Strange regex result

ThreeMonks has asked for the wisdom of the Perl Monks concerning the following question:

I wrote the following code, trying to match balanced parentheses with a given ending. It produces a strange result of '(1, 2)', while I am expecting '(4, 5)'. Would someone tell me what is going wrong? I am using ActivePerl 5.10.0 build 1004.

use strict;

my $a = '(0,((1,2),3,(4,5),6),7,8)';
my $tail = qr{5\)};
my $matching_parens = qr{
    (                         # paren group 1 (parens)
        \(
            (?:
                (?> [^()]+ )  # Non-parens without backtracking
                |
                (?1)          # Recurse to start of paren group 1
            )*
        \)
    )
    .*
    (?<= $tail )
}x;
$a =~ /$matching_parens/;
print "$1\n";
[download]

Comment on Strange regex result Download Code

Replies are listed 'Best First'.
Re: Strange regex result by moritz (Cardinal) on Jul 08, 2009 at 15:48 UTC
The `.* (?<= $tail )` part (assuming that the `\)` is an error, you already have one in `$tail`) means "grab as many characters as possible, and then backtrack until you find the string `5)` before the current position". So the paren group 1 matches `(1, 2)`, the `.` matches `,3,(4,5)`, and the regex engine is happy to have found a match and buys demerphq and dave_the_m a beer - or so I hope. If you want to match `(4, 5)`, omit the `.`.	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re: Strange regex result
by moritz (Cardinal) on Jul 08, 2009 at 15:48 UTC

.* (?<= $tail )

\)

$tail

5)

So the paren group 1 matches (1, 2), the .* matches ,3,(4,5), and the regex engine is happy to have found a match and buys demerphq and dave_the_m a beer - or so I hope.

If you want to match (4, 5), omit the .*.

[reply]
[d/l]
[select]