/* Extended regular expression matching and search library.
- Copyright (C) 2002,2003,2004,2005,2006,2007,2008,2009
- Free Software Foundation, Inc.
+ Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free
+ Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
#ifdef RE_ENABLE_I18N
else if (type == COMPLEX_BRACKET)
{
- Idx i;
re_charset_t *cset = dfa->nodes[node].opr.mbcset;
- if (cset->non_match || cset->ncoll_syms || cset->nequiv_classes
- || cset->nranges || cset->nchar_classes)
- {
+ Idx i;
+
# ifdef _LIBC
- if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0)
+ /* See if we have to try all bytes which start multiple collation
+ elements.
+ e.g. In da_DK, we want to catch 'a' since "aa" is a valid
+ collation element, and don't catch 'b' since 'b' is
+ the only collation element which starts from 'b' (and
+ it is caught by SIMPLE_BRACKET). */
+ if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0
+ && (cset->ncoll_syms || cset->nranges))
{
- /* In this case we want to catch the bytes which are
- the first byte of any collation elements.
- e.g. In da_DK, we want to catch 'a' since "aa"
- is a valid collation element, and don't catch
- 'b' since 'b' is the only collation element
- which starts from 'b'. */
const int32_t *table = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
for (i = 0; i < SBC_MAX; ++i)
if (table[i] < 0)
re_set_fastmap (fastmap, icase, i);
}
-# else
- if (dfa->mb_cur_max > 1)
- for (i = 0; i < SBC_MAX; ++i)
- if (__btowc (i) == WEOF)
- re_set_fastmap (fastmap, icase, i);
-# endif /* not _LIBC */
+# endif /* _LIBC */
+
+ /* See if we have to start the match at all multibyte characters,
+ i.e. where we would not find an invalid sequence. This only
+ applies to multibyte character sets; for single byte character
+ sets, the SIMPLE_BRACKET again suffices. */
+ if (dfa->mb_cur_max > 1
+ && (cset->nchar_classes || cset->non_match || cset->nranges
+# ifdef _LIBC
+ || cset->nequiv_classes
+# endif /* _LIBC */
+ ))
+ {
+ unsigned char c = 0;
+ do
+ {
+ mbstate_t mbs;
+ memset (&mbs, 0, sizeof (mbs));
+ if (__mbrtowc (NULL, (char *) &c, 1, &mbs) == (size_t) -2)
+ re_set_fastmap (fastmap, false, (int) c);
+ }
+ while (++c != 0);
}
- for (i = 0; i < cset->nmbchars; ++i)
+
+ else
{
- char buf[256];
- mbstate_t state;
- memset (&state, '\0', sizeof (state));
- if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1)
- re_set_fastmap (fastmap, icase, *(unsigned char *) buf);
- if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1)
+ /* ... Else catch all bytes which can start the mbchars. */
+ for (i = 0; i < cset->nmbchars; ++i)
{
- if (__wcrtomb (buf, towlower (cset->mbchars[i]), &state)
- != (size_t) -1)
- re_set_fastmap (fastmap, false, *(unsigned char *) buf);
+ char buf[256];
+ mbstate_t state;
+ memset (&state, '\0', sizeof (state));
+ if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1)
+ re_set_fastmap (fastmap, icase, *(unsigned char *) buf);
+ if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1)
+ {
+ if (__wcrtomb (buf, towlower (cset->mbchars[i]), &state)
+ != (size_t) -1)
+ re_set_fastmap (fastmap, false, *(unsigned char *) buf);
+ }
}
}
}
init_dfa (re_dfa_t *dfa, size_t pat_len)
{
__re_size_t table_size;
+#ifndef _LIBC
+ char *codeset_name;
+#endif
#ifdef RE_ENABLE_I18N
size_t max_i18n_object_size = MAX (sizeof (wchar_t), sizeof (wctype_t));
#else
dfa->map_notascii = (_NL_CURRENT_WORD (LC_CTYPE, _NL_CTYPE_MAP_TO_NONASCII)
!= 0);
#else
- if (strcmp (locale_charset (), "UTF-8") == 0)
+ codeset_name = nl_langinfo (CODESET);
+ if (strcasecmp (codeset_name, "UTF-8") == 0
+ || strcasecmp (codeset_name, "UTF8") == 0)
dfa->is_utf8 = 1;
/* We check exhaustively in the loop below if this charset is a
destination. */
org_dest = dfa->edests[org_node].elems[0];
re_node_set_empty (dfa->edests + clone_node);
- clone_dest = search_duplicated_node (dfa, org_dest, constraint);
/* If the node is root_node itself, it means the epsilon closure
has a loop. Then tie it to the destination of the root_node. */
if (org_node == root_node && clone_node != org_node)
}
else
{
- /* There is a duplicated node which satisfy the constraint,
+ /* There is a duplicated node which satisfies the constraint,
use it to avoid infinite loop. */
ok = re_node_set_insert (dfa->edests + clone_node, clone_dest);
if (BE (! ok, 0))
return elem;
}
- if (BE (end != REG_MISSING && start > end, 0))
+ if (BE ((end != REG_MISSING && start > end)
+ || token->type != OP_CLOSE_DUP_NUM, 0))
{
/* First number greater than second. */
*err = REG_BADBR;
return elem;
}
- /* Local function for parse_bracket_exp used in _LIBC environement.
+ /* Local function for parse_bracket_exp used in _LIBC environment.
Look up the collation sequence value of BR_ELEM.
Return the value if succeeded, UINT_MAX otherwise. */
}
else if (br_elem->type == MB_CHAR)
{
- return __collseq_table_lookup (collseqwc, br_elem->opr.wch);
+ if (nrules != 0)
+ return __collseq_table_lookup (collseqwc, br_elem->opr.wch);
}
else if (br_elem->type == COLL_SYM)
{
/* Build single byte matcing table for this equivalence class. */
char_buf[1] = (unsigned char) '\0';
- len = weights[idx1];
+ len = weights[idx1 & 0xffffff];
for (ch = 0; ch < SBC_MAX; ++ch)
{
char_buf[0] = ch;
if (idx2 == 0)
/* This isn't a valid character. */
continue;
- if (len == weights[idx2])
+ /* Compare only if the length matches and the collation rule
+ index is the same. */
+ if (len == weights[idx2 & 0xffffff] && (idx1 >> 24) == (idx2 >> 24))
{
int cnt = 0;
+
while (cnt <= len &&
- weights[idx1 + 1 + cnt] == weights[idx2 + 1 + cnt])
+ weights[(idx1 & 0xffffff) + 1 + cnt]
+ == weights[(idx2 & 0xffffff) + 1 + cnt])
++cnt;
if (cnt > len)