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