posix/fnmatch.c

   1 /* Copyright (C) 1991-2014 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3
   4    The GNU C Library is free software; you can redistribute it and/or
   5    modify it under the terms of the GNU Lesser General Public
   6    License as published by the Free Software Foundation; either
   7    version 2.1 of the License, or (at your option) any later version.
   8
   9    The GNU C 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    Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser General Public
  15    License along with the GNU C Library; if not, see
  16    <http://www.gnu.org/licenses/>.  */
  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 #include <assert.h>
  28 #include <errno.h>
  29 #include <fnmatch.h>
  30 #include <ctype.h>
  31
  32 #if HAVE_STRING_H || defined _LIBC
  33 # include <string.h>
  34 #else
  35 # include <strings.h>
  36 #endif
  37
  38 #if defined STDC_HEADERS || defined _LIBC
  39 # include <stdlib.h>
  40 #endif
  41
  42 #ifdef _LIBC
  43 # include <alloca.h>
  44 #else
  45 # define alloca_account(size., var) alloca (size)
  46 #endif
  47
  48 /* For platform which support the ISO C amendement 1 functionality we
  49    support user defined character classes.  */
  50 #if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
  51 /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>.  */
  52 # include <wchar.h>
  53 # include <wctype.h>
  54 #endif
  55
  56 /* We need some of the locale data (the collation sequence information)
  57    but there is no interface to get this information in general.  Therefore
  58    we support a correct implementation only in glibc.  */
  59 #ifdef _LIBC
  60 # include "../locale/localeinfo.h"
  61 # include "../locale/elem-hash.h"
  62 # include "../locale/coll-lookup.h"
  63 # include <shlib-compat.h>
  64
  65 # define CONCAT(a,b) __CONCAT(a,b)
  66 # define mbsrtowcs __mbsrtowcs
  67 # define fnmatch __fnmatch
  68 extern int fnmatch (const char *pattern, const char *string, int flags);
  69 #endif
  70
  71 /* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set.  */
  72 #define NO_LEADING_PERIOD(flags) \
  73   ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
  74
  75 /* Comment out all this code if we are using the GNU C Library, and are not
  76    actually compiling the library itself.  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__
  84
  85
  86 # if defined STDC_HEADERS || !defined isascii
  87 #  define ISASCII(c) 1
  88 # else
  89 #  define ISASCII(c) isascii(c)
  90 # endif
  91
  92 # ifdef isblank
  93 #  define ISBLANK(c) (ISASCII (c) && isblank (c))
  94 # else
  95 #  define ISBLANK(c) ((c) == ' ' || (c) == '\t')
  96 # endif
  97 # ifdef isgraph
  98 #  define ISGRAPH(c) (ISASCII (c) && isgraph (c))
  99 # else
 100 #  define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
 101 # endif
 102
 103 # define ISPRINT(c) (ISASCII (c) && isprint (c))
 104 # define ISDIGIT(c) (ISASCII (c) && isdigit (c))
 105 # define ISALNUM(c) (ISASCII (c) && isalnum (c))
 106 # define ISALPHA(c) (ISASCII (c) && isalpha (c))
 107 # define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
 108 # define ISLOWER(c) (ISASCII (c) && islower (c))
 109 # define ISPUNCT(c) (ISASCII (c) && ispunct (c))
 110 # define ISSPACE(c) (ISASCII (c) && isspace (c))
 111 # define ISUPPER(c) (ISASCII (c) && isupper (c))
 112 # define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
 113
 114 # define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
 115
 116 # if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
 117 /* The GNU C library provides support for user-defined character classes
 118    and the functions from ISO C amendement 1.  */
 119 #  ifdef CHARCLASS_NAME_MAX
 120 #   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
 121 #  else
 122 /* This shouldn't happen but some implementation might still have this
 123    problem.  Use a reasonable default value.  */
 124 #   define CHAR_CLASS_MAX_LENGTH 256
 125 #  endif
 126
 127 #  ifdef _LIBC
 128 #   define IS_CHAR_CLASS(string) __wctype (string)
 129 #  else
 130 #   define IS_CHAR_CLASS(string) wctype (string)
 131 #  endif
 132
 133 #  ifdef _LIBC
 134 #   define ISWCTYPE(WC, WT)     __iswctype (WC, WT)
 135 #  else
 136 #   define ISWCTYPE(WC, WT)     iswctype (WC, WT)
 137 #  endif
 138
 139 #  if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
 140 /* In this case we are implementing the multibyte character handling.  */
 141 #   define HANDLE_MULTIBYTE     1
 142 #  endif
 143
 144 # else
 145 #  define CHAR_CLASS_MAX_LENGTH  6 /* Namely, `xdigit'.  */
 146
 147 #  define IS_CHAR_CLASS(string)                                               \
 148    (STREQ (string, "alpha") || STREQ (string, "upper")                        \
 149     || STREQ (string, "lower") || STREQ (string, "digit")                     \
 150     || STREQ (string, "alnum") || STREQ (string, "xdigit")                    \
 151     || STREQ (string, "space") || STREQ (string, "print")                     \
 152     || STREQ (string, "punct") || STREQ (string, "graph")                     \
 153     || STREQ (string, "cntrl") || STREQ (string, "blank"))
 154 # endif
 155
 156 /* Avoid depending on library functions or files
 157    whose names are inconsistent.  */
 158
 159 # if !defined _LIBC && !defined getenv
 160 extern char *getenv ();
 161 # endif
 162
 163 # ifndef errno
 164 extern int errno;
 165 # endif
 166
 167 /* Global variable.  */
 168 static int posixly_correct;
 169
 170 /* This function doesn't exist on most systems.  */
 171
 172 # if !defined HAVE___STRCHRNUL && !defined _LIBC
 173 static char *
 174 __strchrnul (s, c)
 175      const char *s;
 176      int c;
 177 {
 178   char *result = strchr (s, c);
 179   if (result == NULL)
 180     result = strchr (s, '\0');
 181   return result;
 182 }
 183 # endif
 184
 185 # if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
 186 static wchar_t *
 187 __wcschrnul (s, c)
 188      const wchar_t *s;
 189      wint_t c;
 190 {
 191   wchar_t *result = wcschr (s, c);
 192   if (result == NULL)
 193     result = wcschr (s, '\0');
 194   return result;
 195 }
 196 # endif
 197
 198 # ifndef internal_function
 199 /* Inside GNU libc we mark some function in a special way.  In other
 200    environments simply ignore the marking.  */
 201 #  define internal_function
 202 # endif
 203
 204 /* Note that this evaluates C many times.  */
 205 # ifdef _LIBC
 206 #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
 207 # else
 208 #  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
 209 # endif
 210 # define CHAR   char
 211 # define UCHAR  unsigned char
 212 # define INT    int
 213 # define FCT    internal_fnmatch
 214 # define EXT    ext_match
 215 # define END    end_pattern
 216 # define STRUCT fnmatch_struct
 217 # define L(CS)  CS
 218 # ifdef _LIBC
 219 #  define BTOWC(C)      __btowc (C)
 220 # else
 221 #  define BTOWC(C)      btowc (C)
 222 # endif
 223 # define STRLEN(S) strlen (S)
 224 # define STRCAT(D, S) strcat (D, S)
 225 # define MEMPCPY(D, S, N) __mempcpy (D, S, N)
 226 # define MEMCHR(S, C, N) memchr (S, C, N)
 227 # define STRCOLL(S1, S2) strcoll (S1, S2)
 228 # include "fnmatch_loop.c"
 229
 230
 231 # if HANDLE_MULTIBYTE
 232 /* Note that this evaluates C many times.  */
 233 #  ifdef _LIBC
 234 #   define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
 235 #  else
 236 #   define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
 237 #  endif
 238 #  define CHAR  wchar_t
 239 #  define UCHAR wint_t
 240 #  define INT   wint_t
 241 #  define FCT   internal_fnwmatch
 242 #  define EXT   ext_wmatch
 243 #  define END   end_wpattern
 244 #  define STRUCT fnwmatch_struct
 245 #  define L(CS) L##CS
 246 #  define BTOWC(C)      (C)
 247 #  define STRLEN(S) __wcslen (S)
 248 #  define STRCAT(D, S) __wcscat (D, S)
 249 #  define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
 250 #  define MEMCHR(S, C, N) wmemchr (S, C, N)
 251 #  define STRCOLL(S1, S2) wcscoll (S1, S2)
 252 #  define WIDE_CHAR_VERSION 1
 253
 254 #  undef IS_CHAR_CLASS
 255 /* We have to convert the wide character string in a multibyte string.  But
 256    we know that the character class names consist of alphanumeric characters
 257    from the portable character set, and since the wide character encoding
 258    for a member of the portable character set is the same code point as
 259    its single-byte encoding, we can use a simplified method to convert the
 260    string to a multibyte character string.  */
 261 static wctype_t
 262 is_char_class (const wchar_t *wcs)
 263 {
 264   char s[CHAR_CLASS_MAX_LENGTH + 1];
 265   char *cp = s;
 266
 267   do
 268     {
 269       /* Test for a printable character from the portable character set.  */
 270 #  ifdef _LIBC
 271       if (*wcs < 0x20 || *wcs > 0x7e
 272           || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
 273         return (wctype_t) 0;
 274 #  else
 275       switch (*wcs)
 276         {
 277         case L' ': case L'!': case L'"': case L'#': case L'%':
 278         case L'&': case L'\'': case L'(': case L')': case L'*':
 279         case L'+': case L',': case L'-': case L'.': case L'/':
 280         case L'0': case L'1': case L'2': case L'3': case L'4':
 281         case L'5': case L'6': case L'7': case L'8': case L'9':
 282         case L':': case L';': case L'<': case L'=': case L'>':
 283         case L'?':
 284         case L'A': case L'B': case L'C': case L'D': case L'E':
 285         case L'F': case L'G': case L'H': case L'I': case L'J':
 286         case L'K': case L'L': case L'M': case L'N': case L'O':
 287         case L'P': case L'Q': case L'R': case L'S': case L'T':
 288         case L'U': case L'V': case L'W': case L'X': case L'Y':
 289         case L'Z':
 290         case L'[': case L'\\': case L']': case L'^': case L'_':
 291         case L'a': case L'b': case L'c': case L'd': case L'e':
 292         case L'f': case L'g': case L'h': case L'i': case L'j':
 293         case L'k': case L'l': case L'm': case L'n': case L'o':
 294         case L'p': case L'q': case L'r': case L's': case L't':
 295         case L'u': case L'v': case L'w': case L'x': case L'y':
 296         case L'z': case L'{': case L'|': case L'}': case L'~':
 297           break;
 298         default:
 299           return (wctype_t) 0;
 300         }
 301 #  endif
 302
 303       /* Avoid overrunning the buffer.  */
 304       if (cp == s + CHAR_CLASS_MAX_LENGTH)
 305         return (wctype_t) 0;
 306
 307       *cp++ = (char) *wcs++;
 308     }
 309   while (*wcs != L'\0');
 310
 311   *cp = '\0';
 312
 313 #  ifdef _LIBC
 314   return __wctype (s);
 315 #  else
 316   return wctype (s);
 317 #  endif
 318 }
 319 #  define IS_CHAR_CLASS(string) is_char_class (string)
 320
 321 #  include "fnmatch_loop.c"
 322 # endif
 323
 324
 325 int
 326 fnmatch (pattern, string, flags)
 327      const char *pattern;
 328      const char *string;
 329      int flags;
 330 {
 331 # if HANDLE_MULTIBYTE
 332   if (__builtin_expect (MB_CUR_MAX, 1) != 1)
 333     {
 334       mbstate_t ps;
 335       size_t n;
 336       const char *p;
 337       wchar_t *wpattern_malloc = NULL;
 338       wchar_t *wpattern;
 339       wchar_t *wstring_malloc = NULL;
 340       wchar_t *wstring;
 341       size_t alloca_used = 0;
 342
 343       /* Convert the strings into wide characters.  */
 344       memset (&ps, '\0', sizeof (ps));
 345       p = pattern;
 346 #ifdef _LIBC
 347       n = strnlen (pattern, 1024);
 348 #else
 349       n = strlen (pattern);
 350 #endif
 351       if (__glibc_likely (n < 1024))
 352         {
 353           wpattern = (wchar_t *) alloca_account ((n + 1) * sizeof (wchar_t),
 354                                                  alloca_used);
 355           n = mbsrtowcs (wpattern, &p, n + 1, &ps);
 356           if (__glibc_unlikely (n == (size_t) -1))
 357             /* Something wrong.
 358                XXX Do we have to set `errno' to something which mbsrtows hasn't
 359                already done?  */
 360             return -1;
 361           if (p)
 362             {
 363               memset (&ps, '\0', sizeof (ps));
 364               goto prepare_wpattern;
 365             }
 366         }
 367       else
 368         {
 369         prepare_wpattern:
 370           n = mbsrtowcs (NULL, &pattern, 0, &ps);
 371           if (__glibc_unlikely (n == (size_t) -1))
 372             /* Something wrong.
 373                XXX Do we have to set `errno' to something which mbsrtows hasn't
 374                already done?  */
 375             return -1;
 376           if (__glibc_unlikely (n >= (size_t) -1 / sizeof (wchar_t)))
 377             {
 378               __set_errno (ENOMEM);
 379               return -2;
 380             }
 381           wpattern_malloc = wpattern
 382             = (wchar_t *) malloc ((n + 1) * sizeof (wchar_t));
 383           assert (mbsinit (&ps));
 384           if (wpattern == NULL)
 385             return -2;
 386           (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
 387         }
 388
 389       assert (mbsinit (&ps));
 390 #ifdef _LIBC
 391       n = strnlen (string, 1024);
 392 #else
 393       n = strlen (string);
 394 #endif
 395       p = string;
 396       if (__glibc_likely (n < 1024))
 397         {
 398           wstring = (wchar_t *) alloca_account ((n + 1) * sizeof (wchar_t),
 399                                                 alloca_used);
 400           n = mbsrtowcs (wstring, &p, n + 1, &ps);
 401           if (__glibc_unlikely (n == (size_t) -1))
 402             {
 403               /* Something wrong.
 404                  XXX Do we have to set `errno' to something which
 405                  mbsrtows hasn't already done?  */
 406             free_return:
 407               free (wpattern_malloc);
 408               return -1;
 409             }
 410           if (p)
 411             {
 412               memset (&ps, '\0', sizeof (ps));
 413               goto prepare_wstring;
 414             }
 415         }
 416       else
 417         {
 418         prepare_wstring:
 419           n = mbsrtowcs (NULL, &string, 0, &ps);
 420           if (__glibc_unlikely (n == (size_t) -1))
 421             /* Something wrong.
 422                XXX Do we have to set `errno' to something which mbsrtows hasn't
 423                already done?  */
 424             goto free_return;
 425           if (__glibc_unlikely (n >= (size_t) -1 / sizeof (wchar_t)))
 426             {
 427               free (wpattern_malloc);
 428               __set_errno (ENOMEM);
 429               return -2;
 430             }
 431
 432           wstring_malloc = wstring
 433             = (wchar_t *) malloc ((n + 1) * sizeof (wchar_t));
 434           if (wstring == NULL)
 435             {
 436               free (wpattern_malloc);
 437               return -2;
 438             }
 439           assert (mbsinit (&ps));
 440           (void) mbsrtowcs (wstring, &string, n + 1, &ps);
 441         }
 442
 443       int res = internal_fnwmatch (wpattern, wstring, wstring + n,
 444                                    flags & FNM_PERIOD, flags, NULL,
 445                                    alloca_used);
 446
 447       free (wstring_malloc);
 448       free (wpattern_malloc);
 449
 450       return res;
 451     }
 452 # endif  /* mbstate_t and mbsrtowcs or _LIBC.  */
 453
 454   return internal_fnmatch (pattern, string, string + strlen (string),
 455                            flags & FNM_PERIOD, flags, NULL, 0);
 456 }
 457
 458 # ifdef _LIBC
 459 #  undef fnmatch
 460 versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
 461 #  if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
 462 strong_alias (__fnmatch, __fnmatch_old)
 463 compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
 464 #  endif
 465 libc_hidden_ver (__fnmatch, fnmatch)
 466 # endif
 467
 468 #endif  /* _LIBC or not __GNU_LIBRARY__.  */