1 /*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1989, 1993, 1994
5 * The Regents of the University of California. All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * Guido van Rossum.
9 *
10 * Copyright (c) 2011 The FreeBSD Foundation
11 *
12 * Portions of this software were developed by David Chisnall
13 * under sponsorship from the FreeBSD Foundation.
14 *
15 * Redistribution and use in source and binary forms, with or without
16 * modification, are permitted provided that the following conditions
17 * are met:
18 * 1. Redistributions of source code must retain the above copyright
19 * notice, this list of conditions and the following disclaimer.
20 * 2. Redistributions in binary form must reproduce the above copyright
21 * notice, this list of conditions and the following disclaimer in the
22 * documentation and/or other materials provided with the distribution.
23 * 3. Neither the name of the University nor the names of its contributors
24 * may be used to endorse or promote products derived from this software
25 * without specific prior written permission.
26 *
27 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
37 * SUCH DAMAGE.
38 */
39
40 #include <sys/cdefs.h>
41 __SCCSID("@(#)fnmatch.c 8.2 (Berkeley) 4/16/94");
42 /*
43 * Function fnmatch() as specified in POSIX 1003.2-1992, section B.6.
44 * Compares a filename or pathname to a pattern.
45 */
46
47 /*
48 * Some notes on multibyte character support:
49 * 1. Patterns with illegal byte sequences match nothing.
50 * 2. Illegal byte sequences in the "string" argument are handled by treating
51 * them as single-byte characters with a value of the first byte of the
52 * sequence cast to wchar_t.
53 * 3. Multibyte conversion state objects (mbstate_t) are passed around and
54 * used for most, but not all, conversions. Further work will be required
55 * to support state-dependent encodings.
56 */
57
58 #include <fnmatch.h>
59 #include <limits.h>
60 #include <string.h>
61 #include <wchar.h>
62 #include <wctype.h>
63
64 #include "collate.h"
65
66 #define EOS '\0'
67
68 #define RANGE_MATCH 1
69 #define RANGE_NOMATCH 0
70 #define RANGE_ERROR (-1)
71
72 static int rangematch(const char *, wchar_t, int, char **, mbstate_t *);
73 static int fnmatch1(const char *, const char *, const char *, int, mbstate_t,
74 mbstate_t);
75
76 int
fnmatch(const char * pattern,const char * string,int flags)77 fnmatch(const char *pattern, const char *string, int flags)
78 {
79 static const mbstate_t initial;
80
81 return (fnmatch1(pattern, string, string, flags, initial, initial));
82 }
83
84 static int
fnmatch1(const char * pattern,const char * string,const char * stringstart,int flags,mbstate_t patmbs,mbstate_t strmbs)85 fnmatch1(const char *pattern, const char *string, const char *stringstart,
86 int flags, mbstate_t patmbs, mbstate_t strmbs)
87 {
88 const char *bt_pattern, *bt_string;
89 mbstate_t bt_patmbs, bt_strmbs;
90 char *newp;
91 char c;
92 wchar_t pc, sc;
93 size_t pclen, sclen;
94
95 bt_pattern = bt_string = NULL;
96 for (;;) {
97 pclen = mbrtowc(&pc, pattern, MB_LEN_MAX, &patmbs);
98 if (pclen == (size_t)-1 || pclen == (size_t)-2)
99 return (FNM_NOMATCH);
100 pattern += pclen;
101 sclen = mbrtowc(&sc, string, MB_LEN_MAX, &strmbs);
102 if (sclen == (size_t)-1 || sclen == (size_t)-2) {
103 sc = (unsigned char)*string;
104 sclen = 1;
105 memset(&strmbs, 0, sizeof(strmbs));
106 }
107 switch (pc) {
108 case EOS:
109 if ((flags & FNM_LEADING_DIR) && sc == '/')
110 return (0);
111 if (sc == EOS)
112 return (0);
113 goto backtrack;
114 case '?':
115 if (sc == EOS)
116 return (FNM_NOMATCH);
117 if (sc == '/' && (flags & FNM_PATHNAME))
118 goto backtrack;
119 if (sc == '.' && (flags & FNM_PERIOD) &&
120 (string == stringstart ||
121 ((flags & FNM_PATHNAME) && *(string - 1) == '/')))
122 goto backtrack;
123 string += sclen;
124 break;
125 case '*':
126 c = *pattern;
127 /* Collapse multiple stars. */
128 while (c == '*')
129 c = *++pattern;
130
131 if (sc == '.' && (flags & FNM_PERIOD) &&
132 (string == stringstart ||
133 ((flags & FNM_PATHNAME) && *(string - 1) == '/')))
134 goto backtrack;
135
136 /* Optimize for pattern with * at end or before /. */
137 if (c == EOS)
138 if (flags & FNM_PATHNAME)
139 return ((flags & FNM_LEADING_DIR) ||
140 strchr(string, '/') == NULL ?
141 0 : FNM_NOMATCH);
142 else
143 return (0);
144 else if (c == '/' && flags & FNM_PATHNAME) {
145 if ((string = strchr(string, '/')) == NULL)
146 return (FNM_NOMATCH);
147 break;
148 }
149
150 /*
151 * First try the shortest match for the '*' that
152 * could work. We can forget any earlier '*' since
153 * there is no way having it match more characters
154 * can help us, given that we are already here.
155 */
156 bt_pattern = pattern, bt_patmbs = patmbs;
157 bt_string = string, bt_strmbs = strmbs;
158 break;
159 case '[':
160 if (sc == EOS)
161 return (FNM_NOMATCH);
162 if (sc == '/' && (flags & FNM_PATHNAME))
163 goto backtrack;
164 if (sc == '.' && (flags & FNM_PERIOD) &&
165 (string == stringstart ||
166 ((flags & FNM_PATHNAME) && *(string - 1) == '/')))
167 goto backtrack;
168
169 switch (rangematch(pattern, sc, flags, &newp,
170 &patmbs)) {
171 case RANGE_ERROR:
172 goto norm;
173 case RANGE_MATCH:
174 pattern = newp;
175 break;
176 case RANGE_NOMATCH:
177 goto backtrack;
178 }
179 string += sclen;
180 break;
181 case '\\':
182 if (!(flags & FNM_NOESCAPE)) {
183 pclen = mbrtowc(&pc, pattern, MB_LEN_MAX,
184 &patmbs);
185 if (pclen == 0 || pclen == (size_t)-1 ||
186 pclen == (size_t)-2)
187 return (FNM_NOMATCH);
188 pattern += pclen;
189 }
190 /* FALLTHROUGH */
191 default:
192 norm:
193 string += sclen;
194 if (pc == sc)
195 ;
196 else if ((flags & FNM_CASEFOLD) &&
197 (towlower(pc) == towlower(sc)))
198 ;
199 else {
200 backtrack:
201 /*
202 * If we have a mismatch (other than hitting
203 * the end of the string), go back to the last
204 * '*' seen and have it match one additional
205 * character.
206 */
207 if (bt_pattern == NULL)
208 return (FNM_NOMATCH);
209 sclen = mbrtowc(&sc, bt_string, MB_LEN_MAX,
210 &bt_strmbs);
211 if (sclen == (size_t)-1 ||
212 sclen == (size_t)-2) {
213 sc = (unsigned char)*bt_string;
214 sclen = 1;
215 memset(&bt_strmbs, 0,
216 sizeof(bt_strmbs));
217 }
218 if (sc == EOS)
219 return (FNM_NOMATCH);
220 if (sc == '/' && flags & FNM_PATHNAME)
221 return (FNM_NOMATCH);
222 bt_string += sclen;
223 pattern = bt_pattern, patmbs = bt_patmbs;
224 string = bt_string, strmbs = bt_strmbs;
225 }
226 break;
227 }
228 }
229 /* NOTREACHED */
230 }
231
232 static int
rangematch(const char * pattern,wchar_t test,int flags,char ** newp,mbstate_t * patmbs)233 rangematch(const char *pattern, wchar_t test, int flags, char **newp,
234 mbstate_t *patmbs)
235 {
236 int negate, ok;
237 wchar_t c, c2;
238 size_t pclen;
239 const char *origpat;
240 struct xlocale_collate *table =
241 (struct xlocale_collate*)__get_locale()->components[XLC_COLLATE];
242
243 /*
244 * A bracket expression starting with an unquoted circumflex
245 * character produces unspecified results (IEEE 1003.2-1992,
246 * 3.13.2). This implementation treats it like '!', for
247 * consistency with the regular expression syntax.
248 * J.T. Conklin (conklin@ngai.kaleida.com)
249 */
250 if ((negate = (*pattern == '!' || *pattern == '^')))
251 ++pattern;
252
253 if (flags & FNM_CASEFOLD)
254 test = towlower(test);
255
256 /*
257 * A right bracket shall lose its special meaning and represent
258 * itself in a bracket expression if it occurs first in the list.
259 * -- POSIX.2 2.8.3.2
260 */
261 ok = 0;
262 origpat = pattern;
263 for (;;) {
264 if (*pattern == ']' && pattern > origpat) {
265 pattern++;
266 break;
267 } else if (*pattern == '\0') {
268 return (RANGE_ERROR);
269 } else if (*pattern == '/' && (flags & FNM_PATHNAME)) {
270 return (RANGE_NOMATCH);
271 } else if (*pattern == '\\' && !(flags & FNM_NOESCAPE))
272 pattern++;
273 pclen = mbrtowc(&c, pattern, MB_LEN_MAX, patmbs);
274 if (pclen == (size_t)-1 || pclen == (size_t)-2)
275 return (RANGE_NOMATCH);
276 pattern += pclen;
277
278 if (flags & FNM_CASEFOLD)
279 c = towlower(c);
280
281 if (*pattern == '-' && *(pattern + 1) != EOS &&
282 *(pattern + 1) != ']') {
283 if (*++pattern == '\\' && !(flags & FNM_NOESCAPE))
284 if (*pattern != EOS)
285 pattern++;
286 pclen = mbrtowc(&c2, pattern, MB_LEN_MAX, patmbs);
287 if (pclen == (size_t)-1 || pclen == (size_t)-2)
288 return (RANGE_NOMATCH);
289 pattern += pclen;
290 if (c2 == EOS)
291 return (RANGE_ERROR);
292
293 if (flags & FNM_CASEFOLD)
294 c2 = towlower(c2);
295
296 if (table->__collate_load_error ?
297 c <= test && test <= c2 :
298 __wcollate_range_cmp(c, test) <= 0
299 && __wcollate_range_cmp(test, c2) <= 0
300 )
301 ok = 1;
302 } else if (c == test)
303 ok = 1;
304 }
305
306 *newp = (char *)pattern;
307 return (ok == negate ? RANGE_NOMATCH : RANGE_MATCH);
308 }
309