Dogcows Code - chaz/tar/blob - lib/fnmatch.c

   1 /* Copyright (C) 1991, 1992, 1993, 1996, 1997, 1998, 1999, 2000, 2001,
   2    2002 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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17
  18 #if HAVE_CONFIG_H
  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 #ifdef __GNUC__
  28 # define alloca __builtin_alloca
  29 # define HAVE_ALLOCA 1
  30 #else
  31 # if defined HAVE_ALLOCA_H || defined _LIBC
  32 #  include <alloca.h>
  33 # else
  34 #  ifdef _AIX
  35  #  pragma alloca
  36 #  else
  37 #   ifndef alloca
  38 char *alloca ();
  39 #   endif
  40 #  endif
  41 # endif
  42 #endif
  43
  44 #if ! defined __builtin_expect && __GNUC__ < 3
  45 # define __builtin_expect(expr, expected) (expr)
  46 #endif
  47
  48 #include <assert.h>
  49 #include <errno.h>
  50 #include <fnmatch.h>
  51 #include <ctype.h>
  52
  53 #if HAVE_STRING_H || defined _LIBC
  54 # include <string.h>
  55 #else
  56 # if HAVE_STRINGS_H
  57 #  include <strings.h>
  58 # endif
  59 #endif
  60
  61 #if defined STDC_HEADERS || defined _LIBC
  62 # include <stddef.h>
  63 # include <stdlib.h>
  64 #endif
  65
  66 #define WIDE_CHAR_SUPPORT (HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_BTOWC)
  67
  68 /* For platform which support the ISO C amendement 1 functionality we
  69    support user defined character classes.  */
  70 #if defined _LIBC || WIDE_CHAR_SUPPORT
  71 /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>.  */
  72 # include <wchar.h>
  73 # include <wctype.h>
  74 #endif
  75
  76 /* We need some of the locale data (the collation sequence information)
  77    but there is no interface to get this information in general.  Therefore
  78    we support a correct implementation only in glibc.  */
  79 #ifdef _LIBC
  80 # include "../locale/localeinfo.h"
  81 # include "../locale/elem-hash.h"
  82 # include "../locale/coll-lookup.h"
  83 # include <shlib-compat.h>
  84
  85 # define CONCAT(a,b) __CONCAT(a,b)
  86 # define mbsinit __mbsinit
  87 # define mbsrtowcs __mbsrtowcs
  88 # define fnmatch __fnmatch
  89 extern int fnmatch (const char *pattern, const char *string, int flags);
  90 #endif
  91
  92 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set.  */
  93 #define NO_LEADING_PERIOD(flags) \
  94   ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
  95
  96 /* Comment out all this code if we are using the GNU C Library, are not
  97    actually compiling the library itself, and have not detected a bug
  98    in the library.  This code is part of the GNU C
  99    Library, but also included in many other GNU distributions.  Compiling
 100    and linking in this code is a waste when using the GNU C library
 101    (especially if it is a shared library).  Rather than having every GNU
 102    program understand `configure --with-gnu-libc' and omit the object files,
 103    it is simpler to just do this in the source for each such file.  */
 104
 105 #if defined _LIBC || !defined __GNU_LIBRARY__ || !HAVE_FNMATCH_GNU
 106
 107
 108 # if defined STDC_HEADERS || !defined isascii
 109 #  define ISASCII(c) 1
 110 # else
 111 #  define ISASCII(c) isascii(c)
 112 # endif
 113
 114 # ifdef isblank
 115 #  define ISBLANK(c) (ISASCII (c) && isblank (c))
 116 # else
 117 #  define ISBLANK(c) ((c) == ' ' || (c) == '\t')
 118 # endif
 119 # ifdef isgraph
 120 #  define ISGRAPH(c) (ISASCII (c) && isgraph (c))
 121 # else
 122 #  define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
 123 # endif
 124
 125 # define ISPRINT(c) (ISASCII (c) && isprint (c))
 126 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
 127 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
 128 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
 129 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
 130 # define ISLOWER(c) (ISASCII (c) && islower (c))
 131 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
 132 # define ISSPACE(c) (ISASCII (c) && isspace (c))
 133 # define ISUPPER(c) (ISASCII (c) && isupper (c))
 134 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
 135
 136 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
 137
 138 # if defined _LIBC || WIDE_CHAR_SUPPORT
 139 /* The GNU C library provides support for user-defined character classes
 140    and the functions from ISO C amendement 1.  */
 141 #  ifdef CHARCLASS_NAME_MAX
 142 #   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
 143 #  else
 144 /* This shouldn't happen but some implementation might still have this
 145    problem.  Use a reasonable default value.  */
 146 #   define CHAR_CLASS_MAX_LENGTH 256
 147 #  endif
 148
 149 #  ifdef _LIBC
 150 #   define IS_CHAR_CLASS(string) __wctype (string)
 151 #  else
 152 #   define IS_CHAR_CLASS(string) wctype (string)
 153 #  endif
 154
 155 #  ifdef _LIBC
 156 #   define ISWCTYPE(WC, WT)     __iswctype (WC, WT)
 157 #  else
 158 #   define ISWCTYPE(WC, WT)     iswctype (WC, WT)
 159 #  endif
 160
 161 #  if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
 162 /* In this case we are implementing the multibyte character handling.  */
 163 #   define HANDLE_MULTIBYTE     1
 164 #  endif
 165
 166 # else
 167 #  define CHAR_CLASS_MAX_LENGTH  6 /* Namely, `xdigit'.  */
 168
 169 #  define IS_CHAR_CLASS(string)                                               \
 170    (STREQ (string, "alpha") || STREQ (string, "upper")                        \
 171     || STREQ (string, "lower") || STREQ (string, "digit")                     \
 172     || STREQ (string, "alnum") || STREQ (string, "xdigit")                    \
 173     || STREQ (string, "space") || STREQ (string, "print")                     \
 174     || STREQ (string, "punct") || STREQ (string, "graph")                     \
 175     || STREQ (string, "cntrl") || STREQ (string, "blank"))
 176 # endif
 177
 178 /* Avoid depending on library functions or files
 179    whose names are inconsistent.  */
 180
 181 # if !defined _LIBC && !defined getenv && !HAVE_DECL_GETENV
 182 extern char *getenv ();
 183 # endif
 184
 185 # ifndef errno
 186 extern int errno;
 187 # endif
 188
 189 /* Global variable.  */
 190 static int posixly_correct;
 191
 192 # ifndef internal_function
 193 /* Inside GNU libc we mark some function in a special way.  In other
 194    environments simply ignore the marking.  */
 195 #  define internal_function
 196 # endif
 197
 198 /* Note that this evaluates C many times.  */
 199 # ifdef _LIBC
 200 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
 201 # else
 202 #  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
 203 # endif
 204 # define CHAR   char
 205 # define UCHAR  unsigned char
 206 # define INT    int
 207 # define FCT    internal_fnmatch
 208 # define EXT    ext_match
 209 # define END    end_pattern
 210 # define L(CS)  CS
 211 # ifdef _LIBC
 212 #  define BTOWC(C)      __btowc (C)
 213 # else
 214 #  define BTOWC(C)      btowc (C)
 215 # endif
 216 # define STRLEN(S) strlen (S)
 217 # define STRCAT(D, S) strcat (D, S)
 218 # ifdef _LIBC
 219 #  define MEMPCPY(D, S, N) __mempcpy (D, S, N)
 220 # else
 221 #  if HAVE_MEMPCPY
 222 #   define MEMPCPY(D, S, N) mempcpy (D, S, N)
 223 #  else
 224 #   define MEMPCPY(D, S, N) ((void *) ((char *) memcpy (D, S, N) + (N)))
 225 #  endif
 226 # endif
 227 # define MEMCHR(S, C, N) memchr (S, C, N)
 228 # define STRCOLL(S1, S2) strcoll (S1, S2)
 229 # include "fnmatch_loop.c"
 230
 231
 232 # if HANDLE_MULTIBYTE
 233 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
 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 #  ifdef _LIBC
 243 #   define STRLEN(S) __wcslen (S)
 244 #   define STRCAT(D, S) __wcscat (D, S)
 245 #   define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
 246 #  else
 247 #   define STRLEN(S) wcslen (S)
 248 #   define STRCAT(D, S) wcscat (D, S)
 249 #   if HAVE_WMEMPCPY
 250 #    define MEMPCPY(D, S, N) wmempcpy (D, S, N)
 251 #   else
 252 #    define MEMPCPY(D, S, N) (wmemcpy (D, S, N) + (N))
 253 #   endif
 254 #  endif
 255 #  define MEMCHR(S, C, N) wmemchr (S, C, N)
 256 #  define STRCOLL(S1, S2) wcscoll (S1, S2)
 257 #  define WIDE_CHAR_VERSION 1
 258
 259 #  undef IS_CHAR_CLASS
 260 /* We have to convert the wide character string in a multibyte string.  But
 261    we know that the character class names consist of alphanumeric characters
 262    from the portable character set, and since the wide character encoding
 263    for a member of the portable character set is the same code point as
 264    its single-byte encoding, we can use a simplified method to convert the
 265    string to a multibyte character string.  */
 266 static wctype_t
 267 is_char_class (const wchar_t *wcs)
 268 {
 269   char s[CHAR_CLASS_MAX_LENGTH + 1];
 270   char *cp = s;
 271
 272   do
 273     {
 274       /* Test for a printable character from the portable character set.  */
 275 #  ifdef _LIBC
 276       if (*wcs < 0x20 || *wcs > 0x7e
 277           || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
 278         return (wctype_t) 0;
 279 #  else
 280       switch (*wcs)
 281         {
 282         case L' ': case L'!': case L'"': case L'#': case L'%':
 283         case L'&': case L'\'': case L'(': case L')': case L'*':
 284         case L'+': case L',': case L'-': case L'.': case L'/':
 285         case L'0': case L'1': case L'2': case L'3': case L'4':
 286         case L'5': case L'6': case L'7': case L'8': case L'9':
 287         case L':': case L';': case L'<': case L'=': case L'>':
 288         case L'?':
 289         case L'A': case L'B': case L'C': case L'D': case L'E':
 290         case L'F': case L'G': case L'H': case L'I': case L'J':
 291         case L'K': case L'L': case L'M': case L'N': case L'O':
 292         case L'P': case L'Q': case L'R': case L'S': case L'T':
 293         case L'U': case L'V': case L'W': case L'X': case L'Y':
 294         case L'Z':
 295         case L'[': case L'\\': case L']': case L'^': case L'_':
 296         case L'a': case L'b': case L'c': case L'd': case L'e':
 297         case L'f': case L'g': case L'h': case L'i': case L'j':
 298         case L'k': case L'l': case L'm': case L'n': case L'o':
 299         case L'p': case L'q': case L'r': case L's': case L't':
 300         case L'u': case L'v': case L'w': case L'x': case L'y':
 301         case L'z': case L'{': case L'|': case L'}': case L'~':
 302           break;
 303         default:
 304           return (wctype_t) 0;
 305         }
 306 #  endif
 307
 308       /* Avoid overrunning the buffer.  */
 309       if (cp == s + CHAR_CLASS_MAX_LENGTH)
 310         return (wctype_t) 0;
 311
 312       *cp++ = (char) *wcs++;
 313     }
 314   while (*wcs != L'\0');
 315
 316   *cp = '\0';
 317
 318 #  ifdef _LIBC
 319   return __wctype (s);
 320 #  else
 321   return wctype (s);
 322 #  endif
 323 }
 324 #  define IS_CHAR_CLASS(string) is_char_class (string)
 325
 326 #  include "fnmatch_loop.c"
 327 # endif
 328
 329
 330 int
 331 fnmatch (pattern, string, flags)
 332      const char *pattern;
 333      const char *string;
 334      int flags;
 335 {
 336 # if HANDLE_MULTIBYTE
 337   if (__builtin_expect (MB_CUR_MAX, 1) != 1)
 338     {
 339       mbstate_t ps;
 340       size_t n;
 341       wchar_t *wpattern;
 342       wchar_t *wstring;
 343
 344       /* Convert the strings into wide characters.  */
 345       memset (&ps, '\0', sizeof (ps));
 346       n = mbsrtowcs (NULL, &pattern, 0, &ps);
 347       if (__builtin_expect (n, 0) == (size_t) -1)
 348         /* Something wrong.
 349            XXX Do we have to set `errno' to something which mbsrtows hasn't
 350            already done?  */
 351         return -1;
 352       wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 353       assert (mbsinit (&ps));
 354       (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
 355
 356       assert (mbsinit (&ps));
 357       n = mbsrtowcs (NULL, &string, 0, &ps);
 358       if (__builtin_expect (n, 0) == (size_t) -1)
 359         /* Something wrong.
 360            XXX Do we have to set `errno' to something which mbsrtows hasn't
 361            already done?  */
 362         return -1;
 363       wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 364       assert (mbsinit (&ps));
 365       (void) mbsrtowcs (wstring, &string, n + 1, &ps);
 366
 367       return internal_fnwmatch (wpattern, wstring, wstring + n,
 368                                 flags & FNM_PERIOD, flags);
 369     }
 370 # endif  /* mbstate_t and mbsrtowcs or _LIBC.  */
 371
 372   return internal_fnmatch (pattern, string, string + strlen (string),
 373                            flags & FNM_PERIOD, flags);
 374 }
 375
 376 # ifdef _LIBC
 377 #  undef fnmatch
 378 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
 379 #  if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
 380 strong_alias (__fnmatch, __fnmatch_old)
 381 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
 382 #  endif
 383 # endif
 384
 385 #endif  /* _LIBC or not __GNU_LIBRARY__.  */