Well, people, including me, are always saying you shouldn't try to parse HTML with a regexp. It's not because it's impossible. It is possible. But you shouldn't do it because doing it with a regex is non-trivial. The program below will use a regex to extract a div element with a certain id from a piece of limited HTML. I say limited, because the regex doesn't take comments into account, or CDATA declared content. It won't be able to recover from misplaced </div> tags either.

#!/usr/bin/perl

use strict;
use warnings;

$_ = <<'--';
<div id = "foo">
Foo text
  <div id = "iwant">
     Text text.
     <div id = "insideiwant">
       Bla
     </div>
     <div id = "alsoinsideiwant">
       Bla bla <em>Bla</em>!
         <div id = "innerinnerdiv">
           Inner!
         </div>
     </div>
  </div>
</div>
--


my $div;

$div = qr {<div \s+ (?:id \s* = \s*
              (?: "[^"]*" | '[^']*' | [-.\d]+))? \s* >
           (?: (?>[^<]+) | <(?!/?div) | (??{$div}) ) *
           </div>}ix;

my $iwant = qr {<div \s+ id \s* = \s*
                   (?: "iwant" | 'iwant' | iwant) \s* >
                (?: (?>[^<]+) | <(?!/?div) | (??{$div}) ) *
                </div>}ix;

print $&, "\n" if /$iwant/; # Don't try to be the smartass
                            # to point out potential issues
                            # about $&. They are irrelevant 
                            # here.
__END__
<div id = "iwant">
     Text text.
     <div id = "insideiwant">
       Bla
     </div>
     <div id = "alsoinsideiwant">
       Bla bla <em>Bla</em>!
         <div id = "innerinnerdiv">
           Inner!
         </div>
     </div>
  </div>
[download]

Abigail