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  * Copyright (c) 2011 The FreeBSD Foundation
   9  * All rights reserved.
  10  * Portions of this software were developed by David Chisnall
  11  * under sponsorship from the FreeBSD Foundation.
  12  *
  13  * Redistribution and use in source and binary forms, with or without
  14  * modification, are permitted provided that the following conditions
  15  * are met:
  16  * 1. Redistributions of source code must retain the above copyright
  17  *    notice, this list of conditions and the following disclaimer.
  18  * 2. Redistributions in binary form must reproduce the above copyright
  19  *    notice, this list of conditions and the following disclaimer in the
  20  *    documentation and/or other materials provided with the distribution.
  21  * 3. Neither the name of the University nor the names of its contributors
  22  *    may be used to endorse or promote products derived from this software
  23  *    without specific prior written permission.
  24  *
  25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  35  * SUCH DAMAGE.
  36  */
  37
  38 #pragma clang diagnostic push
  39 #pragma clang diagnostic ignored "-Wstrict-prototypes"
  40
  41 #if defined(LIBC_SCCS) && !defined(lint)
  42 static char sccsid[] = "@(#)glob.c      8.3 (Berkeley) 10/13/93";
  43 #endif /* LIBC_SCCS and not lint */
  44 #include <sys/cdefs.h>
  45 __FBSDID("$FreeBSD$");
  46
  47 #include "xlocale_private.h"
  48
  49 /*
  50  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  51  *
  52  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  53  *
  54  * Optional extra services, controlled by flags not defined by POSIX:
  55  *
  56  * GLOB_QUOTE:
  57  *      Escaping convention: \ inhibits any special meaning the following
  58  *      character might have (except \ at end of string is retained).
  59  * GLOB_MAGCHAR:
  60  *      Set in gl_flags if pattern contained a globbing character.
  61  * GLOB_NOMAGIC:
  62  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  63  *      not contain any magic characters.  [Used in csh style globbing]
  64  * GLOB_ALTDIRFUNC:
  65  *      Use alternately specified directory access functions.
  66  * GLOB_TILDE:
  67  *      expand ~user/foo to the /home/dir/of/user/foo
  68  * GLOB_BRACE:
  69  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  70  * gl_matchc:
  71  *      Number of matches in the current invocation of glob.
  72  */
  73
  74 /*
  75  * Some notes on multibyte character support:
  76  * 1. Patterns with illegal byte sequences match nothing - even if
  77  *    GLOB_NOCHECK is specified.
  78  * 2. Illegal byte sequences in filenames are handled by treating them as
  79  *    single-byte characters with a values of such bytes of the sequence
  80  *    cast to wchar_t.
  81  * 3. State-dependent encodings are not currently supported.
  82  */
  83
  84 #include <sys/param.h>
  85 #include <sys/stat.h>
  86
  87 #include <ctype.h>
  88 #include <dirent.h>
  89 #include <errno.h>
  90 #include <glob.h>
  91 #include <limits.h>
  92 #include <pwd.h>
  93 #include <stdint.h>
  94 #include <stdio.h>
  95 #include <stdlib.h>
  96 #include <string.h>
  97 #include <unistd.h>
  98 #include <wchar.h>
  99 #include <malloc_private.h>
 100
 101 #include "collate.h"
 102
 103 /*
 104  * glob(3) expansion limits. Stop the expansion if any of these limits
 105  * is reached. This caps the runtime in the face of DoS attacks. See
 106  * also CVE-2010-2632
 107  */
 108 #define GLOB_LIMIT_BRACE        128     /* number of brace calls */
 109 #define GLOB_LIMIT_PATH         1024    /* number of path elements */
 110 #define GLOB_LIMIT_READDIR      16384   /* number of readdirs */
 111 #define GLOB_LIMIT_STAT         128     /* number of stat system calls */
 112 #define GLOB_LIMIT_STRING       65536   /* maximum total size for paths */
 113
 114 struct glob_limit {
 115         size_t  l_brace_cnt;
 116         size_t  l_path_lim;
 117         size_t  l_readdir_cnt;
 118         size_t  l_stat_cnt;
 119         size_t  l_string_cnt;
 120 };
 121
 122 #define DOT             L'.'
 123 #define EOS             L'\0'
 124 #define LBRACKET        L'['
 125 #define NOT             L'!'
 126 #define QUESTION        L'?'
 127 #define QUOTE           L'\\'
 128 #define RANGE           L'-'
 129 #define RBRACKET        L']'
 130 #define SEP             L'/'
 131 #define STAR            L'*'
 132 #define TILDE           L'~'
 133 #define LBRACE          L'{'
 134 #define RBRACE          L'}'
 135 #define COMMA           L','
 136
 137 #define M_QUOTE         0x8000000000ULL
 138 #define M_PROTECT       0x4000000000ULL
 139 #define M_MASK          0xffffffffffULL
 140 #define M_CHAR          0x00ffffffffULL
 141
 142 typedef uint_fast64_t Char;
 143
 144 #define CHAR(c)         ((Char)((c)&M_CHAR))
 145 #define META(c)         ((Char)((c)|M_QUOTE))
 146 #define UNPROT(c)       ((c) & ~M_PROTECT)
 147 #define M_ALL           META(L'*')
 148 #define M_END           META(L']')
 149 #define M_NOT           META(L'!')
 150 #define M_ONE           META(L'?')
 151 #define M_RNG           META(L'-')
 152 #define M_SET           META(L'[')
 153 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 154 #ifdef DEBUG
 155 #define isprot(c)       (((c)&M_PROTECT) != 0)
 156 #endif
 157
 158
 159 #define compare         __gl_compare
 160 #define g_Ctoc          __gl_g_Ctoc
 161 #define g_strchr        __gl_g_strchr
 162 #define globextend      __gl_globextend
 163 #define globtilde       __gl_globtilde
 164 #define match           __gl_match
 165
 166 __private_extern__ int   compare(const void *, const void *);
 167 __private_extern__ int   g_Ctoc(const Char *, char *, size_t, locale_t);
 168 static int       g_lstat(Char *, struct stat *, glob_t *, locale_t);
 169 static DIR      *g_opendir(Char *, glob_t *, locale_t);
 170 __private_extern__ const Char *g_strchr(const Char *, wchar_t);
 171 #ifdef notdef
 172 static Char     *g_strcat(Char *, const Char *);
 173 #endif
 174 static int       g_stat(Char *, struct stat *, glob_t *, locale_t);
 175 static int       glob0(const Char *, glob_t *, struct glob_limit *,
 176                 const char *,locale_t);
 177 static int       glob1(Char *, glob_t *, struct glob_limit *, locale_t);
 178 static int       glob2(Char *, Char *, Char *, Char *, glob_t *,
 179                 struct glob_limit *, locale_t);
 180 static int       glob3(Char *, Char *, Char *, Char *, Char *, glob_t *,
 181                 struct glob_limit *, locale_t);
 182 __private_extern__ int   globextend(const Char *, glob_t *, struct glob_limit *,
 183                 const char *, locale_t);
 184 __private_extern__ const Char * globtilde(const Char *, Char *, size_t, glob_t *,
 185                 locale_t);
 186 static int       globexp0(const Char *, glob_t *, struct glob_limit *,
 187     const char *, locale_t);
 188 static int       globexp1(const Char *, glob_t *, struct glob_limit *, locale_t);
 189 static int       globexp2(const Char *, const Char *, glob_t *,
 190     struct glob_limit *, locale_t);
 191 static int       globfinal(glob_t *, struct glob_limit *, size_t,
 192     const char *, locale_t);
 193 __private_extern__ int   match(Char *, Char *, Char *, locale_t);
 194 static int       err_nomatch(glob_t *, struct glob_limit *, const char *, locale_t loc);
 195 static int       err_aborted(glob_t *, int, char *);
 196 #ifdef DEBUG
 197 static void      qprintf(const char *, Char *);
 198 #endif
 199
 200 static int
 201 __glob(const char *pattern, glob_t *pglob)
 202 {
 203         struct glob_limit limit = { 0, 0, 0, 0, 0 };
 204         const char *patnext;
 205         Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
 206         mbstate_t mbs;
 207         wchar_t wc;
 208         size_t clen;
 209         int too_long;
 210         locale_t loc = __current_locale();
 211         int mb_cur_max = MB_CUR_MAX_L(loc);
 212
 213         patnext = pattern;
 214         if (!(pglob->gl_flags & GLOB_APPEND)) {
 215                 pglob->gl_pathc = 0;
 216                 pglob->gl_pathv = NULL;
 217                 if (!(pglob->gl_flags & GLOB_DOOFFS))
 218                         pglob->gl_offs = 0;
 219         }
 220         if (pglob->gl_flags & GLOB_LIMIT) {
 221                 limit.l_path_lim = pglob->gl_matchc;
 222                 if (limit.l_path_lim == 0)
 223                         limit.l_path_lim = GLOB_LIMIT_PATH;
 224         }
 225         pglob->gl_matchc = 0;
 226
 227         bufnext = patbuf;
 228         bufend = bufnext + MAXPATHLEN - 1;
 229         too_long = 1;
 230         if (pglob->gl_flags & GLOB_NOESCAPE) {
 231                 memset(&mbs, 0, sizeof(mbs));
 232                 while (bufend - bufnext >= mb_cur_max) {
 233                         clen = mbrtowc_l(&wc, patnext, MB_LEN_MAX, &mbs, loc);
 234                         if (clen == (size_t)-1 || clen == (size_t)-2)
 235                                 return (err_nomatch(pglob, &limit, pattern, loc));
 236                         else if (clen == 0) {
 237                                 too_long = 0;
 238                                 break;
 239                         }
 240                         *bufnext++ = wc;
 241                         patnext += clen;
 242                 }
 243         } else {
 244                 /* Protect the quoted characters. */
 245                 memset(&mbs, 0, sizeof(mbs));
 246                 while (bufend - bufnext >= mb_cur_max) {
 247                         if (*patnext == '\\') {
 248                                 if (*++patnext == '\0') {
 249                                         *bufnext++ = QUOTE;
 250                                         continue;
 251                                 }
 252                                 prot = M_PROTECT;
 253                         } else
 254                                 prot = 0;
 255                         clen = mbrtowc_l(&wc, patnext, MB_LEN_MAX, &mbs, loc);
 256                         if (clen == (size_t)-1 || clen == (size_t)-2)
 257                                 return (err_nomatch(pglob, &limit, pattern, loc));
 258                         else if (clen == 0) {
 259                                 too_long = 0;
 260                                 break;
 261                         }
 262                         *bufnext++ = wc | prot;
 263                         patnext += clen;
 264                 }
 265         }
 266         if (too_long)
 267                 return (err_nomatch(pglob, &limit, pattern, loc));
 268         *bufnext = EOS;
 269
 270         if (pglob->gl_flags & GLOB_BRACE)
 271             return globexp0(patbuf, pglob, &limit, pattern, loc);
 272         else
 273             return glob0(patbuf, pglob, &limit, pattern, loc);
 274 }
 275
 276 int
 277 glob(const char *pattern, int flags, int (*errfunc)(const char *, int), glob_t *pglob)
 278 {
 279 #ifdef __BLOCKS__
 280         pglob->gl_flags = flags & ~(GLOB_MAGCHAR | _GLOB_ERR_BLOCK);
 281 #else /* !__BLOCKS__ */
 282         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 283 #endif /* __BLOCKS__ */
 284         pglob->gl_errfunc = errfunc;
 285         return __glob(pattern, pglob);
 286 }
 287
 288 #ifdef __BLOCKS__
 289 int
 290 glob_b(const char *pattern, int flags, int (^errblk)(const char *, int), glob_t *pglob)
 291 {
 292         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 293         pglob->gl_flags |= _GLOB_ERR_BLOCK;
 294         pglob->gl_errblk = errblk;
 295         return __glob(pattern, pglob);
 296 }
 297 #endif /* __BLOCKS__ */
 298
 299 static int
 300 globexp0(const Char *pattern, glob_t *pglob, struct glob_limit *limit,
 301     const char *origpat, locale_t loc) {
 302         int rv;
 303         size_t oldpathc;
 304
 305         /* Protect a single {}, for find(1), like csh */
 306         if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS) {
 307                 if ((pglob->gl_flags & GLOB_LIMIT) &&
 308                     limit->l_brace_cnt++ >= GLOB_LIMIT_BRACE) {
 309                         errno = E2BIG;
 310                         return (GLOB_NOSPACE);
 311                 }
 312                 return (glob0(pattern, pglob, limit, origpat, loc));
 313         }
 314
 315         oldpathc = pglob->gl_pathc;
 316
 317         if ((rv = globexp1(pattern, pglob, limit, loc)) != 0)
 318                 return rv;
 319
 320         return (globfinal(pglob, limit, oldpathc, origpat, loc));
 321 }
 322
 323 /*
 324  * Expand recursively a glob {} pattern. When there is no more expansion
 325  * invoke the standard globbing routine to glob the rest of the magic
 326  * characters
 327  */
 328 static int
 329 globexp1(const Char *pattern, glob_t *pglob, struct glob_limit *limit, locale_t loc)
 330 {
 331         const Char* ptr;
 332
 333         if ((ptr = g_strchr(pattern, LBRACE)) != NULL) {
 334                 if ((pglob->gl_flags & GLOB_LIMIT) &&
 335                     limit->l_brace_cnt++ >= GLOB_LIMIT_BRACE) {
 336                         errno = E2BIG;
 337                         return (GLOB_NOSPACE);
 338                 }
 339                 return (globexp2(ptr, pattern, pglob, limit, loc));
 340         }
 341
 342         return (glob0(pattern, pglob, limit, NULL, loc));
 343 }
 344
 345
 346 /*
 347  * Recursive brace globbing helper. Tries to expand a single brace.
 348  * If it succeeds then it invokes globexp1 with the new pattern.
 349  * If it fails then it tries to glob the rest of the pattern and returns.
 350  */
 351 static int
 352 globexp2(const Char *ptr, const Char *pattern, glob_t *pglob,
 353                 struct glob_limit *limit, locale_t loc)
 354 {
 355         int     i, rv;
 356         Char   *lm, *ls;
 357         const Char *pe, *pm, *pm1, *pl;
 358         Char    patbuf[MAXPATHLEN];
 359
 360         /* copy part up to the brace */
 361         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 362                 continue;
 363         *lm = EOS;
 364         ls = lm;
 365
 366         /* Find the balanced brace */
 367         for (i = 0, pe = ++ptr; *pe != EOS; pe++)
 368                 if (*pe == LBRACKET) {
 369                         /* Ignore everything between [] */
 370                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
 371                                 continue;
 372                         if (*pe == EOS) {
 373                                 /*
 374                                  * We could not find a matching RBRACKET.
 375                                  * Ignore and just look for RBRACE
 376                                  */
 377                                 pe = pm;
 378                         }
 379                 }
 380                 else if (*pe == LBRACE)
 381                         i++;
 382                 else if (*pe == RBRACE) {
 383                         if (i == 0)
 384                                 break;
 385                         i--;
 386                 }
 387
 388         /* Non matching braces; just glob the pattern */
 389         if (i != 0 || *pe == EOS)
 390                 return (glob0(pattern, pglob, limit, NULL, loc));
 391
 392         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 393                 switch (*pm) {
 394                 case LBRACKET:
 395                         /* Ignore everything between [] */
 396                         for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++)
 397                                 continue;
 398                         if (*pm == EOS) {
 399                                 /*
 400                                  * We could not find a matching RBRACKET.
 401                                  * Ignore and just look for RBRACE
 402                                  */
 403                                 pm = pm1;
 404                         }
 405                         break;
 406
 407                 case LBRACE:
 408                         i++;
 409                         break;
 410
 411                 case RBRACE:
 412                         if (i) {
 413                             i--;
 414                             break;
 415                         }
 416                         /* FALLTHROUGH */
 417                 case COMMA:
 418                         if (i && *pm == COMMA)
 419                                 break;
 420                         else {
 421                                 /* Append the current string */
 422                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 423                                         continue;
 424                                 /*
 425                                  * Append the rest of the pattern after the
 426                                  * closing brace
 427                                  */
 428                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 429                                         continue;
 430
 431                                 /* Expand the current pattern */
 432 #ifdef DEBUG
 433                                 qprintf("globexp2:", patbuf);
 434 #endif
 435                                 rv = globexp1(patbuf, pglob, limit, loc);
 436                                 if (rv)
 437                                         return (rv);
 438
 439                                 /* move after the comma, to the next string */
 440                                 pl = pm + 1;
 441                         }
 442                         break;
 443
 444                 default:
 445                         break;
 446                 }
 447         return (0);
 448 }
 449
 450
 451
 452 #ifndef BUILDING_VARIANT
 453 /*
 454  * expand tilde from the passwd file.
 455  */
 456 __private_extern__ const Char *
 457 globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob, locale_t loc)
 458 {
 459         struct passwd *pwd;
 460         char *h, *sc;
 461         const Char *p;
 462         Char *b, *eb;
 463         wchar_t wc;
 464         wchar_t wbuf[MAXPATHLEN];
 465         wchar_t *wbufend, *dc;
 466         size_t clen;
 467         mbstate_t mbs;
 468         int too_long;
 469
 470         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 471                 return (pattern);
 472
 473         /*
 474          * Copy up to the end of the string or /
 475          */
 476         eb = &patbuf[patbuf_len - 1];
 477         for (p = pattern + 1, b = patbuf;
 478             b < eb && *p != EOS && UNPROT(*p) != SEP; *b++ = *p++)
 479                 continue;
 480
 481         if (*p != EOS && UNPROT(*p) != SEP)
 482                 return (NULL);
 483
 484         *b = EOS;
 485         h = NULL;
 486
 487         if (patbuf[0] == EOS) {
 488                 /*
 489                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 490                  * we're not running setuid or setgid) and then trying
 491                  * the password file
 492                  */
 493                 if (issetugid() != 0 ||
 494                     (h = getenv("HOME")) == NULL) {
 495                         if (((h = getlogin()) != NULL &&
 496                              (pwd = getpwnam(h)) != NULL) ||
 497                             (pwd = getpwuid(getuid())) != NULL)
 498                                 h = pwd->pw_dir;
 499                         else
 500                                 return (pattern);
 501                 }
 502         }
 503         else {
 504                 /*
 505                  * Expand a ~user
 506                  */
 507                 if (g_Ctoc(patbuf, (char *)wbuf, sizeof(wbuf), loc))
 508                         return (NULL);
 509                 if ((pwd = getpwnam((char *)wbuf)) == NULL)
 510                         return (pattern);
 511                 else
 512                         h = pwd->pw_dir;
 513         }
 514
 515         /* Copy the home directory */
 516         dc = wbuf;
 517         sc = h;
 518         wbufend = wbuf + MAXPATHLEN - 1;
 519         too_long = 1;
 520         memset(&mbs, 0, sizeof(mbs));
 521         while (dc <= wbufend) {
 522                 clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs);
 523                 if (clen == (size_t)-1 || clen == (size_t)-2) {
 524                         /* XXX See initial comment #2. */
 525                         wc = (unsigned char)*sc;
 526                         clen = 1;
 527                         memset(&mbs, 0, sizeof(mbs));
 528                 }
 529                 if ((*dc++ = wc) == EOS) {
 530                         too_long = 0;
 531                         break;
 532                 }
 533                 sc += clen;
 534         }
 535         if (too_long)
 536                 return (NULL);
 537
 538         dc = wbuf;
 539         for (b = patbuf; b < eb && *dc != EOS; *b++ = *dc++ | M_PROTECT)
 540                 continue;
 541         if (*dc != EOS)
 542                 return (NULL);
 543
 544         /* Append the rest of the pattern */
 545         if (*p != EOS) {
 546                 too_long = 1;
 547                 while (b <= eb) {
 548                         if ((*b++ = *p++) == EOS) {
 549                                 too_long = 0;
 550                                 break;
 551                         }
 552                 }
 553                 if (too_long)
 554                         return (NULL);
 555         } else
 556                 *b = EOS;
 557
 558         return (patbuf);
 559 }
 560 #endif /* BUILDING_VARIANT */
 561
 562
 563 /*
 564  * The main glob() routine: compiles the pattern (optionally processing
 565  * quotes), calls glob1() to do the real pattern matching, and finally
 566  * sorts the list (unless unsorted operation is requested).  Returns 0
 567  * if things went well, nonzero if errors occurred.
 568  */
 569 static int
 570 glob0(const Char *pattern, glob_t *pglob, struct glob_limit *limit,
 571     const char *origpat, locale_t loc)
 572 {
 573         const Char *qpatnext;
 574         int err;
 575         size_t oldpathc;
 576         Char *bufnext, c, patbuf[MAXPATHLEN];
 577
 578         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob, loc);
 579         if (qpatnext == NULL) {
 580                 errno = E2BIG;
 581                 return (GLOB_NOSPACE);
 582         }
 583         oldpathc = pglob->gl_pathc;
 584         bufnext = patbuf;
 585
 586         /* We don't need to check for buffer overflow any more. */
 587         while ((c = *qpatnext++) != EOS) {
 588                 switch (c) {
 589                 case LBRACKET:
 590                         c = *qpatnext;
 591                         if (c == NOT)
 592                                 ++qpatnext;
 593                         if (*qpatnext == EOS ||
 594                             g_strchr(qpatnext+1, RBRACKET) == NULL) {
 595                                 *bufnext++ = LBRACKET;
 596                                 if (c == NOT)
 597                                         --qpatnext;
 598                                 break;
 599                         }
 600                         *bufnext++ = M_SET;
 601                         if (c == NOT)
 602                                 *bufnext++ = M_NOT;
 603                         c = *qpatnext++;
 604                         do {
 605                                 *bufnext++ = CHAR(c);
 606                                 if (*qpatnext == RANGE &&
 607                                     (c = qpatnext[1]) != RBRACKET) {
 608                                         *bufnext++ = M_RNG;
 609                                         *bufnext++ = CHAR(c);
 610                                         qpatnext += 2;
 611                                 }
 612                         } while ((c = *qpatnext++) != RBRACKET);
 613                         pglob->gl_flags |= GLOB_MAGCHAR;
 614                         *bufnext++ = M_END;
 615                         break;
 616                 case QUESTION:
 617                         pglob->gl_flags |= GLOB_MAGCHAR;
 618                         *bufnext++ = M_ONE;
 619                         break;
 620                 case STAR:
 621                         pglob->gl_flags |= GLOB_MAGCHAR;
 622                         /* collapse adjacent stars to one,
 623                          * to ensure "**" at the end continues to match the
 624                          * empty string
 625                          */
 626                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 627                             *bufnext++ = M_ALL;
 628                         break;
 629                 default:
 630                         *bufnext++ = CHAR(c);
 631                         break;
 632                 }
 633         }
 634         *bufnext = EOS;
 635 #ifdef DEBUG
 636         qprintf("glob0:", patbuf);
 637 #endif
 638
 639         if ((err = glob1(patbuf, pglob, limit, loc)) != 0)
 640                 return(err);
 641
 642         if (origpat != NULL)
 643                 return (globfinal(pglob, limit, oldpathc, origpat, loc));
 644
 645         return (0);
 646 }
 647
 648 static int
 649 globfinal(glob_t *pglob, struct glob_limit *limit, size_t oldpathc,
 650     const char *origpat, locale_t loc) {
 651         if (pglob->gl_pathc == oldpathc)
 652                 return (err_nomatch(pglob, limit, origpat, loc));
 653
 654         if (!(pglob->gl_flags & GLOB_NOSORT))
 655                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 656                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 657
 658         return (0);
 659 }
 660
 661 #ifndef BUILDING_VARIANT
 662 __private_extern__ int
 663 compare(const void *p, const void *q)
 664 {
 665         return(strcoll(*(char **)p, *(char **)q));
 666 }
 667 #endif /* BUILDING_VARIANT */
 668
 669 static int
 670 glob1(Char *pattern, glob_t *pglob, struct glob_limit *limit, locale_t loc)
 671 {
 672         Char pathbuf[MAXPATHLEN];
 673
 674         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 675         if (*pattern == EOS)
 676                 return (0);
 677         return (glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 678             pattern, pglob, limit, loc));
 679 }
 680
 681 /*
 682  * The functions glob2 and glob3 are mutually recursive; there is one level
 683  * of recursion for each segment in the pattern that contains one or more
 684  * meta characters.
 685  */
 686 static int
 687 glob2(Char *pathbuf, Char *pathend, Char *pathend_last, Char *pattern,
 688       glob_t *pglob, struct glob_limit *limit, locale_t loc)
 689 {
 690         struct stat sb;
 691         Char *p, *q;
 692         int anymeta;
 693
 694         /*
 695          * Loop over pattern segments until end of pattern or until
 696          * segment with meta character found.
 697          */
 698         for (anymeta = 0;;) {
 699                 if (*pattern == EOS) {          /* End of pattern? */
 700                         *pathend = EOS;
 701                         if (g_lstat(pathbuf, &sb, pglob, loc))
 702                                 return (0);
 703
 704                         if ((pglob->gl_flags & GLOB_LIMIT) &&
 705                             limit->l_stat_cnt++ >= GLOB_LIMIT_STAT) {
 706                                 errno = E2BIG;
 707                                 return (GLOB_NOSPACE);
 708                         }
 709                         if ((pglob->gl_flags & GLOB_MARK) &&
 710                                         UNPROT(pathend[-1]) != SEP &&
 711                                         (S_ISDIR(sb.st_mode) ||
 712                                         (S_ISLNK(sb.st_mode) &&
 713                                         g_stat(pathbuf, &sb, pglob, loc) == 0 &&
 714                                         S_ISDIR(sb.st_mode)))) {
 715                                 if (pathend + 1 > pathend_last) {
 716                                         errno = E2BIG;
 717                                         return (GLOB_NOSPACE);
 718                                 }
 719                                 *pathend++ = SEP;
 720                                 *pathend = EOS;
 721                         }
 722                         ++pglob->gl_matchc;
 723                         return (globextend(pathbuf, pglob, limit, NULL, loc));
 724                 }
 725
 726                 /* Find end of next segment, copy tentatively to pathend. */
 727                 q = pathend;
 728                 p = pattern;
 729                 while (*p != EOS && UNPROT(*p) != SEP) {
 730                         if (ismeta(*p))
 731                                 anymeta = 1;
 732                         if (q + 1 > pathend_last) {
 733                                 errno = E2BIG;
 734                                 return (GLOB_NOSPACE);
 735                         }
 736                         *q++ = *p++;
 737                 }
 738
 739                 if (!anymeta) {         /* No expansion, do next segment. */
 740                         pathend = q;
 741                         pattern = p;
 742                         while (UNPROT(*pattern) == SEP) {
 743                                 if (pathend + 1 > pathend_last) {
 744                                         errno = E2BIG;
 745                                         return (GLOB_NOSPACE);
 746                                 }
 747                                 *pathend++ = *pattern++;
 748                         }
 749                 } else                  /* Need expansion, recurse. */
 750                         return(glob3(pathbuf, pathend, pathend_last, pattern,
 751                                         p, pglob, limit, loc));
 752         }
 753         /* NOTREACHED */
 754 }
 755
 756 static int
 757 glob3(Char *pathbuf, Char *pathend, Char *pathend_last,
 758       Char *pattern, Char *restpattern,
 759       glob_t *pglob, struct glob_limit *limit, locale_t loc)
 760 {
 761         struct dirent *dp;
 762         DIR *dirp;
 763         int err, too_long, saverrno, saverrno2;
 764         char buf[MAXPATHLEN + MB_LEN_MAX - 1];
 765
 766         /*
 767          * The readdirfunc declaration can't be prototyped, because it is
 768          * assigned, below, to two functions which are prototyped in glob.h
 769          * and dirent.h as taking pointers to differently typed opaque
 770          * structures.
 771          */
 772         struct dirent *(*readdirfunc)();
 773
 774         if (pathend > pathend_last) {
 775                 errno = E2BIG;
 776                 return (GLOB_NOSPACE);
 777         }
 778         *pathend = EOS;
 779         if ((pglob->gl_errfunc != NULL || pglob->gl_errblk != NULL) &&
 780             g_Ctoc(pathbuf, buf, sizeof(buf), loc)) {
 781                 errno = E2BIG;
 782                 return (GLOB_NOSPACE);
 783         }
 784
 785         saverrno = errno;
 786         errno = 0;
 787
 788         if ((dirp = g_opendir(pathbuf, pglob, loc)) == NULL) {
 789                 if (errno == ENOENT || errno == ENOTDIR)
 790                         return (0);
 791
 792                 err = err_aborted(pglob, errno, buf);
 793                 if (errno == 0)
 794                         errno = saverrno;
 795                 return (err);
 796         }
 797
 798         err = 0;
 799
 800         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 801                 readdirfunc = pglob->gl_readdir;
 802         else
 803                 readdirfunc = readdir;
 804
 805         errno = 0;
 806         /* Search directory for matching names. */
 807         while ((dp = (*readdirfunc)(dirp)) != NULL) {
 808                 char *sc;
 809                 Char *dc;
 810                 wchar_t wc;
 811                 size_t clen;
 812                 mbstate_t mbs;
 813
 814                 if ((pglob->gl_flags & GLOB_LIMIT) &&
 815                     limit->l_readdir_cnt++ >= GLOB_LIMIT_READDIR) {
 816                         errno = E2BIG;
 817                         err = GLOB_NOSPACE;
 818                         break;
 819                 }
 820
 821                 /* Initial DOT must be matched literally. */
 822                 if (dp->d_name[0] == '.' && UNPROT(*pattern) != DOT) {
 823                         errno = 0;
 824                         continue;
 825                 }
 826                 memset(&mbs, 0, sizeof(mbs));
 827                 dc = pathend;
 828                 sc = dp->d_name;
 829                 too_long = 1;
 830                 while (dc <= pathend_last) {
 831                         clen = mbrtowc_l(&wc, sc, MB_LEN_MAX, &mbs, loc);
 832                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 833                                 /* XXX See initial comment #2. */
 834                                 wc = (unsigned char)*sc;
 835                                 clen = 1;
 836                                 memset(&mbs, 0, sizeof(mbs));
 837                         }
 838                         if ((*dc++ = wc) == EOS) {
 839                                 too_long = 0;
 840                                 break;
 841                         }
 842                         sc += clen;
 843                 }
 844                 if (too_long && (err = err_aborted(pglob, ENAMETOOLONG,
 845                     buf))) {
 846                         errno = ENAMETOOLONG;
 847                         break;
 848                 }
 849                 if (too_long || !match(pathend, pattern, restpattern, loc)) {
 850                         *pathend = EOS;
 851                         errno = 0;
 852                         continue;
 853                 }
 854                 if (errno == 0)
 855                         errno = saverrno;
 856                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 857                     pglob, limit, loc);
 858                 if (err)
 859                         break;
 860                 errno = 0;
 861         }
 862
 863         saverrno2 = errno;
 864         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 865                 (*pglob->gl_closedir)(dirp);
 866         else
 867                 closedir(dirp);
 868         errno = saverrno2;
 869
 870         if (err)
 871                 return (err);
 872
 873         if (dp == NULL && errno != 0 &&
 874             (err = err_aborted(pglob, errno, buf)))
 875                 return (err);
 876
 877         if (errno == 0)
 878                 errno = saverrno;
 879         return (0);
 880 }
 881
 882
 883 #ifndef BUILDING_VARIANT
 884 /*
 885  * Extend the gl_pathv member of a glob_t structure to accommodate a new item,
 886  * add the new item, and update gl_pathc.
 887  *
 888  * This assumes the BSD realloc, which only copies the block when its size
 889  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 890  * behavior.
 891  *
 892  * Return 0 if new item added, error code if memory couldn't be allocated.
 893  *
 894  * Invariant of the glob_t structure:
 895  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 896  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 897  */
 898 __private_extern__ int
 899 globextend(const Char *path, glob_t *pglob, struct glob_limit *limit,
 900                 const char *origpat, locale_t loc)
 901 {
 902         char **pathv;
 903         size_t i, newn, len;
 904         char *copy;
 905         const Char *p;
 906
 907         if ((pglob->gl_flags & GLOB_LIMIT) &&
 908             pglob->gl_matchc > limit->l_path_lim) {
 909                 errno = E2BIG;
 910                 return (GLOB_NOSPACE);
 911         }
 912
 913         newn = 2 + pglob->gl_pathc + pglob->gl_offs;
 914         /* reallocarray(NULL, newn, size) is equivalent to malloc(newn*size). */
 915         pathv = reallocarray(pglob->gl_pathv, newn, sizeof(*pathv));
 916         if (pathv == NULL)
 917                 return (GLOB_NOSPACE);
 918
 919         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 920                 /* first time around -- clear initial gl_offs items */
 921                 pathv += pglob->gl_offs;
 922                 for (i = pglob->gl_offs + 1; --i > 0; )
 923                         *--pathv = NULL;
 924         }
 925         pglob->gl_pathv = pathv;
 926
 927         if (origpat != NULL)
 928                 copy = strdup(origpat);
 929         else {
 930                 for (p = path; *p++ != EOS;)
 931                         continue;
 932                 len = MB_CUR_MAX_L(loc) * (size_t)(p - path); /* XXX overallocation */
 933                 if ((copy = malloc(len)) != NULL) {
 934                         if (g_Ctoc(path, copy, len, loc)) {
 935                                 free(copy);
 936                                 errno = E2BIG;
 937                                 return (GLOB_NOSPACE);
 938                         }
 939                 }
 940         }
 941         if (copy != NULL) {
 942                 limit->l_string_cnt += strlen(copy) + 1;
 943                 if ((pglob->gl_flags & GLOB_LIMIT) &&
 944                     limit->l_string_cnt >= GLOB_LIMIT_STRING) {
 945                         free(copy);
 946                         errno = E2BIG;
 947                         return (GLOB_NOSPACE);
 948                 }
 949                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 950         }
 951         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 952         return (copy == NULL ? GLOB_NOSPACE : 0);
 953 }
 954
 955 /*
 956  * pattern matching function for filenames.
 957  */
 958 __private_extern__ int
 959 match(Char *name, Char *pat, Char *patend, locale_t loc)
 960 {
 961         int ok, negate_range;
 962         Char c, k, *nextp, *nextn;
 963
 964         nextn = NULL;
 965         nextp = NULL;
 966
 967         while (1) {
 968                 while (pat < patend) {
 969                         c = *pat++;
 970                         switch (c & M_MASK) {
 971                         case M_ALL:
 972                                 if (pat == patend)
 973                                         return (1);
 974                                 if (*name == EOS)
 975                                         return (0);
 976                                 nextn = name + 1;
 977                                 nextp = pat - 1;
 978                                 break;
 979                         case M_ONE:
 980                                 if (*name++ == EOS)
 981                                         goto fail;
 982                                 break;
 983                         case M_SET:
 984                                 ok = 0;
 985                                 if ((k = *name++) == EOS)
 986                                         goto fail;
 987                                 negate_range = ((*pat & M_MASK) == M_NOT);
 988                                 if (negate_range != 0)
 989                                         ++pat;
 990                                 while (((c = *pat++) & M_MASK) != M_END)
 991                                         if ((*pat & M_MASK) == M_RNG) {
 992                                                 if (loc->__collate_load_error ?
 993                                                         CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1]) :
 994                                                            __collate_range_cmp(CHAR(c), CHAR(k), loc) <= 0
 995                                                         && __collate_range_cmp(CHAR(k), CHAR(pat[1]), loc) <= 0
 996                                                    ) {
 997                                                         ok = 1;
 998                                                 }
 999                                                 pat += 2;
1000                                         } else if (c == k)
1001                                                 ok = 1;
1002                                 if (ok == negate_range)
1003                                         goto fail;
1004                                 break;
1005                         default:
1006                                 if (*name++ != c)
1007                                         goto fail;
1008                                 break;
1009                         }
1010                 }
1011                 if (*name == EOS)
1012                         return (1);
1013
1014         fail:
1015                 if (nextn == NULL)
1016                         break;
1017                 pat = nextp;
1018                 name = nextn;
1019         }
1020         return (0);
1021 }
1022
1023 /* Free allocated data belonging to a glob_t structure. */
1024 void
1025 globfree(glob_t *pglob)
1026 {
1027         size_t i;
1028         char **pp;
1029
1030         if (pglob->gl_pathv != NULL) {
1031                 pp = pglob->gl_pathv + pglob->gl_offs;
1032                 for (i = pglob->gl_pathc; i--; ++pp)
1033                         if (*pp)
1034                                 free(*pp);
1035                 free(pglob->gl_pathv);
1036                 pglob->gl_pathv = NULL;
1037         }
1038 }
1039 #endif /* !BUILDING_VARIANT */
1040
1041 static DIR *
1042 g_opendir(Char *str, glob_t *pglob, locale_t loc)
1043 {
1044         char buf[MAXPATHLEN + MB_LEN_MAX - 1];
1045
1046         if (*str == EOS)
1047                 strcpy(buf, ".");
1048         else {
1049                 if (g_Ctoc(str, buf, sizeof(buf), loc)) {
1050                         errno = ENAMETOOLONG;
1051                         return (NULL);
1052                 }
1053         }
1054
1055         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1056                 return ((*pglob->gl_opendir)(buf));
1057
1058         return (opendir(buf));
1059 }
1060
1061 static int
1062 g_lstat(Char *fn, struct stat *sb, glob_t *pglob, locale_t loc)
1063 {
1064         char buf[MAXPATHLEN + MB_LEN_MAX - 1];
1065
1066         if (g_Ctoc(fn, buf, sizeof(buf), loc)) {
1067                 errno = ENAMETOOLONG;
1068                 return (-1);
1069         }
1070         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1071                 return((*pglob->gl_lstat)(buf, sb));
1072         return (lstat(buf, sb));
1073 }
1074
1075 static int
1076 g_stat(Char *fn, struct stat *sb, glob_t *pglob, locale_t loc)
1077 {
1078         char buf[MAXPATHLEN + MB_LEN_MAX - 1];
1079
1080         if (g_Ctoc(fn, buf, sizeof(buf), loc)) {
1081                 errno = ENAMETOOLONG;
1082                 return (-1);
1083         }
1084         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1085                 return ((*pglob->gl_stat)(buf, sb));
1086         return (stat(buf, sb));
1087 }
1088
1089 #ifndef BUILDING_VARIANT
1090 __private_extern__ const Char *
1091 g_strchr(const Char *str, wchar_t ch)
1092 {
1093
1094         do {
1095                 if (*str == ch)
1096                         return (str);
1097         } while (*str++);
1098         return (NULL);
1099 }
1100
1101 __private_extern__ int
1102 g_Ctoc(const Char *str, char *buf, size_t len, locale_t loc)
1103 {
1104         mbstate_t mbs;
1105         size_t clen;
1106         int mb_cur_max = MB_CUR_MAX_L(loc);
1107
1108         memset(&mbs, 0, sizeof(mbs));
1109         while (len >= mb_cur_max) {
1110                 clen = wcrtomb_l(buf, *str, &mbs, loc);
1111                 if (clen == (size_t)-1) {
1112                         /* XXX See initial comment #2. */
1113                         *buf = (char)CHAR(*str);
1114                         clen = 1;
1115                         memset(&mbs, 0, sizeof(mbs));
1116                 }
1117                 if (CHAR(*str) == EOS)
1118                         return (0);
1119                 str++;
1120                 buf += clen;
1121                 len -= clen;
1122         }
1123         return (1);
1124 }
1125 #endif /* !BUILDING_VARIANT */
1126
1127 static int
1128 err_nomatch(glob_t *pglob, struct glob_limit *limit, const char *origpat, locale_t loc) {
1129         /*
1130          * If there was no match we are going to append the origpat
1131          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
1132          * and the origpat did not contain any magic characters
1133          * GLOB_NOMAGIC is there just for compatibility with csh.
1134          */
1135         if ((pglob->gl_flags & GLOB_NOCHECK) ||
1136             ((pglob->gl_flags & GLOB_NOMAGIC) &&
1137             !(pglob->gl_flags & GLOB_MAGCHAR)))
1138                 return (globextend(NULL, pglob, limit, origpat, loc));
1139         return (GLOB_NOMATCH);
1140 }
1141
1142 static int
1143 err_aborted(glob_t *pglob, int err, char *buf) {
1144 #ifdef __BLOCKS__
1145         if (pglob->gl_flags & _GLOB_ERR_BLOCK && pglob->gl_errblk(buf, errno)) {
1146                 return (GLOB_ABORTED);
1147         } else
1148 #endif /* __BLOCKS__ */
1149         if (pglob->gl_errfunc != NULL && pglob->gl_errfunc(buf, errno)) {
1150                 return (GLOB_ABORTED);
1151         } else if (pglob->gl_flags & GLOB_ERR) {
1152                 return (GLOB_ABORTED);
1153         }
1154         return (0);
1155 }
1156
1157 #ifdef DEBUG
1158 static void
1159 qprintf(const char *str, Char *s)
1160 {
1161         Char *p;
1162
1163         (void)printf("%s\n", str);
1164         if (s != NULL) {
1165                 for (p = s; *p != EOS; p++)
1166                         (void)printf("%c", (char)CHAR(*p));
1167                 (void)printf("\n");
1168                 for (p = s; *p != EOS; p++)
1169                         (void)printf("%c", (isprot(*p) ? '\\' : ' '));
1170                 (void)printf("\n");
1171                 for (p = s; *p != EOS; p++)
1172                         (void)printf("%c", (ismeta(*p) ? '_' : ' '));
1173                 (void)printf("\n");
1174         }
1175 }
1176 #endif
1177 #pragma clang diagnostic pop