glob.c revision 159294 
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 * 3. All advertising materials mentioning features or use of this software
17 *    must display the following acknowledgement:
18 *	This product includes software developed by the University of
19 *	California, Berkeley and its contributors.
20 * 4. Neither the name of the University nor the names of its contributors
21 *    may be used to endorse or promote products derived from this software
22 *    without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 */
36
37#if defined(LIBC_SCCS) && !defined(lint)
38static 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: head/lib/libc/gen/glob.c 159294 2006-06-05 18:22:13Z delphij $");
42
43/*
44 * glob(3) -- a superset of the one defined in POSIX 1003.2.
45 *
46 * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
47 *
48 * Optional extra services, controlled by flags not defined by POSIX:
49 *
50 * GLOB_QUOTE:
51 *	Escaping convention: \ inhibits any special meaning the following
52 *	character might have (except \ at end of string is retained).
53 * GLOB_MAGCHAR:
54 *	Set in gl_flags if pattern contained a globbing character.
55 * GLOB_NOMAGIC:
56 *	Same as GLOB_NOCHECK, but it will only append pattern if it did
57 *	not contain any magic characters.  [Used in csh style globbing]
58 * GLOB_ALTDIRFUNC:
59 *	Use alternately specified directory access functions.
60 * GLOB_TILDE:
61 *	expand ~user/foo to the /home/dir/of/user/foo
62 * GLOB_BRACE:
63 *	expand {1,2}{a,b} to 1a 1b 2a 2b
64 * gl_matchc:
65 *	Number of matches in the current invocation of glob.
66 */
67
68/*
69 * Some notes on multibyte character support:
70 * 1. Patterns with illegal byte sequences match nothing - even if
71 *    GLOB_NOCHECK is specified.
72 * 2. Illegal byte sequences in filenames are handled by treating them as
73 *    single-byte characters with a value of the first byte of the sequence
74 *    cast to wchar_t.
75 * 3. State-dependent encodings are not currently supported.
76 */
77
78#include <sys/param.h>
79#include <sys/stat.h>
80
81#include <ctype.h>
82#include <dirent.h>
83#include <errno.h>
84#include <glob.h>
85#include <limits.h>
86#include <pwd.h>
87#include <stdint.h>
88#include <stdio.h>
89#include <stdlib.h>
90#include <string.h>
91#include <unistd.h>
92#include <wchar.h>
93
94#include "collate.h"
95
96#define	DOLLAR		'$'
97#define	DOT		'.'
98#define	EOS		'\0'
99#define	LBRACKET	'['
100#define	NOT		'!'
101#define	QUESTION	'?'
102#define	QUOTE		'\\'
103#define	RANGE		'-'
104#define	RBRACKET	']'
105#define	SEP		'/'
106#define	STAR		'*'
107#define	TILDE		'~'
108#define	UNDERSCORE	'_'
109#define	LBRACE		'{'
110#define	RBRACE		'}'
111#define	SLASH		'/'
112#define	COMMA		','
113
114#ifndef DEBUG
115
116#define	M_QUOTE		0x8000000000ULL
117#define	M_PROTECT	0x4000000000ULL
118#define	M_MASK		0xffffffffffULL
119#define	M_CHAR		0x00ffffffffULL
120
121typedef uint_fast64_t Char;
122
123#else
124
125#define	M_QUOTE		0x80
126#define	M_PROTECT	0x40
127#define	M_MASK		0xff
128#define	M_CHAR		0x7f
129
130typedef char Char;
131
132#endif
133
134
135#define	CHAR(c)		((Char)((c)&M_CHAR))
136#define	META(c)		((Char)((c)|M_QUOTE))
137#define	M_ALL		META('*')
138#define	M_END		META(']')
139#define	M_NOT		META('!')
140#define	M_ONE		META('?')
141#define	M_RNG		META('-')
142#define	M_SET		META('[')
143#define	ismeta(c)	(((c)&M_QUOTE) != 0)
144
145
146static int	 compare(const void *, const void *);
147static int	 g_Ctoc(const Char *, char *, size_t);
148static int	 g_lstat(Char *, struct stat *, glob_t *);
149static DIR	*g_opendir(Char *, glob_t *);
150static Char	*g_strchr(Char *, wchar_t);
151#ifdef notdef
152static Char	*g_strcat(Char *, const Char *);
153#endif
154static int	 g_stat(Char *, struct stat *, glob_t *);
155static int	 glob0(const Char *, glob_t *, size_t *);
156static int	 glob1(Char *, glob_t *, size_t *);
157static int	 glob2(Char *, Char *, Char *, Char *, glob_t *, size_t *);
158static int	 glob3(Char *, Char *, Char *, Char *, Char *, glob_t *, size_t *);
159static int	 globextend(const Char *, glob_t *, size_t *);
160static const Char *
161		 globtilde(const Char *, Char *, size_t, glob_t *);
162static int	 globexp1(const Char *, glob_t *, size_t *);
163static int	 globexp2(const Char *, const Char *, glob_t *, int *, size_t *);
164static int	 match(Char *, Char *, Char *);
165#ifdef DEBUG
166static void	 qprintf(const char *, Char *);
167#endif
168
169int
170glob(const char *pattern, int flags, int (*errfunc)(const char *, int), glob_t *pglob)
171{
172	const char *patnext;
173	size_t limit;
174	Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot;
175	mbstate_t mbs;
176	wchar_t wc;
177	size_t clen;
178
179	patnext = pattern;
180	if (!(flags & GLOB_APPEND)) {
181		pglob->gl_pathc = 0;
182		pglob->gl_pathv = NULL;
183		if (!(flags & GLOB_DOOFFS))
184			pglob->gl_offs = 0;
185	}
186	if (flags & GLOB_LIMIT) {
187		limit = pglob->gl_matchc;
188		if (limit == 0)
189			limit = ARG_MAX;
190	} else
191		limit = 0;
192	pglob->gl_flags = flags & ~GLOB_MAGCHAR;
193	pglob->gl_errfunc = errfunc;
194	pglob->gl_matchc = 0;
195
196	bufnext = patbuf;
197	bufend = bufnext + MAXPATHLEN - 1;
198	if (flags & GLOB_NOESCAPE) {
199		memset(&mbs, 0, sizeof(mbs));
200		while (bufend - bufnext >= MB_CUR_MAX) {
201			clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
202			if (clen == (size_t)-1 || clen == (size_t)-2)
203				return (GLOB_NOMATCH);
204			else if (clen == 0)
205				break;
206			*bufnext++ = wc;
207			patnext += clen;
208		}
209	} else {
210		/* Protect the quoted characters. */
211		memset(&mbs, 0, sizeof(mbs));
212		while (bufend - bufnext >= MB_CUR_MAX) {
213			if (*patnext == QUOTE) {
214				if (*++patnext == EOS) {
215					*bufnext++ = QUOTE | M_PROTECT;
216					continue;
217				}
218				prot = M_PROTECT;
219			} else
220				prot = 0;
221			clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs);
222			if (clen == (size_t)-1 || clen == (size_t)-2)
223				return (GLOB_NOMATCH);
224			else if (clen == 0)
225				break;
226			*bufnext++ = wc | prot;
227			patnext += clen;
228		}
229	}
230	*bufnext = EOS;
231
232	if (flags & GLOB_BRACE)
233	    return globexp1(patbuf, pglob, &limit);
234	else
235	    return glob0(patbuf, pglob, &limit);
236}
237
238/*
239 * Expand recursively a glob {} pattern. When there is no more expansion
240 * invoke the standard globbing routine to glob the rest of the magic
241 * characters
242 */
243static int
244globexp1(const Char *pattern, glob_t *pglob, size_t *limit)
245{
246	const Char* ptr = pattern;
247	int rv;
248
249	/* Protect a single {}, for find(1), like csh */
250	if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
251		return glob0(pattern, pglob, limit);
252
253	while ((ptr = (const Char *) g_strchr((Char *) ptr, LBRACE)) != NULL)
254		if (!globexp2(ptr, pattern, pglob, &rv, limit))
255			return rv;
256
257	return glob0(pattern, pglob, limit);
258}
259
260
261/*
262 * Recursive brace globbing helper. Tries to expand a single brace.
263 * If it succeeds then it invokes globexp1 with the new pattern.
264 * If it fails then it tries to glob the rest of the pattern and returns.
265 */
266static int
267globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv, size_t *limit)
268{
269	int     i;
270	Char   *lm, *ls;
271	const Char *pe, *pm, *pm1, *pl;
272	Char    patbuf[MAXPATHLEN];
273
274	/* copy part up to the brace */
275	for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
276		continue;
277	*lm = EOS;
278	ls = lm;
279
280	/* Find the balanced brace */
281	for (i = 0, pe = ++ptr; *pe; pe++)
282		if (*pe == LBRACKET) {
283			/* Ignore everything between [] */
284			for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
285				continue;
286			if (*pe == EOS) {
287				/*
288				 * We could not find a matching RBRACKET.
289				 * Ignore and just look for RBRACE
290				 */
291				pe = pm;
292			}
293		}
294		else if (*pe == LBRACE)
295			i++;
296		else if (*pe == RBRACE) {
297			if (i == 0)
298				break;
299			i--;
300		}
301
302	/* Non matching braces; just glob the pattern */
303	if (i != 0 || *pe == EOS) {
304		*rv = glob0(patbuf, pglob, limit);
305		return 0;
306	}
307
308	for (i = 0, pl = pm = ptr; pm <= pe; pm++)
309		switch (*pm) {
310		case LBRACKET:
311			/* Ignore everything between [] */
312			for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++)
313				continue;
314			if (*pm == EOS) {
315				/*
316				 * We could not find a matching RBRACKET.
317				 * Ignore and just look for RBRACE
318				 */
319				pm = pm1;
320			}
321			break;
322
323		case LBRACE:
324			i++;
325			break;
326
327		case RBRACE:
328			if (i) {
329			    i--;
330			    break;
331			}
332			/* FALLTHROUGH */
333		case COMMA:
334			if (i && *pm == COMMA)
335				break;
336			else {
337				/* Append the current string */
338				for (lm = ls; (pl < pm); *lm++ = *pl++)
339					continue;
340				/*
341				 * Append the rest of the pattern after the
342				 * closing brace
343				 */
344				for (pl = pe + 1; (*lm++ = *pl++) != EOS;)
345					continue;
346
347				/* Expand the current pattern */
348#ifdef DEBUG
349				qprintf("globexp2:", patbuf);
350#endif
351				*rv = globexp1(patbuf, pglob, limit);
352
353				/* move after the comma, to the next string */
354				pl = pm + 1;
355			}
356			break;
357
358		default:
359			break;
360		}
361	*rv = 0;
362	return 0;
363}
364
365
366
367/*
368 * expand tilde from the passwd file.
369 */
370static const Char *
371globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
372{
373	struct passwd *pwd;
374	char *h;
375	const Char *p;
376	Char *b, *eb;
377
378	if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
379		return pattern;
380
381	/*
382	 * Copy up to the end of the string or /
383	 */
384	eb = &patbuf[patbuf_len - 1];
385	for (p = pattern + 1, h = (char *) patbuf;
386	    h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
387		continue;
388
389	*h = EOS;
390
391	if (((char *) patbuf)[0] == EOS) {
392		/*
393		 * handle a plain ~ or ~/ by expanding $HOME first (iff
394		 * we're not running setuid or setgid) and then trying
395		 * the password file
396		 */
397		if (issetugid() != 0 ||
398		    (h = getenv("HOME")) == NULL) {
399			if (((h = getlogin()) != NULL &&
400			     (pwd = getpwnam(h)) != NULL) ||
401			    (pwd = getpwuid(getuid())) != NULL)
402				h = pwd->pw_dir;
403			else
404				return pattern;
405		}
406	}
407	else {
408		/*
409		 * Expand a ~user
410		 */
411		if ((pwd = getpwnam((char*) patbuf)) == NULL)
412			return pattern;
413		else
414			h = pwd->pw_dir;
415	}
416
417	/* Copy the home directory */
418	for (b = patbuf; b < eb && *h; *b++ = *h++)
419		continue;
420
421	/* Append the rest of the pattern */
422	while (b < eb && (*b++ = *p++) != EOS)
423		continue;
424	*b = EOS;
425
426	return patbuf;
427}
428
429
430/*
431 * The main glob() routine: compiles the pattern (optionally processing
432 * quotes), calls glob1() to do the real pattern matching, and finally
433 * sorts the list (unless unsorted operation is requested).  Returns 0
434 * if things went well, nonzero if errors occurred.
435 */
436static int
437glob0(const Char *pattern, glob_t *pglob, size_t *limit)
438{
439	const Char *qpatnext;
440	int c, err;
441	size_t oldpathc;
442	Char *bufnext, patbuf[MAXPATHLEN];
443
444	qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
445	oldpathc = pglob->gl_pathc;
446	bufnext = patbuf;
447
448	/* We don't need to check for buffer overflow any more. */
449	while ((c = *qpatnext++) != EOS) {
450		switch (c) {
451		case LBRACKET:
452			c = *qpatnext;
453			if (c == NOT)
454				++qpatnext;
455			if (*qpatnext == EOS ||
456			    g_strchr((Char *) qpatnext+1, RBRACKET) == NULL) {
457				*bufnext++ = LBRACKET;
458				if (c == NOT)
459					--qpatnext;
460				break;
461			}
462			*bufnext++ = M_SET;
463			if (c == NOT)
464				*bufnext++ = M_NOT;
465			c = *qpatnext++;
466			do {
467				*bufnext++ = CHAR(c);
468				if (*qpatnext == RANGE &&
469				    (c = qpatnext[1]) != RBRACKET) {
470					*bufnext++ = M_RNG;
471					*bufnext++ = CHAR(c);
472					qpatnext += 2;
473				}
474			} while ((c = *qpatnext++) != RBRACKET);
475			pglob->gl_flags |= GLOB_MAGCHAR;
476			*bufnext++ = M_END;
477			break;
478		case QUESTION:
479			pglob->gl_flags |= GLOB_MAGCHAR;
480			*bufnext++ = M_ONE;
481			break;
482		case STAR:
483			pglob->gl_flags |= GLOB_MAGCHAR;
484			/* collapse adjacent stars to one,
485			 * to avoid exponential behavior
486			 */
487			if (bufnext == patbuf || bufnext[-1] != M_ALL)
488			    *bufnext++ = M_ALL;
489			break;
490		default:
491			*bufnext++ = CHAR(c);
492			break;
493		}
494	}
495	*bufnext = EOS;
496#ifdef DEBUG
497	qprintf("glob0:", patbuf);
498#endif
499
500	if ((err = glob1(patbuf, pglob, limit)) != 0)
501		return(err);
502
503	/*
504	 * If there was no match we are going to append the pattern
505	 * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
506	 * and the pattern did not contain any magic characters
507	 * GLOB_NOMAGIC is there just for compatibility with csh.
508	 */
509	if (pglob->gl_pathc == oldpathc) {
510		if (((pglob->gl_flags & GLOB_NOCHECK) ||
511		    ((pglob->gl_flags & GLOB_NOMAGIC) &&
512			!(pglob->gl_flags & GLOB_MAGCHAR))))
513			return(globextend(pattern, pglob, limit));
514		else
515			return(GLOB_NOMATCH);
516	}
517	if (!(pglob->gl_flags & GLOB_NOSORT))
518		qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
519		    pglob->gl_pathc - oldpathc, sizeof(char *), compare);
520	return(0);
521}
522
523static int
524compare(const void *p, const void *q)
525{
526	return(strcmp(*(char **)p, *(char **)q));
527}
528
529static int
530glob1(Char *pattern, glob_t *pglob, size_t *limit)
531{
532	Char pathbuf[MAXPATHLEN];
533
534	/* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
535	if (*pattern == EOS)
536		return(0);
537	return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1,
538	    pattern, pglob, limit));
539}
540
541/*
542 * The functions glob2 and glob3 are mutually recursive; there is one level
543 * of recursion for each segment in the pattern that contains one or more
544 * meta characters.
545 */
546static int
547glob2(Char *pathbuf, Char *pathend, Char *pathend_last, Char *pattern,
548      glob_t *pglob, size_t *limit)
549{
550	struct stat sb;
551	Char *p, *q;
552	int anymeta;
553
554	/*
555	 * Loop over pattern segments until end of pattern or until
556	 * segment with meta character found.
557	 */
558	for (anymeta = 0;;) {
559		if (*pattern == EOS) {		/* End of pattern? */
560			*pathend = EOS;
561			if (g_lstat(pathbuf, &sb, pglob))
562				return(0);
563
564			if (((pglob->gl_flags & GLOB_MARK) &&
565			    pathend[-1] != SEP) && (S_ISDIR(sb.st_mode)
566			    || (S_ISLNK(sb.st_mode) &&
567			    (g_stat(pathbuf, &sb, pglob) == 0) &&
568			    S_ISDIR(sb.st_mode)))) {
569				if (pathend + 1 > pathend_last)
570					return (GLOB_ABORTED);
571				*pathend++ = SEP;
572				*pathend = EOS;
573			}
574			++pglob->gl_matchc;
575			return(globextend(pathbuf, pglob, limit));
576		}
577
578		/* Find end of next segment, copy tentatively to pathend. */
579		q = pathend;
580		p = pattern;
581		while (*p != EOS && *p != SEP) {
582			if (ismeta(*p))
583				anymeta = 1;
584			if (q + 1 > pathend_last)
585				return (GLOB_ABORTED);
586			*q++ = *p++;
587		}
588
589		if (!anymeta) {		/* No expansion, do next segment. */
590			pathend = q;
591			pattern = p;
592			while (*pattern == SEP) {
593				if (pathend + 1 > pathend_last)
594					return (GLOB_ABORTED);
595				*pathend++ = *pattern++;
596			}
597		} else			/* Need expansion, recurse. */
598			return(glob3(pathbuf, pathend, pathend_last, pattern, p,
599			    pglob, limit));
600	}
601	/* NOTREACHED */
602}
603
604static int
605glob3(Char *pathbuf, Char *pathend, Char *pathend_last,
606      Char *pattern, Char *restpattern,
607      glob_t *pglob, size_t *limit)
608{
609	struct dirent *dp;
610	DIR *dirp;
611	int err;
612	char buf[MAXPATHLEN];
613
614	/*
615	 * The readdirfunc declaration can't be prototyped, because it is
616	 * assigned, below, to two functions which are prototyped in glob.h
617	 * and dirent.h as taking pointers to differently typed opaque
618	 * structures.
619	 */
620	struct dirent *(*readdirfunc)();
621
622	if (pathend > pathend_last)
623		return (GLOB_ABORTED);
624	*pathend = EOS;
625	errno = 0;
626
627	if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
628		/* TODO: don't call for ENOENT or ENOTDIR? */
629		if (pglob->gl_errfunc) {
630			if (g_Ctoc(pathbuf, buf, sizeof(buf)))
631				return (GLOB_ABORTED);
632			if (pglob->gl_errfunc(buf, errno) ||
633			    pglob->gl_flags & GLOB_ERR)
634				return (GLOB_ABORTED);
635		}
636		return(0);
637	}
638
639	err = 0;
640
641	/* Search directory for matching names. */
642	if (pglob->gl_flags & GLOB_ALTDIRFUNC)
643		readdirfunc = pglob->gl_readdir;
644	else
645		readdirfunc = readdir;
646	while ((dp = (*readdirfunc)(dirp))) {
647		char *sc;
648		Char *dc;
649		wchar_t wc;
650		size_t clen;
651		mbstate_t mbs;
652
653		/* Initial DOT must be matched literally. */
654		if (dp->d_name[0] == DOT && *pattern != DOT)
655			continue;
656		memset(&mbs, 0, sizeof(mbs));
657		dc = pathend;
658		sc = dp->d_name;
659		while (dc < pathend_last) {
660			clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs);
661			if (clen == (size_t)-1 || clen == (size_t)-2) {
662				wc = *sc;
663				clen = 1;
664				memset(&mbs, 0, sizeof(mbs));
665			}
666			if ((*dc++ = wc) == EOS)
667				break;
668			sc += clen;
669		}
670		if (!match(pathend, pattern, restpattern)) {
671			*pathend = EOS;
672			continue;
673		}
674		err = glob2(pathbuf, --dc, pathend_last, restpattern,
675		    pglob, limit);
676		if (err)
677			break;
678	}
679
680	if (pglob->gl_flags & GLOB_ALTDIRFUNC)
681		(*pglob->gl_closedir)(dirp);
682	else
683		closedir(dirp);
684	return(err);
685}
686
687
688/*
689 * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
690 * add the new item, and update gl_pathc.
691 *
692 * This assumes the BSD realloc, which only copies the block when its size
693 * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
694 * behavior.
695 *
696 * Return 0 if new item added, error code if memory couldn't be allocated.
697 *
698 * Invariant of the glob_t structure:
699 *	Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
700 *	gl_pathv points to (gl_offs + gl_pathc + 1) items.
701 */
702static int
703globextend(const Char *path, glob_t *pglob, size_t *limit)
704{
705	char **pathv;
706	size_t i, newsize, len;
707	char *copy;
708	const Char *p;
709
710	if (*limit && pglob->gl_pathc > *limit) {
711		errno = 0;
712		return (GLOB_NOSPACE);
713	}
714
715	newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
716	pathv = pglob->gl_pathv ?
717		    realloc((char *)pglob->gl_pathv, newsize) :
718		    malloc(newsize);
719	if (pathv == NULL) {
720		if (pglob->gl_pathv) {
721			free(pglob->gl_pathv);
722			pglob->gl_pathv = NULL;
723		}
724		return(GLOB_NOSPACE);
725	}
726
727	if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
728		/* first time around -- clear initial gl_offs items */
729		pathv += pglob->gl_offs;
730		for (i = pglob->gl_offs + 1; --i > 0; )
731			*--pathv = NULL;
732	}
733	pglob->gl_pathv = pathv;
734
735	for (p = path; *p++;)
736		continue;
737	len = MB_CUR_MAX * (size_t)(p - path);	/* XXX overallocation */
738	if ((copy = malloc(len)) != NULL) {
739		if (g_Ctoc(path, copy, len)) {
740			free(copy);
741			return (GLOB_NOSPACE);
742		}
743		pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
744	}
745	pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
746	return(copy == NULL ? GLOB_NOSPACE : 0);
747}
748
749/*
750 * pattern matching function for filenames.  Each occurrence of the *
751 * pattern causes a recursion level.
752 */
753static int
754match(Char *name, Char *pat, Char *patend)
755{
756	int ok, negate_range;
757	Char c, k;
758
759	while (pat < patend) {
760		c = *pat++;
761		switch (c & M_MASK) {
762		case M_ALL:
763			if (pat == patend)
764				return(1);
765			do
766			    if (match(name, pat, patend))
767				    return(1);
768			while (*name++ != EOS);
769			return(0);
770		case M_ONE:
771			if (*name++ == EOS)
772				return(0);
773			break;
774		case M_SET:
775			ok = 0;
776			if ((k = *name++) == EOS)
777				return(0);
778			if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
779				++pat;
780			while (((c = *pat++) & M_MASK) != M_END)
781				if ((*pat & M_MASK) == M_RNG) {
782					if (__collate_load_error ?
783					    CHAR(c) <= CHAR(k) && CHAR(k) <= CHAR(pat[1]) :
784					       __collate_range_cmp(CHAR(c), CHAR(k)) <= 0
785					    && __collate_range_cmp(CHAR(k), CHAR(pat[1])) <= 0
786					   )
787						ok = 1;
788					pat += 2;
789				} else if (c == k)
790					ok = 1;
791			if (ok == negate_range)
792				return(0);
793			break;
794		default:
795			if (*name++ != c)
796				return(0);
797			break;
798		}
799	}
800	return(*name == EOS);
801}
802
803/* Free allocated data belonging to a glob_t structure. */
804void
805globfree(glob_t *pglob)
806{
807	size_t i;
808	char **pp;
809
810	if (pglob->gl_pathv != NULL) {
811		pp = pglob->gl_pathv + pglob->gl_offs;
812		for (i = pglob->gl_pathc; i--; ++pp)
813			if (*pp)
814				free(*pp);
815		free(pglob->gl_pathv);
816		pglob->gl_pathv = NULL;
817	}
818}
819
820static DIR *
821g_opendir(Char *str, glob_t *pglob)
822{
823	char buf[MAXPATHLEN];
824
825	if (!*str)
826		strcpy(buf, ".");
827	else {
828		if (g_Ctoc(str, buf, sizeof(buf)))
829			return (NULL);
830	}
831
832	if (pglob->gl_flags & GLOB_ALTDIRFUNC)
833		return((*pglob->gl_opendir)(buf));
834
835	return(opendir(buf));
836}
837
838static int
839g_lstat(Char *fn, struct stat *sb, glob_t *pglob)
840{
841	char buf[MAXPATHLEN];
842
843	if (g_Ctoc(fn, buf, sizeof(buf))) {
844		errno = ENAMETOOLONG;
845		return (-1);
846	}
847	if (pglob->gl_flags & GLOB_ALTDIRFUNC)
848		return((*pglob->gl_lstat)(buf, sb));
849	return(lstat(buf, sb));
850}
851
852static int
853g_stat(Char *fn, struct stat *sb, glob_t *pglob)
854{
855	char buf[MAXPATHLEN];
856
857	if (g_Ctoc(fn, buf, sizeof(buf))) {
858		errno = ENAMETOOLONG;
859		return (-1);
860	}
861	if (pglob->gl_flags & GLOB_ALTDIRFUNC)
862		return((*pglob->gl_stat)(buf, sb));
863	return(stat(buf, sb));
864}
865
866static Char *
867g_strchr(Char *str, wchar_t ch)
868{
869
870	do {
871		if (*str == ch)
872			return (str);
873	} while (*str++);
874	return (NULL);
875}
876
877static int
878g_Ctoc(const Char *str, char *buf, size_t len)
879{
880	mbstate_t mbs;
881	size_t clen;
882
883	memset(&mbs, 0, sizeof(mbs));
884	while (len >= MB_CUR_MAX) {
885		clen = wcrtomb(buf, *str, &mbs);
886		if (clen == (size_t)-1)
887			return (1);
888		if (*str == L'\0')
889			return (0);
890		str++;
891		buf += clen;
892		len -= clen;
893	}
894	return (1);
895}
896
897#ifdef DEBUG
898static void
899qprintf(const char *str, Char *s)
900{
901	Char *p;
902
903	(void)printf("%s:\n", str);
904	for (p = s; *p; p++)
905		(void)printf("%c", CHAR(*p));
906	(void)printf("\n");
907	for (p = s; *p; p++)
908		(void)printf("%c", *p & M_PROTECT ? '"' : ' ');
909	(void)printf("\n");
910	for (p = s; *p; p++)
911		(void)printf("%c", ismeta(*p) ? '_' : ' ');
912	(void)printf("\n");
913}
914#endif
915