Comment subject is calculated incorrectly when the comment text contains only HTML tags

I' m using a smiley module. I have entered ':)' as comment body. After all filters are done, ':)' will be replaced with

<p><img class="smiley" src="/emoticon/smile.gif" width="24" height="24" title="Smiling" alt="Smiling" /></p>

Nice.

And the following code will make us a zero-length code due to strip_tags magic.

  if (trim($form_values['subject']) == '') {
    // The body may be in any format, so we:
    // 1) Filter it into HTML
    // 2) Strip out all HTML tags
    // 3) Convert entities back to plain-text.
  // Note: format is checked by check_markup().
    $form_values['subject'] = truncate_utf8(decode_entities(strip_tags(check_markup($form_values['comment'], $form_values['format']))), 29, TRUE);
  }
  return $form_values;
}

Well. I know you can say - its smileys issue. But I think ot only smiley, even some full-working module can break Drupal functionality in such a way.

The solution: the best I think is to ask user to enter a comment title, 'cos it can't be automattically generated.