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