Detect and mummify XHTML in HTML files.
Even when we support general HTML parsing in GUISE-93, we'll still need to detect if the .html file is really using XHTML and parseable as XML. Since we already support parsing XHTML, we can add support for detecting XML in .html files and parsing them as XHTML as well. We can leave support for actual HTML in .html files to GUISE-93.
A comprehensive and nuanced approach would involve checking for namespace declarations and even perhaps multiple parsing passes, but as a practical matter it's much simpler just to make a rule saying that XHTML in an .html file must begin with an XML prolog to be processed as XHTML in Guise Mummy, at least for now.