Commit | Line | Data |
---|---|---|
5e8754f9 | 1 | /* Copyright (C) 1991-1993, 1996-2007, 2009-2016 Free Software Foundation, Inc. |
8690e634 JK |
2 | |
3 | This program is free software; you can redistribute it and/or modify | |
4 | it under the terms of the GNU General Public License as published by | |
5 | the Free Software Foundation; either version 3, or (at your option) | |
6 | any later version. | |
7 | ||
8 | This program is distributed in the hope that it will be useful, | |
9 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
10 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
11 | GNU General Public License for more details. | |
12 | ||
13 | You should have received a copy of the GNU General Public License | |
5e8754f9 | 14 | along with this program; if not, see <http://www.gnu.org/licenses/>. */ |
8690e634 JK |
15 | |
16 | #ifndef _LIBC | |
17 | # include <config.h> | |
18 | #endif | |
19 | ||
20 | /* Enable GNU extensions in fnmatch.h. */ | |
21 | #ifndef _GNU_SOURCE | |
22 | # define _GNU_SOURCE 1 | |
23 | #endif | |
24 | ||
5e8754f9 SDJ |
25 | #if ! defined __builtin_expect && defined __GNUC__ && __GNUC__ < 3 |
26 | # define __builtin_expect(expr, expected) (expr) | |
27 | #endif | |
28 | ||
8690e634 JK |
29 | #include <fnmatch.h> |
30 | ||
31 | #include <alloca.h> | |
32 | #include <assert.h> | |
33 | #include <ctype.h> | |
34 | #include <errno.h> | |
35 | #include <stddef.h> | |
36 | #include <stdbool.h> | |
37 | #include <stdlib.h> | |
38 | #include <string.h> | |
39 | ||
40 | #define WIDE_CHAR_SUPPORT \ | |
41 | (HAVE_WCTYPE_H && HAVE_BTOWC && HAVE_ISWCTYPE \ | |
42 | && HAVE_WMEMCHR && (HAVE_WMEMCPY || HAVE_WMEMPCPY)) | |
43 | ||
44 | /* For platform which support the ISO C amendment 1 functionality we | |
45 | support user defined character classes. */ | |
46 | #if defined _LIBC || WIDE_CHAR_SUPPORT | |
47 | # include <wctype.h> | |
48 | # include <wchar.h> | |
49 | #endif | |
50 | ||
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. */ | |
54 | #ifdef _LIBC | |
55 | # include "../locale/localeinfo.h" | |
56 | # include "../locale/elem-hash.h" | |
57 | # include "../locale/coll-lookup.h" | |
58 | # include <shlib-compat.h> | |
59 | ||
60 | # define CONCAT(a,b) __CONCAT(a,b) | |
61 | # define mbsrtowcs __mbsrtowcs | |
62 | # define fnmatch __fnmatch | |
63 | extern int fnmatch (const char *pattern, const char *string, int flags); | |
64 | #endif | |
65 | ||
66 | #ifndef SIZE_MAX | |
67 | # define SIZE_MAX ((size_t) -1) | |
68 | #endif | |
69 | ||
49e4877c PA |
70 | #include "flexmember.h" |
71 | ||
8690e634 JK |
72 | /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */ |
73 | #define NO_LEADING_PERIOD(flags) \ | |
74 | ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD)) | |
75 | ||
76 | /* Comment out all this code if we are using the GNU C Library, and are not | |
77 | actually compiling the library itself, and have not detected a bug | |
78 | in the library. This code is part of the GNU C | |
79 | Library, but also included in many other GNU distributions. Compiling | |
80 | and linking in this code is a waste when using the GNU C library | |
81 | (especially if it is a shared library). Rather than having every GNU | |
82 | program understand 'configure --with-gnu-libc' and omit the object files, | |
83 | it is simpler to just do this in the source for each such file. */ | |
84 | ||
85 | #if defined _LIBC || !defined __GNU_LIBRARY__ || !HAVE_FNMATCH_GNU | |
86 | ||
87 | ||
88 | # if ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK)) | |
89 | # define isblank(c) ((c) == ' ' || (c) == '\t') | |
90 | # endif | |
91 | ||
92 | # define STREQ(s1, s2) (strcmp (s1, s2) == 0) | |
93 | ||
94 | # if defined _LIBC || WIDE_CHAR_SUPPORT | |
95 | /* The GNU C library provides support for user-defined character classes | |
96 | and the functions from ISO C amendment 1. */ | |
97 | # ifdef CHARCLASS_NAME_MAX | |
98 | # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX | |
99 | # else | |
100 | /* This shouldn't happen but some implementation might still have this | |
101 | problem. Use a reasonable default value. */ | |
102 | # define CHAR_CLASS_MAX_LENGTH 256 | |
103 | # endif | |
104 | ||
105 | # ifdef _LIBC | |
106 | # define IS_CHAR_CLASS(string) __wctype (string) | |
107 | # else | |
108 | # define IS_CHAR_CLASS(string) wctype (string) | |
109 | # endif | |
110 | ||
111 | # ifdef _LIBC | |
112 | # define ISWCTYPE(WC, WT) __iswctype (WC, WT) | |
113 | # else | |
114 | # define ISWCTYPE(WC, WT) iswctype (WC, WT) | |
115 | # endif | |
116 | ||
117 | # if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC | |
118 | /* In this case we are implementing the multibyte character handling. */ | |
119 | # define HANDLE_MULTIBYTE 1 | |
120 | # endif | |
121 | ||
122 | # else | |
123 | # define CHAR_CLASS_MAX_LENGTH 6 /* Namely, 'xdigit'. */ | |
124 | ||
125 | # define IS_CHAR_CLASS(string) \ | |
126 | (STREQ (string, "alpha") || STREQ (string, "upper") \ | |
127 | || STREQ (string, "lower") || STREQ (string, "digit") \ | |
128 | || STREQ (string, "alnum") || STREQ (string, "xdigit") \ | |
129 | || STREQ (string, "space") || STREQ (string, "print") \ | |
130 | || STREQ (string, "punct") || STREQ (string, "graph") \ | |
131 | || STREQ (string, "cntrl") || STREQ (string, "blank")) | |
132 | # endif | |
133 | ||
134 | /* Avoid depending on library functions or files | |
135 | whose names are inconsistent. */ | |
136 | ||
137 | /* Global variable. */ | |
138 | static int posixly_correct; | |
139 | ||
140 | # ifndef internal_function | |
141 | /* Inside GNU libc we mark some function in a special way. In other | |
142 | environments simply ignore the marking. */ | |
143 | # define internal_function | |
144 | # endif | |
145 | ||
146 | /* Note that this evaluates C many times. */ | |
147 | # define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c)) | |
148 | # define CHAR char | |
149 | # define UCHAR unsigned char | |
150 | # define INT int | |
151 | # define FCT internal_fnmatch | |
152 | # define EXT ext_match | |
153 | # define END end_pattern | |
154 | # define L_(CS) CS | |
155 | # ifdef _LIBC | |
156 | # define BTOWC(C) __btowc (C) | |
157 | # else | |
158 | # define BTOWC(C) btowc (C) | |
159 | # endif | |
160 | # define STRLEN(S) strlen (S) | |
161 | # define STRCAT(D, S) strcat (D, S) | |
162 | # ifdef _LIBC | |
163 | # define MEMPCPY(D, S, N) __mempcpy (D, S, N) | |
164 | # else | |
165 | # if HAVE_MEMPCPY | |
166 | # define MEMPCPY(D, S, N) mempcpy (D, S, N) | |
167 | # else | |
168 | # define MEMPCPY(D, S, N) ((void *) ((char *) memcpy (D, S, N) + (N))) | |
169 | # endif | |
170 | # endif | |
171 | # define MEMCHR(S, C, N) memchr (S, C, N) | |
172 | # include "fnmatch_loop.c" | |
173 | ||
174 | ||
175 | # if HANDLE_MULTIBYTE | |
176 | # define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c)) | |
177 | # define CHAR wchar_t | |
178 | # define UCHAR wint_t | |
179 | # define INT wint_t | |
180 | # define FCT internal_fnwmatch | |
181 | # define EXT ext_wmatch | |
182 | # define END end_wpattern | |
183 | # define L_(CS) L##CS | |
184 | # define BTOWC(C) (C) | |
185 | # ifdef _LIBC | |
186 | # define STRLEN(S) __wcslen (S) | |
187 | # define STRCAT(D, S) __wcscat (D, S) | |
188 | # define MEMPCPY(D, S, N) __wmempcpy (D, S, N) | |
189 | # else | |
190 | # define STRLEN(S) wcslen (S) | |
191 | # define STRCAT(D, S) wcscat (D, S) | |
192 | # if HAVE_WMEMPCPY | |
193 | # define MEMPCPY(D, S, N) wmempcpy (D, S, N) | |
194 | # else | |
195 | # define MEMPCPY(D, S, N) (wmemcpy (D, S, N) + (N)) | |
196 | # endif | |
197 | # endif | |
198 | # define MEMCHR(S, C, N) wmemchr (S, C, N) | |
199 | # define WIDE_CHAR_VERSION 1 | |
200 | ||
201 | # undef IS_CHAR_CLASS | |
202 | /* We have to convert the wide character string in a multibyte string. But | |
203 | we know that the character class names consist of alphanumeric characters | |
204 | from the portable character set, and since the wide character encoding | |
205 | for a member of the portable character set is the same code point as | |
206 | its single-byte encoding, we can use a simplified method to convert the | |
207 | string to a multibyte character string. */ | |
208 | static wctype_t | |
209 | is_char_class (const wchar_t *wcs) | |
210 | { | |
211 | char s[CHAR_CLASS_MAX_LENGTH + 1]; | |
212 | char *cp = s; | |
213 | ||
214 | do | |
215 | { | |
216 | /* Test for a printable character from the portable character set. */ | |
217 | # ifdef _LIBC | |
218 | if (*wcs < 0x20 || *wcs > 0x7e | |
219 | || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60) | |
220 | return (wctype_t) 0; | |
221 | # else | |
222 | switch (*wcs) | |
223 | { | |
224 | case L' ': case L'!': case L'"': case L'#': case L'%': | |
225 | case L'&': case L'\'': case L'(': case L')': case L'*': | |
226 | case L'+': case L',': case L'-': case L'.': case L'/': | |
227 | case L'0': case L'1': case L'2': case L'3': case L'4': | |
228 | case L'5': case L'6': case L'7': case L'8': case L'9': | |
229 | case L':': case L';': case L'<': case L'=': case L'>': | |
230 | case L'?': | |
231 | case L'A': case L'B': case L'C': case L'D': case L'E': | |
232 | case L'F': case L'G': case L'H': case L'I': case L'J': | |
233 | case L'K': case L'L': case L'M': case L'N': case L'O': | |
234 | case L'P': case L'Q': case L'R': case L'S': case L'T': | |
235 | case L'U': case L'V': case L'W': case L'X': case L'Y': | |
236 | case L'Z': | |
237 | case L'[': case L'\\': case L']': case L'^': case L'_': | |
238 | case L'a': case L'b': case L'c': case L'd': case L'e': | |
239 | case L'f': case L'g': case L'h': case L'i': case L'j': | |
240 | case L'k': case L'l': case L'm': case L'n': case L'o': | |
241 | case L'p': case L'q': case L'r': case L's': case L't': | |
242 | case L'u': case L'v': case L'w': case L'x': case L'y': | |
243 | case L'z': case L'{': case L'|': case L'}': case L'~': | |
244 | break; | |
245 | default: | |
246 | return (wctype_t) 0; | |
247 | } | |
248 | # endif | |
249 | ||
250 | /* Avoid overrunning the buffer. */ | |
251 | if (cp == s + CHAR_CLASS_MAX_LENGTH) | |
252 | return (wctype_t) 0; | |
253 | ||
254 | *cp++ = (char) *wcs++; | |
255 | } | |
256 | while (*wcs != L'\0'); | |
257 | ||
258 | *cp = '\0'; | |
259 | ||
260 | # ifdef _LIBC | |
261 | return __wctype (s); | |
262 | # else | |
263 | return wctype (s); | |
264 | # endif | |
265 | } | |
266 | # define IS_CHAR_CLASS(string) is_char_class (string) | |
267 | ||
268 | # include "fnmatch_loop.c" | |
269 | # endif | |
270 | ||
271 | ||
272 | int | |
273 | fnmatch (const char *pattern, const char *string, int flags) | |
274 | { | |
275 | # if HANDLE_MULTIBYTE | |
276 | # define ALLOCA_LIMIT 2000 | |
277 | if (__builtin_expect (MB_CUR_MAX, 1) != 1) | |
278 | { | |
279 | mbstate_t ps; | |
280 | size_t patsize; | |
281 | size_t strsize; | |
282 | size_t totsize; | |
283 | wchar_t *wpattern; | |
284 | wchar_t *wstring; | |
285 | int res; | |
286 | ||
287 | /* Calculate the size needed to convert the strings to | |
288 | wide characters. */ | |
289 | memset (&ps, '\0', sizeof (ps)); | |
290 | patsize = mbsrtowcs (NULL, &pattern, 0, &ps) + 1; | |
291 | if (__builtin_expect (patsize != 0, 1)) | |
292 | { | |
293 | assert (mbsinit (&ps)); | |
294 | strsize = mbsrtowcs (NULL, &string, 0, &ps) + 1; | |
295 | if (__builtin_expect (strsize != 0, 1)) | |
296 | { | |
297 | assert (mbsinit (&ps)); | |
298 | totsize = patsize + strsize; | |
299 | if (__builtin_expect (! (patsize <= totsize | |
300 | && totsize <= SIZE_MAX / sizeof (wchar_t)), | |
301 | 0)) | |
302 | { | |
303 | errno = ENOMEM; | |
304 | return -1; | |
305 | } | |
306 | ||
307 | /* Allocate room for the wide characters. */ | |
308 | if (__builtin_expect (totsize < ALLOCA_LIMIT, 1)) | |
309 | wpattern = (wchar_t *) alloca (totsize * sizeof (wchar_t)); | |
310 | else | |
311 | { | |
312 | wpattern = malloc (totsize * sizeof (wchar_t)); | |
313 | if (__builtin_expect (! wpattern, 0)) | |
314 | { | |
315 | errno = ENOMEM; | |
316 | return -1; | |
317 | } | |
318 | } | |
319 | wstring = wpattern + patsize; | |
320 | ||
321 | /* Convert the strings into wide characters. */ | |
322 | mbsrtowcs (wpattern, &pattern, patsize, &ps); | |
323 | assert (mbsinit (&ps)); | |
324 | mbsrtowcs (wstring, &string, strsize, &ps); | |
325 | ||
326 | res = internal_fnwmatch (wpattern, wstring, wstring + strsize - 1, | |
327 | flags & FNM_PERIOD, flags); | |
328 | ||
329 | if (__builtin_expect (! (totsize < ALLOCA_LIMIT), 0)) | |
330 | free (wpattern); | |
331 | return res; | |
332 | } | |
333 | } | |
334 | } | |
335 | ||
336 | # endif /* HANDLE_MULTIBYTE */ | |
337 | ||
338 | return internal_fnmatch (pattern, string, string + strlen (string), | |
339 | flags & FNM_PERIOD, flags); | |
340 | } | |
341 | ||
342 | # ifdef _LIBC | |
343 | # undef fnmatch | |
344 | versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3); | |
345 | # if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3) | |
346 | strong_alias (__fnmatch, __fnmatch_old) | |
347 | compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0); | |
348 | # endif | |
349 | libc_hidden_ver (__fnmatch, fnmatch) | |
350 | # endif | |
351 | ||
352 | #endif /* _LIBC or not __GNU_LIBRARY__. */ |