1/* Copyright (C) 1991,1992,1993,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005,2006,2007 2 Free Software Foundation, Inc. 3 4 This program is free software; you can redistribute it and/or modify 5 it under the terms of the GNU General Public License as published by 6 the Free Software Foundation; either version 2, or (at your option) 7 any later version. 8 9 This program 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 12 GNU General Public License for more details. 13 14 You should have received a copy of the GNU General Public License 15 along with this program; if not, write to the Free Software Foundation, 16 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ 17 18#ifndef _LIBC 19# include <config.h> 20#endif 21 22/* Enable GNU extensions in fnmatch.h. */ 23#ifndef _GNU_SOURCE 24# define _GNU_SOURCE 1 25#endif 26 27#if ! defined __builtin_expect && __GNUC__ < 3 28# define __builtin_expect(expr, expected) (expr) 29#endif 30 31#include <fnmatch.h> 32 33#include <alloca.h> 34#include <assert.h> 35#include <ctype.h> 36#include <errno.h> 37#include <stddef.h> 38#include <stdbool.h> 39#include <stdlib.h> 40#include <string.h> 41 42#define WIDE_CHAR_SUPPORT \ 43 (HAVE_WCTYPE_H && HAVE_BTOWC && HAVE_ISWCTYPE \ 44 && HAVE_WMEMCHR && (HAVE_WMEMCPY || HAVE_WMEMPCPY)) 45 46/* For platform which support the ISO C amendement 1 functionality we 47 support user defined character classes. */ 48#if defined _LIBC || WIDE_CHAR_SUPPORT 49# include <wctype.h> 50# include <wchar.h> 51#endif 52 53/* We need some of the locale data (the collation sequence information) 54 but there is no interface to get this information in general. Therefore 55 we support a correct implementation only in glibc. */ 56#ifdef _LIBC 57# include "../locale/localeinfo.h" 58# include "../locale/elem-hash.h" 59# include "../locale/coll-lookup.h" 60# include <shlib-compat.h> 61 62# define CONCAT(a,b) __CONCAT(a,b) 63# define mbsrtowcs __mbsrtowcs 64# define fnmatch __fnmatch 65extern int fnmatch (const char *pattern, const char *string, int flags); 66#endif 67 68#ifndef SIZE_MAX 69# define SIZE_MAX ((size_t) -1) 70#endif 71 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 amendement 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. */ 138static 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# define STRCOLL(S1, S2) strcoll (S1, S2) 173# include "fnmatch_loop.c" 174 175 176# if HANDLE_MULTIBYTE 177# define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c)) 178# define CHAR wchar_t 179# define UCHAR wint_t 180# define INT wint_t 181# define FCT internal_fnwmatch 182# define EXT ext_wmatch 183# define END end_wpattern 184# define L_(CS) L##CS 185# define BTOWC(C) (C) 186# ifdef _LIBC 187# define STRLEN(S) __wcslen (S) 188# define STRCAT(D, S) __wcscat (D, S) 189# define MEMPCPY(D, S, N) __wmempcpy (D, S, N) 190# else 191# define STRLEN(S) wcslen (S) 192# define STRCAT(D, S) wcscat (D, S) 193# if HAVE_WMEMPCPY 194# define MEMPCPY(D, S, N) wmempcpy (D, S, N) 195# else 196# define MEMPCPY(D, S, N) (wmemcpy (D, S, N) + (N)) 197# endif 198# endif 199# define MEMCHR(S, C, N) wmemchr (S, C, N) 200# define STRCOLL(S1, S2) wcscoll (S1, S2) 201# define WIDE_CHAR_VERSION 1 202 203# undef IS_CHAR_CLASS 204/* We have to convert the wide character string in a multibyte string. But 205 we know that the character class names consist of alphanumeric characters 206 from the portable character set, and since the wide character encoding 207 for a member of the portable character set is the same code point as 208 its single-byte encoding, we can use a simplified method to convert the 209 string to a multibyte character string. */ 210static wctype_t 211is_char_class (const wchar_t *wcs) 212{ 213 char s[CHAR_CLASS_MAX_LENGTH + 1]; 214 char *cp = s; 215 216 do 217 { 218 /* Test for a printable character from the portable character set. */ 219# ifdef _LIBC 220 if (*wcs < 0x20 || *wcs > 0x7e 221 || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60) 222 return (wctype_t) 0; 223# else 224 switch (*wcs) 225 { 226 case L' ': case L'!': case L'"': case L'#': case L'%': 227 case L'&': case L'\'': case L'(': case L')': case L'*': 228 case L'+': case L',': case L'-': case L'.': case L'/': 229 case L'0': case L'1': case L'2': case L'3': case L'4': 230 case L'5': case L'6': case L'7': case L'8': case L'9': 231 case L':': case L';': case L'<': case L'=': case L'>': 232 case L'?': 233 case L'A': case L'B': case L'C': case L'D': case L'E': 234 case L'F': case L'G': case L'H': case L'I': case L'J': 235 case L'K': case L'L': case L'M': case L'N': case L'O': 236 case L'P': case L'Q': case L'R': case L'S': case L'T': 237 case L'U': case L'V': case L'W': case L'X': case L'Y': 238 case L'Z': 239 case L'[': case L'\\': case L']': case L'^': case L'_': 240 case L'a': case L'b': case L'c': case L'd': case L'e': 241 case L'f': case L'g': case L'h': case L'i': case L'j': 242 case L'k': case L'l': case L'm': case L'n': case L'o': 243 case L'p': case L'q': case L'r': case L's': case L't': 244 case L'u': case L'v': case L'w': case L'x': case L'y': 245 case L'z': case L'{': case L'|': case L'}': case L'~': 246 break; 247 default: 248 return (wctype_t) 0; 249 } 250# endif 251 252 /* Avoid overrunning the buffer. */ 253 if (cp == s + CHAR_CLASS_MAX_LENGTH) 254 return (wctype_t) 0; 255 256 *cp++ = (char) *wcs++; 257 } 258 while (*wcs != L'\0'); 259 260 *cp = '\0'; 261 262# ifdef _LIBC 263 return __wctype (s); 264# else 265 return wctype (s); 266# endif 267} 268# define IS_CHAR_CLASS(string) is_char_class (string) 269 270# include "fnmatch_loop.c" 271# endif 272 273 274int 275fnmatch (const char *pattern, const char *string, int flags) 276{ 277# if HANDLE_MULTIBYTE 278# if HAVE_ALLOCA || defined _LIBC 279# define ALLOCA_LIMIT 2000 280# else 281# define ALLOCA_LIMIT 0 282# endif 283 if (__builtin_expect (MB_CUR_MAX, 1) != 1) 284 { 285 mbstate_t ps; 286 size_t patsize; 287 size_t strsize; 288 size_t totsize; 289 wchar_t *wpattern; 290 wchar_t *wstring; 291 int res; 292 293 /* Calculate the size needed to convert the strings to 294 wide characters. */ 295 memset (&ps, '\0', sizeof (ps)); 296 patsize = mbsrtowcs (NULL, &pattern, 0, &ps) + 1; 297 if (__builtin_expect (patsize != 0, 1)) 298 { 299 assert (mbsinit (&ps)); 300 strsize = mbsrtowcs (NULL, &string, 0, &ps) + 1; 301 if (__builtin_expect (strsize != 0, 1)) 302 { 303 assert (mbsinit (&ps)); 304 totsize = patsize + strsize; 305 if (__builtin_expect (! (patsize <= totsize 306 && totsize <= SIZE_MAX / sizeof (wchar_t)), 307 0)) 308 { 309 errno = ENOMEM; 310 return -1; 311 } 312 313 /* Allocate room for the wide characters. */ 314 if (__builtin_expect (totsize < ALLOCA_LIMIT, 1)) 315 wpattern = (wchar_t *) alloca (totsize * sizeof (wchar_t)); 316 else 317 { 318 wpattern = malloc (totsize * sizeof (wchar_t)); 319 if (__builtin_expect (! wpattern, 0)) 320 { 321 errno = ENOMEM; 322 return -1; 323 } 324 } 325 wstring = wpattern + patsize; 326 327 /* Convert the strings into wide characters. */ 328 mbsrtowcs (wpattern, &pattern, patsize, &ps); 329 assert (mbsinit (&ps)); 330 mbsrtowcs (wstring, &string, strsize, &ps); 331 332 res = internal_fnwmatch (wpattern, wstring, wstring + strsize - 1, 333 flags & FNM_PERIOD, flags); 334 335 if (__builtin_expect (! (totsize < ALLOCA_LIMIT), 0)) 336 free (wpattern); 337 return res; 338 } 339 } 340 } 341 342# endif /* HANDLE_MULTIBYTE */ 343 344 return internal_fnmatch (pattern, string, string + strlen (string), 345 flags & FNM_PERIOD, flags); 346} 347 348# ifdef _LIBC 349# undef fnmatch 350versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3); 351# if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3) 352strong_alias (__fnmatch, __fnmatch_old) 353compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0); 354# endif 355libc_hidden_ver (__fnmatch, fnmatch) 356# endif 357 358#endif /* _LIBC or not __GNU_LIBRARY__. */ 359