contrib/diffutils/lib/fnmatch.c

   1 /* Copyright (C) 1991-1993, 1996-2007, 2009-2013 Free Software Foundation, Inc.
   2
   3    This program is free software; you can redistribute it and/or modify
   4    it under the terms of the GNU General Public License as published by
   5    the Free Software Foundation; either version 3, or (at your option)
   6    any later version.
   7
   8    This program is distributed in the hope that it will be useful,
   9    but WITHOUT ANY WARRANTY; without even the implied warranty of
  10    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  11    GNU General Public License for more details.
  12
  13    You should have received a copy of the GNU General Public License
  14    along with this program; if not, see <http://www.gnu.org/licenses/>.  */
  15
  16 #ifndef _LIBC
  17 # include <config.h>
  18 #endif
  19
  20 /* Enable GNU extensions in fnmatch.h.  */
  21 #ifndef _GNU_SOURCE
  22 # define _GNU_SOURCE    1
  23 #endif
  24
  25 #if ! defined __builtin_expect && __GNUC__ < 3
  26 # define __builtin_expect(expr, expected) (expr)
  27 #endif
  28
  29 #include <fnmatch.h>
  30
  31 #include <alloca.h>
  32 #include <assert.h>
  33 #include <ctype.h>
  34 #include <errno.h>
  35 #include <stddef.h>
  36 #include <stdbool.h>
  37 #include <stdlib.h>
  38 #include <string.h>
  39
  40 #define WIDE_CHAR_SUPPORT \
  41   (HAVE_WCTYPE_H && HAVE_BTOWC && HAVE_ISWCTYPE \
  42    && HAVE_WMEMCHR && (HAVE_WMEMCPY || HAVE_WMEMPCPY))
  43
  44 /* For platform which support the ISO C amendment 1 functionality we
  45    support user defined character classes.  */
  46 #if defined _LIBC || WIDE_CHAR_SUPPORT
  47 # include <wctype.h>
  48 # include <wchar.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 # include <shlib-compat.h>
  59
  60 # define CONCAT(a,b) __CONCAT(a,b)
  61 # define mbsrtowcs __mbsrtowcs
  62 # define fnmatch __fnmatch
  63 extern int fnmatch (const char *pattern, const char *string, int flags);
  64 #endif
  65
  66 #ifndef SIZE_MAX
  67 # define SIZE_MAX ((size_t) -1)
  68 #endif
  69
  70 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set.  */
  71 #define NO_LEADING_PERIOD(flags) \
  72   ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
  73
  74 /* Comment out all this code if we are using the GNU C Library, and are not
  75    actually compiling the library itself, and have not detected a bug
  76    in the library.  This code is part of the GNU C
  77    Library, but also included in many other GNU distributions.  Compiling
  78    and linking in this code is a waste when using the GNU C library
  79    (especially if it is a shared library).  Rather than having every GNU
  80    program understand 'configure --with-gnu-libc' and omit the object files,
  81    it is simpler to just do this in the source for each such file.  */
  82
  83 #if defined _LIBC || !defined __GNU_LIBRARY__ || !HAVE_FNMATCH_GNU
  84
  85
  86 # if ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
  87 #  define isblank(c) ((c) == ' ' || (c) == '\t')
  88 # endif
  89
  90 # define STREQ(s1, s2) (strcmp (s1, s2) == 0)
  91
  92 # if defined _LIBC || WIDE_CHAR_SUPPORT
  93 /* The GNU C library provides support for user-defined character classes
  94    and the functions from ISO C amendment 1.  */
  95 #  ifdef CHARCLASS_NAME_MAX
  96 #   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
  97 #  else
  98 /* This shouldn't happen but some implementation might still have this
  99    problem.  Use a reasonable default value.  */
 100 #   define CHAR_CLASS_MAX_LENGTH 256
 101 #  endif
 102
 103 #  ifdef _LIBC
 104 #   define IS_CHAR_CLASS(string) __wctype (string)
 105 #  else
 106 #   define IS_CHAR_CLASS(string) wctype (string)
 107 #  endif
 108
 109 #  ifdef _LIBC
 110 #   define ISWCTYPE(WC, WT)     __iswctype (WC, WT)
 111 #  else
 112 #   define ISWCTYPE(WC, WT)     iswctype (WC, WT)
 113 #  endif
 114
 115 #  if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
 116 /* In this case we are implementing the multibyte character handling.  */
 117 #   define HANDLE_MULTIBYTE     1
 118 #  endif
 119
 120 # else
 121 #  define CHAR_CLASS_MAX_LENGTH  6 /* Namely, 'xdigit'.  */
 122
 123 #  define IS_CHAR_CLASS(string)                                               \
 124    (STREQ (string, "alpha") || STREQ (string, "upper")                        \
 125     || STREQ (string, "lower") || STREQ (string, "digit")                     \
 126     || STREQ (string, "alnum") || STREQ (string, "xdigit")                    \
 127     || STREQ (string, "space") || STREQ (string, "print")                     \
 128     || STREQ (string, "punct") || STREQ (string, "graph")                     \
 129     || STREQ (string, "cntrl") || STREQ (string, "blank"))
 130 # endif
 131
 132 /* Avoid depending on library functions or files
 133    whose names are inconsistent.  */
 134
 135 /* Global variable.  */
 136 static int posixly_correct;
 137
 138 # ifndef internal_function
 139 /* Inside GNU libc we mark some function in a special way.  In other
 140    environments simply ignore the marking.  */
 141 #  define internal_function
 142 # endif
 143
 144 /* Note that this evaluates C many times.  */
 145 # define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
 146 # define CHAR   char
 147 # define UCHAR  unsigned char
 148 # define INT    int
 149 # define FCT    internal_fnmatch
 150 # define EXT    ext_match
 151 # define END    end_pattern
 152 # define L_(CS) CS
 153 # ifdef _LIBC
 154 #  define BTOWC(C)      __btowc (C)
 155 # else
 156 #  define BTOWC(C)      btowc (C)
 157 # endif
 158 # define STRLEN(S) strlen (S)
 159 # define STRCAT(D, S) strcat (D, S)
 160 # ifdef _LIBC
 161 #  define MEMPCPY(D, S, N) __mempcpy (D, S, N)
 162 # else
 163 #  if HAVE_MEMPCPY
 164 #   define MEMPCPY(D, S, N) mempcpy (D, S, N)
 165 #  else
 166 #   define MEMPCPY(D, S, N) ((void *) ((char *) memcpy (D, S, N) + (N)))
 167 #  endif
 168 # endif
 169 # define MEMCHR(S, C, N) memchr (S, C, N)
 170 # include "fnmatch_loop.c"
 171
 172
 173 # if HANDLE_MULTIBYTE
 174 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
 175 #  define CHAR  wchar_t
 176 #  define UCHAR wint_t
 177 #  define INT   wint_t
 178 #  define FCT   internal_fnwmatch
 179 #  define EXT   ext_wmatch
 180 #  define END   end_wpattern
 181 #  define L_(CS)        L##CS
 182 #  define BTOWC(C)      (C)
 183 #  ifdef _LIBC
 184 #   define STRLEN(S) __wcslen (S)
 185 #   define STRCAT(D, S) __wcscat (D, S)
 186 #   define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
 187 #  else
 188 #   define STRLEN(S) wcslen (S)
 189 #   define STRCAT(D, S) wcscat (D, S)
 190 #   if HAVE_WMEMPCPY
 191 #    define MEMPCPY(D, S, N) wmempcpy (D, S, N)
 192 #   else
 193 #    define MEMPCPY(D, S, N) (wmemcpy (D, S, N) + (N))
 194 #   endif
 195 #  endif
 196 #  define MEMCHR(S, C, N) wmemchr (S, C, N)
 197 #  define WIDE_CHAR_VERSION 1
 198
 199 #  undef IS_CHAR_CLASS
 200 /* We have to convert the wide character string in a multibyte string.  But
 201    we know that the character class names consist of alphanumeric characters
 202    from the portable character set, and since the wide character encoding
 203    for a member of the portable character set is the same code point as
 204    its single-byte encoding, we can use a simplified method to convert the
 205    string to a multibyte character string.  */
 206 static wctype_t
 207 is_char_class (const wchar_t *wcs)
 208 {
 209   char s[CHAR_CLASS_MAX_LENGTH + 1];
 210   char *cp = s;
 211
 212   do
 213     {
 214       /* Test for a printable character from the portable character set.  */
 215 #  ifdef _LIBC
 216       if (*wcs < 0x20 || *wcs > 0x7e
 217           || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
 218         return (wctype_t) 0;
 219 #  else
 220       switch (*wcs)
 221         {
 222         case L' ': case L'!': case L'"': case L'#': case L'%':
 223         case L'&': case L'\'': case L'(': case L')': case L'*':
 224         case L'+': case L',': case L'-': case L'.': case L'/':
 225         case L'0': case L'1': case L'2': case L'3': case L'4':
 226         case L'5': case L'6': case L'7': case L'8': case L'9':
 227         case L':': case L';': case L'<': case L'=': case L'>':
 228         case L'?':
 229         case L'A': case L'B': case L'C': case L'D': case L'E':
 230         case L'F': case L'G': case L'H': case L'I': case L'J':
 231         case L'K': case L'L': case L'M': case L'N': case L'O':
 232         case L'P': case L'Q': case L'R': case L'S': case L'T':
 233         case L'U': case L'V': case L'W': case L'X': case L'Y':
 234         case L'Z':
 235         case L'[': case L'\\': case L']': case L'^': case L'_':
 236         case L'a': case L'b': case L'c': case L'd': case L'e':
 237         case L'f': case L'g': case L'h': case L'i': case L'j':
 238         case L'k': case L'l': case L'm': case L'n': case L'o':
 239         case L'p': case L'q': case L'r': case L's': case L't':
 240         case L'u': case L'v': case L'w': case L'x': case L'y':
 241         case L'z': case L'{': case L'|': case L'}': case L'~':
 242           break;
 243         default:
 244           return (wctype_t) 0;
 245         }
 246 #  endif
 247
 248       /* Avoid overrunning the buffer.  */
 249       if (cp == s + CHAR_CLASS_MAX_LENGTH)
 250         return (wctype_t) 0;
 251
 252       *cp++ = (char) *wcs++;
 253     }
 254   while (*wcs != L'\0');
 255
 256   *cp = '\0';
 257
 258 #  ifdef _LIBC
 259   return __wctype (s);
 260 #  else
 261   return wctype (s);
 262 #  endif
 263 }
 264 #  define IS_CHAR_CLASS(string) is_char_class (string)
 265
 266 #  include "fnmatch_loop.c"
 267 # endif
 268
 269
 270 int
 271 fnmatch (const char *pattern, const char *string, int flags)
 272 {
 273 # if HANDLE_MULTIBYTE
 274 #  define ALLOCA_LIMIT 2000
 275   if (__builtin_expect (MB_CUR_MAX, 1) != 1)
 276     {
 277       mbstate_t ps;
 278       size_t patsize;
 279       size_t strsize;
 280       size_t totsize;
 281       wchar_t *wpattern;
 282       wchar_t *wstring;
 283       int res;
 284
 285       /* Calculate the size needed to convert the strings to
 286          wide characters.  */
 287       memset (&ps, '\0', sizeof (ps));
 288       patsize = mbsrtowcs (NULL, &pattern, 0, &ps) + 1;
 289       if (__builtin_expect (patsize != 0, 1))
 290         {
 291           assert (mbsinit (&ps));
 292           strsize = mbsrtowcs (NULL, &string, 0, &ps) + 1;
 293           if (__builtin_expect (strsize != 0, 1))
 294             {
 295               assert (mbsinit (&ps));
 296               totsize = patsize + strsize;
 297               if (__builtin_expect (! (patsize <= totsize
 298                                        && totsize <= SIZE_MAX / sizeof (wchar_t)),
 299                                     0))
 300                 {
 301                   errno = ENOMEM;
 302                   return -1;
 303                 }
 304
 305               /* Allocate room for the wide characters.  */
 306               if (__builtin_expect (totsize < ALLOCA_LIMIT, 1))
 307                 wpattern = (wchar_t *) alloca (totsize * sizeof (wchar_t));
 308               else
 309                 {
 310                   wpattern = malloc (totsize * sizeof (wchar_t));
 311                   if (__builtin_expect (! wpattern, 0))
 312                     {
 313                       errno = ENOMEM;
 314                       return -1;
 315                     }
 316                 }
 317               wstring = wpattern + patsize;
 318
 319               /* Convert the strings into wide characters.  */
 320               mbsrtowcs (wpattern, &pattern, patsize, &ps);
 321               assert (mbsinit (&ps));
 322               mbsrtowcs (wstring, &string, strsize, &ps);
 323
 324               res = internal_fnwmatch (wpattern, wstring, wstring + strsize - 1,
 325                                        flags & FNM_PERIOD, flags);
 326
 327               if (__builtin_expect (! (totsize < ALLOCA_LIMIT), 0))
 328                 free (wpattern);
 329               return res;
 330             }
 331         }
 332     }
 333
 334 # endif /* HANDLE_MULTIBYTE */
 335
 336   return internal_fnmatch (pattern, string, string + strlen (string),
 337                            flags & FNM_PERIOD, flags);
 338 }
 339
 340 # ifdef _LIBC
 341 #  undef fnmatch
 342 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
 343 #  if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
 344 strong_alias (__fnmatch, __fnmatch_old)
 345 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
 346 #  endif
 347 libc_hidden_ver (__fnmatch, fnmatch)
 348 # endif
 349
 350 #endif  /* _LIBC or not __GNU_LIBRARY__.  */