*
 501          * If there was no match we are going to append the pattern
 502          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
 503          * and the pattern did not contain any magic characters
 504          * GLOB_NOMAGIC is there just for compatibility with csh.
 505          */
 506         if (pglob->gl_pathc == oldpathc) {
 507                 if (((pglob->gl_flags & GLOB_NOCHECK) ||
 508                     ((pglob->gl_flags & GLOB_NOMAGIC) &&
 509                         !(pglob->gl_flags & GLOB_MAGCHAR))))
 510                         return(globextend(pattern, pglob, limit));
 511                 else
 512                         return(GLOB_NOMATCH);
 513         }
 514         if (!(pglob->gl_flags & GLOB_NOSORT))
 515                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 516                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 517         return(0);
 518 }
 519
 520 static int
 521 compare(const void *p, const void *q)
 522 {
 523         return(strcmp(*(char **)p, *(char **)q));
 524 }
 525
 526 static int
 527 glob1(Char *pattern, glob_t *pglob, size_t *limit)
 528 {
 529         Char pathbuf[MAXPATHLEN];
 530
 531         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 532         if (*pattern == EOS)
 533                 return(0);
 534         return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 535             pattern, pglob, limit));
 536 }
 537
 538 /*
 539  * The functions glob2 and glob3 are mutually recursive; there is one level
 540  * of recursion for each segment in the pattern that contains one or more
 541  * meta characters.
 542  */
 543 static int
 544 glob2(Char *pathbuf, Char *pathend, Char *pathend_last, Char *pattern,
 545       glob_t *pglob, size_t *limit)
 546 {
 547         struct stat sb;
 548         Char *p, *q;
 549         int anymeta;
 550
 551         /*
 552          * Loop over pattern segments until end of pattern or until
 553          * segment with meta character found.
 554          */
 555         for (anymeta = 0;;) {
 556                 if (*pattern == EOS) {          /* End of pattern? */
 557                         *pathend = EOS;
 558                         if (g_lstat(pathbuf, &sb, pglob))
 559                                 return(0);
 560
 561                         if (((pglob->gl_flags & GLOB_MARK) &&
 562                             pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
 563                             || (S_ISLNK(sb.st_mode) &&
 564                             (g_stat(pathbuf, &sb, pglob) == 0) &&
 565                             S_ISDIR(sb.st_mode)))) {
 566                                 if (pathend + 1 > pathend_last)
 567                                         return (GLOB_ABORTED);
 568                                 *pathend++ = SEP;
 569                                 *pathend = EOS;
 570                         }
 571                         ++pglob->gl_matchc;
 572                         return(globextend(pathbuf, pglob, limit));
 573                 }
 574
 575                 /* Find end of next segment, copy tentatively to pathend. */
 576                 q = pathend;
 577                 p = pattern;
 578                 while (*p != EOS && *p != SEP) {
 579                         if (ismeta(*p))
 580                                 anymeta = 1;
 581                         if (q + 1 > pathend_last)
 582                                 return (GLOB_ABORTED);
 583                         *q++ = *p++;
 584                 }
 585
 586                 if (!anymeta) {         /* No expansion, do next segment. */
 587                         pathend = q;
 588                         pattern = p;
 589                         while (*pattern == SEP) {
 590                                 if (pathend + 1 > pathend_last)
 591                                         return (GLOB_ABORTED);
 592                                 *pathend++ = *pattern++;
 593                         }
 594                 } else                  /* Need expansion, recurse. */
 595                         return(glob3(pathbuf, pathend, pathend_last, pattern, p,
 596                             pglob, limit));
 597         }
 598         /* NOTREACHED */
 599 }
 600
 601 static int
 602 glob3(Char *pathbuf, Char *pathend, Char *pathend_last,
 603       Char *pattern, Char *restpattern,
 604       glob_t *pglob, size_t *limit)
 605 {
 606         struct dirent *dp;
 607         DIR *dirp;
 608         int err;
 609         char buf[MAXPATHLEN];
 610
 611         /*
 612          * The readdirfunc declaration can't be prototyped, because it is
 613          * assigned, below, to two functions which are prototyped in glob.h
 614          * and dirent.h as taking pointers to differently typed opaque
 615          * structures.
 616          */
 617         struct dirent *(*readdirfunc)();
 618
 619         if (pathend > pathend_last)
 620                 return (GLOB_ABORTED);
 621         *pathend = EOS;
 622         errno = 0;
 623
 624         if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
 625                 /* TODO: don't call for ENOENT or ENOTDIR? */
 626                 if (pglob->gl_errfunc) {
 627                         if (g_Ctoc(pathbuf, buf, sizeof(buf)))
 628                                 return (GLOB_ABORTED);
 629                         if (pglob->gl_errfunc(buf, errno) ||
 630                             pglob->gl_flags & GLOB_ERR)
 631                                 return (GLOB_ABORTED);
 632                 }
 633                 return(0);
 634         }
 635
 636         err = 0;
 637
 638         /* Search directory for matching names. */
 639         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 640                 readdirfunc = pglob->gl_readdir;
 641         else
 642                 readdirfunc = readdir;
 643         while ((dp = (*readdirfunc)(dirp))) {
 644                 char *sc;
 645                 Char *dc;
 646                 wchar_t wc;
 647                 size_t clen;
 648                 mbstate_t mbs;
 649
 650                 /* Initial DOT must be matched literally. */
 651                 if (dp->d_name[0] == DOT && *pattern != DOT)
 652                         continue;
 653                 memset(&mbs, 0, sizeof(mbs));
 654                 dc = pathend;
 655                 sc = dp->d_name;
 656                 while (dc < pathend_last) {
 657                         clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs);
 658                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 659                                 wc = *sc;
 660                                 clen = 1;
 661                                 memset(&mbs, 0, sizeof(mbs));
 662                         }
 663                         if ((*dc++ = wc) == EOS)
 664                                 break;
 665                         sc += clen;
 666                 }
 667                 if (!match(pathend, pattern, restpattern)) {
 668                         *pathend = EOS;
 669                         continue;
 670                 }
 671                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 672                     pglob, limit);
 673                 if (err)
 674                         break;
 675         }
 676
 677         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 678                 (*pglob->gl_closedir)(dirp);
 679         else
 680                 closedir(dirp);
 681         return(err);
 682 }
 683
 684
 685 /*
 686  * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
 687  * add the new item, and update gl_pathc.
 688  *
 689  * This assumes the BSD realloc, which only copies the block when its size
 690  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 691  * behavior.
 692  *
 693  * Return 0 if new item added, error code if memory couldn't be allocated.
 694  *
 695  * Invariant of the glob_t structure:
 696  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 697  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 698  */
 699 static int
 700 globextend(const Char *path, glob_t *pglob, size_t *limit)
 701 {
 702         char **pathv;
 703         size_t i, newsize, len;
 704         char *copy;
 705         const Char *p;
 706
 707         if (*limit && pglob->gl_pathc > *limit) {
 708                 errno = 0;
 709                 return (GLOB_NOSPACE);
 710         }
 711
 712         newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
 713         pathv = pglob->gl_pathv ?
 714                     realloc((char *)pglob->gl_pathv, newsize) :
 715                     malloc(newsize);
 716         if (pathv == NULL) {
 717                 if (pglob->gl_pathv) {
 718                         free(pglob->gl_pathv);
 719                         pglob->gl_pathv = NULL;
 720                 }
 721                 return(GLOB_NOSPACE);
 722         }
 723
 724         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 725                 /* first time around -- clear initial gl_offs items */
 726                 pathv += pglob->gl_offs;
 727                 for (i = pglob->gl_offs + 1; --i > 0; )
 728                         *--pathv = NULL;
 729         }
 730         pglob->gl_pathv = pathv;
 731
 732         for (p = path; *p++;)
 733                 continue;
 734         len = MB_CUR_MAX * (size_t)(p - path);  /* XXX overallocation */
 735         if ((copy = malloc(len)) != NULL) {
 736                 if (g_Ctoc(path, copy, len)) {
 737                         free(copy);
 738                         return (GLOB_NOSPACE);
 739                 }
 740                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 741         }
 742         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 743         return(copy == NULL ? GLOB_NOSPACE : 0);
 744 }
 745
 746 /*
 747  * pattern matching function for filenames.  Each occurrence of the *
 748  * pattern causes a recursion level.
 749  */
 750 static int
 751 match(Char *name, Char *pat, Char *patend)
 752 {
 753         int ok, negate_range;
 754         Char c, k;
 755
 756         while (pat < patend) {
 757                 c = *pat++;
 758                 switch (c & M_MASK) {
 759                 case M_ALL:
 760                         if (pat == patend)
 761                                 return(1);
 762                         do
 763                             if (match(name, pat, patend))
 764                                     return(1);
 765                         while (*name++ != EOS);
 766                         return(0);
 767                 case M_ONE:
 768                         if (*name++ == EOS)
 769                                 return(0);
 770                         break;
 771                 case M_SET:
 772                         ok = 0;
 773                         if ((k = *name++) == EOS)
 774                                 return(0);
 775                         if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 776                                 ++pat;
 777                         while (((c = *pat++) & M_MASK) != M_END)
 778                                 if ((*pat & M_MASK) == M_RNG) {
 779                                         if (CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1])) ok = 1;
 780                                         pat += 2;
 781                                 } else if (c == k)
 782                                         ok = 1;
 783                         if (ok == negate_range)
 784                                 return(0);
 785                         break;
 786                 default:
 787                         if (*name++ != c)
 788                                 return(0);
 789                         break;
 790                 }
 791         }
 792         return(*name == EOS);
 793 }
 794
 795 /* Free allocated data belonging to a glob_t structure. */
 796 void
 797 globfree(glob_t *pglob)
 798 {
 799         size_t i;
 800         char **pp;
 801
 802         if (pglob->gl_pathv != NULL) {
 803                 pp = pglob->gl_pathv + pglob->gl_offs;
 804                 for (i = pglob->gl_pathc; i--; ++pp)
 805                         if (*pp)
 806                                 free(*pp);
 807                 free(pglob->gl_pathv);
 808                 pglob->gl_pathv = NULL;
 809         }
 810 }
 811
 812 static DIR *
 813 g_opendir(Char *str, glob_t *pglob)
 814 {
 815         char buf[MAXPATHLEN];
 816
 817         if (!*str)
 818                 strcpy(buf, ".");
 819         else {
 820                 if (g_Ctoc(str, buf, sizeof(buf)))
 821                         return (NULL);
 822         }
 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(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(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
 862         do {
 863                 if (*str == ch)
 864                         return (str);
 865         } while (*str++);
 866         return (NULL);
 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, Char *s)
 892 {
 893         Char *p;
 894
 895         (void)printf("%s:\n", str);
 896         for (p = s; *p; p++)
 897                 (void)printf("%c", CHAR(*p));
 898         (void)printf("\n");
 899         for (p = s; *p; p++)
 900                 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
 901         (void)printf("\n");
 902         for (p = s; *p; p++)
 903                 (void)printf("%c", ismeta(*p) ? '_' : ' ');
 904         (void)printf("\n");
 905 }
 906 #endif