From 540de03af741f329cbc263782177e7fc7b8cf21c Mon Sep 17 00:00:00 2001 From: Eric Bollengier Date: Sun, 2 Sep 2007 18:34:22 +0000 Subject: [PATCH] ebl drop old GPL fnmatch, use LGPL new one instead. fnmatch_loop.c is include in fnmatch.c (no need to update Makefile) git-svn-id: https://bacula.svn.sourceforge.net/svnroot/bacula/trunk@5432 91ce42f0-d328-0410-95d8-f526ca767f89 --- bacula/src/lib/fnmatch.c | 582 +++++++++++----- bacula/src/lib/fnmatch.h | 87 +-- bacula/src/lib/fnmatch_loop.c | 1239 +++++++++++++++++++++++++++++++++ 3 files changed, 1683 insertions(+), 225 deletions(-) create mode 100644 bacula/src/lib/fnmatch_loop.c diff --git a/bacula/src/lib/fnmatch.c b/bacula/src/lib/fnmatch.c index 29dad5ac11..c6cdb88772 100644 --- a/bacula/src/lib/fnmatch.c +++ b/bacula/src/lib/fnmatch.c @@ -1,204 +1,420 @@ -/* Copyright (C) 1991, 1992, 1993, 1996, 1997 Free Software Foundation, Inc. +/* Copyright (C) 1991,1992,1993,1996,1997,1998,1999,2000,2001,2002,2003,2007 + Free Software Foundation, Inc. + This file is part of the GNU C Library. - 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 - the Free Software Foundation; either version 2, 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 program 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 General Public License for more details. + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#if HAVE_CONFIG_H +# include +#endif + +/* Enable GNU extensions in fnmatch.h. */ +#ifndef _GNU_SOURCE +# define _GNU_SOURCE 1 +#endif + +#include +#include +#include +#include + +#if HAVE_STRING_H || defined _LIBC +# include +#else +# include +#endif + +#if defined STDC_HEADERS || defined _LIBC +# include +#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) +/* Solaris 2.5 has a bug: must be included before . */ +# include +# include +#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 + +# 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 + 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__ + + +# if defined STDC_HEADERS || !defined isascii +# define ISASCII(c) 1 +# else +# define ISASCII(c) isascii(c) +# endif - 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. */ +# 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 || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H) +/* 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 -#include "bacula.h" -#include "fnmatch.h" +/* Avoid depending on library functions or files + whose names are inconsistent. */ + +# if !defined _LIBC && !defined getenv +extern char *getenv (); +# endif # ifndef errno extern int errno; # 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) +/* 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 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. */ -# define FOLD(c) ((flags & FNM_CASEFOLD) && B_ISUPPER (c) ? tolower (c) : (c)) +# 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 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) +# include "fnmatch_loop.c" + - while ((c = *p++) != '\0') +# 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 + +# 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; + + do { - c = FOLD (c); - - switch (c) - { - case '?': - if (*n == '\0') - return FNM_NOMATCH; - else if ((flags & FNM_FILE_NAME) && IsPathSeparator(*n)) - return FNM_NOMATCH; - else if ((flags & FNM_PERIOD) && *n == '.' && - (n == string || ((flags & FNM_FILE_NAME) && IsPathSeparator(n[-1])))) - return FNM_NOMATCH; - break; - - 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; - - case '*': - if ((flags & FNM_PERIOD) && *n == '.' && - (n == string || ((flags & FNM_FILE_NAME) && IsPathSeparator(n[-1])))) - return FNM_NOMATCH; - - for (c = *p++; c == '?' || c == '*'; c = *p++) - { - if ((flags & FNM_FILE_NAME) && IsPathSeparator(*n)) - /* 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; - } - } - - if (c == '\0') - return 0; - - { - char c1 = (!(flags & FNM_NOESCAPE) && c == '\\') ? *p : c; - c1 = FOLD (c1); - for (--p; *n != '\0'; ++n) - if ((c == '[' || FOLD ((unsigned char)*n) == c1) && - fnmatch (p, n, flags & ~FNM_PERIOD) == 0) - return 0; - return FNM_NOMATCH; - } - - case '[': - { - /* Nonzero if the sense of the character class is inverted. */ - register int nnot; - - if (*n == '\0') - return FNM_NOMATCH; - - if ((flags & FNM_PERIOD) && *n == '.' && - (n == string || ((flags & FNM_FILE_NAME) && IsPathSeparator(n[-1])))) - return FNM_NOMATCH; - - nnot = (*p == '!' || *p == '^'); - if (nnot) - ++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) && IsPathSeparator(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 (!nnot) - 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 (nnot) - return FNM_NOMATCH; - } - break; - - default: - if (c != FOLD (*n)) - return FNM_NOMATCH; - } - - ++n; + /* 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: + return (wctype_t) 0; + } +# endif + + /* Avoid overrunning the buffer. */ + if (cp == s + CHAR_CLASS_MAX_LENGTH) + return (wctype_t) 0; + + *cp++ = (char) *wcs++; } + while (*wcs != L'\0'); + + *cp = '\0'; - if (*n == '\0') - return 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) && IsPathSeparator(*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; + const char *p; + wchar_t *wpattern; + wchar_t *wstring; + + /* 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 (__builtin_expect (n < 1024, 1)) + { + wpattern = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t)); + n = mbsrtowcs (wpattern, &p, n + 1, &ps); + if (__builtin_expect (n == (size_t) -1, 0)) + /* 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)); + } + if (__builtin_expect (p != NULL, 0)) + { + n = mbsrtowcs (NULL, &pattern, 0, &ps); + if (__builtin_expect (n == (size_t) -1, 0)) + /* 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)); +#ifdef _LIBC + n = strnlen (string, 1024); +#else + n = strlen (string); +#endif + p = string; + if (__builtin_expect (n < 1024, 1)) + { + wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t)); + n = mbsrtowcs (wstring, &p, n + 1, &ps); + if (__builtin_expect (n == (size_t) -1, 0)) + /* 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)); + } + if (__builtin_expect (p != NULL, 0)) + { + n = mbsrtowcs (NULL, &string, 0, &ps); + if (__builtin_expect (n == (size_t) -1, 0)) + /* 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, NULL); + } +# endif /* mbstate_t and mbsrtowcs or _LIBC. */ + + return internal_fnmatch (pattern, string, string + strlen (string), + flags & FNM_PERIOD, flags, NULL); } + +# 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__. */ diff --git a/bacula/src/lib/fnmatch.h b/bacula/src/lib/fnmatch.h index 07a5d44ffe..aefb007fba 100644 --- a/bacula/src/lib/fnmatch.h +++ b/bacula/src/lib/fnmatch.h @@ -1,41 +1,35 @@ -/* Copyright (C) 1991, 1992, 1993 Free Software Foundation, Inc. +/* Copyright (C) 1991-93,96,97,98,99,2001,2004 Free Software Foundation, Inc. + This file is part of the GNU C Library. -NOTE: The canonical source of this file is maintained with the GNU C Library. -Bugs can be reported to bug-glibc@prep.ai.mit.edu. + 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 program is free software; you can redistribute it and/or modify it -under the terms of the GNU General Public License as published by the -Free Software Foundation; either version 2, or (at your option) any -later version. + 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 + Lesser General Public License for more details. -This program 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 General Public License for more details. + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; 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. */ - -#ifndef _FNMATCH_H - -#define _FNMATCH_H 1 +#ifndef _FNMATCH_H +#define _FNMATCH_H 1 #ifdef __cplusplus extern "C" { #endif -#if defined (__cplusplus) || (defined (__STDC__) && __STDC__) -#undef __P -#define __P(protos) protos -#else /* Not C++ or ANSI C. */ -#undef __P -#define __P(protos) () -/* We can get away without defining `const' here only because in this file - it is used only inside the prototype for `fnmatch', which is elided in - non-ANSI C where `const' is problematical. */ -#endif /* C++ or ANSI C. */ - +#ifndef const +# if (defined __STDC__ && __STDC__) || defined __cplusplus +# define __const const +# else +# define __const +# endif +#endif /* We #undef these before defining them because some losing systems (HP-UX A.08.07 for example) define these in . */ @@ -44,23 +38,32 @@ extern "C" { #undef FNM_PERIOD /* Bits set in the FLAGS argument to `fnmatch'. */ -#define FNM_PATHNAME (1 << 0) /* No wildcard can ever match `/'. */ -#define FNM_NOESCAPE (1 << 1) /* Backslashes don't quote special chars. */ -#define FNM_PERIOD (1 << 2) /* Leading `.' is matched only explicitly. */ - -/* #if !defined (_POSIX_C_SOURCE) || _POSIX_C_SOURCE < 2 || defined (_GNU_SOURCE) */ -#define FNM_FILE_NAME FNM_PATHNAME /* Preferred GNU name. */ -#define FNM_LEADING_DIR (1 << 3) /* Ignore `/...' after a match. */ -#define FNM_CASEFOLD (1 << 4) /* Compare without regard to case. */ -/* #endif -- kes 12Jan05 */ +#define FNM_PATHNAME (1 << 0) /* No wildcard can ever match `/'. */ +#define FNM_NOESCAPE (1 << 1) /* Backslashes don't quote special chars. */ +#define FNM_PERIOD (1 << 2) /* Leading `.' is matched only explicitly. */ + +#if !defined _POSIX_C_SOURCE || _POSIX_C_SOURCE < 2 || defined _GNU_SOURCE +# define FNM_FILE_NAME FNM_PATHNAME /* Preferred GNU name. */ +# define FNM_LEADING_DIR (1 << 3) /* Ignore `/...' after a match. */ +# define FNM_CASEFOLD (1 << 4) /* Compare without regard to case. */ +# define FNM_EXTMATCH (1 << 5) /* Use ksh-like extended matching. */ +#endif /* Value returned by `fnmatch' if STRING does not match PATTERN. */ -#define FNM_NOMATCH 1 +#define FNM_NOMATCH 1 + +/* This value is returned if the implementation does not support + `fnmatch'. Since this is not the case here it will never be + returned but the conformance test suites still require the symbol + to be defined. */ +#ifdef _XOPEN_SOURCE +# define FNM_NOSYS (-1) +#endif -/* Match STRING against the filename pattern PATTERN, +/* Match NAME against the filename pattern PATTERN, returning zero if it matches, FNM_NOMATCH if not. */ -extern int fnmatch __P ((const char *__pattern, const char *__string, - int __flags)); +extern int fnmatch (__const char *__pattern, __const char *__name, + int __flags); #ifdef __cplusplus } diff --git a/bacula/src/lib/fnmatch_loop.c b/bacula/src/lib/fnmatch_loop.c new file mode 100644 index 0000000000..2bdd837184 --- /dev/null +++ b/bacula/src/lib/fnmatch_loop.c @@ -0,0 +1,1239 @@ +/* Copyright (C) 1991,1992,1993,1996,1997,1998,1999,2000,2001,2003,2004,2005, + 2007 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + 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. + + 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +struct STRUCT +{ + const CHAR *pattern; + const CHAR *string; + int no_leading_period; +}; + +/* Match STRING against the filename pattern PATTERN, returning zero if + it matches, nonzero if not. */ +static int FCT (const CHAR *pattern, const CHAR *string, + const CHAR *string_end, int no_leading_period, int flags, + struct STRUCT *ends) + internal_function; +static int EXT (INT opt, const CHAR *pattern, const CHAR *string, + const CHAR *string_end, int no_leading_period, int flags) + internal_function; +static const CHAR *END (const CHAR *patternp) internal_function; + +static int +internal_function +FCT (pattern, string, string_end, no_leading_period, flags, ends) + const CHAR *pattern; + const CHAR *string; + const CHAR *string_end; + int no_leading_period; + int flags; + struct STRUCT *ends; +{ + register const CHAR *p = pattern, *n = string; + register UCHAR c; +#ifdef _LIBC +# if WIDE_CHAR_VERSION + const char *collseq = (const char *) + _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQWC); +# else + const UCHAR *collseq = (const UCHAR *) + _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQMB); +# endif +#endif + + while ((c = *p++) != L('\0')) + { + int new_no_leading_period = 0; + c = FOLD (c); + + switch (c) + { + case L('?'): + if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(') + { + int res; + + res = EXT (c, p, n, string_end, no_leading_period, + flags); + if (res != -1) + return res; + } + + if (n == string_end) + return FNM_NOMATCH; + else if (*n == L('/') && (flags & FNM_FILE_NAME)) + return FNM_NOMATCH; + else if (*n == L('.') && no_leading_period) + return FNM_NOMATCH; + break; + + case L('\\'): + if (!(flags & FNM_NOESCAPE)) + { + c = *p++; + if (c == L('\0')) + /* Trailing \ loses. */ + return FNM_NOMATCH; + c = FOLD (c); + } + if (n == string_end || FOLD ((UCHAR) *n) != c) + return FNM_NOMATCH; + break; + + case L('*'): + if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(') + { + int res; + + res = EXT (c, p, n, string_end, no_leading_period, + flags); + if (res != -1) + return res; + } + else if (ends != NULL) + { + ends->pattern = p - 1; + ends->string = n; + ends->no_leading_period = no_leading_period; + return 0; + } + + if (n != string_end && *n == L('.') && no_leading_period) + return FNM_NOMATCH; + + for (c = *p++; c == L('?') || c == L('*'); c = *p++) + { + if (*p == L('(') && (flags & FNM_EXTMATCH) != 0) + { + const CHAR *endp = END (p); + if (endp != p) + { + /* This is a pattern. Skip over it. */ + p = endp; + continue; + } + } + + if (c == L('?')) + { + /* A ? needs to match one character. */ + if (n == string_end) + /* There isn't another character; no match. */ + return FNM_NOMATCH; + else if (*n == L('/') + && __builtin_expect (flags & FNM_FILE_NAME, 0)) + /* A slash does not match a wildcard under + FNM_FILE_NAME. */ + 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 == L('\0')) + /* The wildcard(s) is/are the last element of the pattern. + If the name is a file name and contains another slash + this means it cannot match, unless the FNM_LEADING_DIR + flag is set. */ + { + int result = (flags & FNM_FILE_NAME) == 0 ? 0 : FNM_NOMATCH; + + if (flags & FNM_FILE_NAME) + { + if (flags & FNM_LEADING_DIR) + result = 0; + else + { + if (MEMCHR (n, L('/'), string_end - n) == NULL) + result = 0; + } + } + + return result; + } + else + { + const CHAR *endp; + struct STRUCT end; + + end.pattern = NULL; + endp = MEMCHR (n, (flags & FNM_FILE_NAME) ? L('/') : L('\0'), + string_end - n); + if (endp == NULL) + endp = string_end; + + if (c == L('[') + || (__builtin_expect (flags & FNM_EXTMATCH, 0) != 0 + && (c == L('@') || c == L('+') || c == L('!')) + && *p == L('('))) + { + int flags2 = ((flags & FNM_FILE_NAME) + ? flags : (flags & ~FNM_PERIOD)); + + for (--p; n < endp; ++n, no_leading_period = 0) + if (FCT (p, n, string_end, no_leading_period, flags2, + &end) == 0) + goto found; + } + else if (c == L('/') && (flags & FNM_FILE_NAME)) + { + while (n < string_end && *n != L('/')) + ++n; + if (n < string_end && *n == L('/') + && (FCT (p, n + 1, string_end, flags & FNM_PERIOD, flags, + NULL) == 0)) + return 0; + } + else + { + int flags2 = ((flags & FNM_FILE_NAME) + ? flags : (flags & ~FNM_PERIOD)); + + if (c == L('\\') && !(flags & FNM_NOESCAPE)) + c = *p; + c = FOLD (c); + for (--p; n < endp; ++n, no_leading_period = 0) + if (FOLD ((UCHAR) *n) == c + && (FCT (p, n, string_end, no_leading_period, flags2, + &end) == 0)) + { + found: + if (end.pattern == NULL) + return 0; + break; + } + if (end.pattern != NULL) + { + p = end.pattern; + n = end.string; + no_leading_period = end.no_leading_period; + continue; + } + } + } + + /* If we come here no match is possible with the wildcard. */ + return FNM_NOMATCH; + + case L('['): + { + /* Nonzero if the sense of the character class is inverted. */ + register int not; + CHAR cold; + UCHAR fn; + + if (posixly_correct == 0) + posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1; + + if (n == string_end) + return FNM_NOMATCH; + + if (*n == L('.') && no_leading_period) + return FNM_NOMATCH; + + if (*n == L('/') && (flags & FNM_FILE_NAME)) + /* `/' cannot be matched. */ + return FNM_NOMATCH; + + not = (*p == L('!') || (posixly_correct < 0 && *p == L('^'))); + if (not) + ++p; + + fn = FOLD ((UCHAR) *n); + + c = *p++; + for (;;) + { + if (!(flags & FNM_NOESCAPE) && c == L('\\')) + { + if (*p == L('\0')) + return FNM_NOMATCH; + c = FOLD ((UCHAR) *p); + ++p; + + goto normal_bracket; + } + else if (c == L('[') && *p == L(':')) + { + /* 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; +#endif + const CHAR *startp = p; + + 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 == L(':') && p[1] == L(']')) + { + p += 2; + break; + } + if (c < L('a') || c >= L('z')) + { + /* This cannot possibly be a character class name. + Match it as a normal range. */ + p = startp; + c = L('['); + goto normal_bracket; + } + str[c1++] = c; + } + str[c1] = L('\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 defined _LIBC && ! WIDE_CHAR_VERSION + /* The following code is glibc specific but does + there a good job in speeding up the code since + we can avoid the btowc() call. */ + if (_ISCTYPE ((UCHAR) *n, wt)) + goto matched; +# else + if (ISWCTYPE (BTOWC ((UCHAR) *n), wt)) + goto matched; +# endif +#else + if ((STREQ (str, L("alnum")) && ISALNUM ((UCHAR) *n)) + || (STREQ (str, L("alpha")) && ISALPHA ((UCHAR) *n)) + || (STREQ (str, L("blank")) && ISBLANK ((UCHAR) *n)) + || (STREQ (str, L("cntrl")) && ISCNTRL ((UCHAR) *n)) + || (STREQ (str, L("digit")) && ISDIGIT ((UCHAR) *n)) + || (STREQ (str, L("graph")) && ISGRAPH ((UCHAR) *n)) + || (STREQ (str, L("lower")) && ISLOWER ((UCHAR) *n)) + || (STREQ (str, L("print")) && ISPRINT ((UCHAR) *n)) + || (STREQ (str, L("punct")) && ISPUNCT ((UCHAR) *n)) + || (STREQ (str, L("space")) && ISSPACE ((UCHAR) *n)) + || (STREQ (str, L("upper")) && ISUPPER ((UCHAR) *n)) + || (STREQ (str, L("xdigit")) && ISXDIGIT ((UCHAR) *n))) + goto matched; +#endif + c = *p++; + } +#ifdef _LIBC + else if (c == L('[') && *p == L('=')) + { + UCHAR str[1]; + uint32_t nrules = + _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + const CHAR *startp = p; + + c = *++p; + if (c == L('\0')) + { + p = startp; + c = L('['); + goto normal_bracket; + } + str[0] = c; + + c = *++p; + if (c != L('=') || p[1] != L(']')) + { + p = startp; + c = L('['); + goto normal_bracket; + } + p += 2; + + if (nrules == 0) + { + if ((UCHAR) *n == str[0]) + goto matched; + } + else + { + const int32_t *table; +# if WIDE_CHAR_VERSION + const int32_t *weights; + const int32_t *extra; +# else + const unsigned char *weights; + const unsigned char *extra; +# endif + const int32_t *indirect; + int32_t idx; + const UCHAR *cp = (const UCHAR *) str; + + /* This #include defines a local function! */ +# if WIDE_CHAR_VERSION +# include +# else +# include +# endif + +# if WIDE_CHAR_VERSION + table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC); + weights = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC); + extra = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC); + indirect = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC); +# else + table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + weights = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB); +# endif + + idx = findidx (&cp); + if (idx != 0) + { + /* We found a table entry. Now see whether the + character we are currently at has the same + equivalance class value. */ + int len = weights[idx]; + int32_t idx2; + const UCHAR *np = (const UCHAR *) n; + + idx2 = findidx (&np); + if (idx2 != 0 && len == weights[idx2]) + { + int cnt = 0; + + while (cnt < len + && (weights[idx + 1 + cnt] + == weights[idx2 + 1 + cnt])) + ++cnt; + + if (cnt == len) + goto matched; + } + } + } + + c = *p++; + } +#endif + else if (c == L('\0')) + /* [ (unterminated) loses. */ + return FNM_NOMATCH; + else + { + int is_range = 0; + +#ifdef _LIBC + int is_seqval = 0; + + if (c == L('[') && *p == L('.')) + { + uint32_t nrules = + _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + const CHAR *startp = p; + size_t c1 = 0; + + while (1) + { + c = *++p; + if (c == L('.') && p[1] == L(']')) + { + p += 2; + break; + } + if (c == '\0') + return FNM_NOMATCH; + ++c1; + } + + /* We have to handling the symbols differently in + ranges since then the collation sequence is + important. */ + is_range = *p == L('-') && p[1] != L('\0'); + + if (nrules == 0) + { + /* There are no names defined in the collation + data. Therefore we only accept the trivial + names consisting of the character itself. */ + if (c1 != 1) + return FNM_NOMATCH; + + if (!is_range && *n == startp[1]) + goto matched; + + cold = startp[1]; + c = *p++; + } + else + { + int32_t table_size; + const int32_t *symb_table; +# ifdef WIDE_CHAR_VERSION + char str[c1]; + unsigned int strcnt; +# else +# define str (startp + 1) +# endif + const unsigned char *extra; + int32_t idx; + int32_t elem; + int32_t second; + int32_t hash; + +# ifdef WIDE_CHAR_VERSION + /* We have to convert the name to a single-byte + string. This is possible since the names + consist of ASCII characters and the internal + representation is UCS4. */ + for (strcnt = 0; strcnt < c1; ++strcnt) + str[strcnt] = startp[1 + strcnt]; +#endif + + table_size = + _NL_CURRENT_WORD (LC_COLLATE, + _NL_COLLATE_SYMB_HASH_SIZEMB); + symb_table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_TABLEMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_EXTRAMB); + + /* Locate the character in the hashing table. */ + hash = elem_hash (str, c1); + + idx = 0; + elem = hash % table_size; + if (symb_table[2 * elem] != 0) + { + second = hash % (table_size - 2) + 1; + + do + { + /* First compare the hashing value. */ + if (symb_table[2 * elem] == hash + && (c1 + == extra[symb_table[2 * elem + 1]]) + && memcmp (str, + &extra[symb_table[2 * elem + + 1] + + 1], c1) == 0) + { + /* Yep, this is the entry. */ + idx = symb_table[2 * elem + 1]; + idx += 1 + extra[idx]; + break; + } + + /* Next entry. */ + elem += second; + } + while (symb_table[2 * elem] != 0); + } + + if (symb_table[2 * elem] != 0) + { + /* Compare the byte sequence but only if + this is not part of a range. */ +# ifdef WIDE_CHAR_VERSION + int32_t *wextra; + + idx += 1 + extra[idx]; + /* Adjust for the alignment. */ + idx = (idx + 3) & ~3; + + wextra = (int32_t *) &extra[idx + 4]; +# endif + + if (! is_range) + { +# ifdef WIDE_CHAR_VERSION + for (c1 = 0; + (int32_t) c1 < wextra[idx]; + ++c1) + if (n[c1] != wextra[1 + c1]) + break; + + if ((int32_t) c1 == wextra[idx]) + goto matched; +# else + for (c1 = 0; c1 < extra[idx]; ++c1) + if (n[c1] != extra[1 + c1]) + break; + + if (c1 == extra[idx]) + goto matched; +# endif + } + + /* Get the collation sequence value. */ + is_seqval = 1; +# ifdef WIDE_CHAR_VERSION + cold = wextra[1 + wextra[idx]]; +# else + /* Adjust for the alignment. */ + idx += 1 + extra[idx]; + idx = (idx + 3) & ~4; + cold = *((int32_t *) &extra[idx]); +# endif + + c = *p++; + } + else if (c1 == 1) + { + /* No valid character. Match it as a + single byte. */ + if (!is_range && *n == str[0]) + goto matched; + + cold = str[0]; + c = *p++; + } + else + return FNM_NOMATCH; + } + } + else +# undef str +#endif + { + c = FOLD (c); + normal_bracket: + + /* We have to handling the symbols differently in + ranges since then the collation sequence is + important. */ + is_range = (*p == L('-') && p[1] != L('\0') + && p[1] != L(']')); + + if (!is_range && c == fn) + goto matched; + + /* This is needed if we goto normal_bracket; from + outside of is_seqval's scope. */ + is_seqval = 0; + cold = c; + c = *p++; + } + + if (c == L('-') && *p != L(']')) + { +#if _LIBC + /* We have to find the collation sequence + value for C. Collation sequence is nothing + we can regularly access. The sequence + value is defined by the order in which the + definitions of the collation values for the + various characters appear in the source + file. A strange concept, nowhere + documented. */ + uint32_t fcollseq; + uint32_t lcollseq; + UCHAR cend = *p++; + +# ifdef WIDE_CHAR_VERSION + /* Search in the `names' array for the characters. */ + fcollseq = __collseq_table_lookup (collseq, fn); + if (fcollseq == ~((uint32_t) 0)) + /* XXX We don't know anything about the character + we are supposed to match. This means we are + failing. */ + goto range_not_matched; + + if (is_seqval) + lcollseq = cold; + else + lcollseq = __collseq_table_lookup (collseq, cold); +# else + fcollseq = collseq[fn]; + lcollseq = is_seqval ? cold : collseq[(UCHAR) cold]; +# endif + + is_seqval = 0; + if (cend == L('[') && *p == L('.')) + { + uint32_t nrules = + _NL_CURRENT_WORD (LC_COLLATE, + _NL_COLLATE_NRULES); + const CHAR *startp = p; + size_t c1 = 0; + + while (1) + { + c = *++p; + if (c == L('.') && p[1] == L(']')) + { + p += 2; + break; + } + if (c == '\0') + return FNM_NOMATCH; + ++c1; + } + + if (nrules == 0) + { + /* There are no names defined in the + collation data. Therefore we only + accept the trivial names consisting + of the character itself. */ + if (c1 != 1) + return FNM_NOMATCH; + + cend = startp[1]; + } + else + { + int32_t table_size; + const int32_t *symb_table; +# ifdef WIDE_CHAR_VERSION + char str[c1]; + unsigned int strcnt; +# else +# define str (startp + 1) +# endif + const unsigned char *extra; + int32_t idx; + int32_t elem; + int32_t second; + int32_t hash; + +# ifdef WIDE_CHAR_VERSION + /* We have to convert the name to a single-byte + string. This is possible since the names + consist of ASCII characters and the internal + representation is UCS4. */ + for (strcnt = 0; strcnt < c1; ++strcnt) + str[strcnt] = startp[1 + strcnt]; +# endif + + table_size = + _NL_CURRENT_WORD (LC_COLLATE, + _NL_COLLATE_SYMB_HASH_SIZEMB); + symb_table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_TABLEMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_EXTRAMB); + + /* Locate the character in the hashing + table. */ + hash = elem_hash (str, c1); + + idx = 0; + elem = hash % table_size; + if (symb_table[2 * elem] != 0) + { + second = hash % (table_size - 2) + 1; + + do + { + /* First compare the hashing value. */ + if (symb_table[2 * elem] == hash + && (c1 + == extra[symb_table[2 * elem + 1]]) + && memcmp (str, + &extra[symb_table[2 * elem + 1] + + 1], c1) == 0) + { + /* Yep, this is the entry. */ + idx = symb_table[2 * elem + 1]; + idx += 1 + extra[idx]; + break; + } + + /* Next entry. */ + elem += second; + } + while (symb_table[2 * elem] != 0); + } + + if (symb_table[2 * elem] != 0) + { + /* Compare the byte sequence but only if + this is not part of a range. */ +# ifdef WIDE_CHAR_VERSION + int32_t *wextra; + + idx += 1 + extra[idx]; + /* Adjust for the alignment. */ + idx = (idx + 3) & ~4; + + wextra = (int32_t *) &extra[idx + 4]; +# endif + /* Get the collation sequence value. */ + is_seqval = 1; +# ifdef WIDE_CHAR_VERSION + cend = wextra[1 + wextra[idx]]; +# else + /* Adjust for the alignment. */ + idx += 1 + extra[idx]; + idx = (idx + 3) & ~4; + cend = *((int32_t *) &extra[idx]); +# endif + } + else if (symb_table[2 * elem] != 0 && c1 == 1) + { + cend = str[0]; + c = *p++; + } + else + return FNM_NOMATCH; + } +# undef str + } + else + { + if (!(flags & FNM_NOESCAPE) && cend == L('\\')) + cend = *p++; + if (cend == L('\0')) + return FNM_NOMATCH; + cend = FOLD (cend); + } + + /* XXX It is not entirely clear to me how to handle + characters which are not mentioned in the + collation specification. */ + if ( +# ifdef WIDE_CHAR_VERSION + lcollseq == 0xffffffff || +# endif + lcollseq <= fcollseq) + { + /* We have to look at the upper bound. */ + uint32_t hcollseq; + + if (is_seqval) + hcollseq = cend; + else + { +# ifdef WIDE_CHAR_VERSION + hcollseq = + __collseq_table_lookup (collseq, cend); + if (hcollseq == ~((uint32_t) 0)) + { + /* Hum, no information about the upper + bound. The matching succeeds if the + lower bound is matched exactly. */ + if (lcollseq != fcollseq) + goto range_not_matched; + + goto matched; + } +# else + hcollseq = collseq[cend]; +# endif + } + + if (lcollseq <= hcollseq && fcollseq <= hcollseq) + goto matched; + } +# ifdef WIDE_CHAR_VERSION + range_not_matched: +# endif +#else + /* We use a boring value comparison of the character + values. This is better than comparing using + `strcoll' since the latter would have surprising + and sometimes fatal consequences. */ + UCHAR cend = *p++; + + if (!(flags & FNM_NOESCAPE) && cend == L('\\')) + cend = *p++; + if (cend == L('\0')) + return FNM_NOMATCH; + + /* It is a range. */ + if (cold <= fn && fn <= cend) + goto matched; +#endif + + c = *p++; + } + } + + if (c == L(']')) + break; + } + + if (!not) + return FNM_NOMATCH; + break; + + matched: + /* Skip the rest of the [...] that already matched. */ + do + { + ignore_next: + c = *p++; + + if (c == L('\0')) + /* [... (unterminated) loses. */ + return FNM_NOMATCH; + + if (!(flags & FNM_NOESCAPE) && c == L('\\')) + { + if (*p == L('\0')) + return FNM_NOMATCH; + /* XXX 1003.2d11 is unclear if this is right. */ + ++p; + } + else if (c == L('[') && *p == L(':')) + { + int c1 = 0; + const CHAR *startp = p; + + while (1) + { + c = *++p; + if (++c1 == CHAR_CLASS_MAX_LENGTH) + return FNM_NOMATCH; + + if (*p == L(':') && p[1] == L(']')) + break; + + if (c < L('a') || c >= L('z')) + { + p = startp; + goto ignore_next; + } + } + p += 2; + c = *p++; + } + else if (c == L('[') && *p == L('=')) + { + c = *++p; + if (c == L('\0')) + return FNM_NOMATCH; + c = *++p; + if (c != L('=') || p[1] != L(']')) + return FNM_NOMATCH; + p += 2; + c = *p++; + } + else if (c == L('[') && *p == L('.')) + { + ++p; + while (1) + { + c = *++p; + if (c == '\0') + return FNM_NOMATCH; + + if (*p == L('.') && p[1] == L(']')) + break; + } + p += 2; + c = *p++; + } + } + while (c != L(']')); + if (not) + return FNM_NOMATCH; + } + break; + + case L('+'): + case L('@'): + case L('!'): + if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(') + { + int res; + + res = EXT (c, p, n, string_end, no_leading_period, flags); + if (res != -1) + return res; + } + goto normal_match; + + case L('/'): + if (NO_LEADING_PERIOD (flags)) + { + if (n == string_end || c != (UCHAR) *n) + return FNM_NOMATCH; + + new_no_leading_period = 1; + break; + } + /* FALLTHROUGH */ + default: + normal_match: + if (n == string_end || c != FOLD ((UCHAR) *n)) + return FNM_NOMATCH; + } + + no_leading_period = new_no_leading_period; + ++n; + } + + if (n == string_end) + return 0; + + if ((flags & FNM_LEADING_DIR) && n != string_end && *n == L('/')) + /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */ + return 0; + + return FNM_NOMATCH; +} + + +static const CHAR * +internal_function +END (const CHAR *pattern) +{ + const CHAR *p = pattern; + + while (1) + if (*++p == L('\0')) + /* This is an invalid pattern. */ + return pattern; + else if (*p == L('[')) + { + /* Handle brackets special. */ + if (posixly_correct == 0) + posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1; + + /* Skip the not sign. We have to recognize it because of a possibly + following ']'. */ + if (*++p == L('!') || (posixly_correct < 0 && *p == L('^'))) + ++p; + /* A leading ']' is recognized as such. */ + if (*p == L(']')) + ++p; + /* Skip over all characters of the list. */ + while (*p != L(']')) + if (*p++ == L('\0')) + /* This is no valid pattern. */ + return pattern; + } + else if ((*p == L('?') || *p == L('*') || *p == L('+') || *p == L('@') + || *p == L('!')) && p[1] == L('(')) + p = END (p + 1); + else if (*p == L(')')) + break; + + return p + 1; +} + + +static int +internal_function +EXT (INT opt, const CHAR *pattern, const CHAR *string, const CHAR *string_end, + int no_leading_period, int flags) +{ + const CHAR *startp; + int level; + struct patternlist + { + struct patternlist *next; + CHAR str[0]; + } *list = NULL; + struct patternlist **lastp = &list; + size_t pattern_len = STRLEN (pattern); + const CHAR *p; + const CHAR *rs; + + /* Parse the pattern. Store the individual parts in the list. */ + level = 0; + for (startp = p = pattern + 1; level >= 0; ++p) + if (*p == L('\0')) + /* This is an invalid pattern. */ + return -1; + else if (*p == L('[')) + { + /* Handle brackets special. */ + if (posixly_correct == 0) + posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1; + + /* Skip the not sign. We have to recognize it because of a possibly + following ']'. */ + if (*++p == L('!') || (posixly_correct < 0 && *p == L('^'))) + ++p; + /* A leading ']' is recognized as such. */ + if (*p == L(']')) + ++p; + /* Skip over all characters of the list. */ + while (*p != L(']')) + if (*p++ == L('\0')) + /* This is no valid pattern. */ + return -1; + } + else if ((*p == L('?') || *p == L('*') || *p == L('+') || *p == L('@') + || *p == L('!')) && p[1] == L('(')) + /* Remember the nesting level. */ + ++level; + else if (*p == L(')')) + { + if (level-- == 0) + { + /* This means we found the end of the pattern. */ +#define NEW_PATTERN \ + struct patternlist *newp; \ + \ + if (opt == L('?') || opt == L('@')) \ + newp = alloca (sizeof (struct patternlist) \ + + (pattern_len * sizeof (CHAR))); \ + else \ + newp = alloca (sizeof (struct patternlist) \ + + ((p - startp + 1) * sizeof (CHAR))); \ + *((CHAR *) MEMPCPY (newp->str, startp, p - startp)) = L('\0'); \ + newp->next = NULL; \ + *lastp = newp; \ + lastp = &newp->next + NEW_PATTERN; + } + } + else if (*p == L('|')) + { + if (level == 0) + { + NEW_PATTERN; + startp = p + 1; + } + } + assert (list != NULL); + assert (p[-1] == L(')')); +#undef NEW_PATTERN + + switch (opt) + { + case L('*'): + if (FCT (p, string, string_end, no_leading_period, flags, NULL) == 0) + return 0; + /* FALLTHROUGH */ + + case L('+'): + do + { + for (rs = string; rs <= string_end; ++rs) + /* First match the prefix with the current pattern with the + current pattern. */ + if (FCT (list->str, string, rs, no_leading_period, + flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD, + NULL) == 0 + /* This was successful. Now match the rest with the rest + of the pattern. */ + && (FCT (p, rs, string_end, + rs == string + ? no_leading_period + : rs[-1] == '/' && NO_LEADING_PERIOD (flags) ? 1 : 0, + flags & FNM_FILE_NAME + ? flags : flags & ~FNM_PERIOD, NULL) == 0 + /* This didn't work. Try the whole pattern. */ + || (rs != string + && FCT (pattern - 1, rs, string_end, + rs == string + ? no_leading_period + : (rs[-1] == '/' && NO_LEADING_PERIOD (flags) + ? 1 : 0), + flags & FNM_FILE_NAME + ? flags : flags & ~FNM_PERIOD, NULL) == 0))) + /* It worked. Signal success. */ + return 0; + } + while ((list = list->next) != NULL); + + /* None of the patterns lead to a match. */ + return FNM_NOMATCH; + + case L('?'): + if (FCT (p, string, string_end, no_leading_period, flags, NULL) == 0) + return 0; + /* FALLTHROUGH */ + + case L('@'): + do + /* I cannot believe it but `strcat' is actually acceptable + here. Match the entire string with the prefix from the + pattern list and the rest of the pattern following the + pattern list. */ + if (FCT (STRCAT (list->str, p), string, string_end, + no_leading_period, + flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD, + NULL) == 0) + /* It worked. Signal success. */ + return 0; + while ((list = list->next) != NULL); + + /* None of the patterns lead to a match. */ + return FNM_NOMATCH; + + case L('!'): + for (rs = string; rs <= string_end; ++rs) + { + struct patternlist *runp; + + for (runp = list; runp != NULL; runp = runp->next) + if (FCT (runp->str, string, rs, no_leading_period, + flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD, + NULL) == 0) + break; + + /* If none of the patterns matched see whether the rest does. */ + if (runp == NULL + && (FCT (p, rs, string_end, + rs == string + ? no_leading_period + : rs[-1] == '/' && NO_LEADING_PERIOD (flags) ? 1 : 0, + flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD, + NULL) == 0)) + /* This is successful. */ + return 0; + } + + /* None of the patterns together with the rest of the pattern + lead to a match. */ + return FNM_NOMATCH; + + default: + assert (! "Invalid extended matching operator"); + break; + } + + return -1; +} + + +#undef FOLD +#undef CHAR +#undef UCHAR +#undef INT +#undef FCT +#undef EXT +#undef END +#undef STRUCT +#undef MEMPCPY +#undef MEMCHR +#undef STRCOLL +#undef STRLEN +#undef STRCAT +#undef L +#undef BTOWC -- 2.39.5