How can I get Hpricot to play nice with HTML5?

I am using Hpricot to parse a theme 开发者_JS百科file. I have noticed, however, that if I feed a valid HTML5 document into Hpricot(), it auto-closes HTML5 tags (like <section>), and messes with the DOCTYPE.

Are there any extensions to Hpricot, or perhaps a flag I need to set, that will allow HTML5 documents to be parsed correctly?

I know it kind of works around the direct question but I would suggest you try Nokogiri http://nokogiri.org/ as mentioned in some of the comments on your question post. I've had no issues with it parsing any HTML/XML like structured text, including HTML5.

I think Hpricot's to_original_html method is exactly what you're looking for.

From the docs, to_original_html

Attempts to preserve the original HTML of the document, only outputing new tags for elements which have changed.