------- You are receiving this mail because: -------
You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=1158
--- Comment #1 from Philip Hazel <ph10@???> 2011-09-30 10:28:58 ---
On Fri, 30 Sep 2011, Roberto Bagnara wrote:
> $ cat r.c
> #include <stdio.h>
> #if PCRE
> #include <pcreposix.h>
> #else
> #include <regex.h>
> #endif
>
> int main() {
> char buf[1024];
> regex_t re;
> int r;
> r = regcomp(&re, "^$", REG_EXTENDED);
> if (r != 0) {
> regerror(r, &re, buf, sizeof(buf));
> printf("regcomp: %s\n", buf);
> }
> r = regexec(&re, "\n", 0, 0, 0);
> if (r != 0) {
> regerror(r, &re, buf, sizeof(buf));
> printf("regexec: %s\n", buf);
> }
> return 0;
> }
> $ gcc -W -Wall r.c
> $ ./a.out
> regexec: No match
> $ gcc -W -Wall -DPCRE r.c -lpcreposix
> $ ./a.out
> $
Why do you think that the PCRE match is spurious? It looks OK to me.
Remember, the PCRE POSIX interface is *just an interface*. It does NOT
give POSIX semantics. You still get Perl semantics, where $ matches
either the end of the string or a newline at the end of the string.
Philip
--
Configure bugmail:
http://bugs.exim.org/userprefs.cgi?tab=email
