[Summary view] [Print] [Text view]

   1  <?php
   2  
   3  class HTMLPurifier_HTMLModuleManager
   4  {
   5  
   6      /**
   7       * Instance of HTMLPurifier_DoctypeRegistry
   8       */
   9      public $doctypes;
  10  
  11      /**
  12       * Instance of current doctype
  13       */
  14      public $doctype;
  15  
  16      /**
  17       * Instance of HTMLPurifier_AttrTypes
  18       */
  19      public $attrTypes;
  20  
  21      /**
  22       * Active instances of modules for the specified doctype are
  23       * indexed, by name, in this array.
  24       */
  25      public $modules = array();
  26  
  27      /**
  28       * Array of recognized HTMLPurifier_Module instances, indexed by
  29       * module's class name. This array is usually lazy loaded, but a
  30       * user can overload a module by pre-emptively registering it.
  31       */
  32      public $registeredModules = array();
  33  
  34      /**
  35       * List of extra modules that were added by the user using addModule().
  36       * These get unconditionally merged into the current doctype, whatever
  37       * it may be.
  38       */
  39      public $userModules = array();
  40  
  41      /**
  42       * Associative array of element name to list of modules that have
  43       * definitions for the element; this array is dynamically filled.
  44       */
  45      public $elementLookup = array();
  46  
  47      /** List of prefixes we should use for registering small names */
  48      public $prefixes = array('HTMLPurifier_HTMLModule_');
  49  
  50      public $contentSets;     /**< Instance of HTMLPurifier_ContentSets */
  51      public $attrCollections; /**< Instance of HTMLPurifier_AttrCollections */
  52  
  53      /** If set to true, unsafe elements and attributes will be allowed */
  54      public $trusted = false;
  55  
  56      public function __construct() {
  57  
  58          // editable internal objects
  59          $this->attrTypes = new HTMLPurifier_AttrTypes();
  60          $this->doctypes  = new HTMLPurifier_DoctypeRegistry();
  61  
  62          // setup basic modules
  63          $common = array(
  64              'CommonAttributes', 'Text', 'Hypertext', 'List',
  65              'Presentation', 'Edit', 'Bdo', 'Tables', 'Image',
  66              'StyleAttribute',
  67              // Unsafe:
  68              'Scripting', 'Object',  'Forms',
  69              // Sorta legacy, but present in strict:
  70              'Name',
  71          );
  72          $transitional = array('Legacy', 'Target');
  73          $xml = array('XMLCommonAttributes');
  74          $non_xml = array('NonXMLCommonAttributes');
  75  
  76          // setup basic doctypes
  77          $this->doctypes->register(
  78              'HTML 4.01 Transitional', false,
  79              array_merge($common, $transitional, $non_xml),
  80              array('Tidy_Transitional', 'Tidy_Proprietary'),
  81              array(),
  82              '-//W3C//DTD HTML 4.01 Transitional//EN',
  83              'http://www.w3.org/TR/html4/loose.dtd'
  84          );
  85  
  86          $this->doctypes->register(
  87              'HTML 4.01 Strict', false,
  88              array_merge($common, $non_xml),
  89              array('Tidy_Strict', 'Tidy_Proprietary', 'Tidy_Name'),
  90              array(),
  91              '-//W3C//DTD HTML 4.01//EN',
  92              'http://www.w3.org/TR/html4/strict.dtd'
  93          );
  94  
  95          $this->doctypes->register(
  96              'XHTML 1.0 Transitional', true,
  97              array_merge($common, $transitional, $xml, $non_xml),
  98              array('Tidy_Transitional', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_Name'),
  99              array(),
 100              '-//W3C//DTD XHTML 1.0 Transitional//EN',
 101              'http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd'
 102          );
 103  
 104          $this->doctypes->register(
 105              'XHTML 1.0 Strict', true,
 106              array_merge($common, $xml, $non_xml),
 107              array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Strict', 'Tidy_Proprietary', 'Tidy_Name'),
 108              array(),
 109              '-//W3C//DTD XHTML 1.0 Strict//EN',
 110              'http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd'
 111          );
 112  
 113          $this->doctypes->register(
 114              'XHTML 1.1', true,
 115              array_merge($common, $xml, array('Ruby')),
 116              array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_Strict', 'Tidy_Name'), // Tidy_XHTML1_1
 117              array(),
 118              '-//W3C//DTD XHTML 1.1//EN',
 119              'http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd'
 120          );
 121  
 122      }
 123  
 124      /**
 125       * Registers a module to the recognized module list, useful for
 126       * overloading pre-existing modules.
 127       * @param $module Mixed: string module name, with or without
 128       *                HTMLPurifier_HTMLModule prefix, or instance of
 129       *                subclass of HTMLPurifier_HTMLModule.
 130       * @param $overload Boolean whether or not to overload previous modules.
 131       *                  If this is not set, and you do overload a module,
 132       *                  HTML Purifier will complain with a warning.
 133       * @note This function will not call autoload, you must instantiate
 134       *       (and thus invoke) autoload outside the method.
 135       * @note If a string is passed as a module name, different variants
 136       *       will be tested in this order:
 137       *          - Check for HTMLPurifier_HTMLModule_$name
 138       *          - Check all prefixes with $name in order they were added
 139       *          - Check for literal object name
 140       *          - Throw fatal error
 141       *       If your object name collides with an internal class, specify
 142       *       your module manually. All modules must have been included
 143       *       externally: registerModule will not perform inclusions for you!
 144       */
 145      public function registerModule($module, $overload = false) {
 146          if (is_string($module)) {
 147              // attempt to load the module
 148              $original_module = $module;
 149              $ok = false;
 150              foreach ($this->prefixes as $prefix) {
 151                  $module = $prefix . $original_module;
 152                  if (class_exists($module)) {
 153                      $ok = true;
 154                      break;
 155                  }
 156              }
 157              if (!$ok) {
 158                  $module = $original_module;
 159                  if (!class_exists($module)) {
 160                      trigger_error($original_module . ' module does not exist',
 161                          E_USER_ERROR);
 162                      return;
 163                  }
 164              }
 165              $module = new $module();
 166          }
 167          if (empty($module->name)) {
 168              trigger_error('Module instance of ' . get_class($module) . ' must have name');
 169              return;
 170          }
 171          if (!$overload && isset($this->registeredModules[$module->name])) {
 172              trigger_error('Overloading ' . $module->name . ' without explicit overload parameter', E_USER_WARNING);
 173          }
 174          $this->registeredModules[$module->name] = $module;
 175      }
 176  
 177      /**
 178       * Adds a module to the current doctype by first registering it,
 179       * and then tacking it on to the active doctype
 180       */
 181      public function addModule($module) {
 182          $this->registerModule($module);
 183          if (is_object($module)) $module = $module->name;
 184          $this->userModules[] = $module;
 185      }
 186  
 187      /**
 188       * Adds a class prefix that registerModule() will use to resolve a
 189       * string name to a concrete class
 190       */
 191      public function addPrefix($prefix) {
 192          $this->prefixes[] = $prefix;
 193      }
 194  
 195      /**
 196       * Performs processing on modules, after being called you may
 197       * use getElement() and getElements()
 198       * @param $config Instance of HTMLPurifier_Config
 199       */
 200      public function setup($config) {
 201  
 202          $this->trusted = $config->get('HTML', 'Trusted');
 203  
 204          // generate
 205          $this->doctype = $this->doctypes->make($config);
 206          $modules = $this->doctype->modules;
 207  
 208          // take out the default modules that aren't allowed
 209          $lookup = $config->get('HTML', 'AllowedModules');
 210          $special_cases = $config->get('HTML', 'CoreModules');
 211  
 212          if (is_array($lookup)) {
 213              foreach ($modules as $k => $m) {
 214                  if (isset($special_cases[$m])) continue;
 215                  if (!isset($lookup[$m])) unset($modules[$k]);
 216              }
 217          }
 218  
 219          // add proprietary module (this gets special treatment because
 220          // it is completely removed from doctypes, etc.)
 221          if ($config->get('HTML', 'Proprietary')) {
 222              $modules[] = 'Proprietary';
 223          }
 224  
 225          // add SafeObject/Safeembed modules
 226          if ($config->get('HTML', 'SafeObject')) {
 227              $modules[] = 'SafeObject';
 228          }
 229          if ($config->get('HTML', 'SafeEmbed')) {
 230              $modules[] = 'SafeEmbed';
 231          }
 232  
 233          // merge in custom modules
 234          $modules = array_merge($modules, $this->userModules);
 235  
 236          foreach ($modules as $module) {
 237              $this->processModule($module);
 238              $this->modules[$module]->setup($config);
 239          }
 240  
 241          foreach ($this->doctype->tidyModules as $module) {
 242              $this->processModule($module);
 243              $this->modules[$module]->setup($config);
 244          }
 245  
 246          // prepare any injectors
 247          foreach ($this->modules as $module) {
 248              $n = array();
 249              foreach ($module->info_injector as $i => $injector) {
 250                  if (!is_object($injector)) {
 251                      $class = "HTMLPurifier_Injector_$injector";
 252                      $injector = new $class;
 253                  }
 254                  $n[$injector->name] = $injector;
 255              }
 256              $module->info_injector = $n;
 257          }
 258  
 259          // setup lookup table based on all valid modules
 260          foreach ($this->modules as $module) {
 261              foreach ($module->info as $name => $def) {
 262                  if (!isset($this->elementLookup[$name])) {
 263                      $this->elementLookup[$name] = array();
 264                  }
 265                  $this->elementLookup[$name][] = $module->name;
 266              }
 267          }
 268  
 269          // note the different choice
 270          $this->contentSets = new HTMLPurifier_ContentSets(
 271              // content set assembly deals with all possible modules,
 272              // not just ones deemed to be "safe"
 273              $this->modules
 274          );
 275          $this->attrCollections = new HTMLPurifier_AttrCollections(
 276              $this->attrTypes,
 277              // there is no way to directly disable a global attribute,
 278              // but using AllowedAttributes or simply not including
 279              // the module in your custom doctype should be sufficient
 280              $this->modules
 281          );
 282      }
 283  
 284      /**
 285       * Takes a module and adds it to the active module collection,
 286       * registering it if necessary.
 287       */
 288      public function processModule($module) {
 289          if (!isset($this->registeredModules[$module]) || is_object($module)) {
 290              $this->registerModule($module);
 291          }
 292          $this->modules[$module] = $this->registeredModules[$module];
 293      }
 294  
 295      /**
 296       * Retrieves merged element definitions.
 297       * @return Array of HTMLPurifier_ElementDef
 298       */
 299      public function getElements() {
 300  
 301          $elements = array();
 302          foreach ($this->modules as $module) {
 303              if (!$this->trusted && !$module->safe) continue;
 304              foreach ($module->info as $name => $v) {
 305                  if (isset($elements[$name])) continue;
 306                  $elements[$name] = $this->getElement($name);
 307              }
 308          }
 309  
 310          // remove dud elements, this happens when an element that
 311          // appeared to be safe actually wasn't
 312          foreach ($elements as $n => $v) {
 313              if ($v === false) unset($elements[$n]);
 314          }
 315  
 316          return $elements;
 317  
 318      }
 319  
 320      /**
 321       * Retrieves a single merged element definition
 322       * @param $name Name of element
 323       * @param $trusted Boolean trusted overriding parameter: set to true
 324       *                 if you want the full version of an element
 325       * @return Merged HTMLPurifier_ElementDef
 326       * @note You may notice that modules are getting iterated over twice (once
 327       *       in getElements() and once here). This
 328       *       is because
 329       */
 330      public function getElement($name, $trusted = null) {
 331  
 332          if (!isset($this->elementLookup[$name])) {
 333              return false;
 334          }
 335  
 336          // setup global state variables
 337          $def = false;
 338          if ($trusted === null) $trusted = $this->trusted;
 339  
 340          // iterate through each module that has registered itself to this
 341          // element
 342          foreach($this->elementLookup[$name] as $module_name) {
 343  
 344              $module = $this->modules[$module_name];
 345  
 346              // refuse to create/merge from a module that is deemed unsafe--
 347              // pretend the module doesn't exist--when trusted mode is not on.
 348              if (!$trusted && !$module->safe) {
 349                  continue;
 350              }
 351  
 352              // clone is used because, ideally speaking, the original
 353              // definition should not be modified. Usually, this will
 354              // make no difference, but for consistency's sake
 355              $new_def = clone $module->info[$name];
 356  
 357              if (!$def && $new_def->standalone) {
 358                  $def = $new_def;
 359              } elseif ($def) {
 360                  // This will occur even if $new_def is standalone. In practice,
 361                  // this will usually result in a full replacement.
 362                  $def->mergeIn($new_def);
 363              } else {
 364                  // :TODO:
 365                  // non-standalone definitions that don't have a standalone
 366                  // to merge into could be deferred to the end
 367                  continue;
 368              }
 369  
 370              // attribute value expansions
 371              $this->attrCollections->performInclusions($def->attr);
 372              $this->attrCollections->expandIdentifiers($def->attr, $this->attrTypes);
 373  
 374              // descendants_are_inline, for ChildDef_Chameleon
 375              if (is_string($def->content_model) &&
 376                  strpos($def->content_model, 'Inline') !== false) {
 377                  if ($name != 'del' && $name != 'ins') {
 378                      // this is for you, ins/del
 379                      $def->descendants_are_inline = true;
 380                  }
 381              }
 382  
 383              $this->contentSets->generateChildDef($def, $module);
 384          }
 385  
 386          // This can occur if there is a blank definition, but no base to
 387          // mix it in with
 388          if (!$def) return false;
 389  
 390          // add information on required attributes
 391          foreach ($def->attr as $attr_name => $attr_def) {
 392              if ($attr_def->required) {
 393                  $def->required_attr[] = $attr_name;
 394              }
 395          }
 396  
 397          return $def;
 398  
 399      }
 400  
 401  }
 402  
 403  // vim: et sw=4 sts=4