diff options
Diffstat (limited to 'posix/fnmatch_loop.c')
| -rw-r--r-- | posix/fnmatch_loop.c | 1977 |
1 files changed, 985 insertions, 992 deletions
diff --git a/posix/fnmatch_loop.c b/posix/fnmatch_loop.c index 1b2f6f05e1..7f938af590 100644 --- a/posix/fnmatch_loop.c +++ b/posix/fnmatch_loop.c @@ -15,28 +15,30 @@ License along with the GNU C Library; if not, see <https://www.gnu.org/licenses/>. */ -#include <stdint.h> +#ifdef _LIBC +# include <stdint.h> +#endif struct STRUCT { const CHAR *pattern; const CHAR *string; - int no_leading_period; + bool no_leading_period; }; -/* Match STRING against the filename pattern PATTERN, returning zero if +/* Match STRING against the file name 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, size_t alloca_used); + const CHAR *string_end, bool no_leading_period, int flags, + struct STRUCT *ends, size_t alloca_used); static int EXT (INT opt, const CHAR *pattern, const CHAR *string, - const CHAR *string_end, int no_leading_period, int flags, - size_t alloca_used); + const CHAR *string_end, bool no_leading_period, int flags, + size_t alloca_used); static const CHAR *END (const CHAR *patternp); static int FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end, - int no_leading_period, int flags, struct STRUCT *ends, size_t alloca_used) + bool no_leading_period, int flags, struct STRUCT *ends, size_t alloca_used) { const CHAR *p = pattern, *n = string; UCHAR c; @@ -50,882 +52,868 @@ FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end, # endif #endif - while ((c = *p++) != L('\0')) + while ((c = *p++) != L_('\0')) { - int new_no_leading_period = 0; + bool new_no_leading_period = false; c = FOLD (c); switch (c) - { - case L('?'): - if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(') - { - int res = EXT (c, p, n, string_end, no_leading_period, - flags, alloca_used); - 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 = EXT (c, p, n, string_end, no_leading_period, - flags, alloca_used); - 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, alloca_used) == 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, alloca_used) == 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, alloca_used) == 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. */ - const CHAR *p_init = p; - const CHAR *n_init = n; - 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 + { + case L_('?'): + if (__glibc_unlikely (flags & FNM_EXTMATCH) && *p == '(') + { + int res = EXT (c, p, n, string_end, no_leading_period, + flags, alloca_used); + 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 (__glibc_unlikely (flags & FNM_EXTMATCH) && *p == '(') + { + int res = EXT (c, p, n, string_end, no_leading_period, + flags, alloca_used); + 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_('/') + && __glibc_unlikely (flags & FNM_FILE_NAME)) + /* 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_('[') + || (__glibc_unlikely (flags & FNM_EXTMATCH) + && (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 = false) + if (FCT (p, n, string_end, no_leading_period, flags2, + &end, alloca_used) == 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, alloca_used) == 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 = false) + if (FOLD ((UCHAR) *n) == c + && (FCT (p, n, string_end, no_leading_period, flags2, + &end, alloca_used) == 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. */ + const CHAR *p_init = p; + const CHAR *n_init = n; + bool 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; + wctype_t wt; + 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'); + + 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 ((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; + if (iswctype (BTOWC ((UCHAR) *n), wt)) + goto matched; #endif - c = *p++; - } + c = *p++; + } #ifdef _LIBC - else if (c == L('[') && *p == L('=')) - { - /* It's important that STR be a scalar variable rather - than a one-element array, because GCC (at least 4.9.2 - -O2 on x86-64) can be confused by the array and - diagnose a "used initialized" in a dead branch in the - findidx function. */ - UCHAR str; - 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 = 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) - goto matched; - } - else - { - const int32_t *table; + else if (c == L_('[') && *p == L_('=')) + { + /* It's important that STR be a scalar variable rather + than a one-element array, because GCC (at least 4.9.2 + -O2 on x86-64) can be confused by the array and + diagnose a "used initialized" in a dead branch in the + findidx function. */ + UCHAR str; + 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 = 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) + goto matched; + } + else + { + const int32_t *table; # if WIDE_CHAR_VERSION - const int32_t *weights; - const wint_t *extra; + const int32_t *weights; + const wint_t *extra; # else - const unsigned char *weights; - const unsigned char *extra; + const unsigned char *weights; + const unsigned char *extra; # endif - const int32_t *indirect; - int32_t idx; - const UCHAR *cp = (const UCHAR *) &str; + const int32_t *indirect; + int32_t idx; + const UCHAR *cp = (const UCHAR *) &str; # 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 wint_t *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC); - indirect = (const int32_t *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC); + table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC); + weights = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC); + extra = (const wint_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); + 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 (table, indirect, extra, &cp, 1); - 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 & 0xffffff]; - int32_t idx2; - const UCHAR *np = (const UCHAR *) n; - - idx2 = FINDIDX (table, indirect, extra, - & |