For the following test case, doclint report each instance of illegal text at the position of the
end-tag. This position may be separated from the illegal text by large amount of content.
It would be more user friendly if the diagnostic identified the position of the illegal text.

Also, if a table doesn't have a summary or caption, doclint will not report on illegal text.

/**
 * <dl> <li>item1</li>illegal text <li>item2</li></dl>
 * <ol> <li>item1</li>illegal text <li>item2</li></ol>
 * <ul> <li>item1</li>illegal text <li>item2</li></ul>
 * <table> illegal text
 * <tbody> illegal text
 * <tr> illegal text <td> lots of table content</td></tr>
 * </tbody>
 * <caption>no caption or summary, no warning about illegal text</caption>
 * </table>
 */