HtmlParser's content-type handling code needs to be more flexible

Some servers return a content-type response header that has the charset value in a non-standard position. For example:

Content-Type: charset=utf-8; text/html

The HtmlParser code that attempts to extract the charset needs to be more flexible.