posix/fnmatch.c

   1 /* Copyright (C) 1991-1993, 1996-1999, 2000 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3
   4    This library is free software; you can redistribute it and/or
   5    modify it under the terms of the GNU Library General Public License as
   6    published by the Free Software Foundation; either version 2 of the
   7    License, or (at your option) any later version.
   8
   9    This library 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 GNU
  12    Library General Public License for more details.
  13
  14    You should have received a copy of the GNU Library General Public
  15    License along with this library; see the file COPYING.LIB.  If not,
  16    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  17    Boston, MA 02111-1307, USA.  */
  18
  19 #if HAVE_CONFIG_H
  20 # include <config.h>
  21 #endif
  22
  23 /* Enable GNU extensions in fnmatch.h.  */
  24 #ifndef _GNU_SOURCE
  25 # define _GNU_SOURCE    1
  26 #endif
  27
  28 #include <assert.h>
  29 #include <errno.h>
  30 #include <fnmatch.h>
  31 #include <ctype.h>
  32
  33 #if HAVE_STRING_H || defined _LIBC
  34 # include <string.h>
  35 #else
  36 # include <strings.h>
  37 #endif
  38
  39 #if defined STDC_HEADERS || defined _LIBC
  40 # include <stdlib.h>
  41 #endif
  42
  43 /* For platform which support the ISO C amendement 1 functionality we
  44    support user defined character classes.  */
  45 #if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
  46 /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>.  */
  47 # include <wchar.h>
  48 # include <wctype.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
  59 # define CONCAT(a,b) __CONCAT(a,b)
  60 # define mbsinit __mbsinit
  61 # define mbsrtowcs __mbsrtowcs
  62 #endif
  63
  64 /* Comment out all this code if we are using the GNU C Library, and are not
  65    actually compiling the library itself.  This code is part of the GNU C
  66    Library, but also included in many other GNU distributions.  Compiling
  67    and linking in this code is a waste when using the GNU C library
  68    (especially if it is a shared library).  Rather than having every GNU
  69    program understand `configure --with-gnu-libc' and omit the object files,
  70    it is simpler to just do this in the source for each such file.  */
  71
  72 #if defined _LIBC || !defined __GNU_LIBRARY__
  73
  74
  75 # if defined STDC_HEADERS || !defined isascii
  76 #  define ISASCII(c) 1
  77 # else
  78 #  define ISASCII(c) isascii(c)
  79 # endif
  80
  81 # ifdef isblank
  82 #  define ISBLANK(c) (ISASCII (c) && isblank (c))
  83 # else
  84 #  define ISBLANK(c) ((c) == ' ' || (c) == '\t')
  85 # endif
  86 # ifdef isgraph
  87 #  define ISGRAPH(c) (ISASCII (c) && isgraph (c))
  88 # else
  89 #  define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
  90 # endif
  91
  92 # define ISPRINT(c) (ISASCII (c) && isprint (c))
  93 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
  94 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
  95 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
  96 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
  97 # define ISLOWER(c) (ISASCII (c) && islower (c))
  98 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
  99 # define ISSPACE(c) (ISASCII (c) && isspace (c))
 100 # define ISUPPER(c) (ISASCII (c) && isupper (c))
 101 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
 102
 103 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
 104
 105 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
 106 /* The GNU C library provides support for user-defined character classes
 107    and the functions from ISO C amendement 1.  */
 108 #  ifdef CHARCLASS_NAME_MAX
 109 #   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
 110 #  else
 111 /* This shouldn't happen but some implementation might still have this
 112    problem.  Use a reasonable default value.  */
 113 #   define CHAR_CLASS_MAX_LENGTH 256
 114 #  endif
 115
 116 #  ifdef _LIBC
 117 #   define IS_CHAR_CLASS(string) __wctype (string)
 118 #  else
 119 #   define IS_CHAR_CLASS(string) wctype (string)
 120 #  endif
 121
 122 #  ifdef _LIBC
 123 #   define ISWCTYPE(WC, WT)     __iswctype (WC, WT)
 124 #  else
 125 #   define ISWCTYPE(WC, WT)     iswctype (WC, WT)
 126 #  endif
 127
 128 #  if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
 129 /* In this case we are implementing the multibyte character handling.  */
 130 #   define HANDLE_MULTIBYTE     1
 131 #  endif
 132
 133 # else
 134 #  define CHAR_CLASS_MAX_LENGTH  6 /* Namely, `xdigit'.  */
 135
 136 #  define IS_CHAR_CLASS(string)                                               \
 137    (STREQ (string, "alpha") || STREQ (string, "upper")                        \
 138     || STREQ (string, "lower") || STREQ (string, "digit")                     \
 139     || STREQ (string, "alnum") || STREQ (string, "xdigit")                    \
 140     || STREQ (string, "space") || STREQ (string, "print")                     \
 141     || STREQ (string, "punct") || STREQ (string, "graph")                     \
 142     || STREQ (string, "cntrl") || STREQ (string, "blank"))
 143 # endif
 144
 145 /* Avoid depending on library functions or files
 146    whose names are inconsistent.  */
 147
 148 # if !defined _LIBC && !defined getenv
 149 extern char *getenv ();
 150 # endif
 151
 152 # ifndef errno
 153 extern int errno;
 154 # endif
 155
 156 /* This function doesn't exist on most systems.  */
 157
 158 # if !defined HAVE___STRCHRNUL && !defined _LIBC
 159 static char *
 160 __strchrnul (s, c)
 161      const char *s;
 162      int c;
 163 {
 164   char *result = strchr (s, c);
 165   if (result == NULL)
 166     result = strchr (s, '\0');
 167   return result;
 168 }
 169 # endif
 170
 171 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
 172 static wchar_t *
 173 __wcschrnul (s, c)
 174      const wchar_t *s;
 175      wint_t c;
 176 {
 177   wchar_t *result = wcschr (s, c);
 178   if (result == NULL)
 179     result = wcschr (s, '\0');
 180   return result;
 181 }
 182 # endif
 183
 184 # ifndef internal_function
 185 /* Inside GNU libc we mark some function in a special way.  In other
 186    environments simply ignore the marking.  */
 187 #  define internal_function
 188 # endif
 189
 190 /* Note that this evaluates C many times.  */
 191 # ifdef _LIBC
 192 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
 193 # else
 194 #  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
 195 # endif
 196 # define CHAR   char
 197 # define UCHAR  unsigned char
 198 # define FCT    internal_fnmatch
 199 # define L(CS)  CS
 200 # ifdef _LIBC
 201 #  define BTOWC(C)      __btowc (C)
 202 # else
 203 #  define BTOWC(C)      btowc (C)
 204 # endif
 205 # define STRCHR(S, C)   strchr (S, C)
 206 # define STRCHRNUL(S, C) __strchrnul (S, C)
 207 # define STRCOLL(S1, S2) strcoll (S1, S2)
 208 # include "fnmatch_loop.c"
 209
 210
 211 # if HANDLE_MULTIBYTE
 212 /* Note that this evaluates C many times.  */
 213 #  ifdef _LIBC
 214 #   define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
 215 #  else
 216 #   define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
 217 #  endif
 218 #  define CHAR  wchar_t
 219 #  define UCHAR wint_t
 220 #  define FCT   internal_fnwmatch
 221 #  define L(CS) L##CS
 222 #  define BTOWC(C)      (C)
 223 #  define STRCHR(S, C)  wcschr (S, C)
 224 #  define STRCHRNUL(S, C) __wcschrnul (S, C)
 225 #  define STRCOLL(S1, S2) wcscoll (S1, S2)
 226 #  define WIDE_CHAR_VERSION 1
 227
 228 #  undef IS_CHAR_CLASS
 229 /* We have to convert the wide character string in a multibyte string.  But
 230    we know that the character class names consist of alphanumeric characters
 231    from the portable character set, and since the wide character encoding
 232    for a member of the portable character set is the same code point as
 233    its single-byte encoding, we can use a simplified method to convert the
 234    string to a multibyte character string.  */
 235 static wctype_t
 236 is_char_class (const wchar_t *wcs)
 237 {
 238   char s[CHAR_CLASS_MAX_LENGTH + 1];
 239   char *cp = s;
 240
 241   do
 242     {
 243       /* Test for a printable character from the portable character set.  */
 244 #  ifdef _LIBC
 245       if (*wcs < 0x20 || *wcs > 0x7e
 246           || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
 247         return (wctype_t) 0;
 248 #  else
 249       switch (*wcs)
 250         {
 251         case L' ': case L'!': case L'"': case L'#': case L'%':
 252         case L'&': case L'\'': case L'(': case L')': case L'*':
 253         case L'+': case L',': case L'-': case L'.': case L'/':
 254         case L'0': case L'1': case L'2': case L'3': case L'4':
 255         case L'5': case L'6': case L'7': case L'8': case L'9':
 256         case L':': case L';': case L'<': case L'=': case L'>':
 257         case L'?':
 258         case L'A': case L'B': case L'C': case L'D': case L'E':
 259         case L'F': case L'G': case L'H': case L'I': case L'J':
 260         case L'K': case L'L': case L'M': case L'N': case L'O':
 261         case L'P': case L'Q': case L'R': case L'S': case L'T':
 262         case L'U': case L'V': case L'W': case L'X': case L'Y':
 263         case L'Z':
 264         case L'[': case L'\\': case L']': case L'^': case L'_':
 265         case L'a': case L'b': case L'c': case L'd': case L'e':
 266         case L'f': case L'g': case L'h': case L'i': case L'j':
 267         case L'k': case L'l': case L'm': case L'n': case L'o':
 268         case L'p': case L'q': case L'r': case L's': case L't':
 269         case L'u': case L'v': case L'w': case L'x': case L'y':
 270         case L'z': case L'{': case L'|': case L'}': case L'~':
 271           break;
 272         default:
 273           return (wctype_t) 0;
 274         }
 275 #  endif
 276
 277       /* Avoid overrunning the buffer.  */
 278       if (cp == s + CHAR_CLASS_MAX_LENGTH)
 279         return (wctype_t) 0;
 280
 281       *cp++ = (char) *wcs++;
 282     }
 283   while (*wcs != L'\0');
 284
 285   *cp = '\0';
 286
 287 #  ifdef _LIBC
 288   return __wctype (s);
 289 #  else
 290   return wctype (s);
 291 #  endif
 292 }
 293 #  define IS_CHAR_CLASS(string) is_char_class (string)
 294
 295 #  include "fnmatch_loop.c"
 296 # endif
 297
 298
 299 int
 300 fnmatch (pattern, string, flags)
 301      const char *pattern;
 302      const char *string;
 303      int flags;
 304 {
 305 # if HANDLE_MULTIBYTE
 306   mbstate_t ps;
 307   size_t n;
 308   wchar_t *wpattern;
 309   wchar_t *wstring;
 310
 311   if (MB_CUR_MAX == 1)
 312     /* This is an optimization for 8-bit character set.  */
 313     return internal_fnmatch (pattern, string, flags & FNM_PERIOD, flags);
 314
 315   /* Convert the strings into wide characters.  */
 316   memset (&ps, '\0', sizeof (ps));
 317   n = mbsrtowcs (NULL, &pattern, 0, &ps);
 318   if (n == (size_t) -1)
 319     /* Something wrong.
 320        XXX Do we have to set `errno' to something which mbsrtows hasn't
 321        already done?  */
 322     return -1;
 323   wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 324   assert (mbsinit (&ps));
 325   (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
 326
 327   assert (mbsinit (&ps));
 328   n = mbsrtowcs (NULL, &string, 0, &ps);
 329   if (n == (size_t) -1)
 330     /* Something wrong.
 331        XXX Do we have to set `errno' to something which mbsrtows hasn't
 332        already done?  */
 333     return -1;
 334   wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 335   assert (mbsinit (&ps));
 336   (void) mbsrtowcs (wstring, &string, n + 1, &ps);
 337
 338   return internal_fnwmatch (wpattern, wstring, flags & FNM_PERIOD, flags);
 339 # else
 340   return internal_fnmatch (pattern, string, flags & FNM_PERIOD, flags);
 341 # endif  /* mbstate_t and mbsrtowcs or _LIBC.  */
 342 }
 343
 344 #endif  /* _LIBC or not __GNU_LIBRARY__.  */