HTML filter that is HTML5 compliant

Question

Is there a simple approach to add a HTML5 ruleset for HTMLPurifier?

HP can be configured to recognize new tags with:

// setup configurable HP instance
$config = HTMLPurifier_Config::createDefault();
$config->set('HTML.DefinitionID', 'html5 draft');
$config->set('HTML.DefinitionRev', 1);
$config->set('Cache.DefinitionImpl', null); // no caching
$def = $config->getHTMLDefinition(true);

// add a new tag
$form = $def->addElement(
  'article',   // name
  'Block',     // content set
  'Flow',      // allowed children
  'Common',    // attribute collection
  array(       // attributes
  )
);

// add a new attribute
$def->addAttribute('a', 'contextmenu', "ID");

However this is clearly a bit of work. Since there are a lot of new HTML5 tags and attributes that had to be registered. And new global attributes should be combinable even with existing HTML 4 tags. (It's difficult to judge from the docs how to augment core rules). So, is there a more useful config format/array structure to feed new and updated tag+attribute+context configuration (inline/block/empty/flow/..) into HTMLPurifier?

# mostly confused about how to extend existing tags:
$def->addAttribute('input', 'type', "...|...|...");

# or how to allow data-* attributes (if I actually wanted that):
$def->addAttribute("data-*", ...

And of course not all new HTML5 tags are fit for unrestricted allowance. HTMLPurifier is all about content filtering. Defining value constraints is where it's at. -- <canvas> for example might not be that big of a deal when it appears in user content. Because it's useless at best without Javascript (which HP already filters out). But other tags and attributes might be undesirable; so a flexible configuration structure is imperative for enabling/disabling tags and their associated attributes.

(Guess I should update some research...). But there's still no practical compendium/specification (no, XML DTDs aren't) that suits a HP configuration.

• http://simon.html5.org/html-elements
• http://www.w3.org/TR/html5-diff/#new-elements
• http://www.w3.org/TR/html5-diff/#new-attributes

(Uh, and HTML5 is no longer a draft.)

Answer 1

The php tidy extension can be configured to recognize html5 tags. http://tidy.sourceforge.net/docs/quickref.html#new-blocklevel-tags