Because I didn't know of this feature =). I'm just copying together scripts and twist them. I'm not a real programmer. For everyone who is struggeling with the same problem. Here is the piece of code.

//new_from_file parse it instantly so i have to make a new 
my $tree = HTML::TreeBuilder->new();
//set ignore_unkown to false
$tree->ignore_unknown(0);
//than parse the content
$tree->parse_content($webcrawler->content());

print CONTENT $tree;

if (my $div = $tree->look_down(_tag => "article" , class=>"article hen
+try")) {
         print $div->as_text(), "\n";
     } else {
         print "Not found";
     }
     $tree->delete();
[download]

This gives me the expected result. Thanks ^_^ Didn't know there was a ignore_unknown which is by dafault true.