winsup/cygwin/glob.cc

   1 /*
   2  * Copyright (c) 1989, 1993
   3  *      The Regents of the University of California.  All rights reserved.
   4  *
   5  * This code is derived from software contributed to Berkeley by
   6  * Guido van Rossum.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 4. Neither the name of the University nor the names of its contributors
  17  *    may be used to endorse or promote products derived from this software
  18  *    without specific prior written permission.
  19  *
  20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  30  * SUCH DAMAGE.
  31  */
  32
  33 #ifdef __CYGWIN__
  34 #include "winsup.h"
  35 #endif
  36
  37 #if defined(LIBC_SCCS) && !defined(lint)
  38 static char sccsid[] = "@(#)glob.c      8.3 (Berkeley) 10/13/93";
  39 #endif /* LIBC_SCCS and not lint */
  40 #include <sys/cdefs.h>
  41 __FBSDID("$FreeBSD: src/lib/libc/gen/glob.c,v 1.28 2010/05/12 17:44:00 gordon Exp $");
  42
  43 /*
  44  * glob(3) -- a superset of the one defined in POSIX 1003.2.
  45  *
  46  * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
  47  *
  48  * Optional extra services, controlled by flags not defined by POSIX:
  49  *
  50  * GLOB_QUOTE:
  51  *      Escaping convention: \ inhibits any special meaning the following
  52  *      character might have (except \ at end of string is retained).
  53  * GLOB_MAGCHAR:
  54  *      Set in gl_flags if pattern contained a globbing character.
  55  * GLOB_NOMAGIC:
  56  *      Same as GLOB_NOCHECK, but it will only append pattern if it did
  57  *      not contain any magic characters.  [Used in csh style globbing]
  58  * GLOB_ALTDIRFUNC:
  59  *      Use alternately specified directory access functions.
  60  * GLOB_TILDE:
  61  *      expand ~user/foo to the /home/dir/of/user/foo
  62  * GLOB_BRACE:
  63  *      expand {1,2}{a,b} to 1a 1b 2a 2b
  64  * gl_matchc:
  65  *      Number of matches in the current invocation of glob.
  66  */
  67
  68 /*
  69  * Some notes on multibyte character support:
  70  * 1. Patterns with illegal byte sequences match nothing - even if
  71  *    GLOB_NOCHECK is specified.
  72  * 2. Illegal byte sequences in filenames are handled by treating them as
  73  *    single-byte characters with a value of the first byte of the sequence
  74  *    cast to wint_t.
  75  * 3. State-dependent encodings are not currently supported.
  76  */
  77
  78 #include <sys/param.h>
  79 #include <sys/stat.h>
  80
  81 #include <ctype.h>
  82 #include <dirent.h>
  83 #include <errno.h>
  84 #include <glob.h>
  85 #include <limits.h>
  86 #include <pwd.h>
  87 #include <stdint.h>
  88 #include <stdio.h>
  89 #include <stdlib.h>
  90 #include <string.h>
  91 #include <unistd.h>
  92 #include <wchar.h>
  93 #include <wctype.h>
  94
  95 #include "collate.h"
  96
  97 #ifdef __CYGWIN__
  98 #define Cchar(c)        (ignore_case_with_glob ? towlower (c) : (c))
  99 #endif
 100
 101 #undef MAXPATHLEN
 102 #define MAXPATHLEN 8192
 103
 104 #define DOLLAR          '$'
 105 #define DOT             '.'
 106 #define COLON           ':'
 107 #define EQUALS          '='
 108 #define EOS             '\0'
 109 #define LBRACKET        '['
 110 #define NOT             '!'
 111 #define QUESTION        '?'
 112 #define QUOTE           '\\'
 113 #define RANGE           '-'
 114 #define RBRACKET        ']'
 115 #define SEP             '/'
 116 #define STAR            '*'
 117 #define TILDE           '~'
 118 #define UNDERSCORE      '_'
 119 #define LBRACE          '{'
 120 #define RBRACE          '}'
 121 #define SLASH           '/'
 122 #define COMMA           ','
 123
 124 #ifndef DEBUG
 125
 126 #define M_QUOTE         0x40000000U
 127 #define M_PROTECT       0x20000000U
 128 #define M_MASK          0x70ffffffU
 129 #define M_COLL_MASK     0x700000ffU
 130 #define M_CHAR          0x00ffffffU
 131
 132 typedef wint_t Char;
 133
 134 #else
 135
 136 #define M_QUOTE         0x80
 137 #define M_PROTECT       0x40
 138 #define M_MASK          0xff
 139 #define M_CHAR          0x7f
 140
 141 typedef char Char;
 142
 143 #endif
 144
 145
 146 #define CHAR(c)         ((Char)((c)&M_CHAR))
 147 #define META(c)         ((Char)((c)|M_QUOTE))
 148 #define M_ALL           META('*')
 149 #define M_END           META(']')
 150 #define M_NOT           META('!')
 151 #define M_ONE           META('?')
 152 #define M_RNG           META('-')
 153 #define M_SET           META('[')
 154 #define M_NAMED         META(':')
 155 #define M_EQUIV         META('=')
 156 #define M_COLL(_ccnt)   META('.' | ((_ccnt) << 8))
 157 #define M_COLL_P(_c)    (((_c) & M_COLL_MASK) == META('.'))
 158 #define M_COLL_CNT(_c)  (((_c) & ~M_COLL_MASK) >> 8)
 159 #define ismeta(c)       (((c)&M_QUOTE) != 0)
 160
 161 static int       compare(const void *, const void *);
 162 static int       g_Ctoc(const Char *, char *, size_t);
 163 static int       g_lstat(Char *, struct stat *, glob_t *);
 164 static DIR      *g_opendir(Char *, glob_t *);
 165 static const Char *g_strchr(const Char *, wint_t);
 166 #ifdef notdef
 167 static Char     *g_strcat(Char *, const Char *);
 168 #endif
 169 static int       g_stat(Char *, struct stat *, glob_t *);
 170 static int       glob0(const Char *, glob_t *, size_t *);
 171 static int       glob1(Char *, glob_t *, size_t *);
 172 static int       glob2(Char *, Char *, Char *, Char *, glob_t *, size_t *);
 173 static int       glob3(Char *, Char *, Char *, Char *, Char *, glob_t *, size_t *);
 174 static int       globextend(const Char *, glob_t *, size_t *);
 175 static const Char *
 176                  globtilde(const Char *, Char *, size_t, glob_t *);
 177 static int       globexp1(const Char *, glob_t *, size_t *);
 178 static int       globexp2(const Char *, const Char *, glob_t *, int *, size_t *);
 179 static int       match(Char *, Char *, Char *);
 180 #ifdef DEBUG
 181 static void      qprintf(const char *, Char *);
 182 #endif
 183
 184 /* Return value is either EOS, COLON, DOT, EQUALS, or LBRACKET if no class
 185    expression found. */
 186 static inline Char
 187 check_classes_expr(const Char *&cptr, wint_t *classbuf = NULL,
 188                    size_t classbufsize = 0)
 189 {
 190         const Char *ctype = NULL;
 191
 192         if (*cptr == LBRACKET &&
 193             (cptr[1] == COLON || cptr[1] == DOT || cptr[1] == EQUALS)) {
 194                 ctype = ++cptr;
 195                 while (*++cptr != EOS &&
 196                        (*cptr != *ctype || cptr[1] != RBRACKET))
 197                         ;
 198                 if (*cptr == EOS)
 199                         return EOS;
 200                 if (classbuf) {
 201                         const Char *class_p = ctype + 1;
 202                         size_t clen = cptr - class_p;
 203
 204                         if (clen < classbufsize)
 205                                 *wcipncpy (classbuf, class_p, clen) = '\0';
 206                         else
 207                                 ctype = NULL;
 208                 }
 209                 cptr++; /* Advance cptr to closing RBRACKET of class expr */
 210         }
 211         return ctype ? *ctype : LBRACKET;
 212 }
 213
 214 int
 215 glob(const char *__restrict pattern, int flags, int (*errfunc)(const char *, int), glob_t *__restrict pglob)
 216 {
 217         const char *patnext;
 218         size_t limit;
 219         Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
 220         mbstate_t mbs;
 221         wint_t wc;
 222         size_t clen;
 223
 224         patnext = pattern;
 225         if (!(flags & GLOB_APPEND)) {
 226                 pglob->gl_pathc = 0;
 227                 pglob->gl_pathv = NULL;
 228                 if (!(flags & GLOB_DOOFFS))
 229                         pglob->gl_offs = 0;
 230         }
 231         if (flags & GLOB_LIMIT) {
 232                 limit = pglob->gl_matchc;
 233                 if (limit == 0)
 234                         limit = ARG_MAX;
 235         } else
 236                 limit = 0;
 237         pglob->gl_flags = flags & ~GLOB_MAGCHAR;
 238         pglob->gl_errfunc = errfunc;
 239         pglob->gl_matchc = 0;
 240
 241         bufnext = patbuf;
 242         bufend = bufnext + MAXPATHLEN - 1;
 243         if (flags & GLOB_NOESCAPE) {
 244                 memset(&mbs, 0, sizeof(mbs));
 245                 while (bufend - bufnext >= MB_CUR_MAX) {
 246                         clen = mbrtowi(&wc, patnext, MB_LEN_MAX, &mbs);
 247                         if (clen == (size_t)-1 || clen == (size_t)-2)
 248                                 return (GLOB_NOMATCH);
 249                         else if (clen == 0)
 250                                 break;
 251                         *bufnext++ = wc;
 252                         patnext += clen;
 253                 }
 254         } else {
 255                 /* Protect the quoted characters. */
 256                 memset(&mbs, 0, sizeof(mbs));
 257                 while (bufend - bufnext >= MB_CUR_MAX) {
 258                         if (*patnext == QUOTE) {
 259                                 if (*++patnext == EOS) {
 260                                         *bufnext++ = QUOTE | M_PROTECT;
 261                                         continue;
 262                                 }
 263                                 prot = M_PROTECT;
 264                         } else
 265                                 prot = 0;
 266                         clen = mbrtowi(&wc, patnext, MB_LEN_MAX, &mbs);
 267                         if (clen == (size_t)-1 || clen == (size_t)-2)
 268                                 return (GLOB_NOMATCH);
 269                         else if (clen == 0)
 270                                 break;
 271                         *bufnext++ = wc | prot;
 272                         patnext += clen;
 273                 }
 274         }
 275         *bufnext = EOS;
 276
 277         if (flags & GLOB_BRACE)
 278             return globexp1(patbuf, pglob, &limit);
 279         else
 280             return glob0(patbuf, pglob, &limit);
 281 }
 282
 283 /*
 284  * Expand recursively a glob {} pattern. When there is no more expansion
 285  * invoke the standard globbing routine to glob the rest of the magic
 286  * characters
 287  */
 288 static int
 289 globexp1(const Char *pattern, glob_t *pglob, size_t *limit)
 290 {
 291         const Char* ptr = pattern;
 292         int rv;
 293
 294         /* Protect a single {}, for find(1), like csh */
 295         if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
 296                 return glob0(pattern, pglob, limit);
 297
 298         while ((ptr = g_strchr(ptr, LBRACE)) != NULL)
 299                 if (!globexp2(ptr, pattern, pglob, &rv, limit))
 300                         return rv;
 301
 302         return glob0(pattern, pglob, limit);
 303 }
 304
 305
 306 /*
 307  * Recursive brace globbing helper. Tries to expand a single brace.
 308  * If it succeeds then it invokes globexp1 with the new pattern.
 309  * If it fails then it tries to glob the rest of the pattern and returns.
 310  */
 311 static int
 312 globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv, size_t *limit)
 313 {
 314         int     i;
 315         Char   *lm, *ls;
 316         const Char *pe, *pm, *pm1, *pl;
 317         Char    patbuf[MAXPATHLEN];
 318
 319         /* copy part up to the brace */
 320         for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
 321                 continue;
 322         *lm = EOS;
 323         ls = lm;
 324
 325         /* Find the balanced brace */
 326         for (i = 0, pe = ++ptr; *pe; pe++)
 327                 if (*pe == LBRACKET) {
 328                         /* Ignore everything between [] */
 329                         for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++) {
 330                                 if (check_classes_expr (pe) == EOS)
 331                                         break;
 332                         }
 333                         if (*pe == EOS) {
 334                                 /*
 335                                  * We could not find a matching RBRACKET.
 336                                  * Ignore and just look for RBRACE
 337                                  */
 338                                 pe = pm;
 339                         }
 340                 }
 341                 else if (*pe == LBRACE)
 342                         i++;
 343                 else if (*pe == RBRACE) {
 344                         if (i == 0)
 345                                 break;
 346                         i--;
 347                 }
 348
 349         /* Non matching braces; just glob the pattern */
 350         if (i != 0 || *pe == EOS) {
 351                 *rv = glob0(patbuf, pglob, limit);
 352                 return 0;
 353         }
 354
 355         for (i = 0, pl = pm = ptr; pm <= pe; pm++)
 356                 switch (*pm) {
 357                 case LBRACKET:
 358                         /* Ignore everything between [] */
 359                         for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++) {
 360                                 if (check_classes_expr (pm) == EOS)
 361                                         break;
 362                         }
 363                         if (*pm == EOS) {
 364                                 /*
 365                                  * We could not find a matching RBRACKET.
 366                                  * Ignore and just look for RBRACE
 367                                  */
 368                                 pm = pm1;
 369                         }
 370                         break;
 371
 372                 case LBRACE:
 373                         i++;
 374                         break;
 375
 376                 case RBRACE:
 377                         if (i) {
 378                             i--;
 379                             break;
 380                         }
 381                         fallthrough;
 382                 case COMMA:
 383                         if (i && *pm == COMMA)
 384                                 break;
 385                         else {
 386                                 /* Append the current string */
 387                                 for (lm = ls; (pl < pm); *lm++ = *pl++)
 388                                         continue;
 389                                 /*
 390                                  * Append the rest of the pattern after the
 391                                  * closing brace
 392                                  */
 393                                 for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
 394                                         continue;
 395
 396                                 /* Expand the current pattern */
 397 #ifdef DEBUG
 398                                 qprintf("globexp2:", patbuf);
 399 #endif
 400                                 *rv = globexp1(patbuf, pglob, limit);
 401
 402                                 /* move after the comma, to the next string */
 403                                 pl = pm + 1;
 404                         }
 405                         break;
 406
 407                 default:
 408                         break;
 409                 }
 410         *rv = 0;
 411         return 0;
 412 }
 413
 414
 415
 416 /*
 417  * expand tilde from the passwd file.
 418  */
 419 static const Char *
 420 globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
 421 {
 422         struct passwd *pwd;
 423         char *h;
 424         const Char *p;
 425         Char *b, *eb;
 426
 427         if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
 428                 return pattern;
 429
 430         /*
 431          * Copy up to the end of the string or /
 432          */
 433         eb = &patbuf[patbuf_len - 1];
 434         for (p = pattern + 1, h = (char *) patbuf;
 435             h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
 436                 continue;
 437
 438         *h = EOS;
 439
 440         if (((char *) patbuf)[0] == EOS) {
 441                 /*
 442                  * handle a plain ~ or ~/ by expanding $HOME first (iff
 443                  * we're not running setuid or setgid) and then trying
 444                  * the password file
 445                  */
 446                 if (issetugid() != 0 ||
 447                     (h = getenv("HOME")) == NULL) {
 448                         if (((h = getlogin()) != NULL &&
 449                              (pwd = getpwnam(h)) != NULL) ||
 450                             (pwd = getpwuid(getuid())) != NULL)
 451                                 h = pwd->pw_dir;
 452                         else
 453                                 return pattern;
 454                 }
 455         }
 456         else {
 457                 /*
 458                  * Expand a ~user
 459                  */
 460                 if ((pwd = getpwnam((char*) patbuf)) == NULL)
 461                         return pattern;
 462                 else
 463                         h = pwd->pw_dir;
 464         }
 465
 466         /* Copy the home directory */
 467         for (b = patbuf; b < eb && *h; *b++ = *h++)
 468                 continue;
 469
 470         /* Append the rest of the pattern */
 471         while (b < eb && (*b++ = *p++) != EOS)
 472                 continue;
 473         *b = EOS;
 474
 475         return patbuf;
 476 }
 477
 478 /*
 479  * The main glob() routine: compiles the pattern (optionally processing
 480  * quotes), calls glob1() to do the real pattern matching, and finally
 481  * sorts the list (unless unsorted operation is requested).  Returns 0
 482  * if things went well, nonzero if errors occurred.
 483  */
 484 static int
 485 glob0(const Char *pattern, glob_t *pglob, size_t *limit)
 486 {
 487         const Char *qpatnext, *qpatrbsrch;
 488         int err;
 489         size_t oldpathc;
 490         Char *bufnext, c, patbuf[MAXPATHLEN];
 491
 492         qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
 493         oldpathc = pglob->gl_pathc;
 494         bufnext = patbuf;
 495
 496         /* We don't need to check for buffer overflow any more. */
 497         while ((c = *qpatnext++) != EOS) {
 498                 switch (c) {
 499                 case LBRACKET:
 500                         c = *qpatnext;
 501                         if (c == NOT)
 502                                 ++qpatnext;
 503                         for (qpatrbsrch = qpatnext;
 504                              *qpatrbsrch != RBRACKET && *qpatrbsrch != EOS;
 505                              ++qpatrbsrch) {
 506                                 if (check_classes_expr (qpatrbsrch) == EOS)
 507                                         break;
 508                         }
 509                         if (*qpatrbsrch == EOS) {
 510                                 *bufnext++ = LBRACKET;
 511                                 if (c == NOT)
 512                                         --qpatnext;
 513                                 break;
 514                         }
 515                         *bufnext++ = M_SET;
 516                         if (c == NOT)
 517                                 *bufnext++ = M_NOT;
 518                         c = *qpatnext++;
 519                         do {
 520                                 wint_t wclass[64];
 521                                 Char ctype;
 522
 523                                 ctype = check_classes_expr(--qpatnext, wclass,
 524                                                            64);
 525                                 ++qpatnext;
 526                                 if (ctype == COLON) {
 527                                         wctype_t type;
 528                                         char cclass[64];
 529
 530                                         /* No worries, char classes are
 531                                            ASCII-only anyway */
 532                                         wcitoascii (cclass, wclass);
 533                                         if ((type = wctype (cclass))) {
 534                                             *bufnext++ = M_NAMED;
 535                                             *bufnext++ = CHAR (type);
 536                                         }
 537                                         continue;
 538                                 }
 539                                 if (ctype == EQUALS) {
 540                                         if (wclass[0] && !wclass[1]) {
 541                                             *bufnext++ = M_EQUIV;
 542                                             *bufnext++ = CHAR (wclass[0]);
 543                                         }
 544                                         continue;
 545                                 }
 546                                 if (ctype == DOT &&
 547                                     is_unicode_coll_elem (wclass)) {
 548                                         *bufnext++ =
 549                                             M_COLL (wcilen (wclass));
 550                                         wint_t *wcp = wclass;
 551                                         while ((*bufnext++ = *wcp++))
 552                                             ;
 553                                         --bufnext; /* drop NUL */
 554                                 } else
 555                                         *bufnext++ = CHAR(c);
 556                                 if (*qpatnext == RANGE &&
 557                                     (c = qpatnext[1]) != RBRACKET) {
 558                                         *bufnext++ = M_RNG;
 559
 560                                         ctype = check_classes_expr(++qpatnext,
 561                                                                    wclass, 64);
 562                                         if (ctype == DOT &&
 563                                             is_unicode_coll_elem (wclass)) {
 564                                                 *bufnext++ =
 565                                                     M_COLL (wcilen (wclass));
 566                                                 wint_t *wcp = wclass;
 567                                                 while ((*bufnext++ = *wcp++))
 568                                                     ;
 569                                                 --bufnext; /* drop NUL */
 570                                         } else
 571                                                 *bufnext++ = CHAR(c);
 572                                         ++qpatnext;
 573                                 }
 574                         } while ((c = *qpatnext++) != RBRACKET);
 575                         pglob->gl_flags |= GLOB_MAGCHAR;
 576                         *bufnext++ = M_END;
 577                         break;
 578                 case QUESTION:
 579                         pglob->gl_flags |= GLOB_MAGCHAR;
 580                         *bufnext++ = M_ONE;
 581                         break;
 582                 case STAR:
 583                         pglob->gl_flags |= GLOB_MAGCHAR;
 584                         /* collapse adjacent stars to one,
 585                          * to avoid exponential behavior
 586                          */
 587                         if (bufnext == patbuf || bufnext[-1] != M_ALL)
 588                             *bufnext++ = M_ALL;
 589                         break;
 590                 default:
 591                         *bufnext++ = CHAR(c);
 592                         break;
 593                 }
 594         }
 595         *bufnext = EOS;
 596 #ifdef DEBUG
 597         qprintf("glob0:", patbuf);
 598 #endif
 599
 600         if ((err = glob1(patbuf, pglob, limit)) != 0)
 601                 return(err);
 602
 603         /*
 604          * If there was no match we are going to append the pattern
 605          * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
 606          * and the pattern did not contain any magic characters
 607          * GLOB_NOMAGIC is there just for compatibility with csh.
 608          */
 609         if (pglob->gl_pathc == oldpathc) {
 610                 if (((pglob->gl_flags & GLOB_NOCHECK) ||
 611                     ((pglob->gl_flags & GLOB_NOMAGIC) &&
 612                         !(pglob->gl_flags & GLOB_MAGCHAR))))
 613                         return(globextend(pattern, pglob, limit));
 614                 else
 615                         return(GLOB_NOMATCH);
 616         }
 617         if (!(pglob->gl_flags & GLOB_NOSORT))
 618                 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
 619                     pglob->gl_pathc - oldpathc, sizeof(char *), compare);
 620         return(0);
 621 }
 622
 623 static int
 624 compare(const void *p, const void *q)
 625 {
 626         return(strcoll(*(char **)p, *(char **)q));
 627 }
 628
 629 static int
 630 glob1(Char *pattern, glob_t *pglob, size_t *limit)
 631 {
 632         Char pathbuf[MAXPATHLEN];
 633
 634         /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
 635         if (*pattern == EOS)
 636                 return(0);
 637         return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
 638             pattern, pglob, limit));
 639 }
 640
 641 /*
 642  * The functions glob2 and glob3 are mutually recursive; there is one level
 643  * of recursion for each segment in the pattern that contains one or more
 644  * meta characters.
 645  */
 646 static int
 647 glob2(Char *pathbuf, Char *pathend, Char *pathend_last, Char *pattern,
 648       glob_t *pglob, size_t *limit)
 649 {
 650         struct stat sb;
 651         Char *p, *q;
 652         int anymeta;
 653
 654         /*
 655          * Loop over pattern segments until end of pattern or until
 656          * segment with meta character found.
 657          */
 658         for (anymeta = 0;;) {
 659                 if (*pattern == EOS) {          /* End of pattern? */
 660                         *pathend = EOS;
 661                         if (g_lstat(pathbuf, &sb, pglob))
 662                                 return(0);
 663
 664                         if (((pglob->gl_flags & GLOB_MARK) &&
 665                             pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
 666                             || (S_ISLNK(sb.st_mode) &&
 667                             (g_stat(pathbuf, &sb, pglob) == 0) &&
 668                             S_ISDIR(sb.st_mode)))) {
 669                                 if (pathend + 1 > pathend_last)
 670                                         return (GLOB_ABORTED);
 671                                 *pathend++ = SEP;
 672                                 *pathend = EOS;
 673                         }
 674                         ++pglob->gl_matchc;
 675                         return(globextend(pathbuf, pglob, limit));
 676                 }
 677
 678                 /* Find end of next segment, copy tentatively to pathend. */
 679                 q = pathend;
 680                 p = pattern;
 681                 while (*p != EOS && *p != SEP) {
 682                         if (ismeta(*p))
 683                                 anymeta = 1;
 684                         if (q + 1 > pathend_last)
 685                                 return (GLOB_ABORTED);
 686                         *q++ = *p++;
 687                 }
 688
 689                 if (!anymeta) {         /* No expansion, do next segment. */
 690                         pathend = q;
 691                         pattern = p;
 692                         while (*pattern == SEP) {
 693                                 if (pathend + 1 > pathend_last)
 694                                         return (GLOB_ABORTED);
 695                                 *pathend++ = *pattern++;
 696                         }
 697                 } else                  /* Need expansion, recurse. */
 698                         return(glob3(pathbuf, pathend, pathend_last, pattern, p,
 699                             pglob, limit));
 700         }
 701         /* NOTREACHED */
 702 }
 703
 704 static int
 705 glob3(Char *pathbuf, Char *pathend, Char *pathend_last,
 706       Char *pattern, Char *restpattern,
 707       glob_t *pglob, size_t *limit)
 708 {
 709         struct dirent *dp;
 710         DIR *dirp;
 711         int err;
 712         char buf[MAXPATHLEN];
 713
 714         /*
 715          * The readdirfunc declaration can't be prototyped, because it is
 716          * assigned, below, to two functions which are prototyped in glob.h
 717          * and dirent.h as taking pointers to differently typed opaque
 718          * structures.
 719          * CYGWIN: Needs prototype and subsequently wild casting to avoid
 720          * compiler error.
 721          */
 722         struct dirent *(*readdirfunc)(void *);
 723
 724         if (pathend > pathend_last)
 725                 return (GLOB_ABORTED);
 726         *pathend = EOS;
 727         errno = 0;
 728
 729         if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
 730                 /* TODO: don't call for ENOENT or ENOTDIR? */
 731                 if (pglob->gl_errfunc) {
 732                         if (g_Ctoc(pathbuf, buf, sizeof(buf)))
 733                                 return (GLOB_ABORTED);
 734                         if (pglob->gl_errfunc(buf, errno) ||
 735                             pglob->gl_flags & GLOB_ERR)
 736                                 return (GLOB_ABORTED);
 737                 }
 738                 return(0);
 739         }
 740
 741         err = 0;
 742
 743         /* Search directory for matching names. */
 744         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 745                 readdirfunc = pglob->gl_readdir;
 746         else
 747                 readdirfunc = (dirent*(*)(void*)) readdir;
 748         while ((dp = (*readdirfunc)(dirp))) {
 749                 char *sc;
 750                 Char *dc;
 751                 wint_t wc;
 752                 size_t clen;
 753                 mbstate_t mbs;
 754
 755                 /* Initial DOT must be matched literally. */
 756                 if (dp->d_name[0] == DOT && *pattern != DOT)
 757                         continue;
 758                 memset(&mbs, 0, sizeof(mbs));
 759                 dc = pathend;
 760                 sc = dp->d_name;
 761                 while (dc < pathend_last) {
 762                         clen = mbrtowi(&wc, sc, MB_LEN_MAX, &mbs);
 763                         if (clen == (size_t)-1 || clen == (size_t)-2) {
 764                                 wc = *sc;
 765                                 clen = 1;
 766                                 memset(&mbs, 0, sizeof(mbs));
 767                         }
 768                         if ((*dc++ = wc) == EOS)
 769                                 break;
 770                         sc += clen;
 771                 }
 772                 if (!match(pathend, pattern, restpattern)) {
 773                         *pathend = EOS;
 774                         continue;
 775                 }
 776                 err = glob2(pathbuf, --dc, pathend_last, restpattern,
 777                     pglob, limit);
 778                 if (err)
 779                         break;
 780         }
 781
 782         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 783                 (*pglob->gl_closedir)(dirp);
 784         else
 785                 closedir(dirp);
 786         return(err);
 787 }
 788
 789
 790 /*
 791  * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
 792  * add the new item, and update gl_pathc.
 793  *
 794  * This assumes the BSD realloc, which only copies the block when its size
 795  * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
 796  * behavior.
 797  *
 798  * Return 0 if new item added, error code if memory couldn't be allocated.
 799  *
 800  * Invariant of the glob_t structure:
 801  *      Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
 802  *      gl_pathv points to (gl_offs + gl_pathc + 1) items.
 803  */
 804 static int
 805 globextend(const Char *path, glob_t *pglob, size_t *limit)
 806 {
 807         char **pathv;
 808         size_t i, newsize, len;
 809         char *copy;
 810         const Char *p;
 811
 812         if (*limit && pglob->gl_pathc > *limit) {
 813                 errno = 0;
 814                 return (GLOB_NOSPACE);
 815         }
 816
 817         newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
 818         pathv = pglob->gl_pathv ?
 819                     (char **) realloc((char *)pglob->gl_pathv, newsize) :
 820                     (char **) malloc(newsize);
 821         if (pathv == NULL) {
 822                 if (pglob->gl_pathv) {
 823                         free(pglob->gl_pathv);
 824                         pglob->gl_pathv = NULL;
 825                 }
 826                 return(GLOB_NOSPACE);
 827         }
 828
 829         if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 830                 /* first time around -- clear initial gl_offs items */
 831                 pathv += pglob->gl_offs;
 832                 for (i = pglob->gl_offs + 1; --i > 0; )
 833                         *--pathv = NULL;
 834         }
 835         pglob->gl_pathv = pathv;
 836
 837         for (p = path; *p++;)
 838                 continue;
 839         len = MB_CUR_MAX * (size_t)(p - path);  /* XXX overallocation */
 840         if ((copy = (char *) malloc(len)) != NULL) {
 841                 if (g_Ctoc(path, copy, len)) {
 842                         free(copy);
 843                         return (GLOB_NOSPACE);
 844                 }
 845                 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
 846         }
 847         pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 848         return(copy == NULL ? GLOB_NOSPACE : 0);
 849 }
 850
 851 /*
 852  * pattern matching function for filenames.  Each occurrence of the *
 853  * pattern causes a recursion level.
 854  */
 855 static int
 856 match(Char *name, Char *pat, Char *patend)
 857 {
 858         int ok, negate_range;
 859         Char *c, *k;
 860         size_t k_len;
 861
 862         while (pat < patend) {
 863                 c = pat++;
 864                 switch (*c & M_MASK) {
 865                 case M_ALL:
 866                         if (pat == patend)
 867                                 return(1);
 868                         do
 869                             if (match(name, pat, patend))
 870                                     return(1);
 871                         while (*name++ != EOS);
 872                         return(0);
 873                 case M_ONE:
 874                         if (*name++ == EOS)
 875                                 return(0);
 876                         break;
 877                 case M_SET:
 878                         ok = 0;
 879                         if (*(k = name) == EOS)
 880                                 return(0);
 881                         k_len = next_unicode_char (k);
 882                         name += k_len;
 883                         if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 884                                 ++pat;
 885                         while ((*(c = pat++) & M_MASK) != M_END) {
 886                                 size_t len1 = 1, len2 = 1;
 887
 888                                 if ((*c & M_MASK) == M_NAMED) {
 889                                         if (iswctype (*k, *pat++))
 890                                                 ok = 1;
 891                                         continue;
 892                                 }
 893                                 if ((*c & M_MASK) == M_EQUIV) {
 894                                         if (is_unicode_equiv (*k, *pat++))
 895                                                 ok = 1;
 896                                         continue;
 897                                 }
 898                                 if (M_COLL_P(*c)) {
 899                                         len1 = M_COLL_CNT(*c);
 900                                         ++c;
 901                                         pat += len1;
 902                                 }
 903                                 if ((*pat & M_MASK) == M_RNG) {
 904                                         if (M_COLL_P(pat[1]))
 905                                                 len2 = M_COLL_CNT(*++pat);
 906 #ifdef __CYGWIN__
 907                                         if ((!__get_current_collate_locale ()->win_locale[0]) ?
 908 #else
 909                                         if (__collate_load_error ?
 910 #endif
 911                                             *c <= *k && *k <= pat[1] :
 912                                                __wscollate_range_cmp(c, k, len1, k_len) <= 0
 913                                             && __wscollate_range_cmp(k, pat + 1, k_len, len2) <= 0
 914                                            )
 915                                                 ok = 1;
 916                                         pat += len2 + 1;
 917                                 } else if (len1 == k_len &&
 918                                            wcincmp (c, k, len1) == 0)
 919                                         ok = 1;
 920                         }
 921                         if (ok == negate_range)
 922                                 return(0);
 923                         break;
 924                 default:
 925                         if (Cchar(*name++) != Cchar(*c))
 926                                 return(0);
 927                         break;
 928                 }
 929         }
 930         return(*name == EOS);
 931 }
 932
 933 /* Free allocated data belonging to a glob_t structure. */
 934 void
 935 globfree(glob_t *pglob)
 936 {
 937         size_t i;
 938         char **pp;
 939
 940         if (pglob->gl_pathv != NULL) {
 941                 pp = pglob->gl_pathv + pglob->gl_offs;
 942                 for (i = pglob->gl_pathc; i--; ++pp)
 943                         if (*pp)
 944                                 free(*pp);
 945                 free(pglob->gl_pathv);
 946                 pglob->gl_pathv = NULL;
 947         }
 948 }
 949
 950 static DIR *
 951 g_opendir(Char *str, glob_t *pglob)
 952 {
 953         char buf[MAXPATHLEN];
 954
 955         if (!*str)
 956                 strcpy(buf, ".");
 957         else {
 958                 if (g_Ctoc(str, buf, sizeof(buf)))
 959                         return (NULL);
 960         }
 961
 962         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 963                 return (DIR *) ((*pglob->gl_opendir)((const char *) buf));
 964
 965         return(opendir(buf));
 966 }
 967
 968 #define CYGWIN_gl_stat(sfptr) ((*pglob->sfptr) (buf, sb))
 969
 970 static int
 971 g_lstat(Char *fn, struct stat *sb, glob_t *pglob)
 972 {
 973         char buf[MAXPATHLEN];
 974
 975         if (g_Ctoc(fn, buf, sizeof(buf))) {
 976                 errno = ENAMETOOLONG;
 977                 return (-1);
 978         }
 979         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 980                 return CYGWIN_gl_stat (gl_lstat);
 981         return(lstat(buf, sb));
 982 }
 983
 984 static int
 985 g_stat(Char *fn, struct stat *sb, glob_t *pglob)
 986 {
 987         char buf[MAXPATHLEN];
 988
 989         if (g_Ctoc(fn, buf, sizeof(buf))) {
 990                 errno = ENAMETOOLONG;
 991                 return (-1);
 992         }
 993         if (pglob->gl_flags & GLOB_ALTDIRFUNC)
 994                 return CYGWIN_gl_stat (gl_stat);
 995         return(stat(buf, sb));
 996 }
 997
 998 static const Char *
 999 g_strchr(const Char *str, wint_t ch)
1000 {
1001
1002         do {
1003                 if (*str == ch)
1004                         return (str);
1005         } while (*str++);
1006         return (NULL);
1007 }
1008
1009 static int
1010 g_Ctoc(const Char *str, char *buf, size_t len)
1011 {
1012         mbstate_t mbs;
1013         size_t clen;
1014
1015         memset(&mbs, 0, sizeof(mbs));
1016         while (len >= (size_t) MB_CUR_MAX) {
1017                 clen = wirtomb(buf, CHAR (*str), &mbs);
1018                 if (clen == (size_t)-1)
1019                         return (1);
1020                 if (*str == L'\0')
1021                         return (0);
1022                 str++;
1023                 buf += clen;
1024                 len -= clen;
1025         }
1026         return (1);
1027 }
1028
1029 #ifdef DEBUG
1030 static void
1031 qprintf(const char *str, Char *s)
1032 {
1033         Char *p;
1034
1035         (void)printf("%s:\n", str);
1036         for (p = s; *p; p++)
1037                 (void)printf("%c", CHAR(*p));
1038         (void)printf("\n");
1039         for (p = s; *p; p++)
1040                 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
1041         (void)printf("\n");
1042         for (p = s; *p; p++)
1043                 (void)printf("%c", ismeta(*p) ? '_' : ' ');
1044         (void)printf("\n");
1045 }
1046 #endif