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
  57 # define CONCAT(a,b) __CONCAT(a,b)
  58 #endif
  59
  60 /* Comment out all this code if we are using the GNU C Library, and are not
  61    actually compiling the library itself.  This code is part of the GNU C
  62    Library, but also included in many other GNU distributions.  Compiling
  63    and linking in this code is a waste when using the GNU C library
  64    (especially if it is a shared library).  Rather than having every GNU
  65    program understand `configure --with-gnu-libc' and omit the object files,
  66    it is simpler to just do this in the source for each such file.  */
  67
  68 #if defined _LIBC || !defined __GNU_LIBRARY__
  69
  70
  71 # if defined STDC_HEADERS || !defined isascii
  72 #  define ISASCII(c) 1
  73 # else
  74 #  define ISASCII(c) isascii(c)
  75 # endif
  76
  77 # ifdef isblank
  78 #  define ISBLANK(c) (ISASCII (c) && isblank (c))
  79 # else
  80 #  define ISBLANK(c) ((c) == ' ' || (c) == '\t')
  81 # endif
  82 # ifdef isgraph
  83 #  define ISGRAPH(c) (ISASCII (c) && isgraph (c))
  84 # else
  85 #  define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
  86 # endif
  87
  88 # define ISPRINT(c) (ISASCII (c) && isprint (c))
  89 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
  90 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
  91 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
  92 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
  93 # define ISLOWER(c) (ISASCII (c) && islower (c))
  94 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
  95 # define ISSPACE(c) (ISASCII (c) && isspace (c))
  96 # define ISUPPER(c) (ISASCII (c) && isupper (c))
  97 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
  98
  99 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
 100
 101 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
 102 /* The GNU C library provides support for user-defined character classes
 103    and the functions from ISO C amendement 1.  */
 104 #  ifdef CHARCLASS_NAME_MAX
 105 #   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
 106 #  else
 107 /* This shouldn't happen but some implementation might still have this
 108    problem.  Use a reasonable default value.  */
 109 #   define CHAR_CLASS_MAX_LENGTH 256
 110 #  endif
 111
 112 #  ifdef _LIBC
 113 #   define IS_CHAR_CLASS(string) __wctype (string)
 114 #  else
 115 #   define IS_CHAR_CLASS(string) wctype (string)
 116 #  endif
 117
 118 #  ifdef _LIBC
 119 #   define ISWCTYPE(WC, WT)     __iswctype (WC, WT)
 120 #  else
 121 #   define ISWCTYPE(WC, WT)     iswctype (WC, WT)
 122 #  endif
 123
 124 #  if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
 125 /* In this case we are implementing the multibyte character handling.  */
 126 #   define HANDLE_MULTIBYTE     1
 127 #  endif
 128
 129 # else
 130 #  define CHAR_CLASS_MAX_LENGTH  6 /* Namely, `xdigit'.  */
 131
 132 #  define IS_CHAR_CLASS(string)                                               \
 133    (STREQ (string, "alpha") || STREQ (string, "upper")                        \
 134     || STREQ (string, "lower") || STREQ (string, "digit")                     \
 135     || STREQ (string, "alnum") || STREQ (string, "xdigit")                    \
 136     || STREQ (string, "space") || STREQ (string, "print")                     \
 137     || STREQ (string, "punct") || STREQ (string, "graph")                     \
 138     || STREQ (string, "cntrl") || STREQ (string, "blank"))
 139 # endif
 140
 141 /* Avoid depending on library functions or files
 142    whose names are inconsistent.  */
 143
 144 # if !defined _LIBC && !defined getenv
 145 extern char *getenv ();
 146 # endif
 147
 148 # ifndef errno
 149 extern int errno;
 150 # endif
 151
 152 /* This function doesn't exist on most systems.  */
 153
 154 # if !defined HAVE___STRCHRNUL && !defined _LIBC
 155 static char *
 156 __strchrnul (s, c)
 157      const char *s;
 158      int c;
 159 {
 160   char *result = strchr (s, c);
 161   if (result == NULL)
 162     result = strchr (s, '\0');
 163   return result;
 164 }
 165 # endif
 166
 167 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
 168 static wchar_t *
 169 __wcschrnul (s, c)
 170      const wchar_t *s;
 171      wint_t c;
 172 {
 173   wchar_t *result = wcschr (s, c);
 174   if (result == NULL)
 175     result = wcschr (s, '\0');
 176   return result;
 177 }
 178 # endif
 179
 180 # ifndef internal_function
 181 /* Inside GNU libc we mark some function in a special way.  In other
 182    environments simply ignore the marking.  */
 183 #  define internal_function
 184 # endif
 185
 186 /* Note that this evaluates C many times.  */
 187 # ifdef _LIBC
 188 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
 189 # else
 190 #  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
 191 # endif
 192 # define CHAR   char
 193 # define UCHAR  unsigned char
 194 # define FCT    internal_fnmatch
 195 # define L(CS)  CS
 196 # ifdef _LIBC
 197 #  define BTOWC(C)      __btowc (C)
 198 # else
 199 #  define BTOWC(C)      btowc (C)
 200 # endif
 201 # define STRCHR(S, C)   strchr (S, C)
 202 # define STRCHRNUL(S, C) __strchrnul (S, C)
 203 # define STRCOLL(S1, S2) strcoll (S1, S2)
 204 # define SUFFIX MB
 205 # include "fnmatch_loop.c"
 206
 207
 208 # if HANDLE_MULTIBYTE
 209 /* Note that this evaluates C many times.  */
 210 #  ifdef _LIBC
 211 #   define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
 212 #  else
 213 #   define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
 214 #  endif
 215 #  define CHAR  wchar_t
 216 #  define UCHAR wint_t
 217 #  define FCT   internal_fnwmatch
 218 #  define L(CS) L##CS
 219 #  define BTOWC(C)      (C)
 220 #  define STRCHR(S, C)  wcschr (S, C)
 221 #  define STRCHRNUL(S, C) __wcschrnul (S, C)
 222 #  define STRCOLL(S1, S2) wcscoll (S1, S2)
 223 #  define SUFFIX WC
 224 #  define WIDE_CHAR_VERSION 1
 225
 226
 227 #  undef IS_CHAR_CLASS
 228 #  ifdef _LIBC
 229 /* We have to convert the wide character string in a multibyte string.  But
 230    we know that the character class names are ASCII strings and since the
 231    internal wide character encoding is UCS4 we can use a simplified method
 232    to convert the string to a multibyte character string.  */
 233 static wctype_t
 234 is_char_class (const wchar_t *wcs)
 235 {
 236   char s[CHAR_CLASS_MAX_LENGTH + 1];
 237   char *cp = s;
 238
 239   do
 240     {
 241       if (*wcs < 0x20 || *wcs >= 0x7f)
 242         return 0;
 243
 244       *cp++ = (char) *wcs;
 245     }
 246   while (*wcs++ != L'\0');
 247
 248   return __wctype (s);
 249 }
 250 #  else
 251 /* Since we cannot assume anything about the internal encoding we have to
 252    convert the string back to multibyte representation the hard way.  */
 253 static wctype_t
 254 is_char_class (const wchar_t *wcs)
 255 {
 256   mbstate_t ps;
 257   const wchar_t *pwc;
 258   char *s;
 259   size_t n;
 260
 261   memset (&ps, '\0', sizeof (ps));
 262
 263   pwc = wcs;
 264   n = wcsrtombs (NULL, &pwc, 0, &ps);
 265   if (n == (size_t) -1)
 266     /* Something went wrong.  */
 267     return 0;
 268
 269   s = alloca (n + 1);
 270   assert (mbsinit (&ps));
 271   pwc = wcs;
 272   (void) wcsrtombs (s, &pwc, n + 1, &ps);
 273
 274   return wctype (s);
 275 }
 276 #  endif
 277 #  define IS_CHAR_CLASS(string) is_char_class (string)
 278
 279 #  include "fnmatch_loop.c"
 280 # endif
 281
 282 int
 283 fnmatch (pattern, string, flags)
 284      const char *pattern;
 285      const char *string;
 286      int flags;
 287 {
 288 # if HANDLE_MULTIBYTE
 289   mbstate_t ps;
 290   size_t n;
 291   wchar_t *wpattern;
 292   wchar_t *wstring;
 293
 294   if (MB_CUR_MAX == 1)
 295     /* This is an optimization for 8-bit character set.  */
 296     return internal_fnmatch (pattern, string, flags & FNM_PERIOD, flags);
 297
 298   /* Convert the strings into wide characters.  */
 299   memset (&ps, '\0', sizeof (ps));
 300   n = mbsrtowcs (NULL, &pattern, 0, &ps);
 301   if (n == (size_t) -1)
 302     /* Something wrong.
 303        XXX Do we have to set `errno' to something which mbsrtows hasn't
 304        already done?  */
 305     return -1;
 306   wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 307   assert (mbsinit (&ps));
 308   (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
 309
 310   assert (mbsinit (&ps));
 311   n = mbsrtowcs (NULL, &string, 0, &ps);
 312   if (n == (size_t) -1)
 313     /* Something wrong.
 314        XXX Do we have to set `errno' to something which mbsrtows hasn't
 315        already done?  */
 316     return -1;
 317   wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
 318   assert (mbsinit (&ps));
 319   (void) mbsrtowcs (wstring, &string, n + 1, &ps);
 320
 321   return internal_fnwmatch (wpattern, wstring, flags & FNM_PERIOD, flags);
 322 # else
 323   return internal_fnmatch (pattern, string, flags & FNM_PERIOD, flags);
 324 # endif  /* mbstate_t and mbsrtowcs or _LIBC.  */
 325 }
 326
 327 #endif  /* _LIBC or not __GNU_LIBRARY__.  */