+2003-09-26 Paolo Bonzini <bonzini@gnu.org>
+
+ * posix/regcomp.c (parse_sub_exp): Pass RE_CARET_ANCHORS_HERE
+ for the first token in a subexpression as well.
+
+2003-10-02 Jakub Jelinek <jakub@redhat.com>
+
+ * posix/regcomp.c (peek_token): Add 2003-09-20 changes for anchor
+ handling again.
+ (parse_reg_exp): Likewise.
+ * posix/regex.h (RE_CARET_ANCHORS_HERE): Define.
+
+ * posix/bug-regex11.c (tests): Add new tests.
+ * posix/bug-regex12.c (tests): Add new test.
+
2003-10-01 Thorsten Kukuk <kukuk@suse.de>
* elf/dl-reloc.c (_dl_allocate_static_tls): Move definition of
+2003-10-02 Ulrich Drepper <drepper@redhat.com>
+
+ * sysdeps/unix/sysv/linux/i386/sysdep-cancel.h (DOCARGS_1): Use
+ correct offset.
+
2003-10-02 Jakub Jelinek <jakub@redhat.com>
* Makefile (tests): Add tst-cancel8.
# define _POPCARGS_0 /* No arguments to pop. */
# define PUSHCARGS_1 movl %ebx, %edx; PUSHCARGS_0
-# define DOCARGS_1 _DOARGS_1 (4)
+# define DOCARGS_1 _DOARGS_1 (8)
# define POPCARGS_1 POPCARGS_0; movl %edx, %ebx
# define _PUSHCARGS_1 pushl %ebx; L(PUSHBX2): _PUSHCARGS_0
# define _POPCARGS_1 _POPCARGS_0; popl %ebx; L(POPBX2):
+2003-10-02 Ulrich Drepper <drepper@redhat.com>
+
+ * sysdeps/unix/sysv/linux/i386/sysdep-cancel.h (DOCARGS_1): Use
+ correct offset.
+
2003-10-02 Jakub Jelinek <jakub@redhat.com>
* Makefile (tests): Add tst-cancel19.
# define _POPCARGS_0 /* No arguments to pop. */
# define PUSHCARGS_1 movl %ebx, %edx; L(SAVEBX2): PUSHCARGS_0
-# define DOCARGS_1 _DOARGS_1 (4)
+# define DOCARGS_1 _DOARGS_1 (8)
# define POPCARGS_1 POPCARGS_0; movl %edx, %ebx; L(RESTBX2):
# define _PUSHCARGS_1 pushl %ebx; L(PUSHBX2): _PUSHCARGS_0
# define _POPCARGS_1 _POPCARGS_0; popl %ebx; L(POPBX2):
/* Regular expression tests.
- Copyright (C) 2002 Free Software Foundation, Inc.
+ Copyright (C) 2002, 2003 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jakub@redhat.com>, 2002.
{ { 0, 21 }, { 8, 9 }, { 9, 10 } } },
{ "^\\(a*\\)\\1\\{9\\}\\(a\\{0,9\\}\\)\\([0-9]*;.*[^a]\\2\\([0-9]\\)\\)",
"a1;;0a1aa2aaa3aaaa4aaaaa5aaaaaa6aaaaaaa7aaaaaaaa8aaaaaaaaa9aa2aa1a0", 0,
- 5, { { 0, 67 }, { 0, 0 }, { 0, 1 }, { 1, 67 }, { 66, 67 } } }
+ 5, { { 0, 67 }, { 0, 0 }, { 0, 1 }, { 1, 67 }, { 66, 67 } } },
+ /* Test for BRE expression anchoring. POSIX says just that this may match;
+ in glibc regex it always matched, so avoid changing it. */
+ { "\\(^\\|foo\\)bar", "bar", 0, 2, { { 0, 3 }, { -1, -1 } } },
+ { "\\(foo\\|^\\)bar", "bar", 0, 2, { { 0, 3 }, { -1, -1 } } },
+ /* In ERE this must be treated as an anchor. */
+ { "(^|foo)bar", "bar", REG_EXTENDED, 2, { { 0, 3 }, { -1, -1 } } },
+ { "(foo|^)bar", "bar", REG_EXTENDED, 2, { { 0, 3 }, { -1, -1 } } },
+ /* Here ^ cannot be treated as an anchor according to POSIX. */
+ { "(^|foo)bar", "(^|foo)bar", 0, 2, { { 0, 10 }, { -1, -1 } } },
+ { "(foo|^)bar", "(foo|^)bar", 0, 2, { { 0, 10 }, { -1, -1 } } },
};
int
/* Regular expression tests.
- Copyright (C) 2002 Free Software Foundation, Inc.
+ Copyright (C) 2002, 2003 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jakub@redhat.com>, 2002.
int flags, nmatch;
} tests[] = {
{ "^<\\([^~]*\\)\\([^~]\\)[^~]*~\\1\\(.\\).*|=.*\\3.*\\2",
- "<,.8~2,~so-|=-~.0,123456789<><", REG_NOSUB, 0, }
+ "<,.8~2,~so-|=-~.0,123456789<><", REG_NOSUB, 0 },
+ /* In ERE, all carets must be treated as anchors. */
+ { "a^b", "a^b", REG_EXTENDED, 0 }
};
int
reg_syntax_t syntax);
#endif /* not RE_ENABLE_I18N */
static bin_tree_t *build_charclass_op (re_dfa_t *dfa, RE_TRANSLATE_TYPE trans,
- const unsigned char *class_name,
+ const unsigned char *class_name,
const unsigned char *extra, int not,
reg_errcode_t *err);
static void free_bin_tree (bin_tree_t *tree);
token->type = OP_PERIOD;
break;
case '^':
- if (!(syntax & RE_CONTEXT_INDEP_ANCHORS) &&
+ if (!(syntax & (RE_CONTEXT_INDEP_ANCHORS | RE_CARET_ANCHORS_HERE)) &&
re_string_cur_idx (input) != 0)
{
char prev = re_string_peek_byte (input, -1);
- if (prev != '|' && prev != '(' &&
- (!(syntax & RE_NEWLINE_ALT) || prev != '\n'))
+ if (!(syntax & RE_NEWLINE_ALT) || prev != '\n')
break;
}
token->type = ANCHOR;
bin_tree_t *tree, *eor, *root;
re_token_t current_token;
int new_idx;
- current_token = fetch_token (regexp, syntax);
+ current_token = fetch_token (regexp, syntax | RE_CARET_ANCHORS_HERE);
tree = parse_reg_exp (regexp, preg, ¤t_token, syntax, 0, err);
if (BE (*err != REG_NOERROR && tree == NULL, 0))
return NULL;
{
re_token_t alt_token = *token;
new_idx = re_dfa_add_node (dfa, alt_token, 0);
- *token = fetch_token (regexp, syntax);
+ *token = fetch_token (regexp, syntax | RE_CARET_ANCHORS_HERE);
if (token->type != OP_ALT && token->type != END_OF_RE
&& (nest == 0 || token->type != OP_CLOSE_SUBEXP))
{
return NULL;
}
dfa->nodes[new_idx].opr.idx = cur_nsub;
- *token = fetch_token (regexp, syntax);
+ *token = fetch_token (regexp, syntax | RE_CARET_ANCHORS_HERE);
/* The subexpression may be a null string. */
if (token->type == OP_CLOSE_SUBEXP)
If not set, then case is significant. */
#define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1)
+/* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only
+ for ^, because it is difficult to scan the regex backwards to find
+ whether ^ should be special. */
+#define RE_CARET_ANCHORS_HERE (RE_ICASE << 1)
+
/* This global variable defines the particular regexp syntax to use (for
some interfaces). When a regexp is compiled, the syntax used is
stored in the pattern buffer, so changing this does not affect