gen/glob-fbsd.c

   1 /*
   2  * Copyright (c) 1989, 1993
   3  *      The Regents of the University of California.  All rights reserved.
   4  *
   5  * This code is derived from software contributed to Berkeley by
   6  * Guido van Rossum.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 3. All advertising materials mentioning features or use of this software
  17  *    must display the following acknowledgement:
  18  *      This product includes software developed by the University of
  19  *      California, Berkeley and its contributors.
  20  * 4. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  */
  36
  37 #if defined(LIBC_SCCS) && !defined(lint)
  38 static char sccsid[] = "@(#)glob.c      8.3 (Berkeley) 10/13/93";
  39 #endif /* LIBC_SCCS and not lint */
  40 #include <sys/cdefs.h>
  41 __FBSDID("$FreeBSD: src/lib/libc/gen/glob.c,v 1.22 2004/07/29 03:48:52 tjr Exp $");
  42
  43 #include "xlocale_private.h"
  44
  45 /*
  46  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  47  *
  48  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  49  *
  50  * Optional extra services, controlled by flags not defined by POSIX:
  51  *
  52  * GLOB_QUOTE:
  53  *      Escaping convention: \ inhibits any special meaning the following
  54  *      character might have (except \ at end of string is retained).
  55  * GLOB_MAGCHAR:
  56  *      Set in gl_flags if pattern contained a globbing character.
  57  * GLOB_NOMAGIC:
  58  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  59  *      not contain any magic characters.  [Used in csh style globbing]
  60  * GLOB_ALTDIRFUNC:
  61  *      Use alternately specified directory access functions.
  62  * GLOB_TILDE:
  63  *      expand ~user/foo to the /home/dir/of/user/foo
  64  * GLOB_BRACE:
  65  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  66  * gl_matchc:
  67  *      Number of matches in the current invocation of glob.
  68  */
  69
  70 /*
  71  * Some notes on multibyte character support:
  72  * 1. Patterns with illegal byte sequences match nothing - even if
  73  *    GLOB_NOCHECK is specified.
  74  * 2. Illegal byte sequences in filenames are handled by treating them as
  75  *    single-byte characters with a value of the first byte of the sequence
  76  *    cast to wchar_t.
  77  * 3. State-dependent encodings are not currently supported.
  78  */
  79
  80 #include <sys/param.h>
  81 #include <sys/stat.h>
  82
  83 #include <ctype.h>
  84 #include <dirent.h>
  85 #include <errno.h>
  86 #include <glob.h>
  87 #include <limits.h>
  88 #include <pwd.h>
  89 #include <stdint.h>
  90 #include <stdio.h>
  91 #include <stdlib.h>
  92 #include <string.h>
  93 #include <unistd.h>
  94 #include <wchar.h>
  95
  96 #include "collate.h"
  97
  98 #define DOLLAR          '$'
  99 #define DOT             '.'
 100 #define EOS             '\0'
 101 #define LBRACKET        '['
 102 #define NOT             '!'
 103 #define QUESTION        '?'
 104 #define QUOTE           '\\'
 105 #define RANGE           '-'
 106 #define RBRACKET        ']'
 107 #define SEP             '/'
 108 #define STAR            '*'
 109 #define TILDE           '~'
 110 #define UNDERSCORE      '_'
 111 #define LBRACE          '{'
 112 #define RBRACE          '}'
 113 #define SLASH           '/'
 114 #define COMMA           ','
 115
 116 #ifndef DEBUG
 117
 118 #define M_QUOTE         0x8000000000ULL
 119 #define M_PROTECT       0x4000000000ULL
 120 #define M_MASK          0xffffffffffULL
 121 #define M_CHAR          0x00ffffffffULL
 122
 123 typedef uint_fast64_t Char;
 124
 125 #else
 126
 127 #define M_QUOTE         0x80
 128 #define M_PROTECT       0x40
 129 #define M_MASK          0xff
 130 #define M_CHAR          0x7f
 131
 132 typedef char Char;
 133
 134 #endif
 135
 136
 137 #define CHAR(c)         ((Char)((c)&M_CHAR))
 138 #define META(c)         ((Char)((c)|M_QUOTE))
 139 #define M_ALL           META('*')
 140 #define M_END           META(']')
 141 #define M_NOT           META('!')
 142 #define M_ONE           META('?')
 143 #define M_RNG           META('-')
 144 #define M_SET           META('[')
 145 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 146
 147 static int       g_lstat(Char *, struct stat *, glob_t *, locale_t);
 148 static int       g_stat(Char *, struct stat *, glob_t *, locale_t);
 149
 150 #define g_Ctoc          __gl_g_Ctoc
 151 #define glob0           __gl_glob0
 152 #define glob2_32        __gl_glob0_32
 153 #define glob2_64        __gl_glob0_64
 154 #define glob3           __gl_glob3
 155 #define globexp1        __gl_globexp1
 156 #define globextend      __gl_globextend
 157 __private_extern__ int g_Ctoc(const Char *, char *, u_int, locale_t);
 158 __private_extern__ int glob0(const Char *, glob_t *, int *, locale_t);
 159 __private_extern__ int glob2_32(Char *, Char *, Char *, Char *, glob_t *, int *, locale_t);
 160 __private_extern__ int glob2_64(Char *, Char *, Char *, Char *, glob_t *, int *, locale_t);
 161 __private_extern__ int glob3(Char *, Char *, Char *, Char *, Char *, glob_t *, int *, locale_t);
 162 __private_extern__ int globexp1(const Char *, glob_t *, int *, locale_t);
 163 __private_extern__ int globextend(const Char *, glob_t *, int *, locale_t);
 164
 165 #ifndef BUILDING_VARIANT
 166 #define glob2(a,b,c,d,e,f,g)    (((e)->gl_flags & GLOB_INODE64) ? glob2_64((a),(b),(c),(d),(e),(f),(g)) : glob2_32((a),(b),(c),(d),(e),(f),(g)))
 167
 168 static int       compare(const void *, const void *);
 169 static DIR      *g_opendir(Char *, glob_t *, locale_t);
 170 static Char     *g_strchr(Char *, wchar_t);
 171 #ifdef notdef
 172 static Char     *g_strcat(Char *, const Char *);
 173 #endif
 174 static int       glob1(Char *, glob_t *, int *, locale_t);
 175 static const Char *
 176                  globtilde(const Char *, Char *, size_t, glob_t *);
 177 static int       globexp2(const Char *, const Char *, glob_t *, int *, int *, locale_t);
 178 static int       match(Char *, Char *, Char *, locale_t);
 179 #ifdef DEBUG
 180 static void      qprintf(const char *, Char *);
 181 #endif
 182 #endif /* !BUILDING_VARIANT */
 183
 184 int
 185 glob(pattern, flags, errfunc, pglob)
 186         const char *pattern;
 187         int flags, (*errfunc)(const char *, int);
 188         glob_t *pglob;
 189 {
 190         const u_char *patnext;
 191         int limit;
 192         Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
 193         mbstate_t mbs;
 194         wchar_t wc;
 195         size_t clen;
 196         locale_t loc = __current_locale();
 197         int mb_cur_max = MB_CUR_MAX_L(loc);
 198
 199         patnext = (u_char *) pattern;
 200         if (!(flags & GLOB_APPEND)) {
 201                 pglob->gl_pathc = 0;
 202                 pglob->gl_pathv = NULL;
 203                 if (!(flags & GLOB_DOOFFS))
 204                         pglob->gl_offs = 0;
 205         }
 206         if (flags & GLOB_LIMIT) {
 207                 limit = pglob->gl_matchc;
 208                 if (limit == 0)
 209                         limit = ARG_MAX;
 210         } else
 211                 limit = 0;
 212 #if __DARWIN_64_BIT_INO_T
 213         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 214         pglob->gl_flags |= GLOB_INODE64;
 215 #else /* !__DARWIN_64_BIT_INO_T */
 216         pglob->gl_flags = flags & ~(GLOB_MAGCHAR | GLOB_INODE64);
 217 #endif /* __DARWIN_64_BIT_INO_T */
 218         pglob->gl_errfunc = errfunc;
 219         pglob->gl_matchc = 0;
 220
 221         bufnext = patbuf;
 222         bufend = bufnext + MAXPATHLEN - 1;
 223         if (flags & GLOB_NOESCAPE) {
 224                 memset(&mbs, 0, sizeof(mbs));
 225                 while (bufend - bufnext >= mb_cur_max) {
 226                         clen = mbrtowc_l(&wc, (const char *)patnext, MB_LEN_MAX, &mbs, loc);
 227                         if (clen == (size_t)-1 || clen == (size_t)-2)
 228                                 return (GLOB_NOMATCH);
 229                         else if (clen == 0)
 230                                 break;
 231                         *bufnext++ = wc;
 232                         patnext += clen;
 233                 }
 234         } else {
 235                 /* Protect the quoted characters. */
 236                 memset(&mbs, 0, sizeof(mbs));
 237                 while (bufend - bufnext >= mb_cur_max) {
 238                         if (*patnext == QUOTE) {
 239                                 if (*++patnext == EOS) {
 240                                         *bufnext++ = QUOTE | M_PROTECT;
 241                                         continue;
 242                                 }
 243                                 prot = M_PROTECT;
 244                         } else
 245                                 prot = 0;
 246                         clen = mbrtowc_l(&wc, (const char *)patnext, MB_LEN_MAX, &mbs, loc);
 247                         if (clen == (size_t)-1 || clen == (size_t)-2)
 248                                 return (GLOB_NOMATCH);
 249                         else if (clen == 0)
 250                                 break;
 251                         *bufnext++ = wc | prot;
 252                         patnext += clen;
 253                 }
 254         }
 255         *bufnext = EOS;
 256
 257         if (flags & GLOB_BRACE)
 258             return globexp1(patbuf, pglob, &limit, loc);
 259         else
 260             return glob0(patbuf, pglob, &limit, loc);
 261 }
 262
 263 #ifndef BUILDING_VARIANT
 264 /*
 265  * Expand recursively a glob {} pattern. When there is no more expansion
 266  * invoke the standard globbing routine to glob the rest of the magic
 267  * characters
 268  */
 269 __private_extern__ int
 270 globexp1(pattern, pglob, limit, loc)
 271         const Char *pattern;
 272         glob_t *pglob;
 273         int *limit;
 274         locale_t loc;
 275 {
 276         const Char* ptr = pattern;
 277         int rv;
 278
 279         /* Protect a single {}, for find(1), like csh */
 280         if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
 281                 return glob0(pattern, pglob, limit, loc);
 282
 283         while ((ptr = (const Char *) g_strchr((Char *) ptr, LBRACE)) != NULL)
 284                 if (!globexp2(ptr, pattern, pglob, &rv, limit, loc))
 285                         return rv;
 286
 287         return glob0(pattern, pglob, limit, loc);
 288 }
 289
 290
 291 /*
 292  * Recursive brace globbing helper. Tries to expand a single brace.
 293  * If it succeeds then it invokes globexp1 with the new pattern.
 294  * If it fails then it tries to glob the rest of the pattern and returns.
 295  */
 296 static int
 297 globexp2(ptr, pattern, pglob, rv, limit, loc)
 298         const Char *ptr, *pattern;
 299         glob_t *pglob;
 300         int *rv, *limit;
 301         locale_t loc;
 302 {
 303         int     i;
 304         Char   *lm, *ls;
 305         const Char *pe, *pm, *pl;
 306         Char    patbuf[MAXPATHLEN];
 307
 308         /* copy part up to the brace */
 309         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 310                 continue;
 311         *lm = EOS;
 312         ls = lm;
 313
 314         /* Find the balanced brace */
 315         for (i = 0, pe = ++ptr; *pe; pe++)
 316                 if (*pe == LBRACKET) {
 317                         /* Ignore everything between [] */
 318                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
 319                                 continue;
 320                         if (*pe == EOS) {
 321                                 /*
 322                                  * We could not find a matching RBRACKET.
 323                                  * Ignore and just look for RBRACE
 324                                  */
 325                                 pe = pm;
 326                         }
 327                 }
 328                 else if (*pe == LBRACE)
 329                         i++;
 330                 else if (*pe == RBRACE) {
 331                         if (i == 0)
 332                                 break;
 333                         i--;
 334                 }
 335
 336         /* Non matching braces; just glob the pattern */
 337         if (i != 0 || *pe == EOS) {
 338                 *rv = glob0(patbuf, pglob, limit, loc);
 339                 return 0;
 340         }
 341
 342         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 343                 switch (*pm) {
 344                 case LBRACKET:
 345                         /* Ignore everything between [] */
 346                         for (pl = pm++; *pm != RBRACKET && *pm != EOS; pm++)
 347                                 continue;
 348                         if (*pm == EOS) {
 349                                 /*
 350                                  * We could not find a matching RBRACKET.
 351                                  * Ignore and just look for RBRACE
 352                                  */
 353                                 pm = pl;
 354                         }
 355                         break;
 356
 357                 case LBRACE:
 358                         i++;
 359                         break;
 360
 361                 case RBRACE:
 362                         if (i) {
 363                             i--;
 364                             break;
 365                         }
 366                         /* FALLTHROUGH */
 367                 case COMMA:
 368                         if (i && *pm == COMMA)
 369                                 break;
 370                         else {
 371                                 /* Append the current string */
 372                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 373                                         continue;
 374                                 /*
 375                                  * Append the rest of the pattern after the
 376                                  * closing brace
 377                                  */
 378                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 379                                         continue;
 380
 381                                 /* Expand the current pattern */
 382 #ifdef DEBUG
 383                                 qprintf("globexp2:", patbuf);
 384 #endif
 385                                 *rv = globexp1(patbuf, pglob, limit, loc);
 386
 387                                 /* move after the comma, to the next string */
 388                                 pl = pm + 1;
 389                         }
 390                         break;
 391
 392                 default:
 393                         break;
 394                 }
 395         *rv = 0;
 396         return 0;
 397 }
 398
 399
 400
 401 /*
 402  * expand tilde from the passwd file.
 403  */
 404 static const Char *
 405 globtilde(pattern, patbuf, patbuf_len, pglob)
 406         const Char *pattern;
 407         Char *patbuf;
 408         size_t patbuf_len;
 409         glob_t *pglob;
 410 {
 411         struct passwd *pwd;
 412         char *h;
 413         const Char *p;
 414         Char *b, *eb;
 415
 416         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 417                 return pattern;
 418
 419         /*
 420          * Copy up to the end of the string or /
 421          */
 422         eb = &patbuf[patbuf_len - 1];
 423         for (p = pattern + 1, h = (char *) patbuf;
 424             h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
 425                 continue;
 426
 427         *h = EOS;
 428
 429         if (((char *) patbuf)[0] == EOS) {
 430                 /*
 431                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 432                  * we're not running setuid or setgid) and then trying
 433                  * the password file
 434                  */
 435                 if (issetugid() != 0 ||
 436                     (h = getenv("HOME")) == NULL) {
 437                         if (((h = getlogin()) != NULL &&
 438                              (pwd = getpwnam(h)) != NULL) ||
 439                             (pwd = getpwuid(getuid())) != NULL)
 440                                 h = pwd->pw_dir;
 441                         else
 442                                 return pattern;
 443                 }
 444         }
 445         else {
 446                 /*
 447                  * Expand a ~user
 448                  */
 449                 if ((pwd = getpwnam((char*) patbuf)) == NULL)
 450                         return pattern;
 451                 else
 452                         h = pwd->pw_dir;
 453         }
 454
 455         /* Copy the home directory */
 456         for (b = patbuf; b < eb && *h; *b++ = *h++)
 457                 continue;
 458
 459         /* Append the rest of the pattern */
 460         while (b < eb && (*b++ = *p++) != EOS)
 461                 continue;
 462         *b = EOS;
 463
 464         return patbuf;
 465 }
 466
 467
 468 /*
 469  * The main glob() routine: compiles the pattern (optionally processing
 470  * quotes), calls glob1() to do the real pattern matching, and finally
 471  * sorts the list (unless unsorted operation is requested).  Returns 0
 472  * if things went well, nonzero if errors occurred.
 473  */
 474 __private_extern__ int
 475 glob0(pattern, pglob, limit, loc)
 476         const Char *pattern;
 477         glob_t *pglob;
 478         int *limit;
 479         locale_t loc;
 480 {
 481         const Char *qpatnext;
 482         Char c;
 483         int err, oldpathc;
 484         Char *bufnext, patbuf[MAXPATHLEN];
 485
 486         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
 487         oldpathc = pglob->gl_pathc;
 488         bufnext = patbuf;
 489
 490         /* We don't need to check for buffer overflow any more. */
 491         while ((c = *qpatnext++) != EOS) {
 492                 if (c & M_PROTECT) {
 493                         *bufnext++ = CHAR(c);
 494                         continue;
 495                 } /* else */
 496                 switch (c) {
 497                 case LBRACKET:
 498                         c = *qpatnext;
 499                         if (c == NOT)
 500                                 ++qpatnext;
 501                         if (*qpatnext == EOS ||
 502                             g_strchr((Char *) qpatnext+1, RBRACKET) == NULL) {
 503                                 *bufnext++ = LBRACKET;
 504                                 if (c == NOT)
 505                                         --qpatnext;
 506                                 break;
 507                         }
 508                         *bufnext++ = M_SET;
 509                         if (c == NOT)
 510                                 *bufnext++ = M_NOT;
 511                         c = *qpatnext++;
 512                         do {
 513                                 *bufnext++ = CHAR(c);
 514                                 if (*qpatnext == RANGE &&
 515                                     (c = qpatnext[1]) != RBRACKET) {
 516                                         *bufnext++ = M_RNG;
 517                                         *bufnext++ = CHAR(c);
 518                                         qpatnext += 2;
 519                                 }
 520                         } while ((c = *qpatnext++) != RBRACKET);
 521                         pglob->gl_flags |= GLOB_MAGCHAR;
 522                         *bufnext++ = M_END;
 523                         break;
 524                 case QUESTION:
 525                         pglob->gl_flags |= GLOB_MAGCHAR;
 526                         *bufnext++ = M_ONE;
 527                         break;
 528                 case STAR:
 529                         pglob->gl_flags |= GLOB_MAGCHAR;
 530                         /* collapse adjacent stars to one,
 531                          * to avoid exponential behavior
 532                          */
 533                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 534                             *bufnext++ = M_ALL;
 535                         break;
 536                 default:
 537                         *bufnext++ = CHAR(c);
 538                         break;
 539                 }
 540         }
 541         *bufnext = EOS;
 542 #ifdef DEBUG
 543         qprintf("glob0:", patbuf);
 544 #endif
 545
 546         if ((err = glob1(patbuf, pglob, limit, loc)) != 0)
 547                 return(err);
 548
 549         /*
 550          * If there was no match we are going to append the pattern
 551          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
 552          * and the pattern did not contain any magic characters
 553          * GLOB_NOMAGIC is there just for compatibility with csh.
 554          */
 555         if (pglob->gl_pathc == oldpathc) {
 556                 if (((pglob->gl_flags & GLOB_NOCHECK) ||
 557                     ((pglob->gl_flags & GLOB_NOMAGIC) &&
 558                         !(pglob->gl_flags & GLOB_MAGCHAR))))
 559                         return(globextend(pattern, pglob, limit, loc));
 560                 else
 561                         return(GLOB_NOMATCH);
 562         }
 563         if (!(pglob->gl_flags & GLOB_NOSORT))
 564                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 565                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 566         return(0);
 567 }
 568
 569 static int
 570 compare(p, q)
 571         const void *p, *q;
 572 {
 573         return(strcoll(*(char **)p, *(char **)q));
 574 }
 575
 576 static int
 577 glob1(pattern, pglob, limit, loc)
 578         Char *pattern;
 579         glob_t *pglob;
 580         int *limit;
 581         locale_t loc;
 582 {
 583         Char pathbuf[MAXPATHLEN];
 584
 585         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 586         if (*pattern == EOS)
 587                 return(0);
 588         return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 589             pattern, pglob, limit, loc));
 590 }
 591 #endif /* !BUILDING_VARIANT */
 592
 593 /*
 594  * The functions glob2 and glob3 are mutually recursive; there is one level
 595  * of recursion for each segment in the pattern that contains one or more
 596  * meta characters.
 597  */
 598 __private_extern__ int
 599 #if __DARWIN_64_BIT_INO_T
 600 glob2_64(pathbuf, pathend, pathend_last, pattern, pglob, limit, loc)
 601 #else /* !__DARWIN_64_BIT_INO_T */
 602 glob2_32(pathbuf, pathend, pathend_last, pattern, pglob, limit, loc)
 603 #endif /* __DARWIN_64_BIT_INO_T */
 604         Char *pathbuf, *pathend, *pathend_last, *pattern;
 605         glob_t *pglob;
 606         int *limit;
 607         locale_t loc;
 608 {
 609         struct stat sb;
 610         Char *p, *q;
 611         int anymeta;
 612
 613         /*
 614          * Loop over pattern segments until end of pattern or until
 615          * segment with meta character found.
 616          */
 617         for (anymeta = 0;;) {
 618                 if (*pattern == EOS) {          /* End of pattern? */
 619                         *pathend = EOS;
 620                         if (g_lstat(pathbuf, &sb, pglob, loc))
 621                                 return(0);
 622
 623                         if (((pglob->gl_flags & GLOB_MARK) &&
 624                             pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
 625                             || (S_ISLNK(sb.st_mode) &&
 626                             (g_stat(pathbuf, &sb, pglob, loc) == 0) &&
 627                             S_ISDIR(sb.st_mode)))) {
 628                                 if (pathend + 1 > pathend_last)
 629                                         return (GLOB_ABORTED);
 630                                 *pathend++ = SEP;
 631                                 *pathend = EOS;
 632                         }
 633                         ++pglob->gl_matchc;
 634                         return(globextend(pathbuf, pglob, limit, loc));
 635                 }
 636
 637                 /* Find end of next segment, copy tentatively to pathend. */
 638                 q = pathend;
 639                 p = pattern;
 640                 while (*p != EOS && *p != SEP) {
 641                         if (ismeta(*p))
 642                                 anymeta = 1;
 643                         if (q + 1 > pathend_last)
 644                                 return (GLOB_ABORTED);
 645                         *q++ = *p++;
 646                 }
 647
 648                 if (!anymeta) {         /* No expansion, do next segment. */
 649                         pathend = q;
 650                         pattern = p;
 651                         while (*pattern == SEP) {
 652                                 if (pathend + 1 > pathend_last)
 653                                         return (GLOB_ABORTED);
 654                                 *pathend++ = *pattern++;
 655                         }
 656                 } else                  /* Need expansion, recurse. */
 657                         return(glob3(pathbuf, pathend, pathend_last, pattern, p,
 658                             pglob, limit, loc));
 659         }
 660         /* NOTREACHED */
 661 }
 662
 663 #ifndef BUILDING_VARIANT
 664 __private_extern__ int
 665 glob3(pathbuf, pathend, pathend_last, pattern, restpattern, pglob, limit, loc)
 666         Char *pathbuf, *pathend, *pathend_last, *pattern, *restpattern;
 667         glob_t *pglob;
 668         int *limit;
 669         locale_t loc;
 670 {
 671         struct dirent *dp;
 672         DIR *dirp;
 673         int err;
 674         char buf[MAXPATHLEN];
 675
 676         /*
 677          * The readdirfunc declaration can't be prototyped, because it is
 678          * assigned, below, to two functions which are prototyped in glob.h
 679          * and dirent.h as taking pointers to differently typed opaque
 680          * structures.
 681          */
 682         struct dirent *(*readdirfunc)();
 683
 684         if (pathend > pathend_last)
 685                 return (GLOB_ABORTED);
 686         *pathend = EOS;
 687         errno = 0;
 688
 689         if ((dirp = g_opendir(pathbuf, pglob, loc)) == NULL) {
 690                 /* TODO: don't call for ENOENT or ENOTDIR? */
 691                 if (pglob->gl_errfunc) {
 692                         if (g_Ctoc(pathbuf, buf, sizeof(buf), loc))
 693                                 return (GLOB_ABORTED);
 694                         if (pglob->gl_errfunc(buf, errno))
 695                                 return (GLOB_ABORTED);
 696                 }
 697                 if (pglob->gl_flags & GLOB_ERR)
 698                         return (GLOB_ABORTED);
 699                 return(0);
 700         }
 701
 702         err = 0;
 703
 704         /* Search directory for matching names. */
 705         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 706                 readdirfunc = pglob->gl_readdir;
 707         else
 708                 readdirfunc = readdir;
 709         while ((dp = (*readdirfunc)(dirp))) {
 710                 u_char *sc;
 711                 Char *dc;
 712                 wchar_t wc;
 713                 size_t clen;
 714                 mbstate_t mbs;
 715
 716                 /* Initial DOT must be matched literally. */
 717                 if (dp->d_name[0] == DOT && *pattern != DOT)
 718                         continue;
 719                 memset(&mbs, 0, sizeof(mbs));
 720                 dc = pathend;
 721                 sc = (u_char *) dp->d_name;
 722                 while (dc < pathend_last) {
 723                         clen = mbrtowc_l(&wc, (const char *)sc, MB_LEN_MAX, &mbs, loc);
 724                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 725                                 wc = *sc;
 726                                 clen = 1;
 727                                 memset(&mbs, 0, sizeof(mbs));
 728                         }
 729                         if ((*dc++ = wc) == EOS)
 730                                 break;
 731                         sc += clen;
 732                 }
 733                 if (!match(pathend, pattern, restpattern, loc)) {
 734                         *pathend = EOS;
 735                         continue;
 736                 }
 737                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 738                     pglob, limit, loc);
 739                 if (err)
 740                         break;
 741         }
 742
 743         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 744                 (*pglob->gl_closedir)(dirp);
 745         else
 746                 closedir(dirp);
 747         return(err);
 748 }
 749
 750
 751 /*
 752  * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
 753  * add the new item, and update gl_pathc.
 754  *
 755  * This assumes the BSD realloc, which only copies the block when its size
 756  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 757  * behavior.
 758  *
 759  * Return 0 if new item added, error code if memory couldn't be allocated.
 760  *
 761  * Invariant of the glob_t structure:
 762  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 763  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 764  */
 765 __private_extern__ int
 766 globextend(path, pglob, limit, loc)
 767         const Char *path;
 768         glob_t *pglob;
 769         int *limit;
 770         locale_t loc;
 771 {
 772         char **pathv;
 773         int i;
 774         u_int newsize, len;
 775         char *copy;
 776         const Char *p;
 777
 778         if (*limit && pglob->gl_pathc > *limit) {
 779                 errno = 0;
 780                 return (GLOB_NOSPACE);
 781         }
 782
 783         newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
 784         pathv = pglob->gl_pathv ?
 785                     realloc((char *)pglob->gl_pathv, newsize) :
 786                     malloc(newsize);
 787         if (pathv == NULL) {
 788                 if (pglob->gl_pathv) {
 789                         free(pglob->gl_pathv);
 790                         pglob->gl_pathv = NULL;
 791                 }
 792                 return(GLOB_NOSPACE);
 793         }
 794
 795         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 796                 /* first time around -- clear initial gl_offs items */
 797                 pathv += pglob->gl_offs;
 798                 for (i = pglob->gl_offs; --i >= 0; )
 799                         *--pathv = NULL;
 800         }
 801         pglob->gl_pathv = pathv;
 802
 803         for (p = path; *p++;)
 804                 continue;
 805         len = MB_CUR_MAX_L(loc) * (size_t)(p - path);   /* XXX overallocation */
 806         if ((copy = malloc(len)) != NULL) {
 807                 if (g_Ctoc(path, copy, len, loc)) {
 808                         free(copy);
 809                         return (GLOB_NOSPACE);
 810                 }
 811                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 812         }
 813         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 814         return(copy == NULL ? GLOB_NOSPACE : 0);
 815 }
 816
 817 /*
 818  * pattern matching function for filenames.  Each occurrence of the *
 819  * pattern causes a recursion level.
 820  */
 821 static int
 822 match(name, pat, patend, loc)
 823         Char *name, *pat, *patend;
 824         locale_t loc;
 825 {
 826         int ok, negate_range;
 827         Char c, k;
 828
 829         while (pat < patend) {
 830                 c = *pat++;
 831                 switch (c & M_MASK) {
 832                 case M_ALL:
 833                         if (pat == patend)
 834                                 return(1);
 835                         do
 836                             if (match(name, pat, patend, loc))
 837                                     return(1);
 838                         while (*name++ != EOS);
 839                         return(0);
 840                 case M_ONE:
 841                         if (*name++ == EOS)
 842                                 return(0);
 843                         break;
 844                 case M_SET:
 845                         ok = 0;
 846                         if ((k = *name++) == EOS)
 847                                 return(0);
 848                         if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 849                                 ++pat;
 850                         while (((c = *pat++) & M_MASK) != M_END)
 851                                 if ((*pat & M_MASK) == M_RNG) {
 852                                         if (loc->__collate_load_error ?
 853                                             CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1]) :
 854                                                __collate_range_cmp(CHAR(c), CHAR(k), loc) <= 0
 855                                             && __collate_range_cmp(CHAR(k), CHAR(pat[1]), loc) <= 0
 856                                            )
 857                                                 ok = 1;
 858                                         pat += 2;
 859                                 } else if (c == k)
 860                                         ok = 1;
 861                         if (ok == negate_range)
 862                                 return(0);
 863                         break;
 864                 default:
 865                         if (*name++ != c)
 866                                 return(0);
 867                         break;
 868                 }
 869         }
 870         return(*name == EOS);
 871 }
 872
 873 /* Free allocated data belonging to a glob_t structure. */
 874 void
 875 globfree(pglob)
 876         glob_t *pglob;
 877 {
 878         int i;
 879         char **pp;
 880
 881         if (pglob->gl_pathv != NULL) {
 882                 pp = pglob->gl_pathv + pglob->gl_offs;
 883                 for (i = pglob->gl_pathc; i--; ++pp)
 884                         if (*pp)
 885                                 free(*pp);
 886                 free(pglob->gl_pathv);
 887                 pglob->gl_pathv = NULL;
 888         }
 889 }
 890
 891 static DIR *
 892 g_opendir(str, pglob, loc)
 893         Char *str;
 894         glob_t *pglob;
 895         locale_t loc;
 896 {
 897         char buf[MAXPATHLEN];
 898
 899         if (!*str)
 900                 strcpy(buf, ".");
 901         else {
 902                 if (g_Ctoc(str, buf, sizeof(buf), loc))
 903                         return (NULL);
 904         }
 905
 906         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 907                 return((*pglob->gl_opendir)(buf));
 908
 909         return(opendir(buf));
 910 }
 911 #endif /* !BUILDING_VARIANT */
 912
 913 static int
 914 g_lstat(fn, sb, pglob, loc)
 915         Char *fn;
 916         struct stat *sb;
 917         glob_t *pglob;
 918         locale_t loc;
 919 {
 920         char buf[MAXPATHLEN];
 921
 922         if (g_Ctoc(fn, buf, sizeof(buf), loc)) {
 923                 errno = ENAMETOOLONG;
 924                 return (-1);
 925         }
 926         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 927                 return((*pglob->gl_lstat)(buf, sb));
 928         return(lstat(buf, sb));
 929 }
 930
 931 static int
 932 g_stat(fn, sb, pglob, loc)
 933         Char *fn;
 934         struct stat *sb;
 935         glob_t *pglob;
 936         locale_t loc;
 937 {
 938         char buf[MAXPATHLEN];
 939
 940         if (g_Ctoc(fn, buf, sizeof(buf), loc)) {
 941                 errno = ENAMETOOLONG;
 942                 return (-1);
 943         }
 944         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 945                 return((*pglob->gl_stat)(buf, sb));
 946         return(stat(buf, sb));
 947 }
 948
 949 #ifndef BUILDING_VARIANT
 950 static Char *
 951 g_strchr(str, ch)
 952         Char *str;
 953         wchar_t ch;
 954 {
 955         do {
 956                 if (*str == ch)
 957                         return (str);
 958         } while (*str++);
 959         return (NULL);
 960 }
 961
 962 __private_extern__ int
 963 g_Ctoc(str, buf, len, loc)
 964         const Char *str;
 965         char *buf;
 966         u_int len;
 967         locale_t loc;
 968 {
 969         mbstate_t mbs;
 970         size_t clen;
 971         int mb_cur_max = MB_CUR_MAX_L(loc);
 972
 973         memset(&mbs, 0, sizeof(mbs));
 974         while (len >= mb_cur_max) {
 975                 clen = wcrtomb_l(buf, *str, &mbs, loc);
 976                 if (clen == (size_t)-1)
 977                         return (1);
 978                 if (*str == L'\0')
 979                         return (0);
 980                 str++;
 981                 buf += clen;
 982                 len -= clen;
 983         }
 984         return (1);
 985 }
 986
 987 #ifdef DEBUG
 988 static void
 989 qprintf(str, s)
 990         const char *str;
 991         Char *s;
 992 {
 993         Char *p;
 994
 995         (void)printf("%s:\n", str);
 996         for (p = s; *p; p++)
 997                 (void)printf("%c", CHAR(*p));
 998         (void)printf("\n");
 999         for (p = s; *p; p++)
1000                 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
1001         (void)printf("\n");
1002         for (p = s; *p; p++)
1003                 (void)printf("%c", ismeta(*p) ? '_' : ' ');
1004         (void)printf("\n");
1005 }
1006 #endif
1007 #endif /* !BUILDING_VARIANT */