mirror of
https://github.com/ezyang/htmlpurifier.git
synced 2025-07-13 18:46:34 +02:00
Release 2.0.0, merged in 1026 to HEAD.
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/branches/strict@1179 48356398-32a2-884e-a903-53898d9a118a
This commit is contained in:
Edward Z. Yang
@ -2,6 +2,8 @@
|
||||
|
||||
require_once 'HTMLPurifier/HTMLModule.php';
|
||||
require_once 'HTMLPurifier/ElementDef.php';
|
||||
require_once 'HTMLPurifier/Doctype.php';
|
||||
require_once 'HTMLPurifier/DoctypeRegistry.php';
|
||||
|
||||
require_once 'HTMLPurifier/ContentSets.php';
|
||||
require_once 'HTMLPurifier/AttrTypes.php';
|
||||
@ -23,14 +25,20 @@ require_once 'HTMLPurifier/HTMLModule/Image.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/StyleAttribute.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Legacy.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Target.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Scripting.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/XMLCommonAttributes.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php';
|
||||
|
||||
// proprietary modules
|
||||
require_once 'HTMLPurifier/HTMLModule/TransformToStrict.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/TransformToXHTML11.php';
|
||||
// tidy modules
|
||||
require_once 'HTMLPurifier/HTMLModule/Tidy.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Tidy/XHTMLAndHTML4.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Tidy/XHTML.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Tidy/XHTMLStrict.php';
|
||||
require_once 'HTMLPurifier/HTMLModule/Tidy/Proprietary.php';
|
||||
|
||||
HTMLPurifier_ConfigSchema::define(
|
||||
'HTML', 'Doctype', null, 'string/null',
|
||||
'Doctype to use, valid values are HTML 4.01 Transitional, HTML 4.01 '.
|
||||
'Doctype to use, pre-defined values are HTML 4.01 Transitional, HTML 4.01 '.
|
||||
'Strict, XHTML 1.0 Transitional, XHTML 1.0 Strict, XHTML 1.1. '.
|
||||
'Technically speaking this is not actually a doctype (as it does '.
|
||||
'not identify a corresponding DTD), but we are using this name '.
|
||||
@ -38,173 +46,159 @@ HTMLPurifier_ConfigSchema::define(
|
||||
'like %Core.XHTML or %HTML.Strict.'
|
||||
);
|
||||
|
||||
HTMLPurifier_ConfigSchema::define(
|
||||
'HTML', 'Trusted', false, 'bool',
|
||||
'Indicates whether or not the user input is trusted or not. If the '.
|
||||
'input is trusted, a more expansive set of allowed tags and attributes '.
|
||||
'will be used. This directive has been available since 2.0.0.'
|
||||
);
|
||||
|
||||
HTMLPurifier_ConfigSchema::define(
|
||||
'HTML', 'AllowedModules', null, 'lookup/null', '
|
||||
<p>
|
||||
A doctype comes with a set of usual modules to use. Without having
|
||||
to mucking about with the doctypes, you can quickly activate or
|
||||
disable these modules by specifying which modules you wish to allow
|
||||
with this directive. This is most useful for unit testing specific
|
||||
modules, although end users may find it useful for their own ends.
|
||||
</p>
|
||||
<p>
|
||||
If you specify a module that does not exist, the manager will silently
|
||||
fail to use it, so be careful! User-defined modules are not affected
|
||||
by this directive. Modules defined in %HTML.CoreModules are not
|
||||
affected by this directive. This directive has been available since 2.0.0.
|
||||
</p>
|
||||
');
|
||||
|
||||
HTMLPurifier_ConfigSchema::define(
|
||||
'HTML', 'CoreModules', array(
|
||||
'Structure' => true,
|
||||
'Text' => true,
|
||||
'Hypertext' => true,
|
||||
'List' => true,
|
||||
'NonXMLCommonAttributes' => true,
|
||||
'XMLCommonAttributes' => true,
|
||||
'CommonAttributes' => true
|
||||
), 'lookup', '
|
||||
<p>
|
||||
Certain modularized doctypes (XHTML, namely), have certain modules
|
||||
that must be included for the doctype to be an conforming document
|
||||
type: put those modules here. By default, XHTML\'s core modules
|
||||
are used. You can set this to a blank array to disable core module
|
||||
protection, but this is not recommended. This directive has been
|
||||
available since 2.0.0.
|
||||
</p>
|
||||
');
|
||||
|
||||
class HTMLPurifier_HTMLModuleManager
|
||||
{
|
||||
|
||||
/**
|
||||
* Array of HTMLPurifier_Module instances, indexed by module's class name.
|
||||
* All known modules, regardless of use, are in this array.
|
||||
* Instance of HTMLPurifier_DoctypeRegistry
|
||||
* @public
|
||||
*/
|
||||
var $doctypes;
|
||||
|
||||
/**
|
||||
* Instance of current doctype
|
||||
* @public
|
||||
*/
|
||||
var $doctype;
|
||||
|
||||
/**
|
||||
* Instance of HTMLPurifier_AttrTypes
|
||||
* @public
|
||||
*/
|
||||
var $attrTypes;
|
||||
|
||||
/**
|
||||
* Active instances of modules for the specified doctype are
|
||||
* indexed, by name, in this array.
|
||||
*/
|
||||
var $modules = array();
|
||||
|
||||
/**
|
||||
* String doctype we will validate against. See $validModules for use.
|
||||
*
|
||||
* @note
|
||||
* There is a special doctype '*' that acts both as the "default"
|
||||
* doctype if a customized system only defines one doctype and
|
||||
* also a catch-all doctype that gets merged into all the other
|
||||
* module collections. When possible, use a private collection to
|
||||
* share modules between doctypes: this special doctype is to
|
||||
* make life more convenient for users.
|
||||
* Array of recognized HTMLPurifier_Module instances, indexed by
|
||||
* module's class name. This array is usually lazy loaded, but a
|
||||
* user can overload a module by pre-emptively registering it.
|
||||
*/
|
||||
var $doctype;
|
||||
var $doctypeAliases = array(); /**< Lookup array of strings to real doctypes */
|
||||
var $registeredModules = array();
|
||||
|
||||
/**
|
||||
* Associative array: $collections[$type][$doctype] = list of modules.
|
||||
* This is used to logically separate types of functionality so that
|
||||
* based on the doctype and other configuration settings they may
|
||||
* be easily switched and on and off. Custom setups may not need
|
||||
* to use this abstraction, opting to have only one big collection
|
||||
* with one valid doctype.
|
||||
* List of extra modules that were added by the user using addModule().
|
||||
* These get unconditionally merged into the current doctype, whatever
|
||||
* it may be.
|
||||
*/
|
||||
var $collections = array();
|
||||
var $userModules = array();
|
||||
|
||||
/**
|
||||
* Modules that may be used in a valid doctype of this kind.
|
||||
* Correctional and leniency modules should not be placed in this
|
||||
* array unless the user said so: don't stuff every possible lenient
|
||||
* module for this doctype in here.
|
||||
* Associative array of element name to list of modules that have
|
||||
* definitions for the element; this array is dynamically filled.
|
||||
*/
|
||||
var $validModules = array();
|
||||
var $validCollections = array(); /**< Collections to merge into $validModules */
|
||||
|
||||
/**
|
||||
* Modules that we will allow in input, subset of $validModules. Single
|
||||
* element definitions may result in us consulting validModules.
|
||||
*/
|
||||
var $activeModules = array();
|
||||
var $activeCollections = array(); /**< Collections to merge into $activeModules */
|
||||
|
||||
var $counter = 0; /**< Designates next available integer order for modules. */
|
||||
var $initialized = false; /**< Says whether initialize() was called */
|
||||
|
||||
/**
|
||||
* Specifies what doctype to siphon new modules from addModule() to,
|
||||
* or false to disable the functionality. Must be used in conjunction
|
||||
* with $autoCollection.
|
||||
*/
|
||||
var $autoDoctype = false;
|
||||
/**
|
||||
* Specifies what collection to siphon new modules from addModule() to,
|
||||
* or false to disable the functionality. Must be used in conjunction
|
||||
* with $autoCollection.
|
||||
*/
|
||||
var $autoCollection = false;
|
||||
|
||||
/** Associative array of element name to defining modules (always array) */
|
||||
var $elementLookup = array();
|
||||
|
||||
/** List of prefixes we should use for resolving small names */
|
||||
/** List of prefixes we should use for registering small names */
|
||||
var $prefixes = array('HTMLPurifier_HTMLModule_');
|
||||
|
||||
var $contentSets; /**< Instance of HTMLPurifier_ContentSets */
|
||||
var $attrTypes; /**< Instance of HTMLPurifier_AttrTypes */
|
||||
var $contentSets; /**< Instance of HTMLPurifier_ContentSets */
|
||||
var $attrCollections; /**< Instance of HTMLPurifier_AttrCollections */
|
||||
|
||||
/**
|
||||
* @param $blank If true, don't do any initializing
|
||||
*/
|
||||
function HTMLPurifier_HTMLModuleManager($blank = false) {
|
||||
/** If set to true, unsafe elements and attributes will be allowed */
|
||||
var $trusted = false;
|
||||
|
||||
function HTMLPurifier_HTMLModuleManager() {
|
||||
|
||||
// the only editable internal object. The rest need to
|
||||
// be manipulated through modules
|
||||
// editable internal objects
|
||||
$this->attrTypes = new HTMLPurifier_AttrTypes();
|
||||
$this->doctypes = new HTMLPurifier_DoctypeRegistry();
|
||||
|
||||
if (!$blank) $this->initialize();
|
||||
// setup default HTML doctypes
|
||||
|
||||
}
|
||||
|
||||
function initialize() {
|
||||
$this->initialized = true;
|
||||
|
||||
// load default modules to the recognized modules list (not active)
|
||||
$modules = array(
|
||||
// define
|
||||
'CommonAttributes',
|
||||
'Text', 'Hypertext', 'List', 'Presentation',
|
||||
'Edit', 'Bdo', 'Tables', 'Image', 'StyleAttribute',
|
||||
'Target',
|
||||
// define-redefine
|
||||
'Legacy',
|
||||
// redefine
|
||||
'TransformToStrict', 'TransformToXHTML11'
|
||||
// module reuse
|
||||
$common = array(
|
||||
'CommonAttributes', 'Text', 'Hypertext', 'List',
|
||||
'Presentation', 'Edit', 'Bdo', 'Tables', 'Image',
|
||||
'StyleAttribute', 'Scripting'
|
||||
);
|
||||
foreach ($modules as $module) {
|
||||
$this->addModule($module);
|
||||
}
|
||||
$transitional = array('Legacy', 'Target');
|
||||
$xml = array('XMLCommonAttributes');
|
||||
$non_xml = array('NonXMLCommonAttributes');
|
||||
|
||||
// Safe modules for supported doctypes. These are included
|
||||
// in the valid and active module lists by default
|
||||
$this->collections['Safe'] = array(
|
||||
'_Common' => array( // leading _ indicates private
|
||||
'CommonAttributes', 'Text', 'Hypertext', 'List',
|
||||
'Presentation', 'Edit', 'Bdo', 'Tables', 'Image',
|
||||
'StyleAttribute'
|
||||
),
|
||||
// HTML definitions, defer to XHTML definitions
|
||||
'HTML 4.01 Transitional' => array(array('XHTML 1.0 Transitional')),
|
||||
'HTML 4.01 Strict' => array(array('XHTML 1.0 Strict')),
|
||||
// XHTML definitions
|
||||
'XHTML 1.0 Transitional' => array( array('XHTML 1.0 Strict'), 'Legacy', 'Target' ),
|
||||
'XHTML 1.0 Strict' => array(array('_Common')),
|
||||
'XHTML 1.1' => array(array('_Common')),
|
||||
$this->doctypes->register(
|
||||
'HTML 4.01 Transitional', false,
|
||||
array_merge($common, $transitional, $non_xml),
|
||||
array('Tidy_Transitional', 'Tidy_Proprietary')
|
||||
);
|
||||
|
||||
// Modules that specify elements that are unsafe from untrusted
|
||||
// third-parties. These should be registered in $validModules but
|
||||
// almost never $activeModules unless you really know what you're
|
||||
// doing.
|
||||
$this->collections['Unsafe'] = array();
|
||||
|
||||
// Modules to import if lenient mode (attempt to convert everything
|
||||
// to a valid representation) is on. These must not be in $validModules
|
||||
// unless specified so.
|
||||
$this->collections['Lenient'] = array(
|
||||
'HTML 4.01 Strict' => array(array('XHTML 1.0 Strict')),
|
||||
'XHTML 1.0 Strict' => array('TransformToStrict'),
|
||||
'XHTML 1.1' => array(array('XHTML 1.0 Strict'), 'TransformToXHTML11')
|
||||
$this->doctypes->register(
|
||||
'HTML 4.01 Strict', false,
|
||||
array_merge($common, $non_xml),
|
||||
array('Tidy_Strict', 'Tidy_Proprietary')
|
||||
);
|
||||
|
||||
// Modules to import if correctional mode (correct everything that
|
||||
// is feasible to strict mode) is on. These must not be in $validModules
|
||||
// unless specified so.
|
||||
$this->collections['Correctional'] = array(
|
||||
'HTML 4.01 Transitional' => array(array('XHTML 1.0 Transitional')),
|
||||
'XHTML 1.0 Transitional' => array('TransformToStrict'), // probably want a different one
|
||||
$this->doctypes->register(
|
||||
'XHTML 1.0 Transitional', true,
|
||||
array_merge($common, $transitional, $xml, $non_xml),
|
||||
array('Tidy_Transitional', 'Tidy_XHTML', 'Tidy_Proprietary')
|
||||
);
|
||||
|
||||
// User-space modules, custom code or whatever
|
||||
$this->collections['Extension'] = array();
|
||||
$this->doctypes->register(
|
||||
'XHTML 1.0 Strict', true,
|
||||
array_merge($common, $xml, $non_xml),
|
||||
array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_XHTMLStrict', 'Tidy_Proprietary')
|
||||
);
|
||||
|
||||
// setup active versus valid modules. ORDER IS IMPORTANT!
|
||||
// definition modules
|
||||
$this->makeCollectionActive('Safe');
|
||||
$this->makeCollectionValid('Unsafe');
|
||||
// redefinition modules
|
||||
$this->makeCollectionActive('Lenient');
|
||||
$this->makeCollectionActive('Correctional');
|
||||
|
||||
$this->autoDoctype = '*';
|
||||
$this->autoCollection = 'Extension';
|
||||
$this->doctypes->register(
|
||||
'XHTML 1.1', true,
|
||||
array_merge($common, $xml),
|
||||
array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Proprietary') // Tidy_XHTML1_1
|
||||
);
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Adds a module to the recognized module list. This does not
|
||||
* do anything else: the module must be added to a corresponding
|
||||
* collection to be "activated".
|
||||
* Registers a module to the recognized module list, useful for
|
||||
* overloading pre-existing modules.
|
||||
* @param $module Mixed: string module name, with or without
|
||||
* HTMLPurifier_HTMLModule prefix, or instance of
|
||||
* subclass of HTMLPurifier_HTMLModule.
|
||||
@ -217,10 +211,15 @@ class HTMLPurifier_HTMLModuleManager
|
||||
* - Check for literal object name
|
||||
* - Throw fatal error
|
||||
* If your object name collides with an internal class, specify
|
||||
* your module manually.
|
||||
* your module manually. All modules must have been included
|
||||
* externally: registerModule will not perform inclusions for you!
|
||||
* @warning If your module has the same name as an already loaded
|
||||
* module, your module will overload the old one WITHOUT
|
||||
* warning.
|
||||
*/
|
||||
function addModule($module) {
|
||||
function registerModule($module) {
|
||||
if (is_string($module)) {
|
||||
// attempt to load the module
|
||||
$original_module = $module;
|
||||
$ok = false;
|
||||
foreach ($this->prefixes as $prefix) {
|
||||
@ -240,16 +239,19 @@ class HTMLPurifier_HTMLModuleManager
|
||||
}
|
||||
$module = new $module();
|
||||
}
|
||||
$module->order = $this->counter++; // assign then increment
|
||||
$this->modules[$module->name] = $module;
|
||||
if ($this->autoDoctype !== false && $this->autoCollection !== false) {
|
||||
$this->collections[$this->autoCollection][$this->autoDoctype][] = $module->name;
|
||||
if (empty($module->name)) {
|
||||
trigger_error('Module instance of ' . get_class($module) . ' must have name');
|
||||
return;
|
||||
}
|
||||
$this->registeredModules[$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) {
|
||||
@ -265,55 +267,63 @@ class HTMLPurifier_HTMLModuleManager
|
||||
}
|
||||
|
||||
/**
|
||||
* Makes a collection active, while also making it valid if not
|
||||
* already done so. See $activeModules for the semantics of "active".
|
||||
* @param $collection_name Name of collection to activate
|
||||
* Adds a module to the current doctype by first registering it,
|
||||
* and then tacking it on to the active doctype
|
||||
*/
|
||||
function makeCollectionActive($collection_name) {
|
||||
if (!in_array($collection_name, $this->validCollections)) {
|
||||
$this->makeCollectionValid($collection_name);
|
||||
}
|
||||
$this->activeCollections[] = $collection_name;
|
||||
function addModule($module) {
|
||||
$this->registerModule($module);
|
||||
if (is_object($module)) $module = $module->name;
|
||||
$this->userModules[] = $module;
|
||||
}
|
||||
|
||||
/**
|
||||
* Makes a collection valid. See $validModules for the semantics of "valid"
|
||||
*/
|
||||
function makeCollectionValid($collection_name) {
|
||||
$this->validCollections[] = $collection_name;
|
||||
}
|
||||
|
||||
/**
|
||||
* Adds a class prefix that addModule() will use to resolve a
|
||||
* Adds a class prefix that registerModule() will use to resolve a
|
||||
* string name to a concrete class
|
||||
*/
|
||||
function addPrefix($prefix) {
|
||||
$this->prefixes[] = (string) $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) {
|
||||
|
||||
// load up the autocollection
|
||||
if ($this->autoCollection !== false) {
|
||||
$this->makeCollectionActive($this->autoCollection);
|
||||
$this->trusted = $config->get('HTML', 'Trusted');
|
||||
|
||||
// generate
|
||||
$this->doctype = $this->doctypes->make($config);
|
||||
$modules = $this->doctype->modules;
|
||||
|
||||
// take out the default modules that aren't allowed
|
||||
$lookup = $config->get('HTML', 'AllowedModules');
|
||||
$special_cases = $config->get('HTML', 'CoreModules');
|
||||
|
||||
if (is_array($lookup)) {
|
||||
foreach ($modules as $k => $m) {
|
||||
if (isset($special_cases[$m])) continue;
|
||||
if (!isset($lookup[$m])) unset($modules[$k]);
|
||||
}
|
||||
}
|
||||
|
||||
// retrieve the doctype
|
||||
$this->doctype = $this->getDoctype($config);
|
||||
if (isset($this->doctypeAliases[$this->doctype])) {
|
||||
$this->doctype = $this->doctypeAliases[$this->doctype];
|
||||
// merge in custom modules
|
||||
$modules = array_merge($modules, $this->userModules);
|
||||
|
||||
foreach ($modules as $module) {
|
||||
$this->processModule($module);
|
||||
}
|
||||
|
||||
// process module collections to module name => module instance form
|
||||
foreach ($this->collections as $col_i => $x) {
|
||||
$this->processCollections($this->collections[$col_i]);
|
||||
foreach ($this->doctype->tidyModules as $module) {
|
||||
$this->processModule($module);
|
||||
if (method_exists($this->modules[$module], 'construct')) {
|
||||
$this->modules[$module]->construct($config);
|
||||
}
|
||||
}
|
||||
|
||||
$this->validModules = $this->assembleModules($this->validCollections);
|
||||
$this->activeModules = $this->assembleModules($this->activeCollections);
|
||||
|
||||
// setup lookup table based on all valid modules
|
||||
foreach ($this->validModules as $module) {
|
||||
foreach ($this->modules as $module) {
|
||||
foreach ($module->info as $name => $def) {
|
||||
if (!isset($this->elementLookup[$name])) {
|
||||
$this->elementLookup[$name] = array();
|
||||
@ -324,214 +334,51 @@ class HTMLPurifier_HTMLModuleManager
|
||||
|
||||
// note the different choice
|
||||
$this->contentSets = new HTMLPurifier_ContentSets(
|
||||
// content models that contain non-allowed elements are
|
||||
// harmless because RemoveForeignElements will ensure
|
||||
// they never get in anyway, and there is usually no
|
||||
// reason why you should want to restrict a content
|
||||
// model beyond what is mandated by the doctype.
|
||||
// Note, however, that this means redefinitions of
|
||||
// content models can't be tossed in validModels willy-nilly:
|
||||
// that stuff still is regulated by configuration.
|
||||
$this->validModules
|
||||
// content set assembly deals with all possible modules,
|
||||
// not just ones deemed to be "safe"
|
||||
$this->modules
|
||||
);
|
||||
$this->attrCollections = new HTMLPurifier_AttrCollections(
|
||||
$this->attrTypes,
|
||||
// only explicitly allowed modules are allowed to affect
|
||||
// the global attribute collections. This mean's there's
|
||||
// a distinction between loading the Bdo module, and the
|
||||
// bdo element: Bdo will enable the dir attribute on all
|
||||
// elements, while bdo will only define the bdo element,
|
||||
// which will not have an editable directionality. This might
|
||||
// catch people who are loading only elements by surprise, so
|
||||
// we should consider loading an entire module if all the
|
||||
// elements it defines are requested by the user, especially
|
||||
// if it affects the global attribute collections.
|
||||
$this->activeModules
|
||||
// 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 list of collections and merges together all the defined
|
||||
* modules for the current doctype from those collections.
|
||||
* @param $collections List of collection suffixes we should grab
|
||||
* modules from (like 'Safe' or 'Lenient')
|
||||
* Takes a module and adds it to the active module collection,
|
||||
* registering it if necessary.
|
||||
*/
|
||||
function assembleModules($collections) {
|
||||
$modules = array();
|
||||
$numOfCollectionsUsed = 0;
|
||||
foreach ($collections as $name) {
|
||||
$disable_global = false;
|
||||
if (!isset($this->collections[$name])) {
|
||||
trigger_error("$name collection is undefined", E_USER_ERROR);
|
||||
continue;
|
||||
}
|
||||
$cols = $this->collections[$name];
|
||||
if (isset($cols[$this->doctype])) {
|
||||
if (isset($cols[$this->doctype]['*'])) {
|
||||
unset($cols[$this->doctype]['*']);
|
||||
$disable_global = true;
|
||||
}
|
||||
$modules += $cols[$this->doctype];
|
||||
$numOfCollectionsUsed++;
|
||||
}
|
||||
// accept catch-all doctype
|
||||
if (
|
||||
$this->doctype !== '*' &&
|
||||
isset($cols['*']) &&
|
||||
!$disable_global
|
||||
) {
|
||||
$modules += $cols['*'];
|
||||
}
|
||||
function processModule($module) {
|
||||
if (!isset($this->registeredModules[$module]) || is_object($module)) {
|
||||
$this->registerModule($module);
|
||||
}
|
||||
|
||||
if ($numOfCollectionsUsed < 1) {
|
||||
// possible XSS injection if user-specified doctypes
|
||||
// are allowed
|
||||
trigger_error("Doctype {$this->doctype} does not exist, ".
|
||||
"check for typos (if you desire a doctype that allows ".
|
||||
"no elements, use an empty array collection)", E_USER_ERROR);
|
||||
}
|
||||
return $modules;
|
||||
$this->modules[$module] = $this->registeredModules[$module];
|
||||
}
|
||||
|
||||
/**
|
||||
* Takes a collection and performs inclusions and substitutions for it.
|
||||
* @param $cols Reference to collections class member variable
|
||||
* Retrieves merged element definitions.
|
||||
* @return Array of HTMLPurifier_ElementDef
|
||||
*/
|
||||
function processCollections(&$cols) {
|
||||
|
||||
// $cols is the set of collections
|
||||
// $col_i is the name (index) of a collection
|
||||
// $col is a collection/list of modules
|
||||
|
||||
// perform inclusions
|
||||
foreach ($cols as $col_i => $col) {
|
||||
$seen = array();
|
||||
if (!empty($col[0]) && is_array($col[0])) {
|
||||
$seen[$col_i] = true; // recursion reporting
|
||||
$includes = $col[0];
|
||||
unset($cols[$col_i][0]); // remove inclusions value, recursion guard
|
||||
} else {
|
||||
$includes = array();
|
||||
}
|
||||
if (empty($includes)) continue;
|
||||
for ($i = 0; isset($includes[$i]); $i++) {
|
||||
$inc = $includes[$i];
|
||||
if (isset($seen[$inc])) {
|
||||
trigger_error(
|
||||
"Circular inclusion detected in $col_i collection",
|
||||
E_USER_ERROR
|
||||
);
|
||||
continue;
|
||||
} else {
|
||||
$seen[$inc] = true;
|
||||
}
|
||||
if (!isset($cols[$inc])) {
|
||||
trigger_error(
|
||||
"Collection $col_i tried to include undefined ".
|
||||
"collection $inc", E_USER_ERROR);
|
||||
continue;
|
||||
}
|
||||
foreach ($cols[$inc] as $module) {
|
||||
if (is_array($module)) { // another inclusion!
|
||||
foreach ($module as $inc2) $includes[] = $inc2;
|
||||
continue;
|
||||
}
|
||||
$cols[$col_i][] = $module; // merge in the other modules
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// replace with real modules, invert module from list to
|
||||
// assoc array of module name to module instance
|
||||
foreach ($cols as $col_i => $col) {
|
||||
$ignore_global = false;
|
||||
$order = array();
|
||||
foreach ($col as $module_i => $module) {
|
||||
unset($cols[$col_i][$module_i]);
|
||||
if (is_array($module)) {
|
||||
trigger_error("Illegal inclusion array at index".
|
||||
" $module_i found collection $col_i, inclusion".
|
||||
" arrays must be at start of collection (index 0)",
|
||||
E_USER_ERROR);
|
||||
continue;
|
||||
}
|
||||
if ($module_i === '*' && $module === false) {
|
||||
$ignore_global = true;
|
||||
continue;
|
||||
}
|
||||
if (!isset($this->modules[$module])) {
|
||||
trigger_error(
|
||||
"Collection $col_i references undefined ".
|
||||
"module $module",
|
||||
E_USER_ERROR
|
||||
);
|
||||
continue;
|
||||
}
|
||||
$module = $this->modules[$module];
|
||||
$cols[$col_i][$module->name] = $module;
|
||||
$order[$module->name] = $module->order;
|
||||
}
|
||||
array_multisort(
|
||||
$order, SORT_ASC, SORT_NUMERIC, $cols[$col_i]
|
||||
);
|
||||
if ($ignore_global) $cols[$col_i]['*'] = false;
|
||||
}
|
||||
|
||||
// delete pseudo-collections
|
||||
foreach ($cols as $col_i => $col) {
|
||||
if ($col_i[0] == '_') unset($cols[$col_i]);
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Retrieves the doctype from the configuration object
|
||||
*/
|
||||
function getDoctype($config) {
|
||||
$doctype = $config->get('HTML', 'Doctype');
|
||||
if ($doctype !== null) {
|
||||
return $doctype;
|
||||
}
|
||||
if (!$this->initialized) {
|
||||
// don't do HTML-oriented backwards compatibility stuff
|
||||
// use either the auto-doctype, or the catch-all doctype
|
||||
return $this->autoDoctype ? $this->autoDoctype : '*';
|
||||
}
|
||||
// this is backwards-compatibility stuff
|
||||
if ($config->get('Core', 'XHTML')) {
|
||||
$doctype = 'XHTML 1.0';
|
||||
} else {
|
||||
$doctype = 'HTML 4.01';
|
||||
}
|
||||
if ($config->get('HTML', 'Strict')) {
|
||||
$doctype .= ' Strict';
|
||||
} else {
|
||||
$doctype .= ' Transitional';
|
||||
}
|
||||
return $doctype;
|
||||
}
|
||||
|
||||
/**
|
||||
* Retrieves merged element definitions for all active elements.
|
||||
* @note We may want to generate an elements array during setup
|
||||
* and pass that on, because a specific combination of
|
||||
* elements may trigger the loading of a module.
|
||||
* @param $config Instance of HTMLPurifier_Config, for determining
|
||||
* stray elements.
|
||||
*/
|
||||
function getElements($config) {
|
||||
function getElements() {
|
||||
|
||||
$elements = array();
|
||||
foreach ($this->activeModules as $module) {
|
||||
foreach ($this->modules as $module) {
|
||||
foreach ($module->info as $name => $v) {
|
||||
if (isset($elements[$name])) continue;
|
||||
$elements[$name] = $this->getElement($name, $config);
|
||||
// if element is not safe, don't use it
|
||||
if (!$this->trusted && ($v->safe === false)) continue;
|
||||
$elements[$name] = $this->getElement($name);
|
||||
}
|
||||
}
|
||||
|
||||
// standalone elements now loaded
|
||||
// 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;
|
||||
|
||||
@ -540,13 +387,16 @@ class HTMLPurifier_HTMLModuleManager
|
||||
/**
|
||||
* Retrieves a single merged element definition
|
||||
* @param $name Name of element
|
||||
* @param $config Instance of HTMLPurifier_Config, may not be necessary.
|
||||
* @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, $config) {
|
||||
function getElement($name, $trusted = null) {
|
||||
|
||||
$def = false;
|
||||
if ($trusted === null) $trusted = $this->trusted;
|
||||
|
||||
$modules = $this->validModules;
|
||||
$modules = $this->modules;
|
||||
|
||||
if (!isset($this->elementLookup[$name])) {
|
||||
return false;
|
||||
@ -555,9 +405,23 @@ class HTMLPurifier_HTMLModuleManager
|
||||
foreach($this->elementLookup[$name] as $module_name) {
|
||||
|
||||
$module = $modules[$module_name];
|
||||
$new_def = $module->info[$name];
|
||||
|
||||
// copy is used because, ideally speaking, the original
|
||||
// definition should not be modified. Usually, this will
|
||||
// make no difference, but for consistency's sake
|
||||
$new_def = $module->info[$name]->copy();
|
||||
|
||||
// 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);
|
||||
@ -583,6 +447,13 @@ class HTMLPurifier_HTMLModuleManager
|
||||
|
||||
$this->contentSets->generateChildDef($def, $module);
|
||||
}
|
||||
|
||||
// add information on required attributes
|
||||
foreach ($def->attr as $attr_name => $attr_def) {
|
||||
if ($attr_def->required) {
|
||||
$def->required_attr[] = $attr_name;
|
||||
}
|
||||
}
|
||||
|
||||
return $def;
|
||||
|
||||
|
||||
Reference in New Issue
Block a user