1 /* Copyright (C) 1991-2020 Free Software Foundation, Inc.
2 This file is part of the GNU C Library.
3
4 The GNU C Library is free software; you can redistribute it and/or
5 modify it under the terms of the GNU General Public
6 License as published by the Free Software Foundation; either
7 version 3 of the License, or (at your option) any later version.
8
9 The GNU C Library is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 General Public License for more details.
13
14 You should have received a copy of the GNU General Public
15 License along with the GNU C Library; if not, see
16 <https://www.gnu.org/licenses/>. */
17
18 #ifndef _LIBC
19 # include <libc-config.h>
20 #endif
21
22 /* Enable GNU extensions in fnmatch.h. */
23 #ifndef _GNU_SOURCE
24 # define _GNU_SOURCE 1
25 #endif
26
27 #include <fnmatch.h>
28
29 #include <assert.h>
30 #include <errno.h>
31 #include <ctype.h>
32 #include <string.h>
33 #include <stdlib.h>
34 #if defined _LIBC || HAVE_ALLOCA
35 # include <alloca.h>
36 #endif
37 #include <wchar.h>
38 #include <wctype.h>
39 #include <stddef.h>
40 #include <stdbool.h>
41
42 /* We need some of the locale data (the collation sequence information)
43 but there is no interface to get this information in general. Therefore
44 we support a correct implementation only in glibc. */
45 #ifdef _LIBC
46 # include "../locale/localeinfo.h"
47 # include "../locale/coll-lookup.h"
48 # include <shlib-compat.h>
49
50 # define CONCAT(a,b) __CONCAT(a,b)
51 # define btowc __btowc
52 # define iswctype __iswctype
53 # define mbsrtowcs __mbsrtowcs
54 # define mempcpy __mempcpy
55 # define strnlen __strnlen
56 # define towlower __towlower
57 # define wcscat __wcscat
58 # define wcslen __wcslen
59 # define wctype __wctype
60 # define wmemchr __wmemchr
61 # define wmempcpy __wmempcpy
62 # define fnmatch __fnmatch
63 extern int fnmatch (const char *pattern, const char *string, int flags);
64 #endif
65
66 #ifdef _LIBC
67 # if __GNUC__ < 7
68 # define FALLTHROUGH ((void) 0)
69 # else
70 # define FALLTHROUGH __attribute__ ((__fallthrough__))
71 # endif
72 #else
73 # include "attribute.h"
74 #endif
75
76 #include <intprops.h>
77 #include <flexmember.h>
78
79 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */
80 #define NO_LEADING_PERIOD(flags) \
81 ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
82
83 #ifndef _LIBC
84 # if HAVE_ALLOCA
85 /* The OS usually guarantees only one guard page at the bottom of the stack,
86 and a page size can be as small as 4096 bytes. So we cannot safely
87 allocate anything larger than 4096 bytes. Also care for the possibility
88 of a few compiler-allocated temporary stack slots. */
89 # define __libc_use_alloca(n) ((n) < 4032)
90 # else
91 /* Just use malloc. */
92 # define __libc_use_alloca(n) false
93 # undef alloca
94 # define alloca(n) malloc (n)
95 # endif
96 # define alloca_account(size, avar) ((avar) += (size), alloca (size))
97 #endif
98
99 /* Provide support for user-defined character classes, based on the functions
100 from ISO C 90 amendment 1. */
101 #ifdef CHARCLASS_NAME_MAX
102 # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
103 #else
104 /* This shouldn't happen but some implementation might still have this
105 problem. Use a reasonable default value. */
106 # define CHAR_CLASS_MAX_LENGTH 256
107 #endif
108
109 #define IS_CHAR_CLASS(string) wctype (string)
110
111 /* Avoid depending on library functions or files
112 whose names are inconsistent. */
113
114 /* Global variable. */
115 static int posixly_correct;
116
117 /* Note that this evaluates C many times. */
118 #define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
119 #define CHAR char
120 #define UCHAR unsigned char
121 #define INT int
122 #define FCT internal_fnmatch
123 #define EXT ext_match
124 #define END end_pattern
125 #define STRUCT fnmatch_struct
126 #define L_(CS) CS
127 #define BTOWC(C) btowc (C)
128 #define STRLEN(S) strlen (S)
129 #define STRCAT(D, S) strcat (D, S)
130 #define MEMPCPY(D, S, N) mempcpy (D, S, N)
131 #define MEMCHR(S, C, N) memchr (S, C, N)
132 #define WIDE_CHAR_VERSION 0
133 #ifdef _LIBC
134 # include <locale/weight.h>
135 # define FINDIDX findidx
136 #endif
137 #include "fnmatch_loop.c"
138
139
140 #define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
141 #define CHAR wchar_t
142 #define UCHAR wint_t
143 #define INT wint_t
144 #define FCT internal_fnwmatch
145 #define EXT ext_wmatch
146 #define END end_wpattern
147 #define L_(CS) L##CS
148 #define BTOWC(C) (C)
149 #define STRLEN(S) wcslen (S)
150 #define STRCAT(D, S) wcscat (D, S)
151 #define MEMPCPY(D, S, N) wmempcpy (D, S, N)
152 #define MEMCHR(S, C, N) wmemchr (S, C, N)
153 #define WIDE_CHAR_VERSION 1
154 #ifdef _LIBC
155 /* Change the name the header defines so it doesn't conflict with
156 the <locale/weight.h> version included above. */
157 # define findidx findidxwc
158 # include <locale/weightwc.h>
159 # undef findidx
160 # define FINDIDX findidxwc
161 #endif
162
163 #undef IS_CHAR_CLASS
164 /* We have to convert the wide character string in a multibyte string. But
165 we know that the character class names consist of alphanumeric characters
166 from the portable character set, and since the wide character encoding
167 for a member of the portable character set is the same code point as
168 its single-byte encoding, we can use a simplified method to convert the
169 string to a multibyte character string. */
170 static wctype_t
is_char_class(const wchar_t * wcs)171 is_char_class (const wchar_t *wcs)
172 {
173 char s[CHAR_CLASS_MAX_LENGTH + 1];
174 char *cp = s;
175
176 do
177 {
178 /* Test for a printable character from the portable character set. */
179 #ifdef _LIBC
180 if (*wcs < 0x20 || *wcs > 0x7e
181 || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
182 return (wctype_t) 0;
183 #else
184 switch (*wcs)
185 {
186 case L' ': case L'!': case L'"': case L'#': case L'%':
187 case L'&': case L'\'': case L'(': case L')': case L'*':
188 case L'+': case L',': case L'-': case L'.': case L'/':
189 case L'0': case L'1': case L'2': case L'3': case L'4':
190 case L'5': case L'6': case L'7': case L'8': case L'9':
191 case L':': case L';': case L'<': case L'=': case L'>':
192 case L'?':
193 case L'A': case L'B': case L'C': case L'D': case L'E':
194 case L'F': case L'G': case L'H': case L'I': case L'J':
195 case L'K': case L'L': case L'M': case L'N': case L'O':
196 case L'P': case L'Q': case L'R': case L'S': case L'T':
197 case L'U': case L'V': case L'W': case L'X': case L'Y':
198 case L'Z':
199 case L'[': case L'\\': case L']': case L'^': case L'_':
200 case L'a': case L'b': case L'c': case L'd': case L'e':
201 case L'f': case L'g': case L'h': case L'i': case L'j':
202 case L'k': case L'l': case L'm': case L'n': case L'o':
203 case L'p': case L'q': case L'r': case L's': case L't':
204 case L'u': case L'v': case L'w': case L'x': case L'y':
205 case L'z': case L'{': case L'|': case L'}': case L'~':
206 break;
207 default:
208 return (wctype_t) 0;
209 }
210 #endif
211
212 /* Avoid overrunning the buffer. */
213 if (cp == s + CHAR_CLASS_MAX_LENGTH)
214 return (wctype_t) 0;
215
216 *cp++ = (char) *wcs++;
217 }
218 while (*wcs != L'\0');
219
220 *cp = '\0';
221
222 return wctype (s);
223 }
224 #define IS_CHAR_CLASS(string) is_char_class (string)
225
226 #include "fnmatch_loop.c"
227
228
229 int
fnmatch(const char * pattern,const char * string,int flags)230 fnmatch (const char *pattern, const char *string, int flags)
231 {
232 if (__glibc_unlikely (MB_CUR_MAX != 1))
233 {
234 mbstate_t ps;
235 size_t n;
236 const char *p;
237 wchar_t *wpattern_malloc = NULL;
238 wchar_t *wpattern;
239 wchar_t *wstring_malloc = NULL;
240 wchar_t *wstring;
241 size_t alloca_used = 0;
242
243 /* Convert the strings into wide characters. */
244 memset (&ps, '\0', sizeof (ps));
245 p = pattern;
246 n = strnlen (pattern, 1024);
247 if (__glibc_likely (n < 1024))
248 {
249 wpattern = (wchar_t *) alloca_account ((n + 1) * sizeof (wchar_t),
250 alloca_used);
251 n = mbsrtowcs (wpattern, &p, n + 1, &ps);
252 if (__glibc_unlikely (n == (size_t) -1))
253 /* Something wrong.
254 XXX Do we have to set 'errno' to something which mbsrtows hasn't
255 already done? */
256 return -1;
257 if (p)
258 {
259 memset (&ps, '\0', sizeof (ps));
260 goto prepare_wpattern;
261 }
262 }
263 else
264 {
265 prepare_wpattern:
266 n = mbsrtowcs (NULL, &pattern, 0, &ps);
267 if (__glibc_unlikely (n == (size_t) -1))
268 /* Something wrong.
269 XXX Do we have to set 'errno' to something which mbsrtows hasn't
270 already done? */
271 return -1;
272 if (__glibc_unlikely (n >= (size_t) -1 / sizeof (wchar_t)))
273 {
274 __set_errno (ENOMEM);
275 return -2;
276 }
277 wpattern_malloc = wpattern
278 = (wchar_t *) malloc ((n + 1) * sizeof (wchar_t));
279 assert (mbsinit (&ps));
280 if (wpattern == NULL)
281 return -2;
282 (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
283 }
284
285 assert (mbsinit (&ps));
286 n = strnlen (string, 1024);
287 p = string;
288 if (__glibc_likely (n < 1024))
289 {
290 wstring = (wchar_t *) alloca_account ((n + 1) * sizeof (wchar_t),
291 alloca_used);
292 n = mbsrtowcs (wstring, &p, n + 1, &ps);
293 if (__glibc_unlikely (n == (size_t) -1))
294 {
295 /* Something wrong.
296 XXX Do we have to set 'errno' to something which
297 mbsrtows hasn't already done? */
298 free_return:
299 free (wpattern_malloc);
300 return -1;
301 }
302 if (p)
303 {
304 memset (&ps, '\0', sizeof (ps));
305 goto prepare_wstring;
306 }
307 }
308 else
309 {
310 prepare_wstring:
311 n = mbsrtowcs (NULL, &string, 0, &ps);
312 if (__glibc_unlikely (n == (size_t) -1))
313 /* Something wrong.
314 XXX Do we have to set 'errno' to something which mbsrtows hasn't
315 already done? */
316 goto free_return;
317 if (__glibc_unlikely (n >= (size_t) -1 / sizeof (wchar_t)))
318 {
319 free (wpattern_malloc);
320 __set_errno (ENOMEM);
321 return -2;
322 }
323
324 wstring_malloc = wstring
325 = (wchar_t *) malloc ((n + 1) * sizeof (wchar_t));
326 if (wstring == NULL)
327 {
328 free (wpattern_malloc);
329 return -2;
330 }
331 assert (mbsinit (&ps));
332 (void) mbsrtowcs (wstring, &string, n + 1, &ps);
333 }
334
335 int res = internal_fnwmatch (wpattern, wstring, wstring + n,
336 flags & FNM_PERIOD, flags, NULL,
337 alloca_used);
338
339 free (wstring_malloc);
340 free (wpattern_malloc);
341
342 return res;
343 }
344
345 return internal_fnmatch (pattern, string, string + strlen (string),
346 flags & FNM_PERIOD, flags, NULL, 0);
347 }
348
349 #undef fnmatch
350 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
351 #if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
352 strong_alias (__fnmatch, __fnmatch_old)
353 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
354 #endif
355 libc_hidden_ver (__fnmatch, fnmatch)
356