fnmatch_loop.c 44.9 KB
Newer Older
1
/* Copyright (C) 1991-1993, 1996-2006, 2009-2013 Free Software Foundation, Inc.
2 3 4 5
   This file is part of the GNU C Library.

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
6
   the Free Software Foundation; either version 3, or (at your option)
7 8 9 10 11 12 13 14
   any later version.

   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   You should have received a copy of the GNU General Public License
15
   along with this program; if not, see <http://www.gnu.org/licenses/>.  */
16 17 18 19

/* Match STRING against the file name pattern PATTERN, returning zero if
   it matches, nonzero if not.  */
static int EXT (INT opt, const CHAR *pattern, const CHAR *string,
20
                const CHAR *string_end, bool no_leading_period, int flags)
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
     internal_function;
static const CHAR *END (const CHAR *patternp) internal_function;

static int
internal_function
FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
     bool no_leading_period, int flags)
{
  register const CHAR *p = pattern, *n = string;
  register UCHAR c;
#ifdef _LIBC
# if WIDE_CHAR_VERSION
  const char *collseq = (const char *)
    _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQWC);
# else
  const UCHAR *collseq = (const UCHAR *)
    _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQMB);
# endif
#endif

  while ((c = *p++) != L_('\0'))
    {
      bool new_no_leading_period = false;
      c = FOLD (c);

      switch (c)
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201
        {
        case L_('?'):
          if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
            {
              int res;

              res = EXT (c, p, n, string_end, no_leading_period,
                         flags);
              if (res != -1)
                return res;
            }

          if (n == string_end)
            return FNM_NOMATCH;
          else if (*n == L_('/') && (flags & FNM_FILE_NAME))
            return FNM_NOMATCH;
          else if (*n == L_('.') && no_leading_period)
            return FNM_NOMATCH;
          break;

        case L_('\\'):
          if (!(flags & FNM_NOESCAPE))
            {
              c = *p++;
              if (c == L_('\0'))
                /* Trailing \ loses.  */
                return FNM_NOMATCH;
              c = FOLD (c);
            }
          if (n == string_end || FOLD ((UCHAR) *n) != c)
            return FNM_NOMATCH;
          break;

        case L_('*'):
          if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
            {
              int res;

              res = EXT (c, p, n, string_end, no_leading_period,
                         flags);
              if (res != -1)
                return res;
            }

          if (n != string_end && *n == L_('.') && no_leading_period)
            return FNM_NOMATCH;

          for (c = *p++; c == L_('?') || c == L_('*'); c = *p++)
            {
              if (*p == L_('(') && (flags & FNM_EXTMATCH) != 0)
                {
                  const CHAR *endp = END (p);
                  if (endp != p)
                    {
                      /* This is a pattern.  Skip over it.  */
                      p = endp;
                      continue;
                    }
                }

              if (c == L_('?'))
                {
                  /* A ? needs to match one character.  */
                  if (n == string_end)
                    /* There isn't another character; no match.  */
                    return FNM_NOMATCH;
                  else if (*n == L_('/')
                           && __builtin_expect (flags & FNM_FILE_NAME, 0))
                    /* A slash does not match a wildcard under
                       FNM_FILE_NAME.  */
                    return FNM_NOMATCH;
                  else
                    /* One character of the string is consumed in matching
                       this ? wildcard, so *??? won't match if there are
                       less than three characters.  */
                    ++n;
                }
            }

          if (c == L_('\0'))
            /* The wildcard(s) is/are the last element of the pattern.
               If the name is a file name and contains another slash
               this means it cannot match, unless the FNM_LEADING_DIR
               flag is set.  */
            {
              int result = (flags & FNM_FILE_NAME) == 0 ? 0 : FNM_NOMATCH;

              if (flags & FNM_FILE_NAME)
                {
                  if (flags & FNM_LEADING_DIR)
                    result = 0;
                  else
                    {
                      if (MEMCHR (n, L_('/'), string_end - n) == NULL)
                        result = 0;
                    }
                }

              return result;
            }
          else
            {
              const CHAR *endp;

              endp = MEMCHR (n, (flags & FNM_FILE_NAME) ? L_('/') : L_('\0'),
                             string_end - n);
              if (endp == NULL)
                endp = string_end;

              if (c == L_('[')
                  || (__builtin_expect (flags & FNM_EXTMATCH, 0) != 0
                      && (c == L_('@') || c == L_('+') || c == L_('!'))
                      && *p == L_('(')))
                {
                  int flags2 = ((flags & FNM_FILE_NAME)
                                ? flags : (flags & ~FNM_PERIOD));
                  bool no_leading_period2 = no_leading_period;

                  for (--p; n < endp; ++n, no_leading_period2 = false)
                    if (FCT (p, n, string_end, no_leading_period2, flags2)
                        == 0)
                      return 0;
                }
              else if (c == L_('/') && (flags & FNM_FILE_NAME))
                {
                  while (n < string_end && *n != L_('/'))
                    ++n;
                  if (n < string_end && *n == L_('/')
                      && (FCT (p, n + 1, string_end, flags & FNM_PERIOD, flags)
                          == 0))
                    return 0;
                }
              else
                {
                  int flags2 = ((flags & FNM_FILE_NAME)
                                ? flags : (flags & ~FNM_PERIOD));
                  int no_leading_period2 = no_leading_period;

                  if (c == L_('\\') && !(flags & FNM_NOESCAPE))
                    c = *p;
                  c = FOLD (c);
                  for (--p; n < endp; ++n, no_leading_period2 = false)
                    if (FOLD ((UCHAR) *n) == c
                        && (FCT (p, n, string_end, no_leading_period2, flags2)
                            == 0))
                      return 0;
                }
            }

          /* If we come here no match is possible with the wildcard.  */
          return FNM_NOMATCH;

        case L_('['):
          {
            /* Nonzero if the sense of the character class is inverted.  */
202 203
            const CHAR *p_init = p;
            const CHAR *n_init = n;
204 205 206 207 208 209 210 211 212 213 214 215 216 217
            register bool not;
            CHAR cold;
            UCHAR fn;

            if (posixly_correct == 0)
              posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

            if (n == string_end)
              return FNM_NOMATCH;

            if (*n == L_('.') && no_leading_period)
              return FNM_NOMATCH;

            if (*n == L_('/') && (flags & FNM_FILE_NAME))
218
              /* '/' cannot be matched.  */
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
              return FNM_NOMATCH;

            not = (*p == L_('!') || (posixly_correct < 0 && *p == L_('^')));
            if (not)
              ++p;

            fn = FOLD ((UCHAR) *n);

            c = *p++;
            for (;;)
              {
                if (!(flags & FNM_NOESCAPE) && c == L_('\\'))
                  {
                    if (*p == L_('\0'))
                      return FNM_NOMATCH;
                    c = FOLD ((UCHAR) *p);
                    ++p;

                    goto normal_bracket;
                  }
                else if (c == L_('[') && *p == L_(':'))
                  {
                    /* Leave room for the null.  */
                    CHAR str[CHAR_CLASS_MAX_LENGTH + 1];
                    size_t c1 = 0;
244
#if defined _LIBC || WIDE_CHAR_SUPPORT
245
                    wctype_t wt;
246
#endif
247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272
                    const CHAR *startp = p;

                    for (;;)
                      {
                        if (c1 == CHAR_CLASS_MAX_LENGTH)
                          /* The name is too long and therefore the pattern
                             is ill-formed.  */
                          return FNM_NOMATCH;

                        c = *++p;
                        if (c == L_(':') && p[1] == L_(']'))
                          {
                            p += 2;
                            break;
                          }
                        if (c < L_('a') || c >= L_('z'))
                          {
                            /* This cannot possibly be a character class name.
                               Match it as a normal range.  */
                            p = startp;
                            c = L_('[');
                            goto normal_bracket;
                          }
                        str[c1++] = c;
                      }
                    str[c1] = L_('\0');
273 274

#if defined _LIBC || WIDE_CHAR_SUPPORT
275 276 277 278
                    wt = IS_CHAR_CLASS (str);
                    if (wt == 0)
                      /* Invalid character class name.  */
                      return FNM_NOMATCH;
279 280

# if defined _LIBC && ! WIDE_CHAR_VERSION
281 282 283 284 285
                    /* The following code is glibc specific but does
                       there a good job in speeding up the code since
                       we can avoid the btowc() call.  */
                    if (_ISCTYPE ((UCHAR) *n, wt))
                      goto matched;
286
# else
287 288
                    if (ISWCTYPE (BTOWC ((UCHAR) *n), wt))
                      goto matched;
289 290
# endif
#else
291 292 293 294 295 296 297 298 299 300 301 302 303
                    if ((STREQ (str, L_("alnum")) && isalnum ((UCHAR) *n))
                        || (STREQ (str, L_("alpha")) && isalpha ((UCHAR) *n))
                        || (STREQ (str, L_("blank")) && isblank ((UCHAR) *n))
                        || (STREQ (str, L_("cntrl")) && iscntrl ((UCHAR) *n))
                        || (STREQ (str, L_("digit")) && isdigit ((UCHAR) *n))
                        || (STREQ (str, L_("graph")) && isgraph ((UCHAR) *n))
                        || (STREQ (str, L_("lower")) && islower ((UCHAR) *n))
                        || (STREQ (str, L_("print")) && isprint ((UCHAR) *n))
                        || (STREQ (str, L_("punct")) && ispunct ((UCHAR) *n))
                        || (STREQ (str, L_("space")) && isspace ((UCHAR) *n))
                        || (STREQ (str, L_("upper")) && isupper ((UCHAR) *n))
                        || (STREQ (str, L_("xdigit")) && isxdigit ((UCHAR) *n)))
                      goto matched;
304
#endif
305 306
                    c = *p++;
                  }
307
#ifdef _LIBC
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
                else if (c == L_('[') && *p == L_('='))
                  {
                    UCHAR str[1];
                    uint32_t nrules =
                      _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
                    const CHAR *startp = p;

                    c = *++p;
                    if (c == L_('\0'))
                      {
                        p = startp;
                        c = L_('[');
                        goto normal_bracket;
                      }
                    str[0] = c;

                    c = *++p;
                    if (c != L_('=') || p[1] != L_(']'))
                      {
                        p = startp;
                        c = L_('[');
                        goto normal_bracket;
                      }
                    p += 2;

                    if (nrules == 0)
                      {
                        if ((UCHAR) *n == str[0])
                          goto matched;
                      }
                    else
                      {
                        const int32_t *table;
341
# if WIDE_CHAR_VERSION
342 343
                        const int32_t *weights;
                        const int32_t *extra;
344
# else
345 346
                        const unsigned char *weights;
                        const unsigned char *extra;
347
# endif
348 349 350
                        const int32_t *indirect;
                        int32_t idx;
                        const UCHAR *cp = (const UCHAR *) str;
351

352
                        /* This #include defines a local function!  */
353 354 355 356 357 358 359
# if WIDE_CHAR_VERSION
#  include <locale/weightwc.h>
# else
#  include <locale/weight.h>
# endif

# if WIDE_CHAR_VERSION
360 361 362 363 364 365 366 367
                        table = (const int32_t *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC);
                        weights = (const int32_t *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC);
                        extra = (const int32_t *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC);
                        indirect = (const int32_t *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC);
368
# else
369 370 371 372 373 374 375 376
                        table = (const int32_t *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
                        weights = (const unsigned char *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
                        extra = (const unsigned char *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
                        indirect = (const int32_t *)
                          _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
377 378
# endif

379 380 381 382 383
                        idx = findidx (&cp);
                        if (idx != 0)
                          {
                            /* We found a table entry.  Now see whether the
                               character we are currently at has the same
384
                               equivalence class value.  */
385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
                            int len = weights[idx & 0xffffff];
                            int32_t idx2;
                            const UCHAR *np = (const UCHAR *) n;

                            idx2 = findidx (&np);
                            if (idx2 != 0
                                && (idx >> 24) == (idx2 >> 24)
                                && len == weights[idx2 & 0xffffff])
                              {
                                int cnt = 0;

                                idx &= 0xffffff;
                                idx2 &= 0xffffff;

                                while (cnt < len
                                       && (weights[idx + 1 + cnt]
                                           == weights[idx2 + 1 + cnt]))
                                  ++cnt;

                                if (cnt == len)
                                  goto matched;
                              }
                          }
                      }

                    c = *p++;
                  }
412
#endif
413
                else if (c == L_('\0'))
414 415 416 417 418 419 420
                  {
                    /* [ unterminated, treat as normal character.  */
                    p = p_init;
                    n = n_init;
                    c = L_('[');
                    goto normal_match;
                  }
421 422 423
                else
                  {
                    bool is_range = false;
424 425

#ifdef _LIBC
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470
                    bool is_seqval = false;

                    if (c == L_('[') && *p == L_('.'))
                      {
                        uint32_t nrules =
                          _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
                        const CHAR *startp = p;
                        size_t c1 = 0;

                        while (1)
                          {
                            c = *++p;
                            if (c == L_('.') && p[1] == L_(']'))
                              {
                                p += 2;
                                break;
                              }
                            if (c == '\0')
                              return FNM_NOMATCH;
                            ++c1;
                          }

                        /* We have to handling the symbols differently in
                           ranges since then the collation sequence is
                           important.  */
                        is_range = *p == L_('-') && p[1] != L_('\0');

                        if (nrules == 0)
                          {
                            /* There are no names defined in the collation
                               data.  Therefore we only accept the trivial
                               names consisting of the character itself.  */
                            if (c1 != 1)
                              return FNM_NOMATCH;

                            if (!is_range && *n == startp[1])
                              goto matched;

                            cold = startp[1];
                            c = *p++;
                          }
                        else
                          {
                            int32_t table_size;
                            const int32_t *symb_table;
471
# ifdef WIDE_CHAR_VERSION
472 473
                            char str[c1];
                            size_t strcnt;
474 475 476
# else
#  define str (startp + 1)
# endif
477 478 479 480 481
                            const unsigned char *extra;
                            int32_t idx;
                            int32_t elem;
                            int32_t second;
                            int32_t hash;
482 483

# ifdef WIDE_CHAR_VERSION
484 485 486 487 488 489
                            /* We have to convert the name to a single-byte
                               string.  This is possible since the names
                               consist of ASCII characters and the internal
                               representation is UCS4.  */
                            for (strcnt = 0; strcnt < c1; ++strcnt)
                              str[strcnt] = startp[1 + strcnt];
490 491
# endif

492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537
                            table_size =
                              _NL_CURRENT_WORD (LC_COLLATE,
                                                _NL_COLLATE_SYMB_HASH_SIZEMB);
                            symb_table = (const int32_t *)
                              _NL_CURRENT (LC_COLLATE,
                                           _NL_COLLATE_SYMB_TABLEMB);
                            extra = (const unsigned char *)
                              _NL_CURRENT (LC_COLLATE,
                                           _NL_COLLATE_SYMB_EXTRAMB);

                            /* Locate the character in the hashing table.  */
                            hash = elem_hash (str, c1);

                            idx = 0;
                            elem = hash % table_size;
                            if (symb_table[2 * elem] != 0)
                              {
                                second = hash % (table_size - 2) + 1;

                                do
                                  {
                                    /* First compare the hashing value.  */
                                    if (symb_table[2 * elem] == hash
                                        && (c1
                                            == extra[symb_table[2 * elem + 1]])
                                        && memcmp (str,
                                                   &extra[symb_table[2 * elem
                                                                     + 1]
                                                          + 1], c1) == 0)
                                      {
                                        /* Yep, this is the entry.  */
                                        idx = symb_table[2 * elem + 1];
                                        idx += 1 + extra[idx];
                                        break;
                                      }

                                    /* Next entry.  */
                                    elem += second;
                                  }
                                while (symb_table[2 * elem] != 0);
                              }

                            if (symb_table[2 * elem] != 0)
                              {
                                /* Compare the byte sequence but only if
                                   this is not part of a range.  */
538
# ifdef WIDE_CHAR_VERSION
539
                                int32_t *wextra;
540

541 542 543
                                idx += 1 + extra[idx];
                                /* Adjust for the alignment.  */
                                idx = (idx + 3) & ~3;
544

545
                                wextra = (int32_t *) &extra[idx + 4];
546 547
# endif

548 549
                                if (! is_range)
                                  {
550
# ifdef WIDE_CHAR_VERSION
551 552 553 554 555 556 557 558
                                    for (c1 = 0;
                                         (int32_t) c1 < wextra[idx];
                                         ++c1)
                                      if (n[c1] != wextra[1 + c1])
                                        break;

                                    if ((int32_t) c1 == wextra[idx])
                                      goto matched;
559
# else
560 561 562
                                    for (c1 = 0; c1 < extra[idx]; ++c1)
                                      if (n[c1] != extra[1 + c1])
                                        break;
563

564 565
                                    if (c1 == extra[idx])
                                      goto matched;
566
# endif
567
                                  }
568

569 570
                                /* Get the collation sequence value.  */
                                is_seqval = true;
571
# ifdef WIDE_CHAR_VERSION
572
                                cold = wextra[1 + wextra[idx]];
573
# else
574 575 576 577
                                /* Adjust for the alignment.  */
                                idx += 1 + extra[idx];
                                idx = (idx + 3) & ~4;
                                cold = *((int32_t *) &extra[idx]);
578 579
# endif

580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596
                                c = *p++;
                              }
                            else if (c1 == 1)
                              {
                                /* No valid character.  Match it as a
                                   single byte.  */
                                if (!is_range && *n == str[0])
                                  goto matched;

                                cold = str[0];
                                c = *p++;
                              }
                            else
                              return FNM_NOMATCH;
                          }
                      }
                    else
597 598
# undef str
#endif
599 600 601
                      {
                        c = FOLD (c);
                      normal_bracket:
602

603 604 605 606 607
                        /* We have to handling the symbols differently in
                           ranges since then the collation sequence is
                           important.  */
                        is_range = (*p == L_('-') && p[1] != L_('\0')
                                    && p[1] != L_(']'));
608

609 610
                        if (!is_range && c == fn)
                          goto matched;
611 612

#if _LIBC
613 614 615
                        /* This is needed if we goto normal_bracket; from
                           outside of is_seqval's scope.  */
                        is_seqval = false;
616 617
#endif

618 619 620
                        cold = c;
                        c = *p++;
                      }
621

622 623
                    if (c == L_('-') && *p != L_(']'))
                      {
624
#if _LIBC
625 626 627 628 629 630 631 632 633 634 635
                        /* We have to find the collation sequence
                           value for C.  Collation sequence is nothing
                           we can regularly access.  The sequence
                           value is defined by the order in which the
                           definitions of the collation values for the
                           various characters appear in the source
                           file.  A strange concept, nowhere
                           documented.  */
                        uint32_t fcollseq;
                        uint32_t lcollseq;
                        UCHAR cend = *p++;
636 637

# ifdef WIDE_CHAR_VERSION
638
                        /* Search in the 'names' array for the characters.  */
639 640 641 642 643 644 645 646 647 648 649
                        fcollseq = __collseq_table_lookup (collseq, fn);
                        if (fcollseq == ~((uint32_t) 0))
                          /* XXX We don't know anything about the character
                             we are supposed to match.  This means we are
                             failing.  */
                          goto range_not_matched;

                        if (is_seqval)
                          lcollseq = cold;
                        else
                          lcollseq = __collseq_table_lookup (collseq, cold);
650
# else
651 652
                        fcollseq = collseq[fn];
                        lcollseq = is_seqval ? cold : collseq[(UCHAR) cold];
653 654
# endif

655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
                        is_seqval = false;
                        if (cend == L_('[') && *p == L_('.'))
                          {
                            uint32_t nrules =
                              _NL_CURRENT_WORD (LC_COLLATE,
                                                _NL_COLLATE_NRULES);
                            const CHAR *startp = p;
                            size_t c1 = 0;

                            while (1)
                              {
                                c = *++p;
                                if (c == L_('.') && p[1] == L_(']'))
                                  {
                                    p += 2;
                                    break;
                                  }
                                if (c == '\0')
                                  return FNM_NOMATCH;
                                ++c1;
                              }

                            if (nrules == 0)
                              {
                                /* There are no names defined in the
                                   collation data.  Therefore we only
                                   accept the trivial names consisting
                                   of the character itself.  */
                                if (c1 != 1)
                                  return FNM_NOMATCH;

                                cend = startp[1];
                              }
                            else
                              {
                                int32_t table_size;
                                const int32_t *symb_table;
692
# ifdef WIDE_CHAR_VERSION
693 694
                                char str[c1];
                                size_t strcnt;
695 696 697
# else
#  define str (startp + 1)
# endif
698 699 700 701 702
                                const unsigned char *extra;
                                int32_t idx;
                                int32_t elem;
                                int32_t second;
                                int32_t hash;
703 704

# ifdef WIDE_CHAR_VERSION
705 706 707 708 709 710
                                /* We have to convert the name to a single-byte
                                   string.  This is possible since the names
                                   consist of ASCII characters and the internal
                                   representation is UCS4.  */
                                for (strcnt = 0; strcnt < c1; ++strcnt)
                                  str[strcnt] = startp[1 + strcnt];
711 712
# endif

713 714 715 716 717 718 719 720 721 722 723
                                table_size =
                                  _NL_CURRENT_WORD (LC_COLLATE,
                                                    _NL_COLLATE_SYMB_HASH_SIZEMB);
                                symb_table = (const int32_t *)
                                  _NL_CURRENT (LC_COLLATE,
                                               _NL_COLLATE_SYMB_TABLEMB);
                                extra = (const unsigned char *)
                                  _NL_CURRENT (LC_COLLATE,
                                               _NL_COLLATE_SYMB_EXTRAMB);

                                /* Locate the character in the hashing
724
                                   table.  */
725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
                                hash = elem_hash (str, c1);

                                idx = 0;
                                elem = hash % table_size;
                                if (symb_table[2 * elem] != 0)
                                  {
                                    second = hash % (table_size - 2) + 1;

                                    do
                                      {
                                        /* First compare the hashing value.  */
                                        if (symb_table[2 * elem] == hash
                                            && (c1
                                                == extra[symb_table[2 * elem + 1]])
                                            && memcmp (str,
                                                       &extra[symb_table[2 * elem + 1]
                                                              + 1], c1) == 0)
                                          {
                                            /* Yep, this is the entry.  */
                                            idx = symb_table[2 * elem + 1];
                                            idx += 1 + extra[idx];
                                            break;
                                          }

                                        /* Next entry.  */
                                        elem += second;
                                      }
                                    while (symb_table[2 * elem] != 0);
                                  }

                                if (symb_table[2 * elem] != 0)
                                  {
                                    /* Compare the byte sequence but only if
                                       this is not part of a range.  */
759
# ifdef WIDE_CHAR_VERSION
760
                                    int32_t *wextra;
761

762 763 764
                                    idx += 1 + extra[idx];
                                    /* Adjust for the alignment.  */
                                    idx = (idx + 3) & ~4;
765

766
                                    wextra = (int32_t *) &extra[idx + 4];
767
# endif
768 769
                                    /* Get the collation sequence value.  */
                                    is_seqval = true;
770
# ifdef WIDE_CHAR_VERSION
771
                                    cend = wextra[1 + wextra[idx]];
772
# else
773 774 775 776
                                    /* Adjust for the alignment.  */
                                    idx += 1 + extra[idx];
                                    idx = (idx + 3) & ~4;
                                    cend = *((int32_t *) &extra[idx]);
777
# endif
778 779 780 781 782 783 784 785 786
                                  }
                                else if (symb_table[2 * elem] != 0 && c1 == 1)
                                  {
                                    cend = str[0];
                                    c = *p++;
                                  }
                                else
                                  return FNM_NOMATCH;
                              }
787
# undef str
788 789 790 791 792 793 794 795 796 797 798 799 800 801
                          }
                        else
                          {
                            if (!(flags & FNM_NOESCAPE) && cend == L_('\\'))
                              cend = *p++;
                            if (cend == L_('\0'))
                              return FNM_NOMATCH;
                            cend = FOLD (cend);
                          }

                        /* XXX It is not entirely clear to me how to handle
                           characters which are not mentioned in the
                           collation specification.  */
                        if (
802
# ifdef WIDE_CHAR_VERSION
803
                            lcollseq == 0xffffffff ||
804
# endif
805 806 807 808 809 810 811 812 813
                            lcollseq <= fcollseq)
                          {
                            /* We have to look at the upper bound.  */
                            uint32_t hcollseq;

                            if (is_seqval)
                              hcollseq = cend;
                            else
                              {
814
# ifdef WIDE_CHAR_VERSION
815 816 817 818 819 820 821 822 823 824 825 826
                                hcollseq =
                                  __collseq_table_lookup (collseq, cend);
                                if (hcollseq == ~((uint32_t) 0))
                                  {
                                    /* Hum, no information about the upper
                                       bound.  The matching succeeds if the
                                       lower bound is matched exactly.  */
                                    if (lcollseq != fcollseq)
                                      goto range_not_matched;

                                    goto matched;
                                  }
827
# else
828
                                hcollseq = collseq[cend];
829
# endif
830
                              }
831

832 833 834
                            if (lcollseq <= hcollseq && fcollseq <= hcollseq)
                              goto matched;
                          }
835
# ifdef WIDE_CHAR_VERSION
836
                      range_not_matched:
837 838
# endif
#else
839 840
                        /* We use a boring value comparison of the character
                           values.  This is better than comparing using
841
                           'strcoll' since the latter would have surprising
842 843 844 845 846 847 848 849 850 851 852
                           and sometimes fatal consequences.  */
                        UCHAR cend = *p++;

                        if (!(flags & FNM_NOESCAPE) && cend == L_('\\'))
                          cend = *p++;
                        if (cend == L_('\0'))
                          return FNM_NOMATCH;

                        /* It is a range.  */
                        if (cold <= fn && fn <= cend)
                          goto matched;
853 854
#endif

855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968
                        c = *p++;
                      }
                  }

                if (c == L_(']'))
                  break;
              }

            if (!not)
              return FNM_NOMATCH;
            break;

          matched:
            /* Skip the rest of the [...] that already matched.  */
            do
              {
              ignore_next:
                c = *p++;

                if (c == L_('\0'))
                  /* [... (unterminated) loses.  */
                  return FNM_NOMATCH;

                if (!(flags & FNM_NOESCAPE) && c == L_('\\'))
                  {
                    if (*p == L_('\0'))
                      return FNM_NOMATCH;
                    /* XXX 1003.2d11 is unclear if this is right.  */
                    ++p;
                  }
                else if (c == L_('[') && *p == L_(':'))
                  {
                    int c1 = 0;
                    const CHAR *startp = p;

                    while (1)
                      {
                        c = *++p;
                        if (++c1 == CHAR_CLASS_MAX_LENGTH)
                          return FNM_NOMATCH;

                        if (*p == L_(':') && p[1] == L_(']'))
                          break;

                        if (c < L_('a') || c >= L_('z'))
                          {
                            p = startp;
                            goto ignore_next;
                          }
                      }
                    p += 2;
                    c = *p++;
                  }
                else if (c == L_('[') && *p == L_('='))
                  {
                    c = *++p;
                    if (c == L_('\0'))
                      return FNM_NOMATCH;
                    c = *++p;
                    if (c != L_('=') || p[1] != L_(']'))
                      return FNM_NOMATCH;
                    p += 2;
                    c = *p++;
                  }
                else if (c == L_('[') && *p == L_('.'))
                  {
                    ++p;
                    while (1)
                      {
                        c = *++p;
                        if (c == '\0')
                          return FNM_NOMATCH;

                        if (*p == L_('.') && p[1] == L_(']'))
                          break;
                      }
                    p += 2;
                    c = *p++;
                  }
              }
            while (c != L_(']'));
            if (not)
              return FNM_NOMATCH;
          }
          break;

        case L_('+'):
        case L_('@'):
        case L_('!'):
          if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
            {
              int res;

              res = EXT (c, p, n, string_end, no_leading_period, flags);
              if (res != -1)
                return res;
            }
          goto normal_match;

        case L_('/'):
          if (NO_LEADING_PERIOD (flags))
            {
              if (n == string_end || c != (UCHAR) *n)
                return FNM_NOMATCH;

              new_no_leading_period = true;
              break;
            }
          /* FALLTHROUGH */
        default:
        normal_match:
          if (n == string_end || c != FOLD ((UCHAR) *n))
            return FNM_NOMATCH;
        }
969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996

      no_leading_period = new_no_leading_period;
      ++n;
    }

  if (n == string_end)
    return 0;

  if ((flags & FNM_LEADING_DIR) && n != string_end && *n == L_('/'))
    /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz".  */
    return 0;

  return FNM_NOMATCH;
}


static const CHAR *
internal_function
END (const CHAR *pattern)
{
  const CHAR *p = pattern;

  while (1)
    if (*++p == L_('\0'))
      /* This is an invalid pattern.  */
      return pattern;
    else if (*p == L_('['))
      {
997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
        /* Handle brackets special.  */
        if (posixly_correct == 0)
          posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

        /* Skip the not sign.  We have to recognize it because of a possibly
           following ']'.  */
        if (*++p == L_('!') || (posixly_correct < 0 && *p == L_('^')))
          ++p;
        /* A leading ']' is recognized as such.  */
        if (*p == L_(']'))
          ++p;
        /* Skip over all characters of the list.  */
        while (*p != L_(']'))
          if (*p++ == L_('\0'))
            /* This is no valid pattern.  */
            return pattern;
1013 1014
      }
    else if ((*p == L_('?') || *p == L_('*') || *p == L_('+') || *p == L_('@')
1015
              || *p == L_('!')) && p[1] == L_('('))
1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
      p = END (p + 1);
    else if (*p == L_(')'))
      break;

  return p + 1;
}


static int
internal_function
EXT (INT opt, const CHAR *pattern, const CHAR *string, const CHAR *string_end,
     bool no_leading_period, int flags)
{
  const CHAR *startp;
  size_t level;
  struct patternlist
  {
    struct patternlist *next;
    CHAR str[1];
  } *list = NULL;
  struct patternlist **lastp = &list;
  size_t pattern_len = STRLEN (pattern);
  const CHAR *p;
  const CHAR *rs;
  enum { ALLOCA_LIMIT = 8000 };

  /* Parse the pattern.  Store the individual parts in the list.  */
  level = 0;
  for (startp = p = pattern + 1; ; ++p)
    if (*p == L_('\0'))
      /* This is an invalid pattern.  */
      return -1;
    else if (*p == L_('['))
      {
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
        /* Handle brackets special.  */
        if (posixly_correct == 0)
          posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

        /* Skip the not sign.  We have to recognize it because of a possibly
           following ']'.  */
        if (*++p == L_('!') || (posixly_correct < 0 && *p == L_('^')))
          ++p;
        /* A leading ']' is recognized as such.  */
        if (*p == L_(']'))
          ++p;
        /* Skip over all characters of the list.  */
        while (*p != L_(']'))
          if (*p++ == L_('\0'))
            /* This is no valid pattern.  */
            return -1;
1066 1067
      }
    else if ((*p == L_('?') || *p == L_('*') || *p == L_('+') || *p == L_('@')
1068
              || *p == L_('!')) && p[1] == L_('('))
1069 1070 1071 1072
      /* Remember the nesting level.  */
      ++level;
    else if (*p == L_(')'))
      {
1073 1074 1075
        if (level-- == 0)
          {
            /* This means we found the end of the pattern.  */
1076
#define NEW_PATTERN \
1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098
            struct patternlist *newp;                                         \
            size_t plen;                                                      \
            size_t plensize;                                                  \
            size_t newpsize;                                                  \
                                                                              \
            plen = (opt == L_('?') || opt == L_('@')                          \
                    ? pattern_len                                             \
                    : p - startp + 1UL);                                      \
            plensize = plen * sizeof (CHAR);                                  \
            newpsize = offsetof (struct patternlist, str) + plensize;         \
            if ((size_t) -1 / sizeof (CHAR) < plen                            \
                || newpsize < offsetof (struct patternlist, str)              \
                || ALLOCA_LIMIT <= newpsize)                                  \
              return -1;                                                      \
            newp = (struct patternlist *) alloca (newpsize);                  \
            *((CHAR *) MEMPCPY (newp->str, startp, p - startp)) = L_('\0');    \
            newp->next = NULL;                                                \
            *lastp = newp;                                                    \
            lastp = &newp->next
            NEW_PATTERN;
            break;
          }
1099 1100 1101
      }
    else if (*p == L_('|'))
      {
1102 1103 1104 1105 1106
        if (level == 0)
          {
            NEW_PATTERN;
            startp = p + 1;
          }
1107 1108 1109 1110 1111 1112 1113 1114 1115
      }
  assert (list != NULL);
  assert (p[-1] == L_(')'));
#undef NEW_PATTERN

  switch (opt)
    {
    case L_('*'):
      if (FCT (p, string, string_end, no_leading_period, flags) == 0)
1116
        return 0;
1117 1118 1119 1120
      /* FALLTHROUGH */

    case L_('+'):
      do
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
        {
          for (rs = string; rs <= string_end; ++rs)
            /* First match the prefix with the current pattern with the
               current pattern.  */
            if (FCT (list->str, string, rs, no_leading_period,
                     flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0
                /* This was successful.  Now match the rest with the rest
                   of the pattern.  */
                && (FCT (p, rs, string_end,
                         rs == string
                         ? no_leading_period
                         : rs[-1] == '/' && NO_LEADING_PERIOD (flags),
                         flags & FNM_FILE_NAME
                         ? flags : flags & ~FNM_PERIOD) == 0
                    /* This didn't work.  Try the whole pattern.  */
                    || (rs != string
                        && FCT (pattern - 1, rs, string_end,
                                rs == string
                                ? no_leading_period
                                : rs[-1] == '/' && NO_LEADING_PERIOD (flags),
                                flags & FNM_FILE_NAME
                                ? flags : flags & ~FNM_PERIOD) == 0)))
              /* It worked.  Signal success.  */
              return 0;
        }
1146 1147 1148 1149 1150 1151 1152
      while ((list = list->next) != NULL);

      /* None of the patterns lead to a match.  */
      return FNM_NOMATCH;

    case L_('?'):
      if (FCT (p, string, string_end, no_leading_period, flags) == 0)
1153
        return 0;
1154 1155 1156 1157
      /* FALLTHROUGH */

    case L_('@'):
      do
1158
        /* I cannot believe it but 'strcat' is actually acceptable
1159 1160 1161 1162 1163 1164 1165 1166
           here.  Match the entire string with the prefix from the
           pattern list and the rest of the pattern following the
           pattern list.  */
        if (FCT (STRCAT (list->str, p), string, string_end,
                 no_leading_period,
                 flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0)
          /* It worked.  Signal success.  */
          return 0;
1167 1168 1169 1170 1171 1172 1173
      while ((list = list->next) != NULL);

      /* None of the patterns lead to a match.  */
      return FNM_NOMATCH;

    case L_('!'):
      for (rs = string; rs <= string_end; ++rs)
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
        {
          struct patternlist *runp;

          for (runp = list; runp != NULL; runp = runp->next)
            if (FCT (runp->str, string, rs,  no_leading_period,
                     flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0)
              break;

          /* If none of the patterns matched see whether the rest does.  */
          if (runp == NULL
              && (FCT (p, rs, string_end,
                       rs == string
                       ? no_leading_period
                       : rs[-1] == '/' && NO_LEADING_PERIOD (flags),
                       flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD)
                  == 0))
            /* This is successful.  */
            return 0;
        }
1193 1194

      /* None of the patterns together with the rest of the pattern
1195
         lead to a match.  */
1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219
      return FNM_NOMATCH;

    default:
      assert (! "Invalid extended matching operator");
      break;
    }

  return -1;
}


#undef FOLD
#undef CHAR
#undef UCHAR
#undef INT
#undef FCT
#undef EXT
#undef END
#undef MEMPCPY
#undef MEMCHR
#undef STRLEN
#undef STRCAT
#undef L_
#undef BTOWC