posix/fnmatch.c

   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, write to the Free
  17    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  18    02111-1307 USA.  */
  19
  20 #if HAVE_CONFIG_H
  21 # include <config.h>
  22 #endif
  23
  24 /* Enable GNU extensions in fnmatch.h.  */
  25 #ifndef _GNU_SOURCE
  26 # define _GNU_SOURCE    1
  27 #endif
  28
  29 #include <assert.h>
  30 #include <errno.h>
  31 #include <fnmatch.h>
  32 #include <ctype.h>
  33
  34 #if HAVE_STRING_H || defined _LIBC
  35 # include <string.h>
  36 #else
  37 # include <strings.h>
  38 #endif
  39
  40 #if defined STDC_HEADERS || defined _LIBC
  41 # include <stdlib.h>
  42 #endif
  43
  44 /* For platform which support the ISO C amendement 1 functionality we
  45    support user defined character classes.  */
  46 #if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
  47 /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>.  */
  48 # include <wchar.h>
  49 # include <wctype.h>
  50 #endif
  51
  52 /* We need some of the locale data (the collation sequence information)
  53    but there is no interface to get this information in general.  Therefore
  54    we support a correct implementation only in glibc.  */
  55 #ifdef _LIBC
  56 # include "../locale/localeinfo.h"
  57 # include "../locale/elem-hash.h"
  58 # include "../locale/coll-lookup.h"
  59 # include <shlib-compat.h>
  60
  61 # define CONCAT(a,b) __CONCAT(a,b)
  62 # define mbsinit __mbsinit
  63 # define mbsrtowcs __mbsrtowcs
  64 # define fnmatch __fnmatch
  65 extern int fnmatch (const char *pattern, const char *string, int flags);
  66 #endif
  67
  68 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set.  */
  69 #define NO_LEADING_PERIOD(flags) \
  70   ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
  71
  72 /* Comment out all this code if we are using the GNU C Library, and are not
  73    actually compiling the library itself.  This code is part of the GNU C
  74    Library, but also included in many other GNU distributions.  Compiling
  75    and linking in this code is a waste when using the GNU C library
  76    (especially if it is a shared library).  Rather than having every GNU
  77    program understand `configure --with-gnu-libc' and omit the object files,
  78    it is simpler to just do this in the source for each such file.  */
  79
  80 #if defined _LIBC || !defined __GNU_LIBRARY__
  81
  82
  83 # if defined STDC_HEADERS || !defined isascii
  84 #  define ISASCII(c) 1
  85 # else
  86 #  define ISASCII(c) isascii(c)
  87 # endif
  88
  89 # ifdef isblank
  90 #  define ISBLANK(c) (ISASCII (c) && isblank (c))
  91 # else
  92 #  define ISBLANK(c) ((c) == ' ' || (c) == '\t')
  93 # endif
  94 # ifdef isgraph
  95 #  define ISGRAPH(c) (ISASCII (c) && isgraph (c))
  96 # else
  97 #  define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
  98 # endif
  99
 100 # define ISPRINT(c) (ISASCII (c) && isprint (c))
 101 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
 102 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
 103 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
 104 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
 105 # define ISLOWER(c) (ISASCII (c) && islower (c))
 106 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
 107 # define ISSPACE(c) (ISASCII (c) && isspace (c))
 108 # define ISUPPER(c) (ISASCII (c) && isupper (c))
 109 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
 110
 111 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
 112
 113 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
 114 /* The GNU C library provides support for user-defined character classes
 115    and the functions from ISO C amendement 1.  */
 116 #  ifdef CHARCLASS_NAME_MAX
 117 #   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
 118 #  else
 119 /* This shouldn't happen but some implementation might still have this
 120    problem.  Use a reasonable default value.  */
 121 #   define CHAR_CLASS_MAX_LENGTH 256
 122 #  endif
 123
 124 #  ifdef _LIBC
 125 #   define IS_CHAR_CLASS(string) __wctype (string)
 126 #  else
 127 #   define IS_CHAR_CLASS(string) wctype (string)
 128 #  endif
 129
 130 #  ifdef _LIBC
 131 #   define ISWCTYPE(WC, WT)     __iswctype (WC, WT)
 132 #  else
 133 #   define ISWCTYPE(WC, WT)     iswctype (WC, WT)
 134 #  endif
 135
 136 #  if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
 137 /* In this case we are implementing the multibyte character handling.  */
 138 #   define HANDLE_MULTIBYTE     1
 139 #  endif
 140
 141 # else
 142 #  define CHAR_CLASS_MAX_LENGTH  6 /* Namely, `xdigit'.  */
 143
 144 #  define IS_CHAR_CLASS(string)                                               \
 145    (STREQ (string, "alpha") || STREQ (string, "upper")                        \
 146     || STREQ (string, "lower") || STREQ (string, "digit")                     \
 147     || STREQ (string, "alnum") || STREQ (string, "xdigit")                    \
 148     || STREQ (string, "space") || STREQ (string, "print")                     \
 149     || STREQ (string, "punct") || STREQ (string, "graph")                     \
 150     || STREQ (string, "cntrl") || STREQ (string, "blank"))
 151 # endif
 152
 153 /* Avoid depending on library functions or files
 154    whose names are inconsistent.  */
 155
 156 # if !defined _LIBC && !defined getenv
 157 extern char *getenv ();
 158 # endif
 159
 160 # ifndef errno
 161 extern int errno;
 162 # endif
 163
 164 /* Global variable.  */
 165 static int posixly_correct;
 166
 167 /* This function doesn't exist on most systems.  */
 168
 169 # if !defined HAVE___STRCHRNUL && !defined _LIBC
 170 static char *
 171 __strchrnul (s, c)
 172      const char *s;
 173      int c;
 174 {
 175   char *result = strchr (s, c);
 176   if (result == NULL)
 177     result = strchr (s, '\0');
 178   return result;
 179 }
 180 # endif
 181
 182 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
 183 static wchar_t *
 184 __wcschrnul (s, c)
 185      const wchar_t *s;
 186      wint_t c;
 187 {
 188   wchar_t *result = wcschr (s, c);
 189   if (result == NULL)
 190     result = wcschr (s, '\0');
 191   return result;
 192 }
 193 # endif
 194
 195 # ifndef internal_function
 196 /* Inside GNU libc we mark some function in a special way.  In other
 197    environments simply ignore the marking.  */
 198 #  define internal_function
 199 # endif
 200
 201 /* Note that this evaluates C many times.  */
 202 # ifdef _LIBC
 203 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
 204 # else
 205 #  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
 206 # endif
 207 # define CHAR   char
 208 # define UCHAR  unsigned char
 209 # define INT    int
 210 # define FCT    internal_fnmatch
 211 # define EXT    ext_match
 212 # define END    end_pattern
 213 # define L(CS)  CS
 214 # ifdef _LIBC
 215 #  define BTOWC(C)      __btowc (C)
 216 # else
 217 #  define BTOWC(C)      btowc (C)
 218 # endif
 219 # define STRLEN(S) strlen (S)
 220 # define STRCAT(D, S) strcat (D, S)
 221 # define MEMPCPY(D, S, N) __mempcpy (D, S, N)
 222 # define MEMCHR(S, C, N) memchr (S, C, N)
 223 # define STRCOLL(S1, S2) strcoll (S1, S2)
 224 # include "fnmatch_loop.c"
 225
 226
 227 # if HANDLE_MULTIBYTE
 228 /* Note that this evaluates C many times.  */
 229 #  ifdef _LIBC
 230 #   define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
 231 #  else
 232 #   define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
 233 #  endif
 234 #  define CHAR  wchar_t
 235 #  define UCHAR wint_t
 236 #  define INT   wint_t
 237 #  define FCT   internal_fnwmatch
 238 #  define EXT   ext_wmatch
 239 # define END    end_wpattern
 240 #  define L(CS) L##CS
 241 #  define BTOWC(C)      (C)
 242 #  define STRLEN(S) __wcslen (S)
 243 #  define STRCAT(D, S) __wcscat (D, S)
 244 #  define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
 245 #  define MEMCHR(S, C, N) wmemchr (S, C, N)
 246 #  define STRCOLL(S1, S2) wcscoll (S1, S2)
 247 #  define WIDE_CHAR_VERSION 1
 248
 249 #  undef IS_CHAR_CLASS
 250 /* We have to convert the wide character string in a multibyte string.  But
 251    we know that the character class names consist of alphanumeric characters
 252    from the portable character set, and since the wide character encoding
 253    for a member of the portable character set is the same code point as
 254    its single-byte encoding, we can use a simplified method to convert the
 255    string to a multibyte character string.  */
 256 static wctype_t
 257 is_char_class (const wchar_t *wcs)
 258 {
 259   char s[CHAR_CLASS_MAX_LENGTH + 1];
 260   char *cp = s;
 261
 262   do
 263     {
 264       /* Test for a printable character from the portable character set.  */
 265 #  ifdef _LIBC
 266       if (*wcs < 0x20 || *wcs > 0x7e
 267           || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
 268         return (wctype_t) 0;
 269 #  else
 270       switch (*wcs)
 271         {
 272         case L' ': case L'!': case L'"': case L'#': case L'%':
 273         case L'&': case L'\'': case L'(': case L')': case L'*':
 274         case L'+': case L',': case L'-': case L'.': case L'/':
 275         case L'0': case L'1': case L'2': case L'3': case L'4':
 276         case L'5': case L'6': case L'7': case L'8': case L'9':
 277         case L':': case L';': case L'<': case L'=': case L'>':
 278         case L'?':
 279         case L'A': case L'B': case L'C': case L'D': case L'E':
 280         case L'F': case L'G': case L'H': case L'I': case L'J':
 281         case L'K': case L'L': case L'M': case L'N': case L'O':
 282         case L'P': case L'Q': case L'R': case L'S': case L'T':
 283         case L'U': case L'V': case L'W': case L'X': case L'Y':
 284         case L'Z':
 285         case L'[': case L'\\': case L']': case L'^': case L'_':
 286         case L'a': case L'b': case L'c': case L'd': case L'e':
 287         case L'f': case L'g': case L'h': case L'i': case L'j':
 288         case L'k': case L'l': case L'm': case L'n': case L'o':
 289         case L'p': case L'q': case L'r': case L's': case L't':
 290         case L'u': case L'v': case L'w': case L'x': case L'y':
 291         case L'z': case L'{': case L'|': case L'}': case L'~':
 292           break;
 293         default:
 294           return (wctype_t) 0;
 295         }
 296 #  endif
 297
 298       /* Avoid overrunning the buffer.  */
 299       if (cp == s + CHAR_CLASS_MAX_LENGTH)
 300         return (wctype_t) 0;
 301
 302       *cp++ = (char) *wcs++;
 303     }
 304   while (*wcs != L'\0');
 305
 306   *cp = '\0';
 307
 308 #  ifdef _LIBC
 309   return __wctype (s);
 310 #  else
 311   return wctype (s);
 312 #  endif
 313 }
 314 #  define IS_CHAR_CLASS(string) is_char_class (string)
 315
 316 #  include "fnmatch_loop.c"
 317 # endif
 318
 319
 320 int
 321 fnmatch (pattern, string, flags)
 322      const char *pattern;
 323      const char *string;
 324      int flags;
 325 {
 326 # if HANDLE_MULTIBYTE
 327   if (__builtin_expect (MB_CUR_MAX, 1) != 1)
 328     {
 329       mbstate_t ps;
 330       size_t n;
 331       wchar_t *wpattern;
 332       wchar_t *wstring;
 333
 334       /* Convert the strings into wide characters.  */
 335       memset (&ps, '\0', sizeof (ps));
 336       n = mbsrtowcs (NULL, &pattern, 0, &ps);
 337       if (__builtin_expect (n == (size_t) -1, 0))
 338         /* Something wrong.
 339            XXX Do we have to set `errno' to something which mbsrtows hasn't
 340            already done?  */
 341         return -1;
 342       wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 343       assert (mbsinit (&ps));
 344       (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
 345
 346       assert (mbsinit (&ps));
 347       n = mbsrtowcs (NULL, &string, 0, &ps);
 348       if (__builtin_expect (n == (size_t) -1, 0))
 349         /* Something wrong.
 350            XXX Do we have to set `errno' to something which mbsrtows hasn't
 351            already done?  */
 352         return -1;
 353       wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 354       assert (mbsinit (&ps));
 355       (void) mbsrtowcs (wstring, &string, n + 1, &ps);
 356
 357       return internal_fnwmatch (wpattern, wstring, wstring + n,
 358                                 flags & FNM_PERIOD, flags);
 359     }
 360 # endif  /* mbstate_t and mbsrtowcs or _LIBC.  */
 361
 362   return internal_fnmatch (pattern, string, string + strlen (string),
 363                            flags & FNM_PERIOD, flags);
 364 }
 365
 366 # ifdef _LIBC
 367 #  undef fnmatch
 368 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
 369 #  if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
 370 strong_alias (__fnmatch, __fnmatch_old)
 371 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
 372 #  endif
 373 libc_hidden_ver (__fnmatch, fnmatch)
 374 # endif
 375
 376 #endif  /* _LIBC or not __GNU_LIBRARY__.  */