Why this regexp doesn't match nested parens?

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

I've added an \A ancor to well-known regexp from perldoc perlre which matches nested parens. But it doesn't match what I want.

use strict;
use warnings;
no warnings qw /uninitialized/;

my  $re = qr{\A
             \(
             (?:
                (?> [^()]+ )    # Non-parens without backtracking
              |
                (??{ $re })     # Group with matching parens
             )*
             \)
          }x;

              
while (<DATA>) {
    chomp;
    print "$_: ";
    print $& if /$re/;
    print "\n";
}

__DATA__
(*********)
(***(*)***)
((***))
(((***)))
((**)**(**))
(((())))
[download]

/usr/local/bin/perl -w /home/ccn/ab.pl 
(*********): (*********)
(***(*)***): 
((***)): 
(((***))): 
((**)**(**)): 
(((()))):

Comment on Why this regexp doesn't match nested parens? Download Code

Replies are listed 'Best First'.
Re: Why this regexp doesn't match nested parens? by broquaint (Abbot) on Apr 26, 2004 at 16:40 UTC
Using the `\A` anchor will make the regex fail as soon as it recurses since it will try to match the beginning of the string (hence only the first example string succeeding in your case). If you're matching nested parens then `Regexp::Common::balanced` can help you out e.g `use Regexp::Common; my @examples = qw/ (*******) (()*) (()) ((())) (()()) (((()))) /; for( @examples ) { chomp; print "$_: "; print $& if $_ =~ $RE{balanced}{-parens=>'()'}; print "\n"; } __output__ (*******): (*****) (()*): (()*) (()): (()) ((())): ((())) (()()): (()(**)) (((()))): (((())))` [download] HTH `_________ broquaint`	[reply] [d/l]
Re: Re: Why this regexp doesn't match nested parens? by ccn (Vicar) on Apr 26, 2004 at 16:44 UTC
thanks, I mess that	[reply]
Re: Why this regexp doesn't match nested parens? by japhy (Canon) on Apr 26, 2004 at 16:50 UTC
You also fall prey to using a `my()` variable in the same line that you declared it: `use strict; use warnings; # DON'T turn off the uninit warning my $x = qr/(a(??{$x})?b)/; print "x1($1)\n" if "alphabet" =~ $x; print "x2($1)\n" if "caaabbbc" =~ $x; my $y; $y = qr/(a(??{$y})?b)/; print "y1($1)\n" if "alphabet" =~ $y; print "y2($1)\n" if "caaabbbc" =~ $y; __END__ output: Use of uninitialized value in pattern match (m//) at rx line 2. x1(ab) x2(ab) y1(ab) y2(aaabbb)` [download] See? The `$x` regex is looking for `$main::x` (although it doesn't give you a strict-breakage, since it's not looking for package names in the code that handles `(??{ })`), and it's empty, so you get a warning. Take out the `\A` anchor, but also declare the variable first and then assign to it. _____________________________________________________ Jeff`[japhy]`Pinyan: Perl, regex, and perl hacker, who'd like a job (NYC-area) `s++=END;++y(;-P)}y js++=;shajsj<++y(p-q)}?print:??;`	[reply] [d/l] [select]
Re: Re: Why this regexp doesn't match nested parens? by ccn (Vicar) on Apr 26, 2004 at 16:53 UTC
Thanks, japhy, you are right	[reply]
Re: Why this regexp doesn't match nested parens? by jryan (Vicar) on Apr 26, 2004 at 19:42 UTC
The easy solution to your regex problem (although Text::Balanced is probably an easier solution to your general problem) is to just maintain modularity: `use strict; use warnings; no warnings qw /uninitialized/; # \A anchor removed, so this is the same regex from the perlfaq. my $re; $re = qr{ $ (?: (?> [^()]+ ) # Non-parens without backtracking \| (??{ $re }) # Group with matching parens )* $ }x; while (<DATA>) { chomp; print "$_: "; # here is the proper place for the anchor print $& if /\A$re/; print "\n"; } __DATA__ (*******) (()*) (()) ((())) (()()) (((())))` [download]	[reply] [d/l]