attrTypes = new HTMLPurifier_AttrTypes(); $this->doctypes = new HTMLPurifier_DoctypeRegistry(); // setup default HTML doctypes // module reuse $common = array( 'CommonAttributes', 'Text', 'Hypertext', 'List', 'Presentation', 'Edit', 'Bdo', 'Tables', 'Image', 'StyleAttribute', 'Scripting' ); $transitional = array('Legacy', 'Target'); $this->doctypes->register( 'HTML 4.01 Transitional', false, array_merge($common, $transitional), array('TransformToStrict') // Tidy: Transitional ); $this->doctypes->register( 'HTML 4.01 Strict', false, array_merge($common), array('TransformToStrict') // Tidy: Strict ); $this->doctypes->register( 'XHTML 1.0 Transitional', true, array_merge($common, $transitional), array('TransformToStrict') // Tidy: Transitional, XHTML ); $this->doctypes->register( 'XHTML 1.0 Strict', true, array_merge($common), array('TransformToStrict') // Tidy: Strict, XHTML ); $this->doctypes->register( 'XHTML 1.1', true, array_merge($common), array('TransformToStrict', 'TransformToXHTML11') // Tidy: Strict, XHTML1_1 ); } /** * Registers a module to the recognized module list, useful for * overloading pre-existing modules. * @param $module Mixed: string module name, with or without * HTMLPurifier_HTMLModule prefix, or instance of * subclass of HTMLPurifier_HTMLModule. * @note This function will not call autoload, you must instantiate * (and thus invoke) autoload outside the method. * @note If a string is passed as a module name, different variants * will be tested in this order: * - Check for HTMLPurifier_HTMLModule_$name * - Check all prefixes with $name in order they were added * - Check for literal object name * - Throw fatal error * If your object name collides with an internal class, specify * your module manually. All modules must have been included * externally: registerModule will not perform inclusions for you! * @warning If your module has the same name as an already loaded * module, your module will overload the old one WITHOUT * warning. */ function registerModule($module) { if (is_string($module)) { // attempt to load the module $original_module = $module; $ok = false; foreach ($this->prefixes as $prefix) { $module = $prefix . $original_module; if ($this->_classExists($module)) { $ok = true; break; } } if (!$ok) { $module = $original_module; if (!$this->_classExists($module)) { trigger_error($original_module . ' module does not exist', E_USER_ERROR); return; } } $module = new $module(); } $this->registeredModules[$module->name] = $module; } /** * Safely tests for class existence without invoking __autoload in PHP5 * or greater. * @param $name String class name to test * @note If any other class needs it, we'll need to stash in a * conjectured "compatibility" class * @private */ function _classExists($name) { static $is_php_4 = null; if ($is_php_4 === null) { $is_php_4 = version_compare(PHP_VERSION, '5', '<'); } if ($is_php_4) { return class_exists($name); } else { return class_exists($name, false); } } /** * Adds a module to the current doctype by first registering it, * and then tacking it on to the active doctype */ function addModule($module) { $this->registerModule($module); if (is_object($module)) $module = $module->name; $this->userModules[] = $module; } /** * Adds a class prefix that registerModule() will use to resolve a * string name to a concrete class */ function addPrefix($prefix) { $this->prefixes[] = $prefix; } /** * Performs processing on modules, after being called you may * use getElement() and getElements() * @param $config Instance of HTMLPurifier_Config */ function setup($config) { $this->trusted = $config->get('HTML', 'Trusted'); // generate $doctype = $this->doctypes->make($config); $modules = $doctype->modules; // merge in custom modules $modules = array_merge($modules, $this->userModules); foreach ($modules as $module) { $this->processModule($module); } foreach ($doctype->tidyModules as $module) { $this->processModule($module); // FIXME!!! initialize the tidy modules here } // setup lookup table based on all valid modules foreach ($this->modules as $module) { foreach ($module->info as $name => $def) { if (!isset($this->elementLookup[$name])) { $this->elementLookup[$name] = array(); } $this->elementLookup[$name][] = $module->name; } } // note the different choice $this->contentSets = new HTMLPurifier_ContentSets( // content set assembly deals with all possible modules, // not just ones deemed to be "safe" $this->modules ); $this->attrCollections = new HTMLPurifier_AttrCollections( $this->attrTypes, // there is no way to directly disable a global attribute, // but using AllowedAttributes or simply not including // the module in your custom doctype should be sufficient $this->modules ); } /** * Takes a module and adds it to the active module collection, * registering it if necessary. */ function processModule($module) { if (!isset($this->registeredModules[$module]) || is_object($module)) { $this->registerModule($module); } $this->modules[$module] = $this->registeredModules[$module]; } /** * Retrieves merged element definitions. * @return Array of HTMLPurifier_ElementDef */ function getElements() { $elements = array(); foreach ($this->modules as $module) { foreach ($module->info as $name => $v) { if (isset($elements[$name])) continue; // if element is not safe, don't use it if (!$this->trusted && ($v->safe === false)) continue; $elements[$name] = $this->getElement($name); } } // remove dud elements, this happens when an element that // appeared to be safe actually wasn't foreach ($elements as $n => $v) { if ($v === false) unset($elements[$n]); } return $elements; } /** * Retrieves a single merged element definition * @param $name Name of element * @param $trusted Boolean trusted overriding parameter: set to true * if you want the full version of an element * @return Merged HTMLPurifier_ElementDef */ function getElement($name, $trusted = null) { $def = false; if ($trusted === null) $trusted = $this->trusted; $modules = $this->modules; if (!isset($this->elementLookup[$name])) { return false; } foreach($this->elementLookup[$name] as $module_name) { $module = $modules[$module_name]; $new_def = $module->info[$name]; // refuse to create/merge in a definition that is deemed unsafe if (!$trusted && ($new_def->safe === false)) { $def = false; continue; } if (!$def && $new_def->standalone) { // element with unknown safety is not to be trusted. // however, a merge-in definition with undefined safety // is fine if (!$trusted && !$new_def->safe) continue; $def = $new_def; } elseif ($def) { $def->mergeIn($new_def); } else { // could "save it for another day": // non-standalone definitions that don't have a standalone // to merge into could be deferred to the end continue; } // attribute value expansions $this->attrCollections->performInclusions($def->attr); $this->attrCollections->expandIdentifiers($def->attr, $this->attrTypes); // descendants_are_inline, for ChildDef_Chameleon if (is_string($def->content_model) && strpos($def->content_model, 'Inline') !== false) { if ($name != 'del' && $name != 'ins') { // this is for you, ins/del $def->descendants_are_inline = true; } } $this->contentSets->generateChildDef($def, $module); } return $def; } } ?>