[Summary view] [Print] [Text view]

   1  <?php
   2  
   3  class HTMLPurifier_HTMLModuleManager
   4  {
   5  
   6      /**
   7       * @type HTMLPurifier_DoctypeRegistry
   8       */
   9      public $doctypes;
  10  
  11      /**
  12       * Instance of current doctype.
  13       * @type string
  14       */
  15      public $doctype;
  16  
  17      /**
  18       * @type HTMLPurifier_AttrTypes
  19       */
  20      public $attrTypes;
  21  
  22      /**
  23       * Active instances of modules for the specified doctype are
  24       * indexed, by name, in this array.
  25       * @type HTMLPurifier_HTMLModule[]
  26       */
  27      public $modules = array();
  28  
  29      /**
  30       * Array of recognized HTMLPurifier_HTMLModule instances,
  31       * indexed by module's class name. This array is usually lazy loaded, but a
  32       * user can overload a module by pre-emptively registering it.
  33       * @type HTMLPurifier_HTMLModule[]
  34       */
  35      public $registeredModules = array();
  36  
  37      /**
  38       * List of extra modules that were added by the user
  39       * using addModule(). These get unconditionally merged into the current doctype, whatever
  40       * it may be.
  41       * @type HTMLPurifier_HTMLModule[]
  42       */
  43      public $userModules = array();
  44  
  45      /**
  46       * Associative array of element name to list of modules that have
  47       * definitions for the element; this array is dynamically filled.
  48       * @type array
  49       */
  50      public $elementLookup = array();
  51  
  52      /**
  53       * List of prefixes we should use for registering small names.
  54       * @type array
  55       */
  56      public $prefixes = array('HTMLPurifier_HTMLModule_');
  57  
  58      /**
  59       * @type HTMLPurifier_ContentSets
  60       */
  61      public $contentSets;
  62  
  63      /**
  64       * @type HTMLPurifier_AttrCollections
  65       */
  66      public $attrCollections;
  67  
  68      /**
  69       * If set to true, unsafe elements and attributes will be allowed.
  70       * @type bool
  71       */
  72      public $trusted = false;
  73  
  74      public function __construct()
  75      {
  76          // editable internal objects
  77          $this->attrTypes = new HTMLPurifier_AttrTypes();
  78          $this->doctypes  = new HTMLPurifier_DoctypeRegistry();
  79  
  80          // setup basic modules
  81          $common = array(
  82              'CommonAttributes', 'Text', 'Hypertext', 'List',
  83              'Presentation', 'Edit', 'Bdo', 'Tables', 'Image',
  84              'StyleAttribute',
  85              // Unsafe:
  86              'Scripting', 'Object', 'Forms',
  87              // Sorta legacy, but present in strict:
  88              'Name',
  89          );
  90          $transitional = array('Legacy', 'Target', 'Iframe');
  91          $xml = array('XMLCommonAttributes');
  92          $non_xml = array('NonXMLCommonAttributes');
  93  
  94          // setup basic doctypes
  95          $this->doctypes->register(
  96              'HTML 4.01 Transitional',
  97              false,
  98              array_merge($common, $transitional, $non_xml),
  99              array('Tidy_Transitional', 'Tidy_Proprietary'),
 100              array(),
 101              '-//W3C//DTD HTML 4.01 Transitional//EN',
 102              'http://www.w3.org/TR/html4/loose.dtd'
 103          );
 104  
 105          $this->doctypes->register(
 106              'HTML 4.01 Strict',
 107              false,
 108              array_merge($common, $non_xml),
 109              array('Tidy_Strict', 'Tidy_Proprietary', 'Tidy_Name'),
 110              array(),
 111              '-//W3C//DTD HTML 4.01//EN',
 112              'http://www.w3.org/TR/html4/strict.dtd'
 113          );
 114  
 115          $this->doctypes->register(
 116              'XHTML 1.0 Transitional',
 117              true,
 118              array_merge($common, $transitional, $xml, $non_xml),
 119              array('Tidy_Transitional', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_Name'),
 120              array(),
 121              '-//W3C//DTD XHTML 1.0 Transitional//EN',
 122              'http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd'
 123          );
 124  
 125          $this->doctypes->register(
 126              'XHTML 1.0 Strict',
 127              true,
 128              array_merge($common, $xml, $non_xml),
 129              array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Strict', 'Tidy_Proprietary', 'Tidy_Name'),
 130              array(),
 131              '-//W3C//DTD XHTML 1.0 Strict//EN',
 132              'http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd'
 133          );
 134  
 135          $this->doctypes->register(
 136              'XHTML 1.1',
 137              true,
 138              // Iframe is a real XHTML 1.1 module, despite being
 139              // "transitional"!
 140              array_merge($common, $xml, array('Ruby', 'Iframe')),
 141              array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_Strict', 'Tidy_Name'), // Tidy_XHTML1_1
 142              array(),
 143              '-//W3C//DTD XHTML 1.1//EN',
 144              'http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd'
 145          );
 146  
 147      }
 148  
 149      /**
 150       * Registers a module to the recognized module list, useful for
 151       * overloading pre-existing modules.
 152       * @param $module Mixed: string module name, with or without
 153       *                HTMLPurifier_HTMLModule prefix, or instance of
 154       *                subclass of HTMLPurifier_HTMLModule.
 155       * @param $overload Boolean whether or not to overload previous modules.
 156       *                  If this is not set, and you do overload a module,
 157       *                  HTML Purifier will complain with a warning.
 158       * @note This function will not call autoload, you must instantiate
 159       *       (and thus invoke) autoload outside the method.
 160       * @note If a string is passed as a module name, different variants
 161       *       will be tested in this order:
 162       *          - Check for HTMLPurifier_HTMLModule_$name
 163       *          - Check all prefixes with $name in order they were added
 164       *          - Check for literal object name
 165       *          - Throw fatal error
 166       *       If your object name collides with an internal class, specify
 167       *       your module manually. All modules must have been included
 168       *       externally: registerModule will not perform inclusions for you!
 169       */
 170      public function registerModule($module, $overload = false)
 171      {
 172          if (is_string($module)) {
 173              // attempt to load the module
 174              $original_module = $module;
 175              $ok = false;
 176              foreach ($this->prefixes as $prefix) {
 177                  $module = $prefix . $original_module;
 178                  if (class_exists($module)) {
 179                      $ok = true;
 180                      break;
 181                  }
 182              }
 183              if (!$ok) {
 184                  $module = $original_module;
 185                  if (!class_exists($module)) {
 186                      trigger_error(
 187                          $original_module . ' module does not exist',
 188                          E_USER_ERROR
 189                      );
 190                      return;
 191                  }
 192              }
 193              $module = new $module();
 194          }
 195          if (empty($module->name)) {
 196              trigger_error('Module instance of ' . get_class($module) . ' must have name');
 197              return;
 198          }
 199          if (!$overload && isset($this->registeredModules[$module->name])) {
 200              trigger_error('Overloading ' . $module->name . ' without explicit overload parameter', E_USER_WARNING);
 201          }
 202          $this->registeredModules[$module->name] = $module;
 203      }
 204  
 205      /**
 206       * Adds a module to the current doctype by first registering it,
 207       * and then tacking it on to the active doctype
 208       */
 209      public function addModule($module)
 210      {
 211          $this->registerModule($module);
 212          if (is_object($module)) {
 213              $module = $module->name;
 214          }
 215          $this->userModules[] = $module;
 216      }
 217  
 218      /**
 219       * Adds a class prefix that registerModule() will use to resolve a
 220       * string name to a concrete class
 221       */
 222      public function addPrefix($prefix)
 223      {
 224          $this->prefixes[] = $prefix;
 225      }
 226  
 227      /**
 228       * Performs processing on modules, after being called you may
 229       * use getElement() and getElements()
 230       * @param HTMLPurifier_Config $config
 231       */
 232      public function setup($config)
 233      {
 234          $this->trusted = $config->get('HTML.Trusted');
 235  
 236          // generate
 237          $this->doctype = $this->doctypes->make($config);
 238          $modules = $this->doctype->modules;
 239  
 240          // take out the default modules that aren't allowed
 241          $lookup = $config->get('HTML.AllowedModules');
 242          $special_cases = $config->get('HTML.CoreModules');
 243  
 244          if (is_array($lookup)) {
 245              foreach ($modules as $k => $m) {
 246                  if (isset($special_cases[$m])) {
 247                      continue;
 248                  }
 249                  if (!isset($lookup[$m])) {
 250                      unset($modules[$k]);
 251                  }
 252              }
 253          }
 254  
 255          // custom modules
 256          if ($config->get('HTML.Proprietary')) {
 257              $modules[] = 'Proprietary';
 258          }
 259          if ($config->get('HTML.SafeObject')) {
 260              $modules[] = 'SafeObject';
 261          }
 262          if ($config->get('HTML.SafeEmbed')) {
 263              $modules[] = 'SafeEmbed';
 264          }
 265          if ($config->get('HTML.SafeScripting') !== array()) {
 266              $modules[] = 'SafeScripting';
 267          }
 268          if ($config->get('HTML.Nofollow')) {
 269              $modules[] = 'Nofollow';
 270          }
 271          if ($config->get('HTML.TargetBlank')) {
 272              $modules[] = 'TargetBlank';
 273          }
 274  
 275          // merge in custom modules
 276          $modules = array_merge($modules, $this->userModules);
 277  
 278          foreach ($modules as $module) {
 279              $this->processModule($module);
 280              $this->modules[$module]->setup($config);
 281          }
 282  
 283          foreach ($this->doctype->tidyModules as $module) {
 284              $this->processModule($module);
 285              $this->modules[$module]->setup($config);
 286          }
 287  
 288          // prepare any injectors
 289          foreach ($this->modules as $module) {
 290              $n = array();
 291              foreach ($module->info_injector as $injector) {
 292                  if (!is_object($injector)) {
 293                      $class = "HTMLPurifier_Injector_$injector";
 294                      $injector = new $class;
 295                  }
 296                  $n[$injector->name] = $injector;
 297              }
 298              $module->info_injector = $n;
 299          }
 300  
 301          // setup lookup table based on all valid modules
 302          foreach ($this->modules as $module) {
 303              foreach ($module->info as $name => $def) {
 304                  if (!isset($this->elementLookup[$name])) {
 305                      $this->elementLookup[$name] = array();
 306                  }
 307                  $this->elementLookup[$name][] = $module->name;
 308              }
 309          }
 310  
 311          // note the different choice
 312          $this->contentSets = new HTMLPurifier_ContentSets(
 313              // content set assembly deals with all possible modules,
 314              // not just ones deemed to be "safe"
 315              $this->modules
 316          );
 317          $this->attrCollections = new HTMLPurifier_AttrCollections(
 318              $this->attrTypes,
 319              // there is no way to directly disable a global attribute,
 320              // but using AllowedAttributes or simply not including
 321              // the module in your custom doctype should be sufficient
 322              $this->modules
 323          );
 324      }
 325  
 326      /**
 327       * Takes a module and adds it to the active module collection,
 328       * registering it if necessary.
 329       */
 330      public function processModule($module)
 331      {
 332          if (!isset($this->registeredModules[$module]) || is_object($module)) {
 333              $this->registerModule($module);
 334          }
 335          $this->modules[$module] = $this->registeredModules[$module];
 336      }
 337  
 338      /**
 339       * Retrieves merged element definitions.
 340       * @return Array of HTMLPurifier_ElementDef
 341       */
 342      public function getElements()
 343      {
 344          $elements = array();
 345          foreach ($this->modules as $module) {
 346              if (!$this->trusted && !$module->safe) {
 347                  continue;
 348              }
 349              foreach ($module->info as $name => $v) {
 350                  if (isset($elements[$name])) {
 351                      continue;
 352                  }
 353                  $elements[$name] = $this->getElement($name);
 354              }
 355          }
 356  
 357          // remove dud elements, this happens when an element that
 358          // appeared to be safe actually wasn't
 359          foreach ($elements as $n => $v) {
 360              if ($v === false) {
 361                  unset($elements[$n]);
 362              }
 363          }
 364  
 365          return $elements;
 366  
 367      }
 368  
 369      /**
 370       * Retrieves a single merged element definition
 371       * @param string $name Name of element
 372       * @param bool $trusted Boolean trusted overriding parameter: set to true
 373       *                 if you want the full version of an element
 374       * @return HTMLPurifier_ElementDef Merged HTMLPurifier_ElementDef
 375       * @note You may notice that modules are getting iterated over twice (once
 376       *       in getElements() and once here). This
 377       *       is because
 378       */
 379      public function getElement($name, $trusted = null)
 380      {
 381          if (!isset($this->elementLookup[$name])) {
 382              return false;
 383          }
 384  
 385          // setup global state variables
 386          $def = false;
 387          if ($trusted === null) {
 388              $trusted = $this->trusted;
 389          }
 390  
 391          // iterate through each module that has registered itself to this
 392          // element
 393          foreach ($this->elementLookup[$name] as $module_name) {
 394              $module = $this->modules[$module_name];
 395  
 396              // refuse to create/merge from a module that is deemed unsafe--
 397              // pretend the module doesn't exist--when trusted mode is not on.
 398              if (!$trusted && !$module->safe) {
 399                  continue;
 400              }
 401  
 402              // clone is used because, ideally speaking, the original
 403              // definition should not be modified. Usually, this will
 404              // make no difference, but for consistency's sake
 405              $new_def = clone $module->info[$name];
 406  
 407              if (!$def && $new_def->standalone) {
 408                  $def = $new_def;
 409              } elseif ($def) {
 410                  // This will occur even if $new_def is standalone. In practice,
 411                  // this will usually result in a full replacement.
 412                  $def->mergeIn($new_def);
 413              } else {
 414                  // :TODO:
 415                  // non-standalone definitions that don't have a standalone
 416                  // to merge into could be deferred to the end
 417                  // HOWEVER, it is perfectly valid for a non-standalone
 418                  // definition to lack a standalone definition, even
 419                  // after all processing: this allows us to safely
 420                  // specify extra attributes for elements that may not be
 421                  // enabled all in one place.  In particular, this might
 422                  // be the case for trusted elements.  WARNING: care must
 423                  // be taken that the /extra/ definitions are all safe.
 424                  continue;
 425              }
 426  
 427              // attribute value expansions
 428              $this->attrCollections->performInclusions($def->attr);
 429              $this->attrCollections->expandIdentifiers($def->attr, $this->attrTypes);
 430  
 431              // descendants_are_inline, for ChildDef_Chameleon
 432              if (is_string($def->content_model) &&
 433                  strpos($def->content_model, 'Inline') !== false) {
 434                  if ($name != 'del' && $name != 'ins') {
 435                      // this is for you, ins/del
 436                      $def->descendants_are_inline = true;
 437                  }
 438              }
 439  
 440              $this->contentSets->generateChildDef($def, $module);
 441          }
 442  
 443          // This can occur if there is a blank definition, but no base to
 444          // mix it in with
 445          if (!$def) {
 446              return false;
 447          }
 448  
 449          // add information on required attributes
 450          foreach ($def->attr as $attr_name => $attr_def) {
 451              if ($attr_def->required) {
 452                  $def->required_attr[] = $attr_name;
 453              }
 454          }
 455          return $def;
 456      }
 457  }
 458  
 459  // vim: et sw=4 sts=4