xref: /freebsd/lib/libc/gen/fnmatch.c (revision 6f9c8e5b074419423648ffb89b83fd2f257e90b7)
1 /*
2  * Copyright (c) 1989, 1993, 1994
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 #if defined(LIBC_SCCS) && !defined(lint)
34 static char sccsid[] = "@(#)fnmatch.c	8.2 (Berkeley) 4/16/94";
35 #endif /* LIBC_SCCS and not lint */
36 #include <sys/cdefs.h>
37 __FBSDID("$FreeBSD$");
38 
39 /*
40  * Function fnmatch() as specified in POSIX 1003.2-1992, section B.6.
41  * Compares a filename or pathname to a pattern.
42  */
43 
44 /*
45  * Some notes on multibyte character support:
46  * 1. Patterns with illegal byte sequences match nothing.
47  * 2. Illegal byte sequences in the "string" argument are handled by treating
48  *    them as single-byte characters with a value of the first byte of the
49  *    sequence cast to wchar_t.
50  * 3. Multibyte conversion state objects (mbstate_t) are passed around and
51  *    used for most, but not all, conversions. Further work will be required
52  *    to support state-dependent encodings.
53  */
54 
55 #include <fnmatch.h>
56 #include <limits.h>
57 #include <string.h>
58 #include <wchar.h>
59 #include <wctype.h>
60 
61 #include "collate.h"
62 
63 #define	EOS	'\0'
64 
65 #define RANGE_MATCH     1
66 #define RANGE_NOMATCH   0
67 #define RANGE_ERROR     (-1)
68 
69 static int rangematch(const char *, wchar_t, int, char **, mbstate_t *);
70 static int fnmatch1(const char *, const char *, const char *, int, mbstate_t,
71 		mbstate_t);
72 
73 int
74 fnmatch(pattern, string, flags)
75 	const char *pattern, *string;
76 	int flags;
77 {
78 	static const mbstate_t initial;
79 
80 	return (fnmatch1(pattern, string, string, flags, initial, initial));
81 }
82 
83 static int
84 fnmatch1(pattern, string, stringstart, flags, patmbs, strmbs)
85 	const char *pattern, *string, *stringstart;
86 	int flags;
87 	mbstate_t patmbs, strmbs;
88 {
89 	char *newp;
90 	char c;
91 	wchar_t pc, sc;
92 	size_t pclen, sclen;
93 
94 	for (;;) {
95 		pclen = mbrtowc(&pc, pattern, MB_LEN_MAX, &patmbs);
96 		if (pclen == (size_t)-1 || pclen == (size_t)-2)
97 			return (FNM_NOMATCH);
98 		pattern += pclen;
99 		sclen = mbrtowc(&sc, string, MB_LEN_MAX, &strmbs);
100 		if (sclen == (size_t)-1 || sclen == (size_t)-2) {
101 			sc = (unsigned char)*string;
102 			sclen = 1;
103 			memset(&strmbs, 0, sizeof(strmbs));
104 		}
105 		switch (pc) {
106 		case EOS:
107 			if ((flags & FNM_LEADING_DIR) && sc == '/')
108 				return (0);
109 			return (sc == EOS ? 0 : FNM_NOMATCH);
110 		case '?':
111 			if (sc == EOS)
112 				return (FNM_NOMATCH);
113 			if (sc == '/' && (flags & FNM_PATHNAME))
114 				return (FNM_NOMATCH);
115 			if (sc == '.' && (flags & FNM_PERIOD) &&
116 			    (string == stringstart ||
117 			    ((flags & FNM_PATHNAME) && *(string - 1) == '/')))
118 				return (FNM_NOMATCH);
119 			string += sclen;
120 			break;
121 		case '*':
122 			c = *pattern;
123 			/* Collapse multiple stars. */
124 			while (c == '*')
125 				c = *++pattern;
126 
127 			if (sc == '.' && (flags & FNM_PERIOD) &&
128 			    (string == stringstart ||
129 			    ((flags & FNM_PATHNAME) && *(string - 1) == '/')))
130 				return (FNM_NOMATCH);
131 
132 			/* Optimize for pattern with * at end or before /. */
133 			if (c == EOS)
134 				if (flags & FNM_PATHNAME)
135 					return ((flags & FNM_LEADING_DIR) ||
136 					    strchr(string, '/') == NULL ?
137 					    0 : FNM_NOMATCH);
138 				else
139 					return (0);
140 			else if (c == '/' && flags & FNM_PATHNAME) {
141 				if ((string = strchr(string, '/')) == NULL)
142 					return (FNM_NOMATCH);
143 				break;
144 			}
145 
146 			/* General case, use recursion. */
147 			while (sc != EOS) {
148 				if (!fnmatch1(pattern, string, stringstart,
149 				    flags, patmbs, strmbs))
150 					return (0);
151 				sclen = mbrtowc(&sc, string, MB_LEN_MAX,
152 				    &strmbs);
153 				if (sclen == (size_t)-1 ||
154 				    sclen == (size_t)-2) {
155 					sc = (unsigned char)*string;
156 					sclen = 1;
157 					memset(&strmbs, 0, sizeof(strmbs));
158 				}
159 				if (sc == '/' && flags & FNM_PATHNAME)
160 					break;
161 				string += sclen;
162 			}
163 			return (FNM_NOMATCH);
164 		case '[':
165 			if (sc == EOS)
166 				return (FNM_NOMATCH);
167 			if (sc == '/' && (flags & FNM_PATHNAME))
168 				return (FNM_NOMATCH);
169 			if (sc == '.' && (flags & FNM_PERIOD) &&
170 			    (string == stringstart ||
171 			    ((flags & FNM_PATHNAME) && *(string - 1) == '/')))
172 				return (FNM_NOMATCH);
173 
174 			switch (rangematch(pattern, sc, flags, &newp,
175 			    &patmbs)) {
176 			case RANGE_ERROR:
177 				goto norm;
178 			case RANGE_MATCH:
179 				pattern = newp;
180 				break;
181 			case RANGE_NOMATCH:
182 				return (FNM_NOMATCH);
183 			}
184 			string += sclen;
185 			break;
186 		case '\\':
187 			if (!(flags & FNM_NOESCAPE)) {
188 				pclen = mbrtowc(&pc, pattern, MB_LEN_MAX,
189 				    &patmbs);
190 				if (pclen == (size_t)-1 || pclen == (size_t)-2)
191 					return (FNM_NOMATCH);
192 				if (pclen == 0)
193 					pc = '\\';
194 				pattern += pclen;
195 			}
196 			/* FALLTHROUGH */
197 		default:
198 		norm:
199 			if (pc == sc)
200 				;
201 			else if ((flags & FNM_CASEFOLD) &&
202 				 (towlower(pc) == towlower(sc)))
203 				;
204 			else
205 				return (FNM_NOMATCH);
206 			string += sclen;
207 			break;
208 		}
209 	}
210 	/* NOTREACHED */
211 }
212 
213 static int
214 rangematch(pattern, test, flags, newp, patmbs)
215 	const char *pattern;
216 	wchar_t test;
217 	int flags;
218 	char **newp;
219 	mbstate_t *patmbs;
220 {
221 	int negate, ok;
222 	wchar_t c, c2;
223 	size_t pclen;
224 	const char *origpat;
225 
226 	/*
227 	 * A bracket expression starting with an unquoted circumflex
228 	 * character produces unspecified results (IEEE 1003.2-1992,
229 	 * 3.13.2).  This implementation treats it like '!', for
230 	 * consistency with the regular expression syntax.
231 	 * J.T. Conklin (conklin@ngai.kaleida.com)
232 	 */
233 	if ( (negate = (*pattern == '!' || *pattern == '^')) )
234 		++pattern;
235 
236 	if (flags & FNM_CASEFOLD)
237 		test = towlower(test);
238 
239 	/*
240 	 * A right bracket shall lose its special meaning and represent
241 	 * itself in a bracket expression if it occurs first in the list.
242 	 * -- POSIX.2 2.8.3.2
243 	 */
244 	ok = 0;
245 	origpat = pattern;
246 	for (;;) {
247 		if (*pattern == ']' && pattern > origpat) {
248 			pattern++;
249 			break;
250 		} else if (*pattern == '\0') {
251 			return (RANGE_ERROR);
252 		} else if (*pattern == '/' && (flags & FNM_PATHNAME)) {
253 			return (RANGE_NOMATCH);
254 		} else if (*pattern == '\\' && !(flags & FNM_NOESCAPE))
255 			pattern++;
256 		pclen = mbrtowc(&c, pattern, MB_LEN_MAX, patmbs);
257 		if (pclen == (size_t)-1 || pclen == (size_t)-2)
258 			return (RANGE_NOMATCH);
259 		pattern += pclen;
260 
261 		if (flags & FNM_CASEFOLD)
262 			c = towlower(c);
263 
264 		if (*pattern == '-' && *(pattern + 1) != EOS &&
265 		    *(pattern + 1) != ']') {
266 			if (*++pattern == '\\' && !(flags & FNM_NOESCAPE))
267 				if (*pattern != EOS)
268 					pattern++;
269 			pclen = mbrtowc(&c2, pattern, MB_LEN_MAX, patmbs);
270 			if (pclen == (size_t)-1 || pclen == (size_t)-2)
271 				return (RANGE_NOMATCH);
272 			pattern += pclen;
273 			if (c2 == EOS)
274 				return (RANGE_ERROR);
275 
276 			if (flags & FNM_CASEFOLD)
277 				c2 = towlower(c2);
278 
279 			if (__collate_load_error ?
280 			    c <= test && test <= c2 :
281 			       __collate_range_cmp(c, test) <= 0
282 			    && __collate_range_cmp(test, c2) <= 0
283 			   )
284 				ok = 1;
285 		} else if (c == test)
286 			ok = 1;
287 	}
288 
289 	*newp = (char *)pattern;
290 	return (ok == negate ? RANGE_NOMATCH : RANGE_MATCH);
291 }
292