-/* Copyright (C) 1991, 92, 93, 96, 97, 98, 99 Free Software Foundation, Inc.
+/* Copyright (C) 1991-2015 Free Software Foundation, Inc.
This file is part of the GNU C Library.
- This library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Library General Public License as
- published by the Free Software Foundation; either version 2 of the
- License, or (at your option) any later version.
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
- This library is distributed in the hope that it will be useful,
+ The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- Library General Public License for more details.
+ Lesser General Public License for more details.
- You should have received a copy of the GNU Library General Public
- License along with this library; see the file COPYING.LIB. If not,
- write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
- Boston, MA 02111-1307, USA. */
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
#if HAVE_CONFIG_H
# include <config.h>
# define _GNU_SOURCE 1
#endif
+#include <assert.h>
#include <errno.h>
#include <fnmatch.h>
#include <ctype.h>
-
-#if HAVE_STRING_H || defined _LIBC
-# include <string.h>
-#else
-# include <strings.h>
-#endif
+#include <string.h>
#if defined STDC_HEADERS || defined _LIBC
# include <stdlib.h>
#endif
+#ifdef _LIBC
+# include <alloca.h>
+#else
+# define alloca_account(size., var) alloca (size)
+#endif
+
/* For platform which support the ISO C amendement 1 functionality we
support user defined character classes. */
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
# include <wctype.h>
#endif
+/* We need some of the locale data (the collation sequence information)
+ but there is no interface to get this information in general. Therefore
+ we support a correct implementation only in glibc. */
+#ifdef _LIBC
+# include "../locale/localeinfo.h"
+# include "../locale/elem-hash.h"
+# include "../locale/coll-lookup.h"
+# include <shlib-compat.h>
+
+# define CONCAT(a,b) __CONCAT(a,b)
+# define mbsrtowcs __mbsrtowcs
+# define fnmatch __fnmatch
+extern int fnmatch (const char *pattern, const char *string, int flags);
+#endif
+
+/* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */
+#define NO_LEADING_PERIOD(flags) \
+ ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
+
/* Comment out all this code if we are using the GNU C Library, and are not
actually compiling the library itself. This code is part of the GNU C
Library, but also included in many other GNU distributions. Compiling
# define ISASCII(c) isascii(c)
# endif
-#ifdef isblank
-# define ISBLANK(c) (ISASCII (c) && isblank (c))
-#else
-# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
-#endif
-#ifdef isgraph
-# define ISGRAPH(c) (ISASCII (c) && isgraph (c))
-#else
-# define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
-#endif
+# ifdef isblank
+# define ISBLANK(c) (ISASCII (c) && isblank (c))
+# else
+# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
+# endif
+# ifdef isgraph
+# define ISGRAPH(c) (ISASCII (c) && isgraph (c))
+# else
+# define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
+# endif
-#define ISPRINT(c) (ISASCII (c) && isprint (c))
-#define ISDIGIT(c) (ISASCII (c) && isdigit (c))
-#define ISALNUM(c) (ISASCII (c) && isalnum (c))
-#define ISALPHA(c) (ISASCII (c) && isalpha (c))
-#define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
-#define ISLOWER(c) (ISASCII (c) && islower (c))
-#define ISPUNCT(c) (ISASCII (c) && ispunct (c))
-#define ISSPACE(c) (ISASCII (c) && isspace (c))
-#define ISUPPER(c) (ISASCII (c) && isupper (c))
-#define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
+# define ISPRINT(c) (ISASCII (c) && isprint (c))
+# define ISDIGIT(c) (ISASCII (c) && isdigit (c))
+# define ISALNUM(c) (ISASCII (c) && isalnum (c))
+# define ISALPHA(c) (ISASCII (c) && isalpha (c))
+# define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
+# define ISLOWER(c) (ISASCII (c) && islower (c))
+# define ISPUNCT(c) (ISASCII (c) && ispunct (c))
+# define ISSPACE(c) (ISASCII (c) && isspace (c))
+# define ISUPPER(c) (ISASCII (c) && isupper (c))
+# define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
# define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
# else
# define IS_CHAR_CLASS(string) wctype (string)
# endif
+
+# ifdef _LIBC
+# define ISWCTYPE(WC, WT) __iswctype (WC, WT)
+# else
+# define ISWCTYPE(WC, WT) iswctype (WC, WT)
+# endif
+
+# if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
+/* In this case we are implementing the multibyte character handling. */
+# define HANDLE_MULTIBYTE 1
+# endif
+
# else
# define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
extern int errno;
# endif
-/* Match STRING against the filename pattern PATTERN, returning zero if
- it matches, nonzero if not. */
-int
-fnmatch (pattern, string, flags)
- const char *pattern;
- const char *string;
- int flags;
+/* Global variable. */
+static int posixly_correct;
+
+/* This function doesn't exist on most systems. */
+
+# if !defined HAVE___STRCHRNUL && !defined _LIBC
+static char *
+__strchrnul (s, c)
+ const char *s;
+ int c;
+{
+ char *result = strchr (s, c);
+ if (result == NULL)
+ result = strchr (s, '\0');
+ return result;
+}
+# endif
+
+# if HANDLE_MULTIBYTE && !defined HAVE___STRCHRNUL && !defined _LIBC
+static wchar_t *
+__wcschrnul (s, c)
+ const wchar_t *s;
+ wint_t c;
{
- register const char *p = pattern, *n = string;
- register unsigned char c;
+ wchar_t *result = wcschr (s, c);
+ if (result == NULL)
+ result = wcschr (s, '\0');
+ return result;
+}
+# endif
+
+# ifndef internal_function
+/* Inside GNU libc we mark some function in a special way. In other
+ environments simply ignore the marking. */
+# define internal_function
+# endif
/* Note that this evaluates C many times. */
# ifdef _LIBC
# else
# define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
# endif
+# define CHAR char
+# define UCHAR unsigned char
+# define INT int
+# define FCT internal_fnmatch
+# define EXT ext_match
+# define END end_pattern
+# define STRUCT fnmatch_struct
+# define L(CS) CS
+# ifdef _LIBC
+# define BTOWC(C) __btowc (C)
+# else
+# define BTOWC(C) btowc (C)
+# endif
+# define STRLEN(S) strlen (S)
+# define STRCAT(D, S) strcat (D, S)
+# define MEMPCPY(D, S, N) __mempcpy (D, S, N)
+# define MEMCHR(S, C, N) memchr (S, C, N)
+# define STRCOLL(S1, S2) strcoll (S1, S2)
+# define WIDE_CHAR_VERSION 0
+# include <locale/weight.h>
+# define FINDIDX findidx
+# include "fnmatch_loop.c"
+
+
+# if HANDLE_MULTIBYTE
+/* Note that this evaluates C many times. */
+# ifdef _LIBC
+# define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
+# else
+# define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? towlower (c) : (c))
+# endif
+# define CHAR wchar_t
+# define UCHAR wint_t
+# define INT wint_t
+# define FCT internal_fnwmatch
+# define EXT ext_wmatch
+# define END end_wpattern
+# define STRUCT fnwmatch_struct
+# define L(CS) L##CS
+# define BTOWC(C) (C)
+# define STRLEN(S) __wcslen (S)
+# define STRCAT(D, S) __wcscat (D, S)
+# define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
+# define MEMCHR(S, C, N) wmemchr (S, C, N)
+# define STRCOLL(S1, S2) wcscoll (S1, S2)
+# define WIDE_CHAR_VERSION 1
+/* Change the name the header defines so it doesn't conflict with
+ the <locale/weight.h> version included above. */
+# define findidx findidxwc
+# include <locale/weightwc.h>
+# undef findidx
+# define FINDIDX findidxwc
+
+# undef IS_CHAR_CLASS
+/* We have to convert the wide character string in a multibyte string. But
+ we know that the character class names consist of alphanumeric characters
+ from the portable character set, and since the wide character encoding
+ for a member of the portable character set is the same code point as
+ its single-byte encoding, we can use a simplified method to convert the
+ string to a multibyte character string. */
+static wctype_t
+is_char_class (const wchar_t *wcs)
+{
+ char s[CHAR_CLASS_MAX_LENGTH + 1];
+ char *cp = s;
- while ((c = *p++) != '\0')
+ do
{
- c = FOLD (c);
-
- switch (c)
+ /* Test for a printable character from the portable character set. */
+# ifdef _LIBC
+ if (*wcs < 0x20 || *wcs > 0x7e
+ || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
+ return (wctype_t) 0;
+# else
+ switch (*wcs)
{
- case '?':
- if (*n == '\0')
- return FNM_NOMATCH;
- else if (*n == '/' && (flags & FNM_FILE_NAME))
- return FNM_NOMATCH;
- else if (*n == '.' && (flags & FNM_PERIOD) &&
- (n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/')))
- return FNM_NOMATCH;
+ case L' ': case L'!': case L'"': case L'#': case L'%':
+ case L'&': case L'\'': case L'(': case L')': case L'*':
+ case L'+': case L',': case L'-': case L'.': case L'/':
+ case L'0': case L'1': case L'2': case L'3': case L'4':
+ case L'5': case L'6': case L'7': case L'8': case L'9':
+ case L':': case L';': case L'<': case L'=': case L'>':
+ case L'?':
+ case L'A': case L'B': case L'C': case L'D': case L'E':
+ case L'F': case L'G': case L'H': case L'I': case L'J':
+ case L'K': case L'L': case L'M': case L'N': case L'O':
+ case L'P': case L'Q': case L'R': case L'S': case L'T':
+ case L'U': case L'V': case L'W': case L'X': case L'Y':
+ case L'Z':
+ case L'[': case L'\\': case L']': case L'^': case L'_':
+ case L'a': case L'b': case L'c': case L'd': case L'e':
+ case L'f': case L'g': case L'h': case L'i': case L'j':
+ case L'k': case L'l': case L'm': case L'n': case L'o':
+ case L'p': case L'q': case L'r': case L's': case L't':
+ case L'u': case L'v': case L'w': case L'x': case L'y':
+ case L'z': case L'{': case L'|': case L'}': case L'~':
break;
+ default:
+ return (wctype_t) 0;
+ }
+# endif
- case '\\':
- if (!(flags & FNM_NOESCAPE))
- {
- c = *p++;
- if (c == '\0')
- /* Trailing \ loses. */
- return FNM_NOMATCH;
- c = FOLD (c);
- }
- if (FOLD ((unsigned char) *n) != c)
- return FNM_NOMATCH;
- break;
+ /* Avoid overrunning the buffer. */
+ if (cp == s + CHAR_CLASS_MAX_LENGTH)
+ return (wctype_t) 0;
- case '*':
- if (*n == '.' && (flags & FNM_PERIOD) &&
- (n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/')))
- return FNM_NOMATCH;
+ *cp++ = (char) *wcs++;
+ }
+ while (*wcs != L'\0');
- for (c = *p++; c == '?' || c == '*'; c = *p++)
- {
- if (*n == '/' && (flags & FNM_FILE_NAME))
- /* A slash does not match a wildcard under FNM_FILE_NAME. */
- return FNM_NOMATCH;
- else if (c == '?')
- {
- /* A ? needs to match one character. */
- if (*n == '\0')
- /* There isn't another character; no match. */
- return FNM_NOMATCH;
- else
- /* One character of the string is consumed in matching
- this ? wildcard, so *??? won't match if there are
- less than three characters. */
- ++n;
- }
- }
+ *cp = '\0';
- if (c == '\0')
- /* The wildcard(s) is/are the last element of the pattern.
- If the name is a file name and contains another slash
- this does mean it cannot match. */
- return ((flags & FNM_FILE_NAME) && strchr (n, '/') != NULL
- ? FNM_NOMATCH : 0);
- else
- {
- const char *endp;
-
- if (!(flags & FNM_FILE_NAME) || (endp = strchr (n, '/')) == NULL)
- endp = strchr (n, '\0');
-
- if (c == '[')
- {
- for (--p; n < endp; ++n)
- if (fnmatch (p, n, flags & ~FNM_PERIOD) == 0)
- return 0;
- }
- else
- {
- if (c == '\\' && !(flags & FNM_NOESCAPE))
- c = *p;
- c = FOLD (c);
- for (--p; n < endp; ++n)
- if (FOLD ((unsigned char) *n) == c
- && fnmatch (p, n, flags & ~FNM_PERIOD) == 0)
- return 0;
- }
- }
+# ifdef _LIBC
+ return __wctype (s);
+# else
+ return wctype (s);
+# endif
+}
+# define IS_CHAR_CLASS(string) is_char_class (string)
- /* If we come here no match is possible with the wildcard. */
- return FNM_NOMATCH;
-
- case '[':
- {
- /* Nonzero if the sense of the character class is inverted. */
- static int posixly_correct;
- register int not;
- char cold;
-
- if (posixly_correct == 0)
- posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;
-
- if (*n == '\0')
- return FNM_NOMATCH;
-
- if (*n == '.' && (flags & FNM_PERIOD) &&
- (n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/')))
- return FNM_NOMATCH;
-
- if (*n == '/' && (flags & FNM_FILE_NAME))
- /* `/' cannot be matched. */
- return FNM_NOMATCH;
-
- not = (*p == '!' || (posixly_correct < 0 && *p == '^'));
- if (not)
- ++p;
-
- c = *p++;
- for (;;)
- {
- unsigned char fn = FOLD ((unsigned char) *n);
-
- if (!(flags & FNM_NOESCAPE) && c == '\\')
- {
- if (*p == '\0')
- return FNM_NOMATCH;
- c = FOLD ((unsigned char) *p);
- ++p;
-
- if (c == fn)
- goto matched;
- }
- else if (c == '[' && *p == ':')
- {
- /* Leave room for the null. */
- char str[CHAR_CLASS_MAX_LENGTH + 1];
- size_t c1 = 0;
-# if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
- wctype_t wt;
+# include "fnmatch_loop.c"
# endif
- for (;;)
- {
- if (c1 == CHAR_CLASS_MAX_LENGTH)
- /* The name is too long and therefore the pattern
- is ill-formed. */
- return FNM_NOMATCH;
-
- c = *++p;
- if (c == ':' && p[1] == ']')
- {
- p += 2;
- break;
- }
- str[c1++] = 'c';
- }
- str[c1] = '\0';
-
-# if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
- wt = IS_CHAR_CLASS (str);
- if (wt == 0)
- /* Invalid character class name. */
- return FNM_NOMATCH;
- if (__iswctype (__btowc ((unsigned char) *n), wt))
- goto matched;
-# else
- if ((STREQ (str, "alnum") && ISALNUM ((unsigned char) *n))
- || (STREQ (str, "alpha") && ISALPHA ((unsigned char) *n))
- || (STREQ (str, "blank") && ISBLANK ((unsigned char) *n))
- || (STREQ (str, "cntrl") && ISCNTRL ((unsigned char) *n))
- || (STREQ (str, "digit") && ISDIGIT ((unsigned char) *n))
- || (STREQ (str, "graph") && ISGRAPH ((unsigned char) *n))
- || (STREQ (str, "lower") && ISLOWER ((unsigned char) *n))
- || (STREQ (str, "print") && ISPRINT ((unsigned char) *n))
- || (STREQ (str, "punct") && ISPUNCT ((unsigned char) *n))
- || (STREQ (str, "space") && ISSPACE ((unsigned char) *n))
- || (STREQ (str, "upper") && ISUPPER ((unsigned char) *n))
- || (STREQ (str, "xdigit") && ISXDIGIT ((unsigned char) *n)))
- goto matched;
-# endif
- }
- else if (c == '\0')
- /* [ (unterminated) loses. */
- return FNM_NOMATCH;
- else if (FOLD (c) == fn)
- goto matched;
-
- cold = c;
- c = *p++;
-
- if (c == '-' && *p != ']')
- {
- /* It is a range. */
- unsigned char cend = *p++;
- if (!(flags & FNM_NOESCAPE) && cend == '\\')
- cend = *p++;
- if (cend == '\0')
- return FNM_NOMATCH;
-
- if (cold <= fn && fn <= FOLD (cend))
- goto matched;
-
- c = *p++;
- }
- if (c == ']')
- break;
- }
-
- if (!not)
- return FNM_NOMATCH;
- break;
-
- matched:
- /* Skip the rest of the [...] that already matched. */
- while (c != ']')
- {
- if (c == '\0')
- /* [... (unterminated) loses. */
- return FNM_NOMATCH;
-
- c = *p++;
- if (!(flags & FNM_NOESCAPE) && c == '\\')
- {
- if (*p == '\0')
- return FNM_NOMATCH;
- /* XXX 1003.2d11 is unclear if this is right. */
- ++p;
- }
- else if (c == '[' && *p == ':')
- {
- do
- if (*++p == '\0')
- return FNM_NOMATCH;
- while (*p != ':' || p[1] == ']');
- p += 2;
- c = *p;
- }
- }
- if (not)
- return FNM_NOMATCH;
- }
- break;
+int
+fnmatch (pattern, string, flags)
+ const char *pattern;
+ const char *string;
+ int flags;
+{
+# if HANDLE_MULTIBYTE
+ if (__builtin_expect (MB_CUR_MAX, 1) != 1)
+ {
+ mbstate_t ps;
+ size_t n;
+ const char *p;
+ wchar_t *wpattern_malloc = NULL;
+ wchar_t *wpattern;
+ wchar_t *wstring_malloc = NULL;
+ wchar_t *wstring;
+ size_t alloca_used = 0;
+
+ /* Convert the strings into wide characters. */
+ memset (&ps, '\0', sizeof (ps));
+ p = pattern;
+#ifdef _LIBC
+ n = strnlen (pattern, 1024);
+#else
+ n = strlen (pattern);
+#endif
+ if (__glibc_likely (n < 1024))
+ {
+ wpattern = (wchar_t *) alloca_account ((n + 1) * sizeof (wchar_t),
+ alloca_used);
+ n = mbsrtowcs (wpattern, &p, n + 1, &ps);
+ if (__glibc_unlikely (n == (size_t) -1))
+ /* Something wrong.
+ XXX Do we have to set `errno' to something which mbsrtows hasn't
+ already done? */
+ return -1;
+ if (p)
+ {
+ memset (&ps, '\0', sizeof (ps));
+ goto prepare_wpattern;
+ }
+ }
+ else
+ {
+ prepare_wpattern:
+ n = mbsrtowcs (NULL, &pattern, 0, &ps);
+ if (__glibc_unlikely (n == (size_t) -1))
+ /* Something wrong.
+ XXX Do we have to set `errno' to something which mbsrtows hasn't
+ already done? */
+ return -1;
+ if (__glibc_unlikely (n >= (size_t) -1 / sizeof (wchar_t)))
+ {
+ __set_errno (ENOMEM);
+ return -2;
+ }
+ wpattern_malloc = wpattern
+ = (wchar_t *) malloc ((n + 1) * sizeof (wchar_t));
+ assert (mbsinit (&ps));
+ if (wpattern == NULL)
+ return -2;
+ (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
+ }
- default:
- if (c != FOLD ((unsigned char) *n))
- return FNM_NOMATCH;
+ assert (mbsinit (&ps));
+#ifdef _LIBC
+ n = strnlen (string, 1024);
+#else
+ n = strlen (string);
+#endif
+ p = string;
+ if (__glibc_likely (n < 1024))
+ {
+ wstring = (wchar_t *) alloca_account ((n + 1) * sizeof (wchar_t),
+ alloca_used);
+ n = mbsrtowcs (wstring, &p, n + 1, &ps);
+ if (__glibc_unlikely (n == (size_t) -1))
+ {
+ /* Something wrong.
+ XXX Do we have to set `errno' to something which
+ mbsrtows hasn't already done? */
+ free_return:
+ free (wpattern_malloc);
+ return -1;
+ }
+ if (p)
+ {
+ memset (&ps, '\0', sizeof (ps));
+ goto prepare_wstring;
+ }
}
+ else
+ {
+ prepare_wstring:
+ n = mbsrtowcs (NULL, &string, 0, &ps);
+ if (__glibc_unlikely (n == (size_t) -1))
+ /* Something wrong.
+ XXX Do we have to set `errno' to something which mbsrtows hasn't
+ already done? */
+ goto free_return;
+ if (__glibc_unlikely (n >= (size_t) -1 / sizeof (wchar_t)))
+ {
+ free (wpattern_malloc);
+ __set_errno (ENOMEM);
+ return -2;
+ }
- ++n;
- }
+ wstring_malloc = wstring
+ = (wchar_t *) malloc ((n + 1) * sizeof (wchar_t));
+ if (wstring == NULL)
+ {
+ free (wpattern_malloc);
+ return -2;
+ }
+ assert (mbsinit (&ps));
+ (void) mbsrtowcs (wstring, &string, n + 1, &ps);
+ }
- if (*n == '\0')
- return 0;
+ int res = internal_fnwmatch (wpattern, wstring, wstring + n,
+ flags & FNM_PERIOD, flags, NULL,
+ alloca_used);
- if ((flags & FNM_LEADING_DIR) && *n == '/')
- /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */
- return 0;
+ free (wstring_malloc);
+ free (wpattern_malloc);
- return FNM_NOMATCH;
+ return res;
+ }
+# endif /* mbstate_t and mbsrtowcs or _LIBC. */
-# undef FOLD
+ return internal_fnmatch (pattern, string, string + strlen (string),
+ flags & FNM_PERIOD, flags, NULL, 0);
}
+# ifdef _LIBC
+# undef fnmatch
+versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
+# if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
+strong_alias (__fnmatch, __fnmatch_old)
+compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
+# endif
+libc_hidden_ver (__fnmatch, fnmatch)
+# endif
+
#endif /* _LIBC or not __GNU_LIBRARY__. */