Someone here must be familiar with the Diaryland system. It's one of those weblog thing. Well, I was writing a script to have it shown to me when I startup, and thought that despite me being a novice it should be fairly easy. I've run into a hitch, though.

use strict;
use warnings;
use diagnostics;
use HTML::TokeParser;
use LWP::Simple;

print &parse_diaryland('username', 'password');

sub parse_diaryland
    {
    my($username, $password);
    my($diary_url);
    my($parsee_man);
    my($html, $tag);
    my($body);
    
    $username = shift();
    $password = shift();

    $diary_url = "http://$username:$password\@$username.diaryland.com"
+;    # urls are in the form http://username:password@username.diaryla
+nd.com
    
    $html = get($diary_url);
    
    print $html;
    
    $parsee_man = HTML::TokeParser->new(\$html);

    foreach $tag ($parsee_man->get_tag('TD'))
        {
        if ($tag->[1]{'align'} eq 'left' && $tag->[1]{'vAlign'} eq 'to
+p')        # fails here
            {
            my($secondtag);
            $secondtag = $parsee_man->get_tag();
            if ($secondtag->[0] eq 'FONT')
                {
                $body = get_text('/FONT');
                last();        # got the text body so quit loop
                }
            }
        }
    return($body);
    }
[download]

and we're matching this:

<TD align=left vAlign=top><FONT face="Verdana, Arial, Helvetica, sans-
+serif" size=2>I
This is the text body we wanna grab. foo, bar and angst.
</FONT>
[download]

That fails with "use of uninitialised variable in string eq at line 30". I can vaguely make sense of this: the page we're getting has some TD tags earlier, and they lack the align and valign attributes, so that would be undefined. How can I fix this, though?