X-Git-Url: http://erislabs.net/gitweb/?a=blobdiff_plain;f=lib%2Ffnmatch.c;h=794d974e310c032a6a89bbdc6313c79455a4bf52;hb=23eecb48e39afd0d267d64d40ba6bf97aa865e13;hp=aec3d1aba024309ebfce2e288ed3bfccba72394f;hpb=10d0937c1a5290e9c180b42f532191aafd70ceb4;p=gnulib.git diff --git a/lib/fnmatch.c b/lib/fnmatch.c index aec3d1aba..794d974e3 100644 --- a/lib/fnmatch.c +++ b/lib/fnmatch.c @@ -1,202 +1,350 @@ -/* Copyright (C) 1991, 1992, 1993 Free Software Foundation, Inc. +/* Copyright (C) 1991-1993, 1996-2007, 2009-2013 Free Software Foundation, Inc. -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. + 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. -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. + 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. -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 General Public License + along with this program; if not, see . */ -You should have received a copy of the GNU General Public License -along with this program; if not, write to the Free Software -Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ +#ifndef _LIBC +# include +#endif -#ifdef HAVE_CONFIG_H -#include +/* Enable GNU extensions in fnmatch.h. */ +#ifndef _GNU_SOURCE +# define _GNU_SOURCE 1 +#endif + +#if ! defined __builtin_expect && __GNUC__ < 3 +# define __builtin_expect(expr, expected) (expr) #endif -#include #include + +#include +#include #include +#include +#include +#include +#include +#include + +#define WIDE_CHAR_SUPPORT \ + (HAVE_WCTYPE_H && HAVE_BTOWC && HAVE_ISWCTYPE \ + && HAVE_WMEMCHR && (HAVE_WMEMCPY || HAVE_WMEMPCPY)) + +/* For platform which support the ISO C amendment 1 functionality we + support user defined character classes. */ +#if defined _LIBC || WIDE_CHAR_SUPPORT +# 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 + +#ifndef SIZE_MAX +# define SIZE_MAX ((size_t) -1) +#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 + 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, + 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 _LIBC || !defined __GNU_LIBRARY__ || !HAVE_FNMATCH_GNU + + +# if ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK)) +# define isblank(c) ((c) == ' ' || (c) == '\t') +# endif + +# 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 amendment 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. */ + +/* Global variable. */ +static int posixly_correct; + +# 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) ? tolower (c) : (c)) +# 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) +# 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 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 + { + /* 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'); -#ifndef errno -extern int errno; -#endif + *cp = '\0'; -/* 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; -{ - register const char *p = pattern, *n = string; - register char c; +# ifdef _LIBC + return __wctype (s); +# else + return wctype (s); +# endif +} +# define IS_CHAR_CLASS(string) is_char_class (string) + +# include "fnmatch_loop.c" +# endif -/* Note that this evalutes C many times. */ -#define FOLD(c) ((flags & FNM_CASEFOLD) && isupper (c) ? tolower (c) : (c)) - while ((c = *p++) != '\0') +int +fnmatch (const char *pattern, const char *string, int flags) +{ +# if HANDLE_MULTIBYTE +# define ALLOCA_LIMIT 2000 + if (__builtin_expect (MB_CUR_MAX, 1) != 1) { - c = FOLD (c); - - 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; - - case '\\': - if (!(flags & FNM_NOESCAPE)) - { - c = *p++; - c = FOLD (c); - } - if (FOLD (*n) != c) - return FNM_NOMATCH; - break; - - case '*': - if ((flags & FNM_PERIOD) && *n == '.' && - (n == string || ((flags & FNM_FILE_NAME) && n[-1] == '/'))) - return FNM_NOMATCH; - - for (c = *p++; c == '?' || c == '*'; c = *p++, ++n) - if (((flags & FNM_FILE_NAME) && *n == '/') || - (c == '?' && *n == '\0')) - return FNM_NOMATCH; - - if (c == '\0') - 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; - 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 == '\\') - 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 == '\\') - /* XXX 1003.2d11 is unclear if this is right. */ - ++p; - } - if (not) - return FNM_NOMATCH; - } - break; - - default: - if (c != FOLD (*n)) - return FNM_NOMATCH; - } - - ++n; + mbstate_t ps; + size_t patsize; + size_t strsize; + size_t totsize; + wchar_t *wpattern; + wchar_t *wstring; + int res; + + /* Calculate the size needed to convert the strings to + wide characters. */ + memset (&ps, '\0', sizeof (ps)); + patsize = mbsrtowcs (NULL, &pattern, 0, &ps) + 1; + if (__builtin_expect (patsize != 0, 1)) + { + assert (mbsinit (&ps)); + strsize = mbsrtowcs (NULL, &string, 0, &ps) + 1; + if (__builtin_expect (strsize != 0, 1)) + { + assert (mbsinit (&ps)); + totsize = patsize + strsize; + if (__builtin_expect (! (patsize <= totsize + && totsize <= SIZE_MAX / sizeof (wchar_t)), + 0)) + { + errno = ENOMEM; + return -1; + } + + /* Allocate room for the wide characters. */ + if (__builtin_expect (totsize < ALLOCA_LIMIT, 1)) + wpattern = (wchar_t *) alloca (totsize * sizeof (wchar_t)); + else + { + wpattern = malloc (totsize * sizeof (wchar_t)); + if (__builtin_expect (! wpattern, 0)) + { + errno = ENOMEM; + return -1; + } + } + wstring = wpattern + patsize; + + /* Convert the strings into wide characters. */ + mbsrtowcs (wpattern, &pattern, patsize, &ps); + assert (mbsinit (&ps)); + mbsrtowcs (wstring, &string, strsize, &ps); + + res = internal_fnwmatch (wpattern, wstring, wstring + strsize - 1, + flags & FNM_PERIOD, flags); + + if (__builtin_expect (! (totsize < ALLOCA_LIMIT), 0)) + free (wpattern); + return res; + } + } } - if (*n == '\0') - return 0; - - if ((flags & FNM_LEADING_DIR) && *n == '/') - /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */ - return 0; +# endif /* HANDLE_MULTIBYTE */ - return FNM_NOMATCH; + return internal_fnmatch (pattern, string, string + strlen (string), + flags & FNM_PERIOD, flags); } -#endif /* _LIBC or not __GNU_LIBRARY__. */ +# 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__. */