|
yann@1365
|
1 |
Original patch from: gentoo/src/patchsets/glibc/2.9/0052_all_glibc-2.9-regex-BZ9697.patch
|
|
yann@1365
|
2 |
|
|
yann@1365
|
3 |
-= BEGIN original header =-
|
|
yann@1365
|
4 |
http://sourceware.org/ml/libc-alpha/2009-01/msg00005.html
|
|
yann@1365
|
5 |
|
|
yann@1365
|
6 |
From ea8ca0dfcbf2721bcf2c08ce3c01d5764b827572 Mon Sep 17 00:00:00 2001
|
|
yann@1365
|
7 |
From: Ulrich Drepper <drepper@redhat.com>
|
|
yann@1365
|
8 |
Date: Thu, 8 Jan 2009 00:42:28 +0000
|
|
yann@1365
|
9 |
Subject: [PATCH] (re_compile_fastmap_iter): Rewrite COMPLEX_BRACKET handling.
|
|
yann@1365
|
10 |
|
|
yann@1365
|
11 |
-= END original header =-
|
|
yann@1365
|
12 |
|
|
yann@1365
|
13 |
diff -durN glibc-2_9.orig/posix/regcomp.c glibc-2_9/posix/regcomp.c
|
|
yann@1365
|
14 |
--- glibc-2_9.orig/posix/regcomp.c 2008-05-15 05:07:21.000000000 +0200
|
|
yann@1365
|
15 |
+++ glibc-2_9/posix/regcomp.c 2009-02-02 22:00:41.000000000 +0100
|
|
yann@1365
|
16 |
@@ -350,47 +350,67 @@
|
|
yann@1365
|
17 |
#ifdef RE_ENABLE_I18N
|
|
yann@1365
|
18 |
else if (type == COMPLEX_BRACKET)
|
|
yann@1365
|
19 |
{
|
|
yann@1365
|
20 |
- int i;
|
|
yann@1365
|
21 |
re_charset_t *cset = dfa->nodes[node].opr.mbcset;
|
|
yann@1365
|
22 |
- if (cset->non_match || cset->ncoll_syms || cset->nequiv_classes
|
|
yann@1365
|
23 |
- || cset->nranges || cset->nchar_classes)
|
|
yann@1365
|
24 |
- {
|
|
yann@1365
|
25 |
+ int i;
|
|
yann@1365
|
26 |
+
|
|
yann@1365
|
27 |
# ifdef _LIBC
|
|
yann@1365
|
28 |
- if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0)
|
|
yann@1365
|
29 |
+ /* See if we have to try all bytes which start multiple collation
|
|
yann@1365
|
30 |
+ elements.
|
|
yann@1365
|
31 |
+ e.g. In da_DK, we want to catch 'a' since "aa" is a valid
|
|
yann@1365
|
32 |
+ collation element, and don't catch 'b' since 'b' is
|
|
yann@1365
|
33 |
+ the only collation element which starts from 'b' (and
|
|
yann@1365
|
34 |
+ it is caught by SIMPLE_BRACKET). */
|
|
yann@1365
|
35 |
+ if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0
|
|
yann@1365
|
36 |
+ && (cset->ncoll_syms || cset->nranges))
|
|
yann@1365
|
37 |
{
|
|
yann@1365
|
38 |
- /* In this case we want to catch the bytes which are
|
|
yann@1365
|
39 |
- the first byte of any collation elements.
|
|
yann@1365
|
40 |
- e.g. In da_DK, we want to catch 'a' since "aa"
|
|
yann@1365
|
41 |
- is a valid collation element, and don't catch
|
|
yann@1365
|
42 |
- 'b' since 'b' is the only collation element
|
|
yann@1365
|
43 |
- which starts from 'b'. */
|
|
yann@1365
|
44 |
const int32_t *table = (const int32_t *)
|
|
yann@1365
|
45 |
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
|
|
yann@1365
|
46 |
for (i = 0; i < SBC_MAX; ++i)
|
|
yann@1365
|
47 |
if (table[i] < 0)
|
|
yann@1365
|
48 |
re_set_fastmap (fastmap, icase, i);
|
|
yann@1365
|
49 |
}
|
|
yann@1365
|
50 |
-# else
|
|
yann@1365
|
51 |
- if (dfa->mb_cur_max > 1)
|
|
yann@1365
|
52 |
- for (i = 0; i < SBC_MAX; ++i)
|
|
yann@1365
|
53 |
- if (__btowc (i) == WEOF)
|
|
yann@1365
|
54 |
- re_set_fastmap (fastmap, icase, i);
|
|
yann@1365
|
55 |
-# endif /* not _LIBC */
|
|
yann@1365
|
56 |
- }
|
|
yann@1365
|
57 |
- for (i = 0; i < cset->nmbchars; ++i)
|
|
yann@1365
|
58 |
+# endif /* _LIBC */
|
|
yann@1365
|
59 |
+
|
|
yann@1365
|
60 |
+ /* See if we have to start the match at all multibyte characters,
|
|
yann@1365
|
61 |
+ i.e. where we would not find an invalid sequence. This only
|
|
yann@1365
|
62 |
+ applies to multibyte character sets; for single byte character
|
|
yann@1365
|
63 |
+ sets, the SIMPLE_BRACKET again suffices. */
|
|
yann@1365
|
64 |
+ if (dfa->mb_cur_max > 1
|
|
yann@1365
|
65 |
+ && (cset->nchar_classes || cset->non_match
|
|
yann@1365
|
66 |
+# ifdef _LIBC
|
|
yann@1365
|
67 |
+ || cset->nequiv_classes
|
|
yann@1365
|
68 |
+# endif /* _LIBC */
|
|
yann@1365
|
69 |
+ ))
|
|
yann@1365
|
70 |
{
|
|
yann@1365
|
71 |
- char buf[256];
|
|
yann@1365
|
72 |
- mbstate_t state;
|
|
yann@1365
|
73 |
- memset (&state, '\0', sizeof (state));
|
|
yann@1365
|
74 |
- if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1)
|
|
yann@1365
|
75 |
- re_set_fastmap (fastmap, icase, *(unsigned char *) buf);
|
|
yann@1365
|
76 |
- if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1)
|
|
yann@1365
|
77 |
+ unsigned char c = 0;
|
|
yann@1365
|
78 |
+ do
|
|
yann@1365
|
79 |
{
|
|
yann@1365
|
80 |
- if (__wcrtomb (buf, towlower (cset->mbchars[i]), &state)
|
|
yann@1365
|
81 |
- != (size_t) -1)
|
|
yann@1365
|
82 |
- re_set_fastmap (fastmap, 0, *(unsigned char *) buf);
|
|
yann@1365
|
83 |
+ mbstate_t mbs;
|
|
yann@1365
|
84 |
+ memset (&mbs, 0, sizeof (mbs));
|
|
yann@1365
|
85 |
+ if (__mbrtowc (NULL, (char *) &c, 1, &mbs) == (size_t) -2)
|
|
yann@1365
|
86 |
+ re_set_fastmap (fastmap, false, (int) c);
|
|
yann@1365
|
87 |
}
|
|
yann@1365
|
88 |
+ while (++c != 0);
|
|
yann@1365
|
89 |
}
|
|
yann@1365
|
90 |
+
|
|
yann@1365
|
91 |
+ else
|
|
yann@1365
|
92 |
+ {
|
|
yann@1365
|
93 |
+ /* ... Else catch all bytes which can start the mbchars. */
|
|
yann@1365
|
94 |
+ for (i = 0; i < cset->nmbchars; ++i)
|
|
yann@1365
|
95 |
+ {
|
|
yann@1365
|
96 |
+ char buf[256];
|
|
yann@1365
|
97 |
+ mbstate_t state;
|
|
yann@1365
|
98 |
+ memset (&state, '\0', sizeof (state));
|
|
yann@1365
|
99 |
+ if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1)
|
|
yann@1365
|
100 |
+ re_set_fastmap (fastmap, icase, *(unsigned char *) buf);
|
|
yann@1365
|
101 |
+ if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1)
|
|
yann@1365
|
102 |
+ {
|
|
yann@1365
|
103 |
+ if (__wcrtomb (buf, towlower (cset->mbchars[i]), &state)
|
|
yann@1365
|
104 |
+ != (size_t) -1)
|
|
yann@1365
|
105 |
+ re_set_fastmap (fastmap, false, *(unsigned char *) buf);
|
|
yann@1365
|
106 |
+ }
|
|
yann@1365
|
107 |
+ }
|
|
yann@1365
|
108 |
+ }
|
|
yann@1365
|
109 |
}
|
|
yann@1365
|
110 |
#endif /* RE_ENABLE_I18N */
|
|
yann@1365
|
111 |
else if (type == OP_PERIOD
|