The sanitizer (both in PHP parser and Parsoid) has a number of references to the HTML4 standard and semantics. @Arlolra fixed one part of it (attribute parsing) to HTML5 semantics as part of T108134: Parse bare attributes according to the html5 spec. Other pieces of the sanitizer code (entities, whitelisted attributes, and anything else) should also be migrated over to HTML5 as well.