It's incomplete. Sorry.

    sub _html_tag {
        my ( $tag, $attr, $attrseq) = @_;
        my $html;
        $html = "<$tag";
        if ($attrseq and ref($attrseq) eq 'ARRAY') {
            foreach my $key (@$attrseq) {
                if (defined $attr->{$key}) {
                    $html .= " $key="._arg_escape($attr->{$key});
                } else {
                    $html .= ' '.$key;
                }
            }
        } elsif ($attr and ref($attr)) {
            foreach my $key (keys %$attr) {
                if (defined $attr->{$key}) {
                    $html .= " $key="._arg_escape($attr->{$key});
                } else {
                    $html .= ' '.$key;
                }
            }
        }
        $html .= ">";
        return $html;
    }

    sub _arg_escape {
        my $arg = shift;
        return qq{"$arg"} if ($arg !~ /"/);
        return qq{'$arg'} if ($arg !~ /'/);
        $arg =~ s/"/&dblquote;/g;
        return qq{"$arg"};
    }
[download]

Not sure it's complete like this. I never got around to releasing this as a module. If it's not complete, either download the Jenda.Rex zip from http://jenda.krynicky.cz/#Jenda.Rex, extract the .pm and dissect that (remove all references to Win32::OLE (needed only when the module is wrapped as a COM DLL for use in VB(script)) and Win32::Registry (only used to find out the code page used by the system), remove the whole package JendaRex::CSVParser, ...) or send me a message with your email and I'll send you the module.