]>
Commit | Line | Data |
---|---|---|
dff8da6b | 1 | /* Copyright (C) 1991-2024 Free Software Foundation, Inc. |
c84142e8 | 2 | This file is part of the GNU C Library. |
28f540f4 | 3 | |
41bdb6e2 AJ |
4 | The GNU C Library is free software; you can redistribute it and/or |
5 | modify it under the terms of the GNU Lesser General Public | |
6 | License as published by the Free Software Foundation; either | |
7 | version 2.1 of the License, or (at your option) any later version. | |
28f540f4 | 8 | |
41bdb6e2 | 9 | The GNU C Library is distributed in the hope that it will be useful, |
c84142e8 UD |
10 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
11 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
41bdb6e2 | 12 | Lesser General Public License for more details. |
28f540f4 | 13 | |
41bdb6e2 | 14 | You should have received a copy of the GNU Lesser General Public |
59ba27a6 | 15 | License along with the GNU C Library; if not, see |
5a82c748 | 16 | <https://www.gnu.org/licenses/>. */ |
28f540f4 | 17 | |
25286454 AZ |
18 | #ifndef _LIBC |
19 | # include <libc-config.h> | |
28f540f4 RM |
20 | #endif |
21 | ||
97aa195c RM |
22 | /* Enable GNU extensions in fnmatch.h. */ |
23 | #ifndef _GNU_SOURCE | |
25286454 | 24 | # define _GNU_SOURCE 1 |
97aa195c RM |
25 | #endif |
26 | ||
25286454 AZ |
27 | #include <fnmatch.h> |
28 | ||
1fc82a56 | 29 | #include <assert.h> |
28f540f4 | 30 | #include <errno.h> |
28f540f4 | 31 | #include <ctype.h> |
3f3dd810 | 32 | #include <string.h> |
25286454 | 33 | #include <stdlib.h> |
25286454 AZ |
34 | #include <wchar.h> |
35 | #include <wctype.h> | |
36 | #include <stddef.h> | |
37 | #include <stdbool.h> | |
28f540f4 | 38 | |
acb5ee2e UD |
39 | /* We need some of the locale data (the collation sequence information) |
40 | but there is no interface to get this information in general. Therefore | |
41 | we support a correct implementation only in glibc. */ | |
42 | #ifdef _LIBC | |
43 | # include "../locale/localeinfo.h" | |
04ea3b0f | 44 | # include "../locale/coll-lookup.h" |
821a6bb4 | 45 | # include <shlib-compat.h> |
acb5ee2e UD |
46 | |
47 | # define CONCAT(a,b) __CONCAT(a,b) | |
25286454 AZ |
48 | # define btowc __btowc |
49 | # define iswctype __iswctype | |
4aebaa6b | 50 | # define mbsrtowcs __mbsrtowcs |
25286454 AZ |
51 | # define mempcpy __mempcpy |
52 | # define strnlen __strnlen | |
53 | # define towlower __towlower | |
54 | # define wcscat __wcscat | |
55 | # define wcslen __wcslen | |
56 | # define wctype __wctype | |
57 | # define wmemchr __wmemchr | |
58 | # define wmempcpy __wmempcpy | |
821a6bb4 UD |
59 | # define fnmatch __fnmatch |
60 | extern int fnmatch (const char *pattern, const char *string, int flags); | |
acb5ee2e UD |
61 | #endif |
62 | ||
25286454 | 63 | #ifdef _LIBC |
d7ee6bd8 | 64 | # if __GNUC__ >= 7 |
25286454 AZ |
65 | # define FALLTHROUGH __attribute__ ((__fallthrough__)) |
66 | # else | |
67 | # define FALLTHROUGH ((void) 0) | |
68 | # endif | |
69 | #else | |
70 | # include "attribute.h" | |
71 | #endif | |
28f540f4 | 72 | |
25286454 AZ |
73 | #include <intprops.h> |
74 | #include <flexmember.h> | |
a79328c7 | 75 | #include <scratch_buffer.h> |
28f540f4 | 76 | |
25286454 AZ |
77 | #ifdef _LIBC |
78 | typedef ptrdiff_t idx_t; | |
79 | #else | |
80 | # include "idx.h" | |
81 | #endif | |
28f540f4 | 82 | |
25286454 AZ |
83 | /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */ |
84 | #define NO_LEADING_PERIOD(flags) \ | |
85 | ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD)) | |
ba1ffaa1 | 86 | |
25286454 AZ |
87 | /* Provide support for user-defined character classes, based on the functions |
88 | from ISO C 90 amendment 1. */ | |
89 | #ifdef CHARCLASS_NAME_MAX | |
90 | # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX | |
91 | #else | |
a9ddb793 UD |
92 | /* This shouldn't happen but some implementation might still have this |
93 | problem. Use a reasonable default value. */ | |
25286454 AZ |
94 | # define CHAR_CLASS_MAX_LENGTH 256 |
95 | #endif | |
1fc82a56 | 96 | |
25286454 | 97 | #define IS_CHAR_CLASS(string) wctype (string) |
a9ddb793 UD |
98 | |
99 | /* Avoid depending on library functions or files | |
100 | whose names are inconsistent. */ | |
ba1ffaa1 | 101 | |
955994e1 UD |
102 | /* Global variable. */ |
103 | static int posixly_correct; | |
104 | ||
6d52618b | 105 | /* Note that this evaluates C many times. */ |
25286454 AZ |
106 | #define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c)) |
107 | #define CHAR char | |
108 | #define UCHAR unsigned char | |
109 | #define INT int | |
110 | #define FCT internal_fnmatch | |
111 | #define EXT ext_match | |
112 | #define END end_pattern | |
113 | #define STRUCT fnmatch_struct | |
114 | #define L_(CS) CS | |
115 | #define BTOWC(C) btowc (C) | |
116 | #define STRLEN(S) strlen (S) | |
117 | #define STRCAT(D, S) strcat (D, S) | |
118 | #define MEMPCPY(D, S, N) mempcpy (D, S, N) | |
119 | #define MEMCHR(S, C, N) memchr (S, C, N) | |
120 | #define WIDE_CHAR_VERSION 0 | |
121 | #ifdef _LIBC | |
8c0ab919 RM |
122 | # include <locale/weight.h> |
123 | # define FINDIDX findidx | |
25286454 AZ |
124 | #endif |
125 | #include "fnmatch_loop.c" | |
126 | ||
127 | ||
128 | #define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c)) | |
129 | #define CHAR wchar_t | |
130 | #define UCHAR wint_t | |
131 | #define INT wint_t | |
132 | #define FCT internal_fnwmatch | |
133 | #define EXT ext_wmatch | |
134 | #define END end_wpattern | |
135 | #define L_(CS) L##CS | |
136 | #define BTOWC(C) (C) | |
137 | #define STRLEN(S) wcslen (S) | |
138 | #define STRCAT(D, S) wcscat (D, S) | |
139 | #define MEMPCPY(D, S, N) wmempcpy (D, S, N) | |
140 | #define MEMCHR(S, C, N) wmemchr (S, C, N) | |
141 | #define WIDE_CHAR_VERSION 1 | |
142 | #ifdef _LIBC | |
8c0ab919 RM |
143 | /* Change the name the header defines so it doesn't conflict with |
144 | the <locale/weight.h> version included above. */ | |
25286454 AZ |
145 | # define findidx findidxwc |
146 | # include <locale/weightwc.h> | |
147 | # undef findidx | |
148 | # define FINDIDX findidxwc | |
149 | #endif | |
acb5ee2e | 150 | |
25286454 | 151 | #undef IS_CHAR_CLASS |
1fc82a56 | 152 | /* We have to convert the wide character string in a multibyte string. But |
5e463393 UD |
153 | we know that the character class names consist of alphanumeric characters |
154 | from the portable character set, and since the wide character encoding | |
155 | for a member of the portable character set is the same code point as | |
156 | its single-byte encoding, we can use a simplified method to convert the | |
157 | string to a multibyte character string. */ | |
1fc82a56 UD |
158 | static wctype_t |
159 | is_char_class (const wchar_t *wcs) | |
160 | { | |
161 | char s[CHAR_CLASS_MAX_LENGTH + 1]; | |
162 | char *cp = s; | |
28f540f4 | 163 | |
1fc82a56 UD |
164 | do |
165 | { | |
5e463393 | 166 | /* Test for a printable character from the portable character set. */ |
25286454 | 167 | #ifdef _LIBC |
5e463393 | 168 | if (*wcs < 0x20 || *wcs > 0x7e |
25286454 AZ |
169 | || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60) |
170 | return (wctype_t) 0; | |
171 | #else | |
5e463393 | 172 | switch (*wcs) |
25286454 AZ |
173 | { |
174 | case L' ': case L'!': case L'"': case L'#': case L'%': | |
175 | case L'&': case L'\'': case L'(': case L')': case L'*': | |
176 | case L'+': case L',': case L'-': case L'.': case L'/': | |
177 | case L'0': case L'1': case L'2': case L'3': case L'4': | |
178 | case L'5': case L'6': case L'7': case L'8': case L'9': | |
179 | case L':': case L';': case L'<': case L'=': case L'>': | |
180 | case L'?': | |
181 | case L'A': case L'B': case L'C': case L'D': case L'E': | |
182 | case L'F': case L'G': case L'H': case L'I': case L'J': | |
183 | case L'K': case L'L': case L'M': case L'N': case L'O': | |
184 | case L'P': case L'Q': case L'R': case L'S': case L'T': | |
185 | case L'U': case L'V': case L'W': case L'X': case L'Y': | |
186 | case L'Z': | |
187 | case L'[': case L'\\': case L']': case L'^': case L'_': | |
188 | case L'a': case L'b': case L'c': case L'd': case L'e': | |
189 | case L'f': case L'g': case L'h': case L'i': case L'j': | |
190 | case L'k': case L'l': case L'm': case L'n': case L'o': | |
191 | case L'p': case L'q': case L'r': case L's': case L't': | |
192 | case L'u': case L'v': case L'w': case L'x': case L'y': | |
193 | case L'z': case L'{': case L'|': case L'}': case L'~': | |
194 | break; | |
195 | default: | |
196 | return (wctype_t) 0; | |
197 | } | |
198 | #endif | |
1fc82a56 | 199 | |
5e463393 UD |
200 | /* Avoid overrunning the buffer. */ |
201 | if (cp == s + CHAR_CLASS_MAX_LENGTH) | |
25286454 | 202 | return (wctype_t) 0; |
28f540f4 | 203 | |
5e463393 UD |
204 | *cp++ = (char) *wcs++; |
205 | } | |
206 | while (*wcs != L'\0'); | |
28f540f4 | 207 | |
5e463393 | 208 | *cp = '\0'; |
ba1ffaa1 | 209 | |
ea6eb383 | 210 | return wctype (s); |
5e463393 | 211 | } |
25286454 | 212 | #define IS_CHAR_CLASS(string) is_char_class (string) |
28f540f4 | 213 | |
25286454 | 214 | #include "fnmatch_loop.c" |
d8aaef00 | 215 | |
a79328c7 AZ |
216 | static int |
217 | fnmatch_convert_to_wide (const char *str, struct scratch_buffer *buf, | |
218 | size_t *n) | |
219 | { | |
220 | mbstate_t ps; | |
221 | memset (&ps, '\0', sizeof (ps)); | |
222 | ||
223 | size_t nw = buf->length / sizeof (wchar_t); | |
224 | *n = strnlen (str, nw - 1); | |
225 | if (__glibc_likely (*n < nw)) | |
226 | { | |
227 | const char *p = str; | |
228 | *n = mbsrtowcs (buf->data, &p, *n + 1, &ps); | |
229 | if (__glibc_unlikely (*n == (size_t) -1)) | |
230 | /* Something wrong. | |
231 | XXX Do we have to set 'errno' to something which mbsrtows hasn't | |
232 | already done? */ | |
233 | return -1; | |
234 | if (p == NULL) | |
235 | return 0; | |
236 | memset (&ps, '\0', sizeof (ps)); | |
237 | } | |
238 | ||
239 | *n = mbsrtowcs (NULL, &str, 0, &ps); | |
240 | if (__glibc_unlikely (*n == (size_t) -1)) | |
241 | return -1; | |
242 | if (!scratch_buffer_set_array_size (buf, *n + 1, sizeof (wchar_t))) | |
243 | { | |
244 | __set_errno (ENOMEM); | |
245 | return -2; | |
246 | } | |
247 | assert (mbsinit (&ps)); | |
248 | mbsrtowcs (buf->data, &str, *n + 1, &ps); | |
249 | return 0; | |
250 | } | |
5e463393 | 251 | |
d8aaef00 | 252 | int |
9dd346ff | 253 | fnmatch (const char *pattern, const char *string, int flags) |
d8aaef00 | 254 | { |
25286454 | 255 | if (__glibc_unlikely (MB_CUR_MAX != 1)) |
955994e1 | 256 | { |
955994e1 | 257 | size_t n; |
a79328c7 AZ |
258 | struct scratch_buffer wpattern; |
259 | scratch_buffer_init (&wpattern); | |
260 | struct scratch_buffer wstring; | |
261 | scratch_buffer_init (&wstring); | |
262 | int r; | |
263 | ||
264 | /* Convert the strings into wide characters. Any conversion issue | |
265 | fallback to the ascii version. */ | |
266 | r = fnmatch_convert_to_wide (pattern, &wpattern, &n); | |
267 | if (r == 0) | |
25286454 | 268 | { |
a79328c7 AZ |
269 | r = fnmatch_convert_to_wide (string, &wstring, &n); |
270 | if (r == 0) | |
271 | r = internal_fnwmatch (wpattern.data, wstring.data, | |
272 | (wchar_t *) wstring.data + n, | |
4e32c8f5 | 273 | flags & FNM_PERIOD, flags, NULL); |
a79328c7 | 274 | } |
f15ce4d8 | 275 | |
a79328c7 AZ |
276 | scratch_buffer_free (&wstring); |
277 | scratch_buffer_free (&wpattern); | |
f15ce4d8 | 278 | |
a79328c7 AZ |
279 | if (r == -2 || r == 0) |
280 | return r; | |
955994e1 | 281 | } |
955994e1 UD |
282 | |
283 | return internal_fnmatch (pattern, string, string + strlen (string), | |
4e32c8f5 | 284 | flags & FNM_PERIOD, flags, NULL); |
d8aaef00 UD |
285 | } |
286 | ||
25286454 | 287 | #undef fnmatch |
821a6bb4 | 288 | versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3); |
25286454 | 289 | #if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3) |
821a6bb4 UD |
290 | strong_alias (__fnmatch, __fnmatch_old) |
291 | compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0); | |
25286454 | 292 | #endif |
a14f26ef | 293 | libc_hidden_ver (__fnmatch, fnmatch) |