Adapt collation in th_TH locale to use the iso14651_t1_common file and sync the colla...
[glibc.git] / posix / fnmatch.c
blobd773b2d91e3e27f9c23201735dcfa8ed245e0142
1 /* Copyright (C) 1991-2023 Free Software Foundation, Inc.
2 This file is part of the GNU C Library.
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.
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 Lesser General Public License for more details.
14 You should have received a copy of the GNU Lesser General Public
15 License along with the GNU C Library; if not, see
16 <https://www.gnu.org/licenses/>. */
18 #ifndef _LIBC
19 # include <libc-config.h>
20 #endif
22 /* Enable GNU extensions in fnmatch.h. */
23 #ifndef _GNU_SOURCE
24 # define _GNU_SOURCE 1
25 #endif
27 #include <fnmatch.h>
29 #include <assert.h>
30 #include <errno.h>
31 #include <ctype.h>
32 #include <string.h>
33 #include <stdlib.h>
34 #include <wchar.h>
35 #include <wctype.h>
36 #include <stddef.h>
37 #include <stdbool.h>
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"
44 # include "../locale/coll-lookup.h"
45 # include <shlib-compat.h>
47 # define CONCAT(a,b) __CONCAT(a,b)
48 # define btowc __btowc
49 # define iswctype __iswctype
50 # define mbsrtowcs __mbsrtowcs
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
59 # define fnmatch __fnmatch
60 extern int fnmatch (const char *pattern, const char *string, int flags);
61 #endif
63 #ifdef _LIBC
64 # if __GNUC__ >= 7
65 # define FALLTHROUGH __attribute__ ((__fallthrough__))
66 # else
67 # define FALLTHROUGH ((void) 0)
68 # endif
69 #else
70 # include "attribute.h"
71 #endif
73 #include <intprops.h>
74 #include <flexmember.h>
75 #include <scratch_buffer.h>
77 #ifdef _LIBC
78 typedef ptrdiff_t idx_t;
79 #else
80 # include "idx.h"
81 #endif
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))
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
92 /* This shouldn't happen but some implementation might still have this
93 problem. Use a reasonable default value. */
94 # define CHAR_CLASS_MAX_LENGTH 256
95 #endif
97 #define IS_CHAR_CLASS(string) wctype (string)
99 /* Avoid depending on library functions or files
100 whose names are inconsistent. */
102 /* Global variable. */
103 static int posixly_correct;
105 /* Note that this evaluates C many times. */
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
122 # include <locale/weight.h>
123 # define FINDIDX findidx
124 #endif
125 #include "fnmatch_loop.c"
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
143 /* Change the name the header defines so it doesn't conflict with
144 the <locale/weight.h> version included above. */
145 # define findidx findidxwc
146 # include <locale/weightwc.h>
147 # undef findidx
148 # define FINDIDX findidxwc
149 #endif
151 #undef IS_CHAR_CLASS
152 /* We have to convert the wide character string in a multibyte string. But
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. */
158 static wctype_t
159 is_char_class (const wchar_t *wcs)
161 char s[CHAR_CLASS_MAX_LENGTH + 1];
162 char *cp = s;
166 /* Test for a printable character from the portable character set. */
167 #ifdef _LIBC
168 if (*wcs < 0x20 || *wcs > 0x7e
169 || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
170 return (wctype_t) 0;
171 #else
172 switch (*wcs)
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;
198 #endif
200 /* Avoid overrunning the buffer. */
201 if (cp == s + CHAR_CLASS_MAX_LENGTH)
202 return (wctype_t) 0;
204 *cp++ = (char) *wcs++;
206 while (*wcs != L'\0');
208 *cp = '\0';
210 return wctype (s);
212 #define IS_CHAR_CLASS(string) is_char_class (string)
214 #include "fnmatch_loop.c"
216 static int
217 fnmatch_convert_to_wide (const char *str, struct scratch_buffer *buf,
218 size_t *n)
220 mbstate_t ps;
221 memset (&ps, '\0', sizeof (ps));
223 size_t nw = buf->length / sizeof (wchar_t);
224 *n = strnlen (str, nw - 1);
225 if (__glibc_likely (*n < nw))
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));
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)))
244 __set_errno (ENOMEM);
245 return -2;
247 assert (mbsinit (&ps));
248 mbsrtowcs (buf->data, &str, *n + 1, &ps);
249 return 0;
253 fnmatch (const char *pattern, const char *string, int flags)
255 if (__glibc_unlikely (MB_CUR_MAX != 1))
257 size_t n;
258 struct scratch_buffer wpattern;
259 scratch_buffer_init (&wpattern);
260 struct scratch_buffer wstring;
261 scratch_buffer_init (&wstring);
262 int r;
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)
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,
273 flags & FNM_PERIOD, flags, NULL);
276 scratch_buffer_free (&wstring);
277 scratch_buffer_free (&wpattern);
279 if (r == -2 || r == 0)
280 return r;
283 return internal_fnmatch (pattern, string, string + strlen (string),
284 flags & FNM_PERIOD, flags, NULL);
287 #undef fnmatch
288 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
289 #if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
290 strong_alias (__fnmatch, __fnmatch_old)
291 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
292 #endif
293 libc_hidden_ver (__fnmatch, fnmatch)