1 /* Copyright (C) 1991-1993,1996-1999,2000,2001 Free Software Foundation, Inc.
2 This file is part of the GNU C Library.
4 This library is free software; you can redistribute it and/or
5 modify it under the terms of the GNU Library General Public License as
6 published by the Free Software Foundation; either version 2 of the
7 License, or (at your option) any later version.
9 This 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 Library General Public License for more details.
14 You should have received a copy of the GNU Library General Public
15 License along with this library; see the file COPYING.LIB. If not,
16 write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
17 Boston, MA 02111-1307, USA. */
23 /* Enable GNU extensions in fnmatch.h. */
25 # define _GNU_SOURCE 1
33 #if HAVE_STRING_H || defined _LIBC
39 #if defined STDC_HEADERS || defined _LIBC
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/elem-hash.h"
57 # include "../locale/coll-lookup.h"
59 # define CONCAT(a,b) __CONCAT(a,b)
60 # define mbsinit __mbsinit
61 # define mbsrtowcs __mbsrtowcs
64 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */
65 #define NO_LEADING_PERIOD(flags) \
66 ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
68 /* Comment out all this code if we are using the GNU C Library, and are not
69 actually compiling the library itself. This code is part of the GNU C
70 Library, but also included in many other GNU distributions. Compiling
71 and linking in this code is a waste when using the GNU C library
72 (especially if it is a shared library). Rather than having every GNU
73 program understand `configure --with-gnu-libc' and omit the object files,
74 it is simpler to just do this in the source for each such file. */
76 #if defined _LIBC || !defined __GNU_LIBRARY__
79 # if defined STDC_HEADERS || !defined isascii
82 # define ISASCII(c) isascii(c)
86 # define ISBLANK(c) (ISASCII (c) && isblank (c))
88 # define ISBLANK(c) ((c) == ' ' || (c) == '\t')
91 # define ISGRAPH(c) (ISASCII (c) && isgraph (c))
93 # define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
96 # define ISPRINT(c) (ISASCII (c) && isprint (c))
97 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
98 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
99 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
100 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
101 # define ISLOWER(c) (ISASCII (c) && islower (c))
102 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
103 # define ISSPACE(c) (ISASCII (c) && isspace (c))
104 # define ISUPPER(c) (ISASCII (c) && isupper (c))
105 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
107 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
109 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
110 /* The GNU C library provides support for user-defined character classes
111 and the functions from ISO C amendement 1. */
112 # ifdef CHARCLASS_NAME_MAX
113 # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
115 /* This shouldn't happen but some implementation might still have this
116 problem. Use a reasonable default value. */
117 # define CHAR_CLASS_MAX_LENGTH 256
121 # define IS_CHAR_CLASS(string) __wctype (string)
123 # define IS_CHAR_CLASS(string) wctype (string)
127 # define ISWCTYPE(WC, WT) __iswctype (WC, WT)
129 # define ISWCTYPE(WC, WT) iswctype (WC, WT)
132 # if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
133 /* In this case we are implementing the multibyte character handling. */
134 # define HANDLE_MULTIBYTE 1
138 # define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
140 # define IS_CHAR_CLASS(string) \
141 (STREQ (string, "alpha") || STREQ (string, "upper") \
142 || STREQ (string, "lower") || STREQ (string, "digit") \
143 || STREQ (string, "alnum") || STREQ (string, "xdigit") \
144 || STREQ (string, "space") || STREQ (string, "print") \
145 || STREQ (string, "punct") || STREQ (string, "graph") \
146 || STREQ (string, "cntrl") || STREQ (string, "blank"))
149 /* Avoid depending on library functions or files
150 whose names are inconsistent. */
152 # if !defined _LIBC && !defined getenv
153 extern char *getenv ();
160 /* Global variable. */
161 static int posixly_correct;
163 /* This function doesn't exist on most systems. */
165 # if !defined HAVE___STRCHRNUL && !defined _LIBC
171 char *result = strchr (s, c);
173 result = strchr (s, '\0');
178 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
184 wchar_t *result = wcschr (s, c);
186 result = wcschr (s, '\0');
191 # ifndef internal_function
192 /* Inside GNU libc we mark some function in a special way. In other
193 environments simply ignore the marking. */
194 # define internal_function
197 /* Note that this evaluates C many times. */
199 # define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
201 # define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
204 # define UCHAR unsigned char
206 # define FCT internal_fnmatch
207 # define EXT ext_match
208 # define END end_pattern
211 # define BTOWC(C) __btowc (C)
213 # define BTOWC(C) btowc (C)
215 # define MEMPCPY(D, S, N) __mempcpy (D, S, N)
216 # define MEMCHR(S, C, N) memchr (S, C, N)
217 # define STRCOLL(S1, S2) strcoll (S1, S2)
218 # include "fnmatch_loop.c"
221 # if HANDLE_MULTIBYTE
222 /* Note that this evaluates C many times. */
224 # define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
226 # define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
228 # define CHAR wchar_t
229 # define UCHAR wint_t
231 # define FCT internal_fnwmatch
232 # define EXT ext_wmatch
233 # define END end_wpattern
235 # define BTOWC(C) (C)
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)
239 # define WIDE_CHAR_VERSION 1
241 # undef IS_CHAR_CLASS
242 /* We have to convert the wide character string in a multibyte string. But
243 we know that the character class names consist of alphanumeric characters
244 from the portable character set, and since the wide character encoding
245 for a member of the portable character set is the same code point as
246 its single-byte encoding, we can use a simplified method to convert the
247 string to a multibyte character string. */
249 is_char_class (const wchar_t *wcs)
251 char s[CHAR_CLASS_MAX_LENGTH + 1];
256 /* Test for a printable character from the portable character set. */
258 if (*wcs < 0x20 || *wcs > 0x7e
259 || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
264 case L' ': case L'!': case L'"': case L'#': case L'%':
265 case L'&': case L'\'': case L'(': case L')': case L'*':
266 case L'+': case L',': case L'-': case L'.': case L'/':
267 case L'0': case L'1': case L'2': case L'3': case L'4':
268 case L'5': case L'6': case L'7': case L'8': case L'9':
269 case L':': case L';': case L'<': case L'=': case L'>':
271 case L'A': case L'B': case L'C': case L'D': case L'E':
272 case L'F': case L'G': case L'H': case L'I': case L'J':
273 case L'K': case L'L': case L'M': case L'N': case L'O':
274 case L'P': case L'Q': case L'R': case L'S': case L'T':
275 case L'U': case L'V': case L'W': case L'X': case L'Y':
277 case L'[': case L'\\': case L']': case L'^': case L'_':
278 case L'a': case L'b': case L'c': case L'd': case L'e':
279 case L'f': case L'g': case L'h': case L'i': case L'j':
280 case L'k': case L'l': case L'm': case L'n': case L'o':
281 case L'p': case L'q': case L'r': case L's': case L't':
282 case L'u': case L'v': case L'w': case L'x': case L'y':
283 case L'z': case L'{': case L'|': case L'}': case L'~':
290 /* Avoid overrunning the buffer. */
291 if (cp == s + CHAR_CLASS_MAX_LENGTH)
294 *cp++ = (char) *wcs++;
296 while (*wcs != L'\0');
306 # define IS_CHAR_CLASS(string) is_char_class (string)
308 # include "fnmatch_loop.c"
313 fnmatch (pattern, string, flags)
318 # if HANDLE_MULTIBYTE
319 if (__builtin_expect (MB_CUR_MAX, 1) != 1)
326 /* Convert the strings into wide characters. */
327 memset (&ps, '\0', sizeof (ps));
328 n = mbsrtowcs (NULL, &pattern, 0, &ps);
329 if (__builtin_expect (n, 0) == (size_t) -1)
331 XXX Do we have to set `errno' to something which mbsrtows hasn't
334 wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
335 assert (mbsinit (&ps));
336 (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
338 assert (mbsinit (&ps));
339 n = mbsrtowcs (NULL, &string, 0, &ps);
340 if (__builtin_expect (n, 0) == (size_t) -1)
342 XXX Do we have to set `errno' to something which mbsrtows hasn't
345 wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
346 assert (mbsinit (&ps));
347 (void) mbsrtowcs (wstring, &string, n + 1, &ps);
349 return internal_fnwmatch (wpattern, wstring, wstring + n,
350 flags & FNM_PERIOD, flags);
352 # endif /* mbstate_t and mbsrtowcs or _LIBC. */
354 return internal_fnmatch (pattern, string, string + strlen (string),
355 flags & FNM_PERIOD, flags);
358 #endif /* _LIBC or not __GNU_LIBRARY__. */