import from gnulib

lib/fnmatch.c

@@ -1,4 +1,5 @@
-/* Copyright 1991, 1992, 1993, 1996, 1997, 2000 Free Software Foundation, Inc.
+/* Copyright (C) 1991, 1992, 1993, 1996, 1997, 1998, 1999, 2000, 2001,
+   2002 Free Software Foundation, Inc.
 
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -10,9 +11,9 @@
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.
 
-  You should have received a copy of the GNU General Public License
-  along with this program; 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 General Public License
+   along with this program; if not, write to the Free Software Foundation,
+   Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
 
 #if HAVE_CONFIG_H
 # include <config.h>
@@ -23,208 +24,362 @@
 # define _GNU_SOURCE	1
 #endif
 
+#ifdef __GNUC__
+# define alloca __builtin_alloca
+# define HAVE_ALLOCA 1
+#else
+# if defined HAVE_ALLOCA_H || defined _LIBC
+#  include <alloca.h>
+# else
+#  ifdef _AIX
+ #  pragma alloca
+#  else
+#   ifndef alloca
+char *alloca ();
+#   endif
+#  endif
+# endif
+#endif
+
+#if ! defined __builtin_expect && __GNUC__ < 3
+# define __builtin_expect(expr, expected) (expr)
+#endif
+
+#include <assert.h>
 #include <errno.h>
 #include <fnmatch.h>
 #include <ctype.h>
 
-#if defined STDC_HEADERS || !defined isascii
-# define IN_CTYPE_DOMAIN(c) 1
+#if HAVE_STRING_H || defined _LIBC
+# include <string.h>
 #else
-# define IN_CTYPE_DOMAIN(c) isascii (c)
+# if HAVE_STRINGS_H
+#  include <strings.h>
+# endif
 #endif
 
-#define ISUPPER(c) (IN_CTYPE_DOMAIN (c) && isupper (c))
+#if defined STDC_HEADERS || defined _LIBC
+# include <stddef.h>
+# include <stdlib.h>
+#endif
 
+#define WIDE_CHAR_SUPPORT (HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_BTOWC)
 
-#ifndef errno
-extern int errno;
+/* For platform which support the ISO C amendement 1 functionality we
+   support user defined character classes.  */
+#if defined _LIBC || WIDE_CHAR_SUPPORT
+/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>.  */
+# include <wchar.h>
+# include <wctype.h>
 #endif
 
-/* Match STRING against the filename pattern PATTERN, returning zero if
-   it matches, nonzero if not.  */
-int
-fnmatch (const char *pattern, const char *string, int flags)
-{
-  register const char *p = pattern, *n = string;
-  register char c;
+/* 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 mbsinit __mbsinit
+# define mbsrtowcs __mbsrtowcs
+# define fnmatch __fnmatch
+extern int fnmatch (const char *pattern, const char *string, int flags);
+#endif
 
-/* Note that this evaluates C many times.  */
-#define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER ((unsigned char) (c)) \
-                 ? tolower ((unsigned char) (c)) \
-                 : (c))
+/* 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, are not
+   actually compiling the library itself, and have not detected a bug
+   in the library.  This code is part of the GNU C
+   Library, but also included in many other GNU distributions.  Compiling
+   and linking in this code is a waste when using the GNU C library
+   (especially if it is a shared library).  Rather than having every GNU
+   program understand `configure --with-gnu-libc' and omit the object files,
+   it is simpler to just do this in the source for each such file.  */
+
+#if defined _LIBC || !defined __GNU_LIBRARY__ || !HAVE_FNMATCH_GNU
+
+
+# if defined STDC_HEADERS || !defined isascii
+#  define ISASCII(c) 1
+# else
+#  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
+
+# 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))
+
+# if defined _LIBC || WIDE_CHAR_SUPPORT
+/* The GNU C library provides support for user-defined character classes
+   and the functions from ISO C amendement 1.  */
+#  ifdef CHARCLASS_NAME_MAX
+#   define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
+#  else
+/* This shouldn't happen but some implementation might still have this
+   problem.  Use a reasonable default value.  */
+#   define CHAR_CLASS_MAX_LENGTH 256
+#  endif
+
+#  ifdef _LIBC
+#   define IS_CHAR_CLASS(string) __wctype (string)
+#  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'.  */
+
+#  define IS_CHAR_CLASS(string)						      \
+   (STREQ (string, "alpha") || STREQ (string, "upper")			      \
+    || STREQ (string, "lower") || STREQ (string, "digit")		      \
+    || STREQ (string, "alnum") || STREQ (string, "xdigit")		      \
+    || STREQ (string, "space") || STREQ (string, "print")		      \
+    || STREQ (string, "punct") || STREQ (string, "graph")		      \
+    || STREQ (string, "cntrl") || STREQ (string, "blank"))
+# endif
+
+/* Avoid depending on library functions or files
+   whose names are inconsistent.  */
+
+# if !defined _LIBC && !defined getenv && !HAVE_DECL_GETENV
+extern char *getenv ();
+# endif
+
+# ifndef errno
+extern int errno;
+# endif
 
-  while ((c = *p++) != '\0')
-    {
-      c = FOLD (c);
+/* Global variable.  */
+static int posixly_correct;
 
-      switch (c)
-	{
-	case '?':
-	  if (*n == '\0')
-	    return FNM_NOMATCH;
-	  else if ((flags & FNM_FILE_NAME) && *n == '/')
-	    return FNM_NOMATCH;
-	  else if ((flags & FNM_PERIOD) && *n == '.' &&
-		   (n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/')))
-	    return FNM_NOMATCH;
-	  break;
+# 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
 
-	case '\\':
-	  if (!(flags & FNM_NOESCAPE))
-	    {
-	      c = *p++;
-	      if (c == '\0')
-		/* Trailing \ loses.  */
-		return FNM_NOMATCH;
-	      c = FOLD (c);
-	    }
-	  if (FOLD (*n) != c)
-	    return FNM_NOMATCH;
-	  break;
+/* Note that this evaluates C many times.  */
+# ifdef _LIBC
+#  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
+# 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 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)
+# ifdef _LIBC
+#  define MEMPCPY(D, S, N) __mempcpy (D, S, N)
+# else
+#  if HAVE_MEMPCPY
+#   define MEMPCPY(D, S, N) mempcpy (D, S, N)
+#  else
+#   define MEMPCPY(D, S, N) ((void *) ((char *) memcpy (D, S, N) + (N)))
+#  endif
+# endif
+# define MEMCHR(S, C, N) memchr (S, C, N)
+# define STRCOLL(S1, S2) strcoll (S1, S2)
+# include "fnmatch_loop.c"
+
+
+# if HANDLE_MULTIBYTE
+#  define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
+#  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 L(CS)	L##CS
+#  define BTOWC(C)	(C)
+#  ifdef _LIBC
+#   define STRLEN(S) __wcslen (S)
+#   define STRCAT(D, S) __wcscat (D, S)
+#   define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
+#  else
+#   define STRLEN(S) wcslen (S)
+#   define STRCAT(D, S) wcscat (D, S)
+#   if HAVE_WMEMPCPY
+#    define MEMPCPY(D, S, N) wmempcpy (D, S, N)
+#   else
+#    define MEMPCPY(D, S, N) (wmemcpy (D, S, N) + (N))
+#   endif
+#  endif
+#  define MEMCHR(S, C, N) wmemchr (S, C, N)
+#  define STRCOLL(S1, S2) wcscoll (S1, S2)
+#  define WIDE_CHAR_VERSION 1
+
+#  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;
 
-	case '*':
-	  if ((flags & FNM_PERIOD) && *n == '.' &&
-	      (n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/')))
-	    return FNM_NOMATCH;
-
-	  for (c = *p++; c == '?' || c == '*'; c = *p++)
-	    {
-	      if (c == '?')
-		{
-		  /* A ? needs to match one character.  */
-		  if (*n == '\0' || (*n == '/' && (flags & FNM_FILE_NAME)))
-		    /* 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;
-		}
-	    }
-
-	  if (c == '\0')
-	    {
-	      if ((flags & (FNM_FILE_NAME | FNM_LEADING_DIR)) == FNM_FILE_NAME)
-		for (; *n != '\0'; n++)
-		  if (*n == '/')
-		    return FNM_NOMATCH;
-	      return 0;
-	    }
-
-	  {
-	    char c1 = (!(flags & FNM_NOESCAPE) && c == '\\') ? *p : c;
-	    c1 = FOLD (c1);
-	    for (--p; *n != '\0'; ++n)
-	      if ((c == '[' || FOLD (*n) == c1) &&
-		  fnmatch (p, n, flags & ~FNM_PERIOD) == 0)
-		return 0;
-	      else if (*n == '/' && (flags & FNM_FILE_NAME))
-		break;
-	    return FNM_NOMATCH;
-	  }
-
-	case '[':
-	  {
-	    /* Nonzero if the sense of the character class is inverted.  */
-	    register int not;
-
-	    if (*n == '\0')
-	      return FNM_NOMATCH;
-
-	    if ((flags & FNM_PERIOD) && *n == '.' &&
-		(n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/')))
-	      return FNM_NOMATCH;
-
-	    not = (*p == '!' || *p == '^');
-	    if (not)
-	      ++p;
-
-	    c = *p++;
-	    for (;;)
-	      {
-		register char cstart = c, cend = c;
-
-		if (!(flags & FNM_NOESCAPE) && c == '\\')
-		  {
-		    if (*p == '\0')
-		      return FNM_NOMATCH;
-		    cstart = cend = *p++;
-		  }
-
-		cstart = cend = FOLD (cstart);
-
-		if (c == '\0')
-		  /* [ (unterminated) loses.  */
-		  return FNM_NOMATCH;
-
-		c = *p++;
-		c = FOLD (c);
-
-		if ((flags & FNM_FILE_NAME) && c == '/')
-		  /* [/] can never match.  */
-		  return FNM_NOMATCH;
-
-		if (c == '-' && *p != ']')
-		  {
-		    cend = *p++;
-		    if (!(flags & FNM_NOESCAPE) && cend == '\\')
-		      cend = *p++;
-		    if (cend == '\0')
-		      return FNM_NOMATCH;
-		    cend = FOLD (cend);
-
-		    c = *p++;
-		  }
-
-		if (FOLD (*n) >= cstart && FOLD (*n) <= cend)
-		  goto matched;
-
-		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;
-		  }
-	      }
-	    if (not)
-	      return FNM_NOMATCH;
-	  }
+  do
+    {
+      /* 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 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:
-	  if (c != FOLD (*n))
-	    return FNM_NOMATCH;
+	  return (wctype_t) 0;
 	}
+#  endif
+
+      /* Avoid overrunning the buffer.  */
+      if (cp == s + CHAR_CLASS_MAX_LENGTH)
+	return (wctype_t) 0;
 
-      ++n;
+      *cp++ = (char) *wcs++;
     }
+  while (*wcs != L'\0');
 
-  if (*n == '\0')
-    return 0;
+  *cp = '\0';
+
+#  ifdef _LIBC
+  return __wctype (s);
+#  else
+  return wctype (s);
+#  endif
+}
+#  define IS_CHAR_CLASS(string) is_char_class (string)
 
-  if ((flags & FNM_LEADING_DIR) && *n == '/')
-    /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz".  */
-    return 0;
+#  include "fnmatch_loop.c"
+# endif
 
-  return FNM_NOMATCH;
 
-#undef FOLD
+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;
+      wchar_t *wpattern;
+      wchar_t *wstring;
+
+      /* Convert the strings into wide characters.  */
+      memset (&ps, '\0', sizeof (ps));
+      n = mbsrtowcs (NULL, &pattern, 0, &ps);
+      if (__builtin_expect (n, 0) == (size_t) -1)
+	/* Something wrong.
+	   XXX Do we have to set `errno' to something which mbsrtows hasn't
+	   already done?  */
+	return -1;
+      wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
+      assert (mbsinit (&ps));
+      (void) mbsrtowcs (wpattern, &pattern, n + 1, &ps);
+
+      assert (mbsinit (&ps));
+      n = mbsrtowcs (NULL, &string, 0, &ps);
+      if (__builtin_expect (n, 0) == (size_t) -1)
+	/* Something wrong.
+	   XXX Do we have to set `errno' to something which mbsrtows hasn't
+	   already done?  */
+	return -1;
+      wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));
+      assert (mbsinit (&ps));
+      (void) mbsrtowcs (wstring, &string, n + 1, &ps);
+
+      return internal_fnwmatch (wpattern, wstring, wstring + n,
+				flags & FNM_PERIOD, flags);
+    }
+# endif  /* mbstate_t and mbsrtowcs or _LIBC.  */
+
+  return internal_fnmatch (pattern, string, string + strlen (string),
+			   flags & FNM_PERIOD, flags);
 }
+
+# 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
+# endif
+
+#endif	/* _LIBC or not __GNU_LIBRARY__.  */