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