User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2a2pre) Gecko/20090811 Minefield/3.6a2pre Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2a2pre) Gecko/20090811 Minefield/3.6a2pre The HTML5 parsing algorithm deals with this code in a way that is incompatible with legacy expectations. This came up in Bespin. <div> </div <ul> <li>Text </ul> The ul just disappears from the DOM with the HTML5 parser. Reproducible: Always

This is by design. See the section on tokenization.

As far as I can tell, eating up the <ul> start tag is compatible with legacy expectations when the expectations are based on IE behavior. I guess the reason why this hasn't been discovered with Bespin before is that Bespin doesn't work with IE. Marking as WONTFIX. Please escalate as a spec bug if there's a strong reason to believe that not aligning with IE would be better for overall Web compat.