my $p = HTML::TokeParser::Simple->new( $html );

my $token = $p->get_token;

my $tag = "meta";
my $att = "content";

if ( $token->is_tag($tag) ) { print "There is a $tag  tag\n" };

my $att_value = $token->get_attr($att);
print "The value of the $att attribute of the $tag tag is $att_value \
+n\n\n\n";

my $url = substr($att_value,6);
print $url,"\n";
[download]

Actually, there's no need to write your own parser to get at that URL: LWP::UserAgent, which WWW::Mechanize is an extension of, does it for you!

# UNTESTED; USE AT YOUR OWN PERIL :-)

my $response = $mech->submit_form(
        form_name => 'loginForm',
        # ...
);

  # Follow HTTP or META refreshes.
if (my $url = $response->header('Refresh'))
{
    $url =~ s/^\d+;url=//i;
    $mech->get($url);
}
[download]

There may even be way to get LWP::UA or WWW::Mech to automatically follow along, but I haven't delved into the documentation to see although I did stumble across a WWW::Mechanize wishlist item that you may want to keep your eye on.

    --k.