Improve the URLFilter regular expression

The URLFilter transforms textual URLs inside the document into HTML <a> tags. An URL can already be inside an attribute or an anchor, so this filter must not match any URL.

The current (as of 1.0RC3) regular expression does not avoid all URLs that should not be transformed. For example:

• <a href=""> <b>http://www.w3.org</b></a>
• [See http://www.w3.org for more details|http://www.w3.org]
• <script> domnode.appendChild(document.createTextNode("See http://www.w3.org for more details")) </script>