1 /* Copyright (C) 1991-2020 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/>. */
22 /* Enable GNU extensions in fnmatch.h. */
24 # define _GNU_SOURCE 1
33 #if defined STDC_HEADERS || defined _LIBC
40 # define alloca_account(size., var) alloca (size)
43 /* For platform which support the ISO C amendement 1 functionality we
44 support user defined character classes. */
45 #if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
46 /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */
51 /* We need some of the locale data (the collation sequence information)
52 but there is no interface to get this information in general. Therefore
53 we support a correct implementation only in glibc. */
55 # include "../locale/localeinfo.h"
56 # include "../locale/coll-lookup.h"
57 # include <shlib-compat.h>
59 # define CONCAT(a,b) __CONCAT(a,b)
60 # define mbsrtowcs __mbsrtowcs
61 # define fnmatch __fnmatch
62 extern int fnmatch (const char *pattern
, const char *string
, int flags
);
65 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */
66 #define NO_LEADING_PERIOD(flags) \
67 ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
69 /* Comment out all this code if we are using the GNU C Library, and are not
70 actually compiling the library itself. This code is part of the GNU C
71 Library, but also included in many other GNU distributions. Compiling
72 and linking in this code is a waste when using the GNU C library
73 (especially if it is a shared library). Rather than having every GNU
74 program understand `configure --with-gnu-libc' and omit the object files,
75 it is simpler to just do this in the source for each such file. */
77 #if defined _LIBC || !defined __GNU_LIBRARY__
80 # if defined STDC_HEADERS || !defined isascii
83 # define ISASCII(c) isascii(c)
87 # define ISBLANK(c) (ISASCII (c) && isblank (c))
89 # define ISBLANK(c) ((c) == ' ' || (c) == '\t')
92 # define ISGRAPH(c) (ISASCII (c) && isgraph (c))
94 # define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
97 # define ISPRINT(c) (ISASCII (c) && isprint (c))
98 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
99 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
100 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
101 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
102 # define ISLOWER(c) (ISASCII (c) && islower (c))
103 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
104 # define ISSPACE(c) (ISASCII (c) && isspace (c))
105 # define ISUPPER(c) (ISASCII (c) && isupper (c))
106 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
108 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
110 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
111 /* The GNU C library provides support for user-defined character classes
112 and the functions from ISO C amendement 1. */
113 # ifdef CHARCLASS_NAME_MAX
114 # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
116 /* This shouldn't happen but some implementation might still have this
117 problem. Use a reasonable default value. */
118 # define CHAR_CLASS_MAX_LENGTH 256
122 # define IS_CHAR_CLASS(string) __wctype (string)
124 # define IS_CHAR_CLASS(string) wctype (string)
128 # define ISWCTYPE(WC, WT) __iswctype (WC, WT)
130 # define ISWCTYPE(WC, WT) iswctype (WC, WT)
133 # if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
134 /* In this case we are implementing the multibyte character handling. */
135 # define HANDLE_MULTIBYTE 1
139 # define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
141 # define IS_CHAR_CLASS(string) \
142 (STREQ (string, "alpha") || STREQ (string, "upper") \
143 || STREQ (string, "lower") || STREQ (string, "digit") \
144 || STREQ (string, "alnum") || STREQ (string, "xdigit") \
145 || STREQ (string, "space") || STREQ (string, "print") \
146 || STREQ (string, "punct") || STREQ (string, "graph") \
147 || STREQ (string, "cntrl") || STREQ (string, "blank"))
150 /* Avoid depending on library functions or files
151 whose names are inconsistent. */
153 # if !defined _LIBC && !defined getenv
154 extern char *getenv ();
161 /* Global variable. */
162 static int posixly_correct
;
164 /* This function doesn't exist on most systems. */
166 # if !defined HAVE___STRCHRNUL && !defined _LIBC
168 __strchrnul (const char *s
, int c
)
170 char *result
= strchr (s
, c
);
172 result
= strchr (s
, '\0');
177 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
179 __wcschrnul (const wchar_t *s
, wint_t c
)
181 wchar_t *result
= wcschr (s
, c
);
183 result
= wcschr (s
, '\0');
188 /* Note that this evaluates C many times. */
190 # define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
192 # define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
195 # define UCHAR unsigned char
197 # define FCT internal_fnmatch
198 # define EXT ext_match
199 # define END end_pattern
200 # define STRUCT fnmatch_struct
203 # define BTOWC(C) __btowc (C)
205 # define BTOWC(C) btowc (C)
207 # define STRLEN(S) strlen (S)
208 # define STRCAT(D, S) strcat (D, S)
209 # define MEMPCPY(D, S, N) __mempcpy (D, S, N)
210 # define MEMCHR(S, C, N) memchr (S, C, N)
211 # define STRCOLL(S1, S2) strcoll (S1, S2)
212 # define WIDE_CHAR_VERSION 0
213 # include <locale/weight.h>
214 # define FINDIDX findidx
215 # include "fnmatch_loop.c"
218 # if HANDLE_MULTIBYTE
219 /* Note that this evaluates C many times. */
221 # define FOLD(c) ((flags & FNM_CASEFOLD) ? __towlower (c) : (c))
223 # define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
225 # define CHAR wchar_t
226 # define UCHAR wint_t
228 # define FCT internal_fnwmatch
229 # define EXT ext_wmatch
230 # define END end_wpattern
231 # define STRUCT fnwmatch_struct
233 # define BTOWC(C) (C)
234 # define STRLEN(S) __wcslen (S)
235 # define STRCAT(D, S) __wcscat (D, S)
236 # define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
237 # define MEMCHR(S, C, N) __wmemchr (S, C, N)
238 # define STRCOLL(S1, S2) wcscoll (S1, S2)
240 # define WMEMCMP(S1, S2, N) __wmemcmp (S1, S2, N)
242 # define WMEMCMP(S1, S2, N) wmemcmp (S1, S2, N)
244 # define WIDE_CHAR_VERSION 1
245 /* Change the name the header defines so it doesn't conflict with
246 the <locale/weight.h> version included above. */
247 # define findidx findidxwc
248 # include <locale/weightwc.h>
250 # define FINDIDX findidxwc
252 # undef IS_CHAR_CLASS
253 /* We have to convert the wide character string in a multibyte string. But
254 we know that the character class names consist of alphanumeric characters
255 from the portable character set, and since the wide character encoding
256 for a member of the portable character set is the same code point as
257 its single-byte encoding, we can use a simplified method to convert the
258 string to a multibyte character string. */
260 is_char_class (const wchar_t *wcs
)
262 char s
[CHAR_CLASS_MAX_LENGTH
+ 1];
267 /* Test for a printable character from the portable character set. */
269 if (*wcs
< 0x20 || *wcs
> 0x7e
270 || *wcs
== 0x24 || *wcs
== 0x40 || *wcs
== 0x60)
275 case L
' ': case L
'!': case L
'"': case L
'#': case L
'%':
276 case L
'&': case L
'\'': case L
'(': case L
')': case L
'*':
277 case L
'+': case L
',': case L
'-': case L
'.': case L
'/':
278 case L
'0': case L
'1': case L
'2': case L
'3': case L
'4':
279 case L
'5': case L
'6': case L
'7': case L
'8': case L
'9':
280 case L
':': case L
';': case L
'<': case L
'=': case L
'>':
282 case L
'A': case L
'B': case L
'C': case L
'D': case L
'E':
283 case L
'F': case L
'G': case L
'H': case L
'I': case L
'J':
284 case L
'K': case L
'L': case L
'M': case L
'N': case L
'O':
285 case L
'P': case L
'Q': case L
'R': case L
'S': case L
'T':
286 case L
'U': case L
'V': case L
'W': case L
'X': case L
'Y':
288 case L
'[': case L
'\\': case L
']': case L
'^': case L
'_':
289 case L
'a': case L
'b': case L
'c': case L
'd': case L
'e':
290 case L
'f': case L
'g': case L
'h': case L
'i': case L
'j':
291 case L
'k': case L
'l': case L
'm': case L
'n': case L
'o':
292 case L
'p': case L
'q': case L
'r': case L
's': case L
't':
293 case L
'u': case L
'v': case L
'w': case L
'x': case L
'y':
294 case L
'z': case L
'{': case L
'|': case L
'}': case L
'~':
301 /* Avoid overrunning the buffer. */
302 if (cp
== s
+ CHAR_CLASS_MAX_LENGTH
)
305 *cp
++ = (char) *wcs
++;
307 while (*wcs
!= L
'\0');
317 # define IS_CHAR_CLASS(string) is_char_class (string)
319 # include "fnmatch_loop.c"
324 fnmatch (const char *pattern
, const char *string
, int flags
)
326 # if HANDLE_MULTIBYTE
327 if (__builtin_expect (MB_CUR_MAX
, 1) != 1)
332 wchar_t *wpattern_malloc
= NULL
;
334 wchar_t *wstring_malloc
= NULL
;
336 size_t alloca_used
= 0;
338 /* Convert the strings into wide characters. */
339 memset (&ps
, '\0', sizeof (ps
));
342 n
= __strnlen (pattern
, 1024);
344 n
= strlen (pattern
);
346 if (__glibc_likely (n
< 1024))
348 wpattern
= (wchar_t *) alloca_account ((n
+ 1) * sizeof (wchar_t),
350 n
= mbsrtowcs (wpattern
, &p
, n
+ 1, &ps
);
351 if (__glibc_unlikely (n
== (size_t) -1))
353 XXX Do we have to set `errno' to something which mbsrtows hasn't
358 memset (&ps
, '\0', sizeof (ps
));
359 goto prepare_wpattern
;
365 n
= mbsrtowcs (NULL
, &pattern
, 0, &ps
);
366 if (__glibc_unlikely (n
== (size_t) -1))
368 XXX Do we have to set `errno' to something which mbsrtows hasn't
371 if (__glibc_unlikely (n
>= (size_t) -1 / sizeof (wchar_t)))
373 __set_errno (ENOMEM
);
376 wpattern_malloc
= wpattern
377 = (wchar_t *) malloc ((n
+ 1) * sizeof (wchar_t));
378 assert (mbsinit (&ps
));
379 if (wpattern
== NULL
)
381 (void) mbsrtowcs (wpattern
, &pattern
, n
+ 1, &ps
);
384 assert (mbsinit (&ps
));
386 n
= __strnlen (string
, 1024);
391 if (__glibc_likely (n
< 1024))
393 wstring
= (wchar_t *) alloca_account ((n
+ 1) * sizeof (wchar_t),
395 n
= mbsrtowcs (wstring
, &p
, n
+ 1, &ps
);
396 if (__glibc_unlikely (n
== (size_t) -1))
399 XXX Do we have to set `errno' to something which
400 mbsrtows hasn't already done? */
402 free (wpattern_malloc
);
407 memset (&ps
, '\0', sizeof (ps
));
408 goto prepare_wstring
;
414 n
= mbsrtowcs (NULL
, &string
, 0, &ps
);
415 if (__glibc_unlikely (n
== (size_t) -1))
417 XXX Do we have to set `errno' to something which mbsrtows hasn't
420 if (__glibc_unlikely (n
>= (size_t) -1 / sizeof (wchar_t)))
422 free (wpattern_malloc
);
423 __set_errno (ENOMEM
);
427 wstring_malloc
= wstring
428 = (wchar_t *) malloc ((n
+ 1) * sizeof (wchar_t));
431 free (wpattern_malloc
);
434 assert (mbsinit (&ps
));
435 (void) mbsrtowcs (wstring
, &string
, n
+ 1, &ps
);
438 int res
= internal_fnwmatch (wpattern
, wstring
, wstring
+ n
,
439 flags
& FNM_PERIOD
, flags
, NULL
,
442 free (wstring_malloc
);
443 free (wpattern_malloc
);
447 # endif /* mbstate_t and mbsrtowcs or _LIBC. */
449 return internal_fnmatch (pattern
, string
, string
+ strlen (string
),
450 flags
& FNM_PERIOD
, flags
, NULL
, 0);
455 versioned_symbol (libc
, __fnmatch
, fnmatch
, GLIBC_2_2_3
);
456 # if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
457 strong_alias (__fnmatch
, __fnmatch_old
)
458 compat_symbol (libc
, __fnmatch_old
, fnmatch
, GLIBC_2_0
);
460 libc_hidden_ver (__fnmatch
, fnmatch
)
463 #endif /* _LIBC or not __GNU_LIBRARY__. */