The '.' metacharacter doesn't match newlines by default. You'll need to add the s pattern modifier to make it do that. You may want to add the i modifier as well to make it case-insensitive if you don't know that the HTML tags are all uppercase:

   s{<FORM(.*?)/FORM>}{replacement text}is;
[download]

It may also be worth noting that if the HTML is not well formed you could end up removing a heck of a lot more than you intended using this regexp

Update: Whoops. Quite right davido. I assumed that everything was in a single scalar variable. Additionally, for a quickie list of regexp modifiers you can go to perlreref.

That's actually perlre, must be a small typo there.

Update: No typo, I misread perlreref for perlref.

Ummmmm... no... perlreref is the regex quick ref. If it were a typo then th link wouldn't have worked. ;)

Update: Um... still no. perlreref (perl regex reference faq), perlre (perl regex faq) and perlref (perl references and nested datastructures faq) are quite distinct.

#!/bin/perl5

use strict;
use warnings;
use HTML::TokeParser;

open HTML_FILE, 'form.html' or die;

my $tp = HTML::TokeParser->new( \*HTML_FILE )    or die;
my $html;
my $found_form = 0;

while ( my $t = $tp->get_token ) {
  $found_form++, next if $t->[0] eq 'S' and $t->[1] eq 'form';
  $found_form--, next if $t->[0] eq 'E' and $t->[1] eq 'form';
  next if $found_form;
  $html .= $t->[4] if $t->[0] eq 'S';
  $html .= $t->[1] if $t->[0] eq 'T' or $t->[0] eq 'C';
  $html .= $t->[2] if $t->[0] eq 'E';
}

close    HTML_FILE;
print "$html\n";

# ["S",  $t, $attr, $attrseq, $text]
# ["E",  $t, $text]
# ["T",  $text, $is_data]
# ["C",  $text]
# ["D",  $text]
# ["PI", $token0, $text]
[download]