1 /* Copyright (C) 1991,1992,1993,1996,1997,1998,1999,2000,2001,2002,2003
2 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4
5 The GNU C Library is free software; you can redistribute it and/or
6 modify it under the terms of the GNU Lesser General Public
7 License as published by the Free Software Foundation; either
8 version 2.1 of the License, or (at your option) any later version.
9
10 The GNU C Library is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 Lesser General Public License for more details.
14
15 You should have received a copy of the GNU Lesser General Public
16 License along with the GNU C Library; if not, see
17 <http://www.gnu.org/licenses/>. */
18
19 #ifdef HAVE_CONFIG_H
20 # include <config.h>
21 #endif
22
23 /* unistd.h must be included with _LIBC defined: we need smallint */
24 #include <unistd.h>
25 #include <features.h>
26 #ifdef __UCLIBC__
27 # undef _LIBC
28 # define HAVE_STRING_H 1
29 # define STDC_HEADERS
30 # define HAVE___STRCHRNUL 1
31 # ifdef __UCLIBC_HAS_WCHAR__
32 # define HAVE_WCHAR_H 1
33 # define HAVE_WCTYPE_H 1
34 # ifdef __UCLIBC_HAS_LOCALE__
35 # define HAVE_MBSTATE_T 1
36 # define HAVE_MBSRTOWCS 1
37 # endif
38 # endif
39 #endif
40
41 #include <assert.h>
42 #include <errno.h>
43 #include <fnmatch.h>
44 #include <ctype.h>
45
46 #if HAVE_STRING_H || defined _LIBC
47 # include <string.h>
48 #else
49 # include <strings.h>
50 #endif
51
52 #if defined STDC_HEADERS || defined _LIBC
53 # include <stdlib.h>
54 #endif
55
56 /* For platform which support the ISO C amendement 1 functionality we
57 support user defined character classes. */
58 #if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
59 /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */
60 # include <wchar.h>
61 # include <wctype.h>
62 #endif
63
64 /* We need some of the locale data (the collation sequence information)
65 but there is no interface to get this information in general. Therefore
66 we support a correct implementation only in glibc. */
67 #ifdef _LIBC
68 # include "../locale/localeinfo.h"
69 # include "../locale/elem-hash.h"
70 # include "../locale/coll-lookup.h"
71 # include <shlib-compat.h>
72
73 # define CONCAT(a,b) __CONCAT(a,b)
74 # define mbsrtowcs __mbsrtowcs
75 # define fnmatch __fnmatch
76 extern int fnmatch (const char *pattern, const char *string, int flags);
77 #endif
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 /* Comment out all this code if we are using the GNU C Library, and are not
84 actually compiling the library itself. This code is part of the GNU C
85 Library, but also included in many other GNU distributions. Compiling
86 and linking in this code is a waste when using the GNU C library
87 (especially if it is a shared library). Rather than having every GNU
88 program understand `configure --with-gnu-libc' and omit the object files,
89 it is simpler to just do this in the source for each such file. */
90
91 #if defined _LIBC || !defined __GNU_LIBRARY__
92
93
94 # if defined STDC_HEADERS || !defined isascii
95 # define ISASCII(c) 1
96 # else
97 # define ISASCII(c) isascii(c)
98 # endif
99
100 # ifdef isblank
101 # define ISBLANK(c) (ISASCII (c) && isblank (c))
102 # else
103 # define ISBLANK(c) ((c) == ' ' || (c) == '\t')
104 # endif
105 # ifdef isgraph
106 # define ISGRAPH(c) (ISASCII (c) && isgraph (c))
107 # else
108 # define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
109 # endif
110
111 # define ISPRINT(c) (ISASCII (c) && isprint (c))
112 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
113 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
114 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
115 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
116 # define ISLOWER(c) (ISASCII (c) && islower (c))
117 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
118 # define ISSPACE(c) (ISASCII (c) && isspace (c))
119 # define ISUPPER(c) (ISASCII (c) && isupper (c))
120 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
121
122 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
123
124 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
125 /* The GNU C library provides support for user-defined character classes
126 and the functions from ISO C amendement 1. */
127 # ifdef CHARCLASS_NAME_MAX
128 # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
129 # else
130 /* This shouldn't happen but some implementation might still have this
131 problem. Use a reasonable default value. */
132 # define CHAR_CLASS_MAX_LENGTH 256
133 # endif
134
135 # ifdef _LIBC
136 # define IS_CHAR_CLASS(string) __wctype (string)
137 # else
138 # define IS_CHAR_CLASS(string) wctype (string)
139 # endif
140
141 # ifdef _LIBC
142 # define ISWCTYPE(WC, WT) __iswctype (WC, WT)
143 # else
144 # define ISWCTYPE(WC, WT) iswctype (WC, WT)
145 # endif
146
147 # if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
148 /* In this case we are implementing the multibyte character handling. */
149 # define HANDLE_MULTIBYTE 1
150 # endif
151
152 # else
153 # define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
154
155 # define IS_CHAR_CLASS(string) \
156 (STREQ (string, "alpha") || STREQ (string, "upper") \
157 || STREQ (string, "lower") || STREQ (string, "digit") \
158 || STREQ (string, "alnum") || STREQ (string, "xdigit") \
159 || STREQ (string, "space") || STREQ (string, "print") \
160 || STREQ (string, "punct") || STREQ (string, "graph") \
161 || STREQ (string, "cntrl") || STREQ (string, "blank"))
162 # endif
163
164 /* Avoid depending on library functions or files
165 whose names are inconsistent. */
166
167 # if !defined _LIBC && !defined getenv && !defined __UCLIBC__
168 extern char *getenv ();
169 # endif
170
171 # ifndef errno
172 extern int errno;
173 # endif
174
175 /* Global variable. */
176 static smallint posixly_correct;
177
178 /* This function doesn't exist on most systems. */
179
180 # if !defined HAVE___STRCHRNUL && !defined _LIBC
181 static char *
__strchrnul(s,c)182 __strchrnul (s, c)
183 const char *s;
184 int c;
185 {
186 char *result = strchr (s, c);
187 if (result == NULL)
188 result = strchr (s, '\0');
189 return result;
190 }
191 # endif
192
193 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
194 static wchar_t *
__wcschrnul(s,c)195 __wcschrnul (s, c)
196 const wchar_t *s;
197 wint_t c;
198 {
199 wchar_t *result = wcschr (s, c);
200 if (result == NULL)
201 result = wcschr (s, '\0');
202 return result;
203 }
204 # endif
205
206 # ifndef internal_function
207 /* Inside GNU libc we mark some function in a special way. In other
208 environments simply ignore the marking. */
209 # define internal_function
210 # endif
211
212 /* Note that this evaluates C many times. */
213 # ifdef _LIBC
214 # define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
215 # else
216 # define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
217 # endif
218 # define CHAR char
219 # define UCHAR unsigned char
220 # define INT int
221 # define FCT internal_fnmatch
222 # define EXT ext_match
223 # define END end_pattern
224 # define L(CS) CS
225 # ifdef _LIBC
226 # define BTOWC(C) __btowc (C)
227 # else
228 # define BTOWC(C) btowc (C)
229 # endif
230 # define STRLEN(S) strlen (S)
231 # define STRCAT(D, S) strcat (D, S)
232 # define MEMPCPY(D, S, N) mempcpy (D, S, N)
233 # define MEMCHR(S, C, N) memchr (S, C, N)
234 # define STRCOLL(S1, S2) strcoll (S1, S2)
235 # include "fnmatch_loop.c"
236
237
238 # if HANDLE_MULTIBYTE
239 /* Note that this evaluates C many times. */
240 # ifdef _LIBC
241 # define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
242 # else
243 # define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
244 # endif
245 # define CHAR wchar_t
246 # define UCHAR wint_t
247 # define INT wint_t
248 # define FCT internal_fnwmatch
249 # define EXT ext_wmatch
250 # define END end_wpattern
251 # define L(CS) L##CS
252 # define BTOWC(C) (C)
253 # define STRLEN(S) wcslen (S)
254 # define STRCAT(D, S) wcscat (D, S)
255 # define MEMPCPY(D, S, N) wmempcpy (D, S, N)
256 # define MEMCHR(S, C, N) wmemchr (S, C, N)
257 # define STRCOLL(S1, S2) wcscoll (S1, S2)
258 # ifndef __UCLIBC__
259 # define WIDE_CHAR_VERSION 1
260 # endif
261
262 # undef IS_CHAR_CLASS
263 /* We have to convert the wide character string in a multibyte string. But
264 we know that the character class names consist of alphanumeric characters
265 from the portable character set, and since the wide character encoding
266 for a member of the portable character set is the same code point as
267 its single-byte encoding, we can use a simplified method to convert the
268 string to a multibyte character string. */
269 static wctype_t
is_char_class(const wchar_t * wcs)270 is_char_class (const wchar_t *wcs)
271 {
272 char s[CHAR_CLASS_MAX_LENGTH + 1];
273 char *cp = s;
274
275 do
276 {
277 /* Test for a printable character from the portable character set. */
278 # if defined _LIBC || defined __UCLIBC__
279 if (*wcs < 0x20 || *wcs > 0x7e
280 || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
281 return (wctype_t) 0;
282 # else
283 switch (*wcs)
284 {
285 case L' ': case L'!': case L'"': case L'#': case L'%':
286 case L'&': case L'\'': case L'(': case L')': case L'*':
287 case L'+': case L',': case L'-': case L'.': case L'/':
288 case L'0': case L'1': case L'2': case L'3': case L'4':
289 case L'5': case L'6': case L'7': case L'8': case L'9':
290 case L':': case L';': case L'<': case L'=': case L'>':
291 case L'?':
292 case L'A': case L'B': case L'C': case L'D': case L'E':
293 case L'F': case L'G': case L'H': case L'I': case L'J':
294 case L'K': case L'L': case L'M': case L'N': case L'O':
295 case L'P': case L'Q': case L'R': case L'S': case L'T':
296 case L'U': case L'V': case L'W': case L'X': case L'Y':
297 case L'Z':
298 case L'[': case L'\\': case L']': case L'^': case L'_':
299 case L'a': case L'b': case L'c': case L'd': case L'e':
300 case L'f': case L'g': case L'h': case L'i': case L'j':
301 case L'k': case L'l': case L'm': case L'n': case L'o':
302 case L'p': case L'q': case L'r': case L's': case L't':
303 case L'u': case L'v': case L'w': case L'x': case L'y':
304 case L'z': case L'{': case L'|': case L'}': case L'~':
305 break;
306 default:
307 return (wctype_t) 0;
308 }
309 # endif
310
311 /* Avoid overrunning the buffer. */
312 if (cp == s + CHAR_CLASS_MAX_LENGTH)
313 return (wctype_t) 0;
314
315 *cp++ = (char) *wcs++;
316 }
317 while (*wcs != L'\0');
318
319 *cp = '\0';
320
321 # ifdef _LIBC
322 return __wctype (s);
323 # else
324 return wctype (s);
325 # endif
326 }
327 # define IS_CHAR_CLASS(string) is_char_class (string)
328
329 # include "fnmatch_loop.c"
330 # endif
331
332 int
fnmatch(const char * pattern,const char * string,int flags)333 fnmatch (const char *pattern, const char *string, int flags)
334 {
335 # if HANDLE_MULTIBYTE
336 if (__builtin_expect (MB_CUR_MAX, 1) != 1)
337 {
338 mbstate_t ps;
339 size_t n;
340 const char *p;
341 wchar_t *wpattern = NULL;
342 wchar_t *wstring = NULL;
343
344 /* Convert the strings into wide characters. */
345 memset (&ps, '\0', sizeof (ps));
346 p = pattern;
347 #ifdef _LIBC
348 n = strnlen (pattern, 1024);
349 #else
350 n = strlen (pattern);
351 #endif
352 if (__builtin_expect (n < 1024, 1))
353 {
354 wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
355 n = mbsrtowcs (wpattern, &p, n + 1, &ps);
356 if (__builtin_expect (n == (size_t) -1, 0))
357 /* Something wrong.
358 XXX Do we have to set `errno' to something which mbsrtows hasn't
359 already done? */
360 return -1;
361 if (p)
362 memset (&ps, '\0', sizeof (ps));
363 }
364 if (__builtin_expect (p != NULL, 0))
365 {
366 n = mbsrtowcs (NULL, &pattern, 0, &ps);
367 if (__builtin_expect (n == (size_t) -1, 0))
368 /* Something wrong.
369 XXX Do we have to set `errno' to something which mbsrtows hasn't
370 already done? */
371 return -1;
372 wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
373 assert (mbsinit (&ps));
374 (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
375 }
376
377 assert (mbsinit (&ps));
378 #ifdef _LIBC
379 n = strnlen (string, 1024);
380 #else
381 n = strlen (string);
382 #endif
383 p = string;
384 if (__builtin_expect (n < 1024, 1))
385 {
386 wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
387 n = mbsrtowcs (wstring, &p, n + 1, &ps);
388 if (__builtin_expect (n == (size_t) -1, 0))
389 /* Something wrong.
390 XXX Do we have to set `errno' to something which mbsrtows hasn't
391 already done? */
392 return -1;
393 if (p)
394 memset (&ps, '\0', sizeof (ps));
395 }
396 if (__builtin_expect (p != NULL, 0))
397 {
398 n = mbsrtowcs (NULL, &string, 0, &ps);
399 if (__builtin_expect (n == (size_t) -1, 0))
400 /* Something wrong.
401 XXX Do we have to set `errno' to something which mbsrtows hasn't
402 already done? */
403 return -1;
404 wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
405 assert (mbsinit (&ps));
406 (void) mbsrtowcs (wstring, &string, n + 1, &ps);
407 }
408
409 return internal_fnwmatch (wpattern, wstring, wstring + n,
410 flags & FNM_PERIOD, flags);
411 }
412 # endif /* mbstate_t and mbsrtowcs or _LIBC. */
413
414 return internal_fnmatch (pattern, string, string + strlen (string),
415 flags & FNM_PERIOD, flags);
416 }
417
418 # ifdef _LIBC
419 # undef fnmatch
420 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
421 # if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
422 strong_alias (__fnmatch, __fnmatch_old)
423 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
424 # endif
425 libc_hidden_ver (__fnmatch, fnmatch)
426 # else
427 libc_hidden_def(fnmatch)
428 # endif
429
430 #endif /* _LIBC or not __GNU_LIBRARY__. */
431