2 * Copyright (c) 1989, 1993
3 * The Regents of the University of California. All rights reserved.
5 * This code is derived from software contributed to Berkeley by
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. 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.
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
33 #if defined(LIBC_SCCS) && !defined(lint)
34 static char sccsid[] = "@(#)glob.c 8.3 (Berkeley) 10/13/93";
35 #endif /* LIBC_SCCS and not lint */
38 * glob(3) -- a superset of the one defined in POSIX 1003.2.
40 * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
42 * Optional extra services, controlled by flags not defined by POSIX:
45 * Escaping convention: \ inhibits any special meaning the following
46 * character might have (except \ at end of string is retained).
48 * Set in gl_flags if pattern contained a globbing character.
50 * Same as GLOB_NOCHECK, but it will only append pattern if it did
51 * not contain any magic characters. [Used in csh style globbing]
53 * Use alternately specified directory access functions.
55 * expand ~user/foo to the /home/dir/of/user/foo
57 * expand {1,2}{a,b} to 1a 1b 2a 2b
59 * Number of matches in the current invocation of glob.
61 * sort alphabetically like csh (case doesn't matter) instead of in ASCII
74 struct passwd *getpwnam(char *);
75 struct passwd *getpwuid(Uid_t);
81 # define MAXPATHLEN PATH_MAX
82 # ifdef MACOS_TRADITIONAL
83 # define MAXPATHLEN 255
85 # define MAXPATHLEN 1024
96 # define ARG_MAX (sysconf(_SC_ARG_MAX))
98 # ifdef _POSIX_ARG_MAX
99 # define ARG_MAX _POSIX_ARG_MAX
102 # define ARG_MAX 14500 /* from VC's limits.h */
104 # define ARG_MAX 4096 /* from POSIX, be conservative */
110 #define BG_DOLLAR '$'
113 #define BG_LBRACKET '['
115 #define BG_QUESTION '?'
116 #define BG_QUOTE '\\'
118 #define BG_RBRACKET ']'
119 #ifdef MACOS_TRADITIONAL
129 #define BG_UNDERSCORE '_'
130 #define BG_LBRACE '{'
131 #define BG_RBRACE '}'
137 #define M_QUOTE 0x8000
138 #define M_PROTECT 0x4000
139 #define M_MASK 0xffff
140 #define M_ASCII 0x00ff
147 #define M_PROTECT 0x40
153 #endif /* !GLOB_DEBUG */
156 #define CHAR(c) ((Char)((c)&M_ASCII))
157 #define META(c) ((Char)((c)|M_QUOTE))
158 #define M_ALL META('*')
159 #define M_END META(']')
160 #define M_NOT META('!')
161 #define M_ONE META('?')
162 #define M_RNG META('-')
163 #define M_SET META('[')
164 #define ismeta(c) (((c)&M_QUOTE) != 0)
167 static int compare(const void *, const void *);
168 static int ci_compare(const void *, const void *);
169 static int g_Ctoc(const Char *, char *, STRLEN);
170 static int g_lstat(Char *, Stat_t *, glob_t *);
171 static DIR *g_opendir(Char *, glob_t *);
172 static Char *g_strchr(Char *, int);
173 static int g_stat(Char *, Stat_t *, glob_t *);
174 static int glob0(const Char *, glob_t *);
175 static int glob1(Char *, Char *, glob_t *, size_t *);
176 static int glob2(Char *, Char *, Char *, Char *, Char *, Char *,
178 static int glob3(Char *, Char *, Char *, Char *, Char *, Char *,
179 Char *, Char *, glob_t *, size_t *);
180 static int globextend(const Char *, glob_t *, size_t *);
182 globtilde(const Char *, Char *, size_t, glob_t *);
183 static int globexp1(const Char *, glob_t *);
184 static int globexp2(const Char *, const Char *, glob_t *, int *);
185 static int match(Char *, Char *, Char *, int);
187 static void qprintf(const char *, Char *);
188 #endif /* GLOB_DEBUG */
190 #ifdef PERL_IMPLICIT_CONTEXT
191 static Direntry_t * my_readdir(DIR*);
196 return PerlDir_read(d);
199 #define my_readdir readdir
203 bsd_glob(const char *pattern, int flags,
204 int (*errfunc)(const char *, int), glob_t *pglob)
208 Char *bufnext, *bufend, patbuf[MAXPATHLEN];
210 patnext = (U8 *) pattern;
211 if (!(flags & GLOB_APPEND)) {
213 pglob->gl_pathv = NULL;
214 if (!(flags & GLOB_DOOFFS))
217 pglob->gl_flags = flags & ~GLOB_MAGCHAR;
218 pglob->gl_errfunc = errfunc;
219 pglob->gl_matchc = 0;
222 bufend = bufnext + MAXPATHLEN - 1;
224 /* Nasty hack to treat patterns like "C:*" correctly. In this
225 * case, the * should match any file in the current directory
226 * on the C: drive. However, the glob code does not treat the
227 * colon specially, so it looks for files beginning "C:" in
228 * the current directory. To fix this, change the pattern to
229 * add an explicit "./" at the start (just after the drive
230 * letter and colon - ie change to "C:./*").
232 if (isalpha(pattern[0]) && pattern[1] == ':' &&
233 pattern[2] != BG_SEP && pattern[2] != BG_SEP2 &&
234 bufend - bufnext > 4) {
235 *bufnext++ = pattern[0];
242 if (flags & GLOB_QUOTE) {
243 /* Protect the quoted characters. */
244 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
247 /* To avoid backslashitis on Win32,
248 * we only treat \ as a quoting character
249 * if it precedes one of the
250 * metacharacters []-{}~\
252 if ((c = *patnext++) != '[' && c != ']' &&
253 c != '-' && c != '{' && c != '}' &&
254 c != '~' && c != '\\') {
256 if ((c = *patnext++) == BG_EOS) {
261 *bufnext++ = c | M_PROTECT;
265 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
269 if (flags & GLOB_BRACE)
270 return globexp1(patbuf, pglob);
272 return glob0(patbuf, pglob);
276 * Expand recursively a glob {} pattern. When there is no more expansion
277 * invoke the standard globbing routine to glob the rest of the magic
281 globexp1(const Char *pattern, glob_t *pglob)
283 const Char* ptr = pattern;
286 /* Protect a single {}, for find(1), like csh */
287 if (pattern[0] == BG_LBRACE && pattern[1] == BG_RBRACE && pattern[2] == BG_EOS)
288 return glob0(pattern, pglob);
290 while ((ptr = (const Char *) g_strchr((Char *) ptr, BG_LBRACE)) != NULL)
291 if (!globexp2(ptr, pattern, pglob, &rv))
294 return glob0(pattern, pglob);
299 * Recursive brace globbing helper. Tries to expand a single brace.
300 * If it succeeds then it invokes globexp1 with the new pattern.
301 * If it fails then it tries to glob the rest of the pattern and returns.
304 globexp2(const Char *ptr, const Char *pattern,
305 glob_t *pglob, int *rv)
309 const Char *pe, *pm, *pl;
310 Char patbuf[MAXPATHLEN];
312 /* copy part up to the brace */
313 for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
318 /* Find the balanced brace */
319 for (i = 0, pe = ++ptr; *pe; pe++)
320 if (*pe == BG_LBRACKET) {
321 /* Ignore everything between [] */
322 for (pm = pe++; *pe != BG_RBRACKET && *pe != BG_EOS; pe++)
326 * We could not find a matching BG_RBRACKET.
327 * Ignore and just look for BG_RBRACE
331 } else if (*pe == BG_LBRACE)
333 else if (*pe == BG_RBRACE) {
339 /* Non matching braces; just glob the pattern */
340 if (i != 0 || *pe == BG_EOS) {
341 *rv = glob0(patbuf, pglob);
345 for (i = 0, pl = pm = ptr; pm <= pe; pm++) {
348 /* Ignore everything between [] */
349 for (pl = pm++; *pm != BG_RBRACKET && *pm != BG_EOS; pm++)
353 * We could not find a matching BG_RBRACKET.
354 * Ignore and just look for BG_RBRACE
371 if (i && *pm == BG_COMMA)
374 /* Append the current string */
375 for (lm = ls; (pl < pm); *lm++ = *pl++)
379 * Append the rest of the pattern after the
382 for (pl = pe + 1; (*lm++ = *pl++) != BG_EOS; )
385 /* Expand the current pattern */
387 qprintf("globexp2:", patbuf);
388 #endif /* GLOB_DEBUG */
389 *rv = globexp1(patbuf, pglob);
391 /* move after the comma, to the next string */
407 * expand tilde from the passwd file.
410 globtilde(const Char *pattern, Char *patbuf, size_t limit, glob_t *pglob)
417 if (*pattern != BG_TILDE || !(pglob->gl_flags & GLOB_TILDE))
420 /* Copy up to the end of the string or / */
421 for (p = pattern + 1, h = (char *) patbuf; *p && *p != BG_SLASH;
432 if (((char *) patbuf)[0] == BG_EOS) {
434 * handle a plain ~ or ~/ by expanding $HOME
435 * first and then trying the password file
437 if ((h = getenv("HOME")) == NULL) {
439 if ((pwd = getpwuid(getuid())) == NULL)
452 if ((pwd = getpwnam((char*) patbuf)) == NULL)
461 /* Copy the home directory */
462 for (b = patbuf; *h; *b++ = *h++)
465 /* Append the rest of the pattern */
466 while ((*b++ = *p++) != BG_EOS)
474 * The main glob() routine: compiles the pattern (optionally processing
475 * quotes), calls glob1() to do the real pattern matching, and finally
476 * sorts the list (unless unsorted operation is requested). Returns 0
477 * if things went well, nonzero if errors occurred. It is not an error
478 * to find no matches.
481 glob0(const Char *pattern, glob_t *pglob)
483 const Char *qpat, *qpatnext;
484 int c, err, oldflags, oldpathc;
485 Char *bufnext, patbuf[MAXPATHLEN];
488 #ifdef MACOS_TRADITIONAL
489 if ( (*pattern == BG_TILDE) && (pglob->gl_flags & GLOB_TILDE) ) {
490 return(globextend(pattern, pglob));
494 qpat = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
496 oldflags = pglob->gl_flags;
497 oldpathc = pglob->gl_pathc;
500 /* We don't need to check for buffer overflow any more. */
501 while ((c = *qpatnext++) != BG_EOS) {
507 if (*qpatnext == BG_EOS ||
508 g_strchr((Char *) qpatnext+1, BG_RBRACKET) == NULL) {
509 *bufnext++ = BG_LBRACKET;
519 *bufnext++ = CHAR(c);
520 if (*qpatnext == BG_RANGE &&
521 (c = qpatnext[1]) != BG_RBRACKET) {
523 *bufnext++ = CHAR(c);
526 } while ((c = *qpatnext++) != BG_RBRACKET);
527 pglob->gl_flags |= GLOB_MAGCHAR;
531 pglob->gl_flags |= GLOB_MAGCHAR;
535 pglob->gl_flags |= GLOB_MAGCHAR;
536 /* collapse adjacent stars to one,
537 * to avoid exponential behavior
539 if (bufnext == patbuf || bufnext[-1] != M_ALL)
543 *bufnext++ = CHAR(c);
549 qprintf("glob0:", patbuf);
550 #endif /* GLOB_DEBUG */
552 if ((err = glob1(patbuf, patbuf+MAXPATHLEN-1, pglob, &limit)) != 0) {
553 pglob->gl_flags = oldflags;
558 * If there was no match we are going to append the pattern
559 * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
560 * and the pattern did not contain any magic characters
561 * GLOB_NOMAGIC is there just for compatibility with csh.
563 if (pglob->gl_pathc == oldpathc &&
564 ((pglob->gl_flags & GLOB_NOCHECK) ||
565 ((pglob->gl_flags & GLOB_NOMAGIC) &&
566 !(pglob->gl_flags & GLOB_MAGCHAR))))
569 printf("calling globextend from glob0\n");
570 #endif /* GLOB_DEBUG */
571 pglob->gl_flags = oldflags;
572 return(globextend(qpat, pglob, &limit));
574 else if (!(pglob->gl_flags & GLOB_NOSORT))
575 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
576 pglob->gl_pathc - oldpathc, sizeof(char *),
577 (pglob->gl_flags & (GLOB_ALPHASORT|GLOB_NOCASE))
578 ? ci_compare : compare);
579 pglob->gl_flags = oldflags;
584 ci_compare(const void *p, const void *q)
586 const char *pp = *(const char **)p;
587 const char *qq = *(const char **)q;
590 if (tolower(*pp) != tolower(*qq))
595 ci = tolower(*pp) - tolower(*qq);
597 return compare(p, q);
602 compare(const void *p, const void *q)
604 return(strcmp(*(char **)p, *(char **)q));
608 glob1(Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
610 Char pathbuf[MAXPATHLEN];
612 /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
613 if (*pattern == BG_EOS)
615 return(glob2(pathbuf, pathbuf+MAXPATHLEN-1,
616 pathbuf, pathbuf+MAXPATHLEN-1,
617 pattern, pattern_last, pglob, limitp));
621 * The functions glob2 and glob3 are mutually recursive; there is one level
622 * of recursion for each segment in the pattern that contains one or more
626 glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
627 Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
634 * Loop over pattern segments until end of pattern or until
635 * segment with meta character found.
637 for (anymeta = 0;;) {
638 if (*pattern == BG_EOS) { /* End of pattern? */
641 if (g_lstat(pathbuf, &sb, pglob))
644 if (((pglob->gl_flags & GLOB_MARK) &&
645 pathend[-1] != BG_SEP
647 && pathend[-1] != BG_SEP2
649 ) && (S_ISDIR(sb.st_mode) ||
650 (S_ISLNK(sb.st_mode) &&
651 (g_stat(pathbuf, &sb, pglob) == 0) &&
652 S_ISDIR(sb.st_mode)))) {
653 if (pathend+1 > pathend_last)
660 printf("calling globextend from glob2\n");
661 #endif /* GLOB_DEBUG */
662 return(globextend(pathbuf, pglob, limitp));
665 /* Find end of next segment, copy tentatively to pathend. */
668 while (*p != BG_EOS && *p != BG_SEP
675 if (q+1 > pathend_last)
680 if (!anymeta) { /* No expansion, do next segment. */
683 while (*pattern == BG_SEP
685 || *pattern == BG_SEP2
688 if (pathend+1 > pathend_last)
690 *pathend++ = *pattern++;
693 /* Need expansion, recurse. */
694 return(glob3(pathbuf, pathbuf_last, pathend,
695 pathend_last, pattern, pattern_last,
696 p, pattern_last, pglob, limitp));
702 glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
703 Char *pattern, Char *pattern_last,
704 Char *restpattern, Char *restpattern_last, glob_t *pglob, size_t *limitp)
706 register Direntry_t *dp;
710 char buf[MAXPATHLEN];
713 * The readdirfunc declaration can't be prototyped, because it is
714 * assigned, below, to two functions which are prototyped in glob.h
715 * and dirent.h as taking pointers to differently typed opaque
718 Direntry_t *(*readdirfunc)(DIR*);
720 if (pathend > pathend_last)
728 if (q - pathbuf > 5) {
730 if (q[0] == '.' && tolower(q[1]) == 'd' && tolower(q[2]) == 'i'
731 && tolower(q[3]) == 'r' && q[4] == '/')
740 if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
741 /* TODO: don't call for ENOENT or ENOTDIR? */
742 if (pglob->gl_errfunc) {
743 if (g_Ctoc(pathbuf, buf, sizeof(buf)))
745 if (pglob->gl_errfunc(buf, errno) ||
746 (pglob->gl_flags & GLOB_ERR))
753 nocase = ((pglob->gl_flags & GLOB_NOCASE) != 0);
755 /* Search directory for matching names. */
756 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
757 readdirfunc = (Direntry_t *(*)(DIR *))pglob->gl_readdir;
759 readdirfunc = my_readdir;
760 while ((dp = (*readdirfunc)(dirp))) {
764 /* Initial BG_DOT must be matched literally. */
765 if (dp->d_name[0] == BG_DOT && *pattern != BG_DOT)
768 sc = (U8 *) dp->d_name;
769 while (dc < pathend_last && (*dc++ = *sc++) != BG_EOS)
771 if (dc >= pathend_last) {
777 if (!match(pathend, pattern, restpattern, nocase)) {
781 err = glob2(pathbuf, pathbuf_last, --dc, pathend_last,
782 restpattern, restpattern_last, pglob, limitp);
787 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
788 (*pglob->gl_closedir)(dirp);
796 * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
797 * add the new item, and update gl_pathc.
799 * This assumes the BSD realloc, which only copies the block when its size
800 * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
803 * Return 0 if new item added, error code if memory couldn't be allocated.
805 * Invariant of the glob_t structure:
806 * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
807 * gl_pathv points to (gl_offs + gl_pathc + 1) items.
810 globextend(const Char *path, glob_t *pglob, size_t *limitp)
812 register char **pathv;
820 for (p = path; *p; p++)
821 (void)printf("%c", CHAR(*p));
823 #endif /* GLOB_DEBUG */
825 newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
827 pathv = Renew(pglob->gl_pathv,newsize,char*);
829 New(0,pathv,newsize,char*);
831 if (pglob->gl_pathv) {
832 Safefree(pglob->gl_pathv);
833 pglob->gl_pathv = NULL;
835 return(GLOB_NOSPACE);
838 if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
839 /* first time around -- clear initial gl_offs items */
840 pathv += pglob->gl_offs;
841 for (i = pglob->gl_offs; --i >= 0; )
844 pglob->gl_pathv = pathv;
846 for (p = path; *p++;)
848 len = (STRLEN)(p - path);
849 New(0, copy, p-path, char);
851 if (g_Ctoc(path, copy, len)) {
853 return(GLOB_NOSPACE);
855 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
857 pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
859 if ((pglob->gl_flags & GLOB_LIMIT) &&
860 newsize + *limitp >= ARG_MAX) {
862 return(GLOB_NOSPACE);
864 return(copy == NULL ? GLOB_NOSPACE : 0);
869 * pattern matching function for filenames. Each occurrence of the *
870 * pattern causes a recursion level.
873 match(register Char *name, register Char *pat, register Char *patend, int nocase)
875 int ok, negate_range;
878 while (pat < patend) {
880 switch (c & M_MASK) {
885 if (match(name, pat, patend, nocase))
887 while (*name++ != BG_EOS)
891 if (*name++ == BG_EOS)
896 if ((k = *name++) == BG_EOS)
898 if ((negate_range = ((*pat & M_MASK) == M_NOT)) != BG_EOS)
900 while (((c = *pat++) & M_MASK) != M_END)
901 if ((*pat & M_MASK) == M_RNG) {
903 if (tolower(c) <= tolower(k) && tolower(k) <= tolower(pat[1]))
906 if (c <= k && k <= pat[1])
910 } else if (nocase ? (tolower(c) == tolower(k)) : (c == k))
912 if (ok == negate_range)
917 if (nocase ? (tolower(k) != tolower(c)) : (k != c))
922 return(*name == BG_EOS);
925 /* Free allocated data belonging to a glob_t structure. */
927 bsd_globfree(glob_t *pglob)
932 if (pglob->gl_pathv != NULL) {
933 pp = pglob->gl_pathv + pglob->gl_offs;
934 for (i = pglob->gl_pathc; i--; ++pp)
937 Safefree(pglob->gl_pathv);
938 pglob->gl_pathv = NULL;
943 g_opendir(register Char *str, glob_t *pglob)
945 char buf[MAXPATHLEN];
948 #ifdef MACOS_TRADITIONAL
954 if (g_Ctoc(str, buf, sizeof(buf)))
958 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
959 return((*pglob->gl_opendir)(buf));
961 return(PerlDir_open(buf));
965 g_lstat(register Char *fn, Stat_t *sb, glob_t *pglob)
967 char buf[MAXPATHLEN];
969 if (g_Ctoc(fn, buf, sizeof(buf)))
971 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
972 return((*pglob->gl_lstat)(buf, sb));
974 return(PerlLIO_lstat(buf, sb));
976 return(PerlLIO_stat(buf, sb));
977 #endif /* HAS_LSTAT */
981 g_stat(register Char *fn, Stat_t *sb, glob_t *pglob)
983 char buf[MAXPATHLEN];
985 if (g_Ctoc(fn, buf, sizeof(buf)))
987 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
988 return((*pglob->gl_stat)(buf, sb));
989 return(PerlLIO_stat(buf, sb));
993 g_strchr(Char *str, int ch)
1003 g_Ctoc(register const Char *str, char *buf, STRLEN len)
1006 if ((*buf++ = *str++) == BG_EOS)
1014 qprintf(const char *str, register Char *s)
1018 (void)printf("%s:\n", str);
1019 for (p = s; *p; p++)
1020 (void)printf("%c", CHAR(*p));
1022 for (p = s; *p; p++)
1023 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
1025 for (p = s; *p; p++)
1026 (void)printf("%c", ismeta(*p) ? '_' : ' ');
1029 #endif /* GLOB_DEBUG */