MDL-14898 upgraded html purifier to 3.1.0 in HEAD

This commit is contained in:
skodak
2008-06-16 16:39:09 +00:00
parent b553527e99
commit eb203ee412
290 changed files with 4780 additions and 3432 deletions
-9
View File
@@ -1,9 +0,0 @@
<?php
/**
* This is a stub include that automatically configures the include path.
*/
set_include_path(dirname(__FILE__) . PATH_SEPARATOR . get_include_path() );
require_once 'HTMLPurifier.php';
-20
View File
@@ -1,20 +0,0 @@
<?php
/**
* Function wrapper for HTML Purifier for quick use.
* @note This function only includes the library when it is called. While
* this is efficient for instances when you only use HTML Purifier
* on a few of your pages, it murders bytecode caching. You still
* need to add HTML Purifier to your path.
* @note ''HTMLPurifier()'' is NOT the same as ''new HTMLPurifier()''
*/
function HTMLPurifier($html, $config = null) {
static $purifier = false;
if (!$purifier) {
require_once 'HTMLPurifier.php';
$purifier = new HTMLPurifier();
}
return $purifier->purify($html, $config);
}
+65 -66
View File
@@ -1,7 +1,6 @@
<?php
/*!
* @mainpage
/*! @mainpage
*
* HTML Purifier is an HTML filter that will take an arbitrary snippet of
* HTML and rigorously test, validate and filter it into a version that
@@ -16,14 +15,12 @@
* -# Generating HTML from the purified tokens.
*
* However, most users will only need to interface with the HTMLPurifier
* class, so this massive amount of infrastructure is usually concealed.
* If you plan on working with the internals, be sure to include
* HTMLPurifier_ConfigSchema and HTMLPurifier_Config.
* and HTMLPurifier_Config.
*/
/*
HTML Purifier 2.1.4 - Standards Compliant HTML Filtering
Copyright (C) 2006-2007 Edward Z. Yang
HTML Purifier 3.1.0 - Standards Compliant HTML Filtering
Copyright (C) 2006-2008 Edward Z. Yang
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
@@ -40,32 +37,6 @@
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
// constants are slow, but we'll make one exception
define('HTMLPURIFIER_PREFIX', dirname(__FILE__));
// every class has an undocumented dependency to these, must be included!
require_once 'HTMLPurifier/ConfigSchema.php'; // fatal errors if not included
require_once 'HTMLPurifier/Config.php';
require_once 'HTMLPurifier/Context.php';
require_once 'HTMLPurifier/Lexer.php';
require_once 'HTMLPurifier/Generator.php';
require_once 'HTMLPurifier/Strategy/Core.php';
require_once 'HTMLPurifier/Encoder.php';
require_once 'HTMLPurifier/ErrorCollector.php';
require_once 'HTMLPurifier/LanguageFactory.php';
HTMLPurifier_ConfigSchema::define(
'Core', 'CollectErrors', false, 'bool', '
Whether or not to collect errors found while filtering the document. This
is a useful way to give feedback to your users. <strong>Warning:</strong>
Currently this feature is very patchy and experimental, with lots of
possible error messages not yet implemented. It will not cause any problems,
but it may not help your users either. This directive has been available
since 2.0.0.
');
/**
* Facade that coordinates HTML Purifier's subsystems in order to purify HTML.
*
@@ -75,27 +46,36 @@ since 2.0.0.
* -# Instance: new HTMLPurifier($config)
* -# Invocation: purify($html, $config)
* These configurations are entirely independent of each other and
* are *not* merged.
* are *not* merged (this behavior may change in the future).
*
* @todo We need an easier way to inject strategies, it'll probably end
* up getting done through config though.
* @todo We need an easier way to inject strategies using the configuration
* object.
*/
class HTMLPurifier
{
var $version = '2.1.4';
/** Version of HTML Purifier */
public $version = '3.1.0';
var $config;
var $filters = array();
/** Constant with version of HTML Purifier */
const VERSION = '3.1.0';
var $strategy, $generator;
/** Global configuration object */
public $config;
/** Array of extra HTMLPurifier_Filter objects to run on HTML, for backwards compatibility */
private $filters = array();
/** Single instance of HTML Purifier */
private static $instance;
protected $strategy, $generator;
/**
* Resultant HTMLPurifier_Context of last run purification. Is an array
* of contexts if the last called method was purifyArray().
* @public
*/
var $context;
public $context;
/**
* Initializes the purifier.
@@ -105,12 +85,11 @@ class HTMLPurifier
* The parameter can also be any type that
* HTMLPurifier_Config::create() supports.
*/
function HTMLPurifier($config = null) {
public function __construct($config = null) {
$this->config = HTMLPurifier_Config::create($config);
$this->strategy = new HTMLPurifier_Strategy_Core();
$this->generator = new HTMLPurifier_Generator();
}
@@ -118,7 +97,8 @@ class HTMLPurifier
* Adds a filter to process the output. First come first serve
* @param $filter HTMLPurifier_Filter object
*/
function addFilter($filter) {
public function addFilter($filter) {
trigger_error('HTMLPurifier->addFilter() is deprecated, use configuration directives in the Filter namespace or Filter.Custom', E_USER_WARNING);
$this->filters[] = $filter;
}
@@ -132,8 +112,9 @@ class HTMLPurifier
* that HTMLPurifier_Config::create() supports.
* @return Purified HTML
*/
function purify($html, $config = null) {
public function purify($html, $config = null) {
// :TODO: make the config merge in, instead of replace
$config = $config ? HTMLPurifier_Config::create($config) : $this->config;
// implementation is partially environment dependant, partially
@@ -142,8 +123,8 @@ class HTMLPurifier
$context = new HTMLPurifier_Context();
// our friendly neighborhood generator, all primed with configuration too!
$this->generator->generateFromTokens(array(), $config, $context);
// setup HTML generator
$this->generator = new HTMLPurifier_Generator($config, $context);
$context->register('Generator', $this->generator);
// set up global context variables
@@ -164,8 +145,25 @@ class HTMLPurifier
$html = HTMLPurifier_Encoder::convertToUTF8($html, $config, $context);
for ($i = 0, $size = count($this->filters); $i < $size; $i++) {
$html = $this->filters[$i]->preFilter($html, $config, $context);
// setup filters
$filter_flags = $config->getBatch('Filter');
$custom_filters = $filter_flags['Custom'];
unset($filter_flags['Custom']);
$filters = array();
foreach ($filter_flags as $filter => $flag) {
if (!$flag) continue;
$class = "HTMLPurifier_Filter_$filter";
$filters[] = new $class;
}
foreach ($custom_filters as $filter) {
// maybe "HTMLPurifier_Filter_$filter", but be consistent with AutoFormat
$filters[] = $filter;
}
$filters = array_merge($filters, $this->filters);
// maybe prepare(), but later
for ($i = 0, $filter_size = count($filters); $i < $filter_size; $i++) {
$html = $filters[$i]->preFilter($html, $config, $context);
}
// purified HTML
@@ -179,12 +177,11 @@ class HTMLPurifier
$html, $config, $context
),
$config, $context
),
$config, $context
)
);
for ($i = $size - 1; $i >= 0; $i--) {
$html = $this->filters[$i]->postFilter($html, $config, $context);
for ($i = $filter_size - 1; $i >= 0; $i--) {
$html = $filters[$i]->postFilter($html, $config, $context);
}
$html = HTMLPurifier_Encoder::convertFromUTF8($html, $config, $context);
@@ -198,7 +195,7 @@ class HTMLPurifier
* See HTMLPurifier::purify() for more details.
* @return Array of purified HTML
*/
function purifyArray($array_of_html, $config = null) {
public function purifyArray($array_of_html, $config = null) {
$context_array = array();
foreach ($array_of_html as $key => $html) {
$array_of_html[$key] = $this->purify($html, $config);
@@ -211,25 +208,27 @@ class HTMLPurifier
/**
* Singleton for enforcing just one HTML Purifier in your system
* @param $prototype Optional prototype HTMLPurifier instance to
* overload singleton with.
* overload singleton with, or HTMLPurifier_Config
* instance to configure the generated version with.
*/
function &instance($prototype = null) {
static $htmlpurifier;
if (!$htmlpurifier || $prototype) {
if (is_a($prototype, 'HTMLPurifier')) {
$htmlpurifier = $prototype;
public static function instance($prototype = null) {
if (!self::$instance || $prototype) {
if ($prototype instanceof HTMLPurifier) {
self::$instance = $prototype;
} elseif ($prototype) {
$htmlpurifier = new HTMLPurifier($prototype);
self::$instance = new HTMLPurifier($prototype);
} else {
$htmlpurifier = new HTMLPurifier();
self::$instance = new HTMLPurifier();
}
}
return $htmlpurifier;
return self::$instance;
}
function &getInstance($prototype = null) {
/**
* @note Backwards compatibility, see instance()
*/
public static function getInstance($prototype = null) {
return HTMLPurifier::instance($prototype);
}
}
@@ -0,0 +1,183 @@
<?php
/**
* @file
* This file was auto-generated by generate-includes.php and includes all of
* the core files required by HTML Purifier. This is a convenience stub that
* includes all files using dirname(__FILE__) and require_once. PLEASE DO NOT
* EDIT THIS FILE, changes will be overwritten the next time the script is run.
*
* Changes to include_path are not necessary.
*/
$__dir = dirname(__FILE__);
require_once $__dir . '/HTMLPurifier.php';
require_once $__dir . '/HTMLPurifier/AttrCollections.php';
require_once $__dir . '/HTMLPurifier/AttrDef.php';
require_once $__dir . '/HTMLPurifier/AttrTransform.php';
require_once $__dir . '/HTMLPurifier/AttrTypes.php';
require_once $__dir . '/HTMLPurifier/AttrValidator.php';
require_once $__dir . '/HTMLPurifier/Bootstrap.php';
require_once $__dir . '/HTMLPurifier/Definition.php';
require_once $__dir . '/HTMLPurifier/CSSDefinition.php';
require_once $__dir . '/HTMLPurifier/ChildDef.php';
require_once $__dir . '/HTMLPurifier/Config.php';
require_once $__dir . '/HTMLPurifier/ConfigDef.php';
require_once $__dir . '/HTMLPurifier/ConfigSchema.php';
require_once $__dir . '/HTMLPurifier/ContentSets.php';
require_once $__dir . '/HTMLPurifier/Context.php';
require_once $__dir . '/HTMLPurifier/DefinitionCache.php';
require_once $__dir . '/HTMLPurifier/DefinitionCacheFactory.php';
require_once $__dir . '/HTMLPurifier/Doctype.php';
require_once $__dir . '/HTMLPurifier/DoctypeRegistry.php';
require_once $__dir . '/HTMLPurifier/ElementDef.php';
require_once $__dir . '/HTMLPurifier/Encoder.php';
require_once $__dir . '/HTMLPurifier/EntityLookup.php';
require_once $__dir . '/HTMLPurifier/EntityParser.php';
require_once $__dir . '/HTMLPurifier/ErrorCollector.php';
require_once $__dir . '/HTMLPurifier/Exception.php';
require_once $__dir . '/HTMLPurifier/Filter.php';
require_once $__dir . '/HTMLPurifier/Generator.php';
require_once $__dir . '/HTMLPurifier/HTMLDefinition.php';
require_once $__dir . '/HTMLPurifier/HTMLModule.php';
require_once $__dir . '/HTMLPurifier/HTMLModuleManager.php';
require_once $__dir . '/HTMLPurifier/IDAccumulator.php';
require_once $__dir . '/HTMLPurifier/Injector.php';
require_once $__dir . '/HTMLPurifier/Language.php';
require_once $__dir . '/HTMLPurifier/LanguageFactory.php';
require_once $__dir . '/HTMLPurifier/Lexer.php';
require_once $__dir . '/HTMLPurifier/PercentEncoder.php';
require_once $__dir . '/HTMLPurifier/Strategy.php';
require_once $__dir . '/HTMLPurifier/StringHash.php';
require_once $__dir . '/HTMLPurifier/StringHashParser.php';
require_once $__dir . '/HTMLPurifier/TagTransform.php';
require_once $__dir . '/HTMLPurifier/Token.php';
require_once $__dir . '/HTMLPurifier/TokenFactory.php';
require_once $__dir . '/HTMLPurifier/URI.php';
require_once $__dir . '/HTMLPurifier/URIDefinition.php';
require_once $__dir . '/HTMLPurifier/URIFilter.php';
require_once $__dir . '/HTMLPurifier/URIParser.php';
require_once $__dir . '/HTMLPurifier/URIScheme.php';
require_once $__dir . '/HTMLPurifier/URISchemeRegistry.php';
require_once $__dir . '/HTMLPurifier/VarParser.php';
require_once $__dir . '/HTMLPurifier/VarParserException.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS.php';
require_once $__dir . '/HTMLPurifier/AttrDef/Enum.php';
require_once $__dir . '/HTMLPurifier/AttrDef/Integer.php';
require_once $__dir . '/HTMLPurifier/AttrDef/Lang.php';
require_once $__dir . '/HTMLPurifier/AttrDef/Text.php';
require_once $__dir . '/HTMLPurifier/AttrDef/URI.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Number.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/AlphaValue.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Background.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/BackgroundPosition.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Border.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Color.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Composite.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/DenyElementDecorator.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Filter.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Font.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/FontFamily.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/ImportantDecorator.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Length.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/ListStyle.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Multiple.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/Percentage.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/TextDecoration.php';
require_once $__dir . '/HTMLPurifier/AttrDef/CSS/URI.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/Bool.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/Color.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/FrameTarget.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/ID.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/Pixels.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/Length.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/LinkTypes.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/MultiLength.php';
require_once $__dir . '/HTMLPurifier/AttrDef/HTML/Nmtokens.php';
require_once $__dir . '/HTMLPurifier/AttrDef/URI/Email.php';
require_once $__dir . '/HTMLPurifier/AttrDef/URI/Host.php';
require_once $__dir . '/HTMLPurifier/AttrDef/URI/IPv4.php';
require_once $__dir . '/HTMLPurifier/AttrDef/URI/IPv6.php';
require_once $__dir . '/HTMLPurifier/AttrDef/URI/Email/SimpleCheck.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/BdoDir.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/BgColor.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/BoolToCSS.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/Border.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/EnumToCSS.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/ImgRequired.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/ImgSpace.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/Lang.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/Length.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/Name.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/ScriptRequired.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Chameleon.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Custom.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Empty.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Required.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Optional.php';
require_once $__dir . '/HTMLPurifier/ChildDef/StrictBlockquote.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Table.php';
require_once $__dir . '/HTMLPurifier/ConfigDef/Directive.php';
require_once $__dir . '/HTMLPurifier/ConfigDef/DirectiveAlias.php';
require_once $__dir . '/HTMLPurifier/ConfigDef/Namespace.php';
require_once $__dir . '/HTMLPurifier/DefinitionCache/Decorator.php';
require_once $__dir . '/HTMLPurifier/DefinitionCache/Null.php';
require_once $__dir . '/HTMLPurifier/DefinitionCache/Serializer.php';
require_once $__dir . '/HTMLPurifier/DefinitionCache/Decorator/Cleanup.php';
require_once $__dir . '/HTMLPurifier/DefinitionCache/Decorator/Memory.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Bdo.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/CommonAttributes.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Edit.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Hypertext.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Image.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Legacy.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/List.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Object.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Presentation.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Proprietary.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Ruby.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Scripting.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/StyleAttribute.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tables.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Target.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Text.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/XMLCommonAttributes.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy/Proprietary.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy/XHTMLAndHTML4.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy/Strict.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy/Transitional.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy/XHTML.php';
require_once $__dir . '/HTMLPurifier/Injector/AutoParagraph.php';
require_once $__dir . '/HTMLPurifier/Injector/Linkify.php';
require_once $__dir . '/HTMLPurifier/Injector/PurifierLinkify.php';
require_once $__dir . '/HTMLPurifier/Lexer/DOMLex.php';
require_once $__dir . '/HTMLPurifier/Lexer/DirectLex.php';
require_once $__dir . '/HTMLPurifier/Strategy/Composite.php';
require_once $__dir . '/HTMLPurifier/Strategy/Core.php';
require_once $__dir . '/HTMLPurifier/Strategy/FixNesting.php';
require_once $__dir . '/HTMLPurifier/Strategy/MakeWellFormed.php';
require_once $__dir . '/HTMLPurifier/Strategy/RemoveForeignElements.php';
require_once $__dir . '/HTMLPurifier/Strategy/ValidateAttributes.php';
require_once $__dir . '/HTMLPurifier/TagTransform/Font.php';
require_once $__dir . '/HTMLPurifier/TagTransform/Simple.php';
require_once $__dir . '/HTMLPurifier/Token/Comment.php';
require_once $__dir . '/HTMLPurifier/Token/Tag.php';
require_once $__dir . '/HTMLPurifier/Token/Empty.php';
require_once $__dir . '/HTMLPurifier/Token/End.php';
require_once $__dir . '/HTMLPurifier/Token/Start.php';
require_once $__dir . '/HTMLPurifier/Token/Text.php';
require_once $__dir . '/HTMLPurifier/URIFilter/DisableExternal.php';
require_once $__dir . '/HTMLPurifier/URIFilter/DisableExternalResources.php';
require_once $__dir . '/HTMLPurifier/URIFilter/HostBlacklist.php';
require_once $__dir . '/HTMLPurifier/URIFilter/MakeAbsolute.php';
require_once $__dir . '/HTMLPurifier/URIScheme/ftp.php';
require_once $__dir . '/HTMLPurifier/URIScheme/http.php';
require_once $__dir . '/HTMLPurifier/URIScheme/https.php';
require_once $__dir . '/HTMLPurifier/URIScheme/mailto.php';
require_once $__dir . '/HTMLPurifier/URIScheme/news.php';
require_once $__dir . '/HTMLPurifier/URIScheme/nntp.php';
require_once $__dir . '/HTMLPurifier/VarParser/Flexible.php';
require_once $__dir . '/HTMLPurifier/VarParser/Native.php';
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrTypes.php';
/**
* Defines common attribute collections that modules reference
*/
@@ -12,7 +10,7 @@ class HTMLPurifier_AttrCollections
/**
* Associative array of attribute collections, indexed by name
*/
var $info = array();
public $info = array();
/**
* Performs all expansions on internal data for use by other inclusions
@@ -21,7 +19,7 @@ class HTMLPurifier_AttrCollections
* @param $attr_types HTMLPurifier_AttrTypes instance
* @param $modules Hash array of HTMLPurifier_HTMLModule members
*/
function HTMLPurifier_AttrCollections($attr_types, $modules) {
public function __construct($attr_types, $modules) {
// load extensions from the modules
foreach ($modules as $module) {
foreach ($module->attr_collections as $coll_i => $coll) {
@@ -53,7 +51,7 @@ class HTMLPurifier_AttrCollections
* all inclusions specified by the zero index.
* @param &$attr Reference to attribute array
*/
function performInclusions(&$attr) {
public function performInclusions(&$attr) {
if (!isset($attr[0])) return;
$merge = $attr[0];
$seen = array(); // recursion guard
@@ -81,7 +79,7 @@ class HTMLPurifier_AttrCollections
* @param &$attr Reference to attribute array
* @param $attr_types HTMLPurifier_AttrTypes instance
*/
function expandIdentifiers(&$attr, $attr_types) {
public function expandIdentifiers(&$attr, $attr_types) {
// because foreach will process new elements we add, make sure we
// skip duplicates
+11 -16
View File
@@ -10,32 +10,29 @@
* subclasses are also responsible for cleaning the code if possible.
*/
class HTMLPurifier_AttrDef
abstract class HTMLPurifier_AttrDef
{
/**
* Tells us whether or not an HTML attribute is minimized. Has no
* meaning in other contexts.
*/
var $minimized = false;
public $minimized = false;
/**
* Tells us whether or not an HTML attribute is required. Has no
* meaning in other contexts
*/
var $required = false;
public $required = false;
/**
* Validates and cleans passed string according to a definition.
*
* @public
* @param $string String to be validated and cleaned.
* @param $config Mandatory HTMLPurifier_Config object.
* @param $context Mandatory HTMLPurifier_AttrContext object.
*/
function validate($string, $config, &$context) {
trigger_error('Cannot call abstract function', E_USER_ERROR);
}
abstract public function validate($string, $config, $context);
/**
* Convenience method that parses a string as if it were CDATA.
@@ -59,10 +56,8 @@ class HTMLPurifier_AttrDef
* function. Trim and whitespace collapsing are supposed to only
* occur in NMTOKENs. However, note that we are NOT necessarily
* parsing XML, thus, this behavior may still be correct.
*
* @public
*/
function parseCDATA($string) {
public function parseCDATA($string) {
$string = trim($string);
$string = str_replace("\n", '', $string);
$string = str_replace(array("\r", "\t"), ' ', $string);
@@ -73,12 +68,12 @@ class HTMLPurifier_AttrDef
* Factory method for creating this class from a string.
* @param $string String construction info
* @return Created AttrDef object corresponding to $string
* @public
*/
function make($string) {
// default implementation, return flyweight of this object
// if overloaded, it is *necessary* for you to clone the
// object (usually by instantiating a new copy) and return that
public function make($string) {
// default implementation, return a flyweight of this object.
// If $string has an effect on the returned object (i.e. you
// need to overload this method), it is best
// to clone or instantiate new copies. (Instantiation is safer.)
return $this;
}
@@ -86,7 +81,7 @@ class HTMLPurifier_AttrDef
* Removes spaces from rgb(0, 0, 0) so that shorthand CSS properties work
* properly. THIS IS A HACK!
*/
function mungeRgb($string) {
protected function mungeRgb($string) {
return preg_replace('/rgb\((\d+)\s*,\s*(\d+)\s*,\s*(\d+)\)/', 'rgb(\1,\2,\3)', $string);
}
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/CSSDefinition.php';
/**
* Validates the HTML attribute style, otherwise known as CSS.
* @note We don't implement the whole CSS specification, so it might be
@@ -17,7 +14,7 @@ require_once 'HTMLPurifier/CSSDefinition.php';
class HTMLPurifier_AttrDef_CSS extends HTMLPurifier_AttrDef
{
function validate($css, $config, &$context) {
public function validate($css, $config, $context) {
$css = $this->parseCDATA($css);
@@ -0,0 +1,19 @@
<?php
class HTMLPurifier_AttrDef_CSS_AlphaValue extends HTMLPurifier_AttrDef_CSS_Number
{
public function __construct() {
parent::__construct(false); // opacity is non-negative, but we will clamp it
}
public function validate($number, $config, $context) {
$result = parent::validate($number, $config, $context);
if ($result === false) return $result;
$float = (float) $result;
if ($float < 0.0) $result = '0';
if ($float > 1.0) $result = '1';
return $result;
}
}
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/CSSDefinition.php';
/**
* Validates shorthand CSS property background.
* @warning Does not support url tokens that have internal spaces.
@@ -14,9 +11,9 @@ class HTMLPurifier_AttrDef_CSS_Background extends HTMLPurifier_AttrDef
* Local copy of component validators.
* @note See HTMLPurifier_AttrDef_Font::$info for a similar impl.
*/
var $info;
protected $info;
function HTMLPurifier_AttrDef_CSS_Background($config) {
public function __construct($config) {
$def = $config->getCSSDefinition();
$this->info['background-color'] = $def->info['background-color'];
$this->info['background-image'] = $def->info['background-image'];
@@ -25,7 +22,7 @@ class HTMLPurifier_AttrDef_CSS_Background extends HTMLPurifier_AttrDef
$this->info['background-position'] = $def->info['background-position'];
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
// regular pre-processing
$string = $this->parseCDATA($string);
@@ -1,9 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/CSS/Length.php';
require_once 'HTMLPurifier/AttrDef/CSS/Percentage.php';
/* W3C says:
[ // adjective and number must be in correct order, even if
// you could switch them without introducing ambiguity.
@@ -48,15 +44,15 @@ require_once 'HTMLPurifier/AttrDef/CSS/Percentage.php';
class HTMLPurifier_AttrDef_CSS_BackgroundPosition extends HTMLPurifier_AttrDef
{
var $length;
var $percentage;
protected $length;
protected $percentage;
function HTMLPurifier_AttrDef_CSS_BackgroundPosition() {
public function __construct() {
$this->length = new HTMLPurifier_AttrDef_CSS_Length();
$this->percentage = new HTMLPurifier_AttrDef_CSS_Percentage();
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = $this->parseCDATA($string);
$bits = explode(' ', $string);
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates the border property as defined by CSS.
*/
@@ -11,16 +9,16 @@ class HTMLPurifier_AttrDef_CSS_Border extends HTMLPurifier_AttrDef
/**
* Local copy of properties this property is shorthand for.
*/
var $info = array();
protected $info = array();
function HTMLPurifier_AttrDef_CSS_Border($config) {
public function __construct($config) {
$def = $config->getCSSDefinition();
$this->info['border-width'] = $def->info['border-width'];
$this->info['border-style'] = $def->info['border-style'];
$this->info['border-top-color'] = $def->info['border-top-color'];
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = $this->parseCDATA($string);
$string = $this->mungeRgb($string);
$bits = explode(' ', $string);
@@ -1,39 +1,12 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
HTMLPurifier_ConfigSchema::define(
'Core', 'ColorKeywords', array(
'maroon' => '#800000',
'red' => '#FF0000',
'orange' => '#FFA500',
'yellow' => '#FFFF00',
'olive' => '#808000',
'purple' => '#800080',
'fuchsia' => '#FF00FF',
'white' => '#FFFFFF',
'lime' => '#00FF00',
'green' => '#008000',
'navy' => '#000080',
'blue' => '#0000FF',
'aqua' => '#00FFFF',
'teal' => '#008080',
'black' => '#000000',
'silver' => '#C0C0C0',
'gray' => '#808080'
), 'hash', '
Lookup array of color names to six digit hexadecimal number corresponding
to color, with preceding hash mark. Used when parsing colors.
This directive has been available since 2.0.0.
');
/**
* Validates Color as defined by CSS.
*/
class HTMLPurifier_AttrDef_CSS_Color extends HTMLPurifier_AttrDef
{
function validate($color, $config, &$context) {
public function validate($color, $config, $context) {
static $colors = null;
if ($colors === null) $colors = $config->get('Core', 'ColorKeywords');
@@ -14,18 +14,18 @@ class HTMLPurifier_AttrDef_CSS_Composite extends HTMLPurifier_AttrDef
/**
* List of HTMLPurifier_AttrDef objects that may process strings
* @protected
* @todo Make protected
*/
var $defs;
public $defs;
/**
* @param $defs List of HTMLPurifier_AttrDef objects
*/
function HTMLPurifier_AttrDef_CSS_Composite($defs) {
public function __construct($defs) {
$this->defs = $defs;
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
foreach ($this->defs as $i => $def) {
$result = $this->defs[$i]->validate($string, $config, $context);
if ($result !== false) return $result;
@@ -5,20 +5,20 @@
*/
class HTMLPurifier_AttrDef_CSS_DenyElementDecorator extends HTMLPurifier_AttrDef
{
var $def, $element;
protected $def, $element;
/**
* @param $def Definition to wrap
* @param $element Element to deny
*/
function HTMLPurifier_AttrDef_CSS_DenyElementDecorator(&$def, $element) {
$this->def =& $def;
public function __construct($def, $element) {
$this->def = $def;
$this->element = $element;
}
/**
* Checks if CurrentToken is set and equal to $this->element
*/
function validate($string, $config, $context) {
public function validate($string, $config, $context) {
$token = $context->get('CurrentToken', true);
if ($token && $token->name == $this->element) return false;
return $this->def->validate($string, $config, $context);
@@ -0,0 +1,52 @@
<?php
/**
* Microsoft's proprietary filter: CSS property
* @note Currently supports the alpha filter. In the future, this will
* probably need an extensible framework
*/
class HTMLPurifier_AttrDef_CSS_Filter extends HTMLPurifier_AttrDef
{
protected $intValidator;
public function __construct() {
$this->intValidator = new HTMLPurifier_AttrDef_Integer();
}
public function validate($value, $config, $context) {
$value = $this->parseCDATA($value);
if ($value === 'none') return $value;
// if we looped this we could support multiple filters
$function_length = strcspn($value, '(');
$function = trim(substr($value, 0, $function_length));
if ($function !== 'alpha' &&
$function !== 'Alpha' &&
$function !== 'progid:DXImageTransform.Microsoft.Alpha'
) return false;
$cursor = $function_length + 1;
$parameters_length = strcspn($value, ')', $cursor);
$parameters = substr($value, $cursor, $parameters_length);
$params = explode(',', $parameters);
$ret_params = array();
$lookup = array();
foreach ($params as $param) {
list($key, $value) = explode('=', $param);
$key = trim($key);
$value = trim($value);
if (isset($lookup[$key])) continue;
if ($key !== 'opacity') continue;
$value = $this->intValidator->validate($value, $config, $context);
if ($value === false) continue;
$int = (int) $value;
if ($int > 100) $value = '100';
if ($int < 0) $value = '0';
$ret_params[] = "$key=$value";
$lookup[$key] = true;
}
$ret_parameters = implode(',', $ret_params);
$ret_function = "$function($ret_parameters)";
return $ret_function;
}
}
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates shorthand CSS property font.
*/
@@ -16,9 +14,9 @@ class HTMLPurifier_AttrDef_CSS_Font extends HTMLPurifier_AttrDef
* CSSDefinition, this wouldn't be necessary. We'd instantiate
* our own copies.
*/
var $info = array();
protected $info = array();
function HTMLPurifier_AttrDef_CSS_Font($config) {
public function __construct($config) {
$def = $config->getCSSDefinition();
$this->info['font-style'] = $def->info['font-style'];
$this->info['font-variant'] = $def->info['font-variant'];
@@ -28,7 +26,7 @@ class HTMLPurifier_AttrDef_CSS_Font extends HTMLPurifier_AttrDef
$this->info['font-family'] = $def->info['font-family'];
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
static $system_fonts = array(
'caption' => true,
@@ -1,16 +1,13 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
// whitelisting allowed fonts would be nice
/**
* Validates a font family list according to CSS spec
* @todo whitelisting allowed fonts would be nice
*/
class HTMLPurifier_AttrDef_CSS_FontFamily extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
static $generic_names = array(
'serif' => true,
'sans-serif' => true,
@@ -0,0 +1,38 @@
<?php
/**
* Decorator which enables !important to be used in CSS values.
*/
class HTMLPurifier_AttrDef_CSS_ImportantDecorator extends HTMLPurifier_AttrDef
{
protected $def, $allow;
/**
* @param $def Definition to wrap
* @param $allow Whether or not to allow !important
*/
public function __construct($def, $allow = false) {
$this->def = $def;
$this->allow = $allow;
}
/**
* Intercepts and removes !important if necessary
*/
public function validate($string, $config, $context) {
// test for ! and important tokens
$string = trim($string);
$is_important = false;
// :TODO: optimization: test directly for !important and ! important
if (strlen($string) >= 9 && substr($string, -9) === 'important') {
$temp = rtrim(substr($string, 0, -9));
// use a temp, because we might want to restore important
if (strlen($temp) >= 1 && substr($temp, -1) === '!') {
$string = rtrim(substr($temp, 0, -1));
$is_important = true;
}
}
$string = $this->def->validate($string, $config, $context);
if ($this->allow && $is_important) $string .= ' !important';
return $string;
}
}
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/CSS/Number.php';
/**
* Represents a Length as defined by CSS.
*/
@@ -14,22 +11,22 @@ class HTMLPurifier_AttrDef_CSS_Length extends HTMLPurifier_AttrDef
* @warning The code assumes all units are two characters long. Be careful
* if we have to change this behavior!
*/
var $units = array('em' => true, 'ex' => true, 'px' => true, 'in' => true,
protected $units = array('em' => true, 'ex' => true, 'px' => true, 'in' => true,
'cm' => true, 'mm' => true, 'pt' => true, 'pc' => true);
/**
* Instance of HTMLPurifier_AttrDef_Number to defer number validation to
*/
var $number_def;
protected $number_def;
/**
* @param $non_negative Bool indication whether or not negative values are
* allowed.
*/
function HTMLPurifier_AttrDef_CSS_Length($non_negative = false) {
public function __construct($non_negative = false) {
$this->number_def = new HTMLPurifier_AttrDef_CSS_Number($non_negative);
}
function validate($length, $config, &$context) {
public function validate($length, $config, $context) {
$length = $this->parseCDATA($length);
if ($length === '') return false;
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates shorthand CSS property list-style.
* @warning Does not support url tokens that have internal spaces.
@@ -13,16 +11,16 @@ class HTMLPurifier_AttrDef_CSS_ListStyle extends HTMLPurifier_AttrDef
* Local copy of component validators.
* @note See HTMLPurifier_AttrDef_CSS_Font::$info for a similar impl.
*/
var $info;
protected $info;
function HTMLPurifier_AttrDef_CSS_ListStyle($config) {
public function __construct($config) {
$def = $config->getCSSDefinition();
$this->info['list-style-type'] = $def->info['list-style-type'];
$this->info['list-style-position'] = $def->info['list-style-position'];
$this->info['list-style-image'] = $def->info['list-style-image'];
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
// regular pre-processing
$string = $this->parseCDATA($string);
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Framework class for strings that involve multiple values.
*
@@ -18,24 +16,26 @@ class HTMLPurifier_AttrDef_CSS_Multiple extends HTMLPurifier_AttrDef
/**
* Instance of component definition to defer validation to.
* @todo Make protected
*/
var $single;
public $single;
/**
* Max number of values allowed.
* @todo Make protected
*/
var $max;
public $max;
/**
* @param $single HTMLPurifier_AttrDef to multiply
* @param $max Max number of values allowed (usually four)
*/
function HTMLPurifier_AttrDef_CSS_Multiple($single, $max = 4) {
public function __construct($single, $max = 4) {
$this->single = $single;
$this->max = $max;
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = $this->parseCDATA($string);
if ($string === '') return false;
$parts = explode(' ', $string); // parseCDATA replaced \r, \t and \n
@@ -9,20 +9,21 @@ class HTMLPurifier_AttrDef_CSS_Number extends HTMLPurifier_AttrDef
/**
* Bool indicating whether or not only positive values allowed.
*/
var $non_negative = false;
protected $non_negative = false;
/**
* @param $non_negative Bool indicating whether negatives are forbidden
*/
function HTMLPurifier_AttrDef_CSS_Number($non_negative = false) {
public function __construct($non_negative = false) {
$this->non_negative = $non_negative;
}
function validate($number, $config, &$context) {
public function validate($number, $config, $context) {
$number = $this->parseCDATA($number);
if ($number === '') return false;
if ($number === '0') return '0';
$sign = '';
switch ($number[0]) {
@@ -37,13 +38,16 @@ class HTMLPurifier_AttrDef_CSS_Number extends HTMLPurifier_AttrDef
$number = ltrim($number, '0');
return $number ? $sign . $number : '0';
}
if (!strpos($number, '.')) return false;
// Period is the only non-numeric character allowed
if (strpos($number, '.') === false) return false;
list($left, $right) = explode('.', $number, 2);
if (!ctype_digit($left)) return false;
$left = ltrim($left, '0');
if ($left === '' && $right === '') return false;
if ($left !== '' && !ctype_digit($left)) return false;
$left = ltrim($left, '0');
$right = rtrim($right, '0');
if ($right === '') {
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/CSS/Number.php';
/**
* Validates a Percentage as defined by the CSS spec.
*/
@@ -12,16 +9,16 @@ class HTMLPurifier_AttrDef_CSS_Percentage extends HTMLPurifier_AttrDef
/**
* Instance of HTMLPurifier_AttrDef_CSS_Number to defer number validation
*/
var $number_def;
protected $number_def;
/**
* @param Bool indicating whether to forbid negative values
*/
function HTMLPurifier_AttrDef_CSS_Percentage($non_negative = false) {
public function __construct($non_negative = false) {
$this->number_def = new HTMLPurifier_AttrDef_CSS_Number($non_negative);
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = $this->parseCDATA($string);
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates the value for the CSS property text-decoration
* @note This class could be generalized into a version that acts sort of
@@ -10,7 +8,7 @@ require_once 'HTMLPurifier/AttrDef.php';
class HTMLPurifier_AttrDef_CSS_TextDecoration extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
static $allowed_values = array(
'line-through' => true,
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef/URI.php';
/**
* Validates a URI in CSS syntax, which uses url('http://example.com')
* @note While theoretically speaking a URI in a CSS document could
@@ -14,11 +12,11 @@ require_once 'HTMLPurifier/AttrDef/URI.php';
class HTMLPurifier_AttrDef_CSS_URI extends HTMLPurifier_AttrDef_URI
{
function HTMLPurifier_AttrDef_CSS_URI() {
parent::HTMLPurifier_AttrDef_URI(true); // always embedded
public function __construct() {
parent::__construct(true); // always embedded
}
function validate($uri_string, $config, &$context) {
public function validate($uri_string, $config, $context) {
// parse the URI out of the string and then pass it onto
// the parent object
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
// Enum = Enumerated
/**
* Validates a keyword against a list of valid values.
@@ -14,27 +12,28 @@ class HTMLPurifier_AttrDef_Enum extends HTMLPurifier_AttrDef
/**
* Lookup table of valid values.
* @todo Make protected
*/
var $valid_values = array();
public $valid_values = array();
/**
* Bool indicating whether or not enumeration is case sensitive.
* @note In general this is always case insensitive.
*/
var $case_sensitive = false; // values according to W3C spec
protected $case_sensitive = false; // values according to W3C spec
/**
* @param $valid_values List of valid values
* @param $case_sensitive Bool indicating whether or not case sensitive
*/
function HTMLPurifier_AttrDef_Enum(
public function __construct(
$valid_values = array(), $case_sensitive = false
) {
$this->valid_values = array_flip($valid_values);
$this->case_sensitive = $case_sensitive;
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = trim($string);
if (!$this->case_sensitive) {
// we may want to do full case-insensitive libraries
@@ -50,7 +49,7 @@ class HTMLPurifier_AttrDef_Enum extends HTMLPurifier_AttrDef
* valid values. Example: "foo,bar,baz". Prepend "s:" to make
* case sensitive
*/
function make($string) {
public function make($string) {
if (strlen($string) > 2 && $string[0] == 's' && $string[1] == ':') {
$string = substr($string, 2);
$sensitive = true;
@@ -1,19 +1,17 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates a boolean attribute
*/
class HTMLPurifier_AttrDef_HTML_Bool extends HTMLPurifier_AttrDef
{
var $name;
var $minimized = true;
protected $name;
public $minimized = true;
function HTMLPurifier_AttrDef_HTML_Bool($name = false) {$this->name = $name;}
public function __construct($name = false) {$this->name = $name;}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
if (empty($string)) return false;
return $this->name;
}
@@ -21,7 +19,7 @@ class HTMLPurifier_AttrDef_HTML_Bool extends HTMLPurifier_AttrDef
/**
* @param $string Name of attribute
*/
function make($string) {
public function make($string) {
return new HTMLPurifier_AttrDef_HTML_Bool($string);
}
@@ -1,15 +1,12 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/CSS/Color.php'; // for %Core.ColorKeywords
/**
* Validates a color according to the HTML spec.
*/
class HTMLPurifier_AttrDef_HTML_Color extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
static $colors = null;
if ($colors === null) $colors = $config->get('Core', 'ColorKeywords');
@@ -1,30 +1,17 @@
<?php
HTMLPurifier_ConfigSchema::define(
'Attr', 'AllowedFrameTargets', array(), 'lookup',
'Lookup table of all allowed link frame targets. Some commonly used '.
'link targets include _blank, _self, _parent and _top. Values should '.
'be lowercase, as validation will be done in a case-sensitive manner '.
'despite W3C\'s recommendation. XHTML 1.0 Strict does not permit '.
'the target attribute so this directive will have no effect in that '.
'doctype. XHTML 1.1 does not enable the Target module by default, you '.
'will have to manually enable it (see the module documentation for more details.)'
);
require_once 'HTMLPurifier/AttrDef/Enum.php';
/**
* Special-case enum attribute definition that lazy loads allowed frame targets
*/
class HTMLPurifier_AttrDef_HTML_FrameTarget extends HTMLPurifier_AttrDef_Enum
{
var $valid_values = false; // uninitialized value
var $case_sensitive = false;
public $valid_values = false; // uninitialized value
protected $case_sensitive = false;
function HTMLPurifier_AttrDef_HTML_FrameTarget() {}
public function __construct() {}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
if ($this->valid_values === false) $this->valid_values = $config->get('Attr', 'AllowedFrameTargets');
return parent::validate($string, $config, $context);
}
@@ -1,56 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/IDAccumulator.php';
HTMLPurifier_ConfigSchema::define(
'Attr', 'EnableID', false, 'bool',
'Allows the ID attribute in HTML. This is disabled by default '.
'due to the fact that without proper configuration user input can '.
'easily break the validation of a webpage by specifying an ID that is '.
'already on the surrounding HTML. If you don\'t mind throwing caution to '.
'the wind, enable this directive, but I strongly recommend you also '.
'consider blacklisting IDs you use (%Attr.IDBlacklist) or prefixing all '.
'user supplied IDs (%Attr.IDPrefix). This directive has been available '.
'since 1.2.0, and when set to true reverts to the behavior of pre-1.2.0 '.
'versions.'
);
HTMLPurifier_ConfigSchema::defineAlias(
'HTML', 'EnableAttrID', 'Attr', 'EnableID'
);
HTMLPurifier_ConfigSchema::define(
'Attr', 'IDPrefix', '', 'string',
'String to prefix to IDs. If you have no idea what IDs your pages '.
'may use, you may opt to simply add a prefix to all user-submitted ID '.
'attributes so that they are still usable, but will not conflict with '.
'core page IDs. Example: setting the directive to \'user_\' will result in '.
'a user submitted \'foo\' to become \'user_foo\' Be sure to set '.
'%HTML.EnableAttrID to true before using '.
'this. This directive was available since 1.2.0.'
);
HTMLPurifier_ConfigSchema::define(
'Attr', 'IDPrefixLocal', '', 'string',
'Temporary prefix for IDs used in conjunction with %Attr.IDPrefix. If '.
'you need to allow multiple sets of '.
'user content on web page, you may need to have a seperate prefix that '.
'changes with each iteration. This way, seperately submitted user content '.
'displayed on the same page doesn\'t clobber each other. Ideal values '.
'are unique identifiers for the content it represents (i.e. the id of '.
'the row in the database). Be sure to add a seperator (like an underscore) '.
'at the end. Warning: this directive will not work unless %Attr.IDPrefix '.
'is set to a non-empty value! This directive was available since 1.2.0.'
);
HTMLPurifier_ConfigSchema::define(
'Attr', 'IDBlacklistRegexp', null, 'string/null',
'PCRE regular expression to be matched against all IDs. If the expression '.
'is matches, the ID is rejected. Use this with care: may cause '.
'significant degradation. ID matching is done after all other '.
'validation. This directive was available since 1.6.0.'
);
/**
* Validates the HTML attribute ID.
* @warning Even though this is the id processor, it
@@ -66,7 +15,7 @@ class HTMLPurifier_AttrDef_HTML_ID extends HTMLPurifier_AttrDef
// ref functionality disabled, since we also have to verify
// whether or not the ID it refers to exists
function validate($id, $config, &$context) {
public function validate($id, $config, $context) {
if (!$config->get('Attr', 'EnableID')) return false;
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/HTML/Pixels.php';
/**
* Validates the HTML type length (not to be confused with CSS's length).
*
@@ -13,7 +10,7 @@ require_once 'HTMLPurifier/AttrDef/HTML/Pixels.php';
class HTMLPurifier_AttrDef_HTML_Length extends HTMLPurifier_AttrDef_HTML_Pixels
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = trim($string);
if ($string === '') return false;
@@ -1,22 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
HTMLPurifier_ConfigSchema::define(
'Attr', 'AllowedRel', array(), 'lookup',
'List of allowed forward document relationships in the rel attribute. '.
'Common values may be nofollow or print. By default, this is empty, '.
'meaning that no document relationships are allowed. This directive '.
'was available since 1.6.0.'
);
HTMLPurifier_ConfigSchema::define(
'Attr', 'AllowedRev', array(), 'lookup',
'List of allowed reverse document relationships in the rev attribute. '.
'This attribute is a bit of an edge-case; if you don\'t know what it '.
'is for, stay away. This directive was available since 1.6.0.'
);
/**
* Validates a rel/rev link attribute against a directive of allowed values
* @note We cannot use Enum because link types allow multiple
@@ -27,9 +10,9 @@ class HTMLPurifier_AttrDef_HTML_LinkTypes extends HTMLPurifier_AttrDef
{
/** Name config attribute to pull. */
var $name;
protected $name;
function HTMLPurifier_AttrDef_HTML_LinkTypes($name) {
public function __construct($name) {
$configLookup = array(
'rel' => 'AllowedRel',
'rev' => 'AllowedRev'
@@ -42,7 +25,7 @@ class HTMLPurifier_AttrDef_HTML_LinkTypes extends HTMLPurifier_AttrDef
$this->name = $configLookup[$name];
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$allowed = $config->get('Attr', $this->name);
if (empty($allowed)) return false;
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/HTML/Length.php';
/**
* Validates a MultiLength as defined by the HTML spec.
*
@@ -12,7 +9,7 @@ require_once 'HTMLPurifier/AttrDef/HTML/Length.php';
class HTMLPurifier_AttrDef_HTML_MultiLength extends HTMLPurifier_AttrDef_HTML_Length
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = trim($string);
if ($string === '') return false;
@@ -1,8 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/Config.php';
/**
* Validates contents based on NMTOKENS attribute type.
* @note The only current use for this is the class attribute in HTML
@@ -13,7 +10,7 @@ require_once 'HTMLPurifier/Config.php';
class HTMLPurifier_AttrDef_HTML_Nmtokens extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = trim($string);
@@ -1,14 +1,12 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates an integer representation of pixels according to the HTML spec.
*/
class HTMLPurifier_AttrDef_HTML_Pixels extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$string = trim($string);
if ($string === '0') return $string;
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates an integer.
* @note While this class was modeled off the CSS definition, no currently
@@ -15,24 +13,24 @@ class HTMLPurifier_AttrDef_Integer extends HTMLPurifier_AttrDef
/**
* Bool indicating whether or not negative values are allowed
*/
var $negative = true;
protected $negative = true;
/**
* Bool indicating whether or not zero is allowed
*/
var $zero = true;
protected $zero = true;
/**
* Bool indicating whether or not positive values are allowed
*/
var $positive = true;
protected $positive = true;
/**
* @param $negative Bool indicating whether or not negative values are allowed
* @param $zero Bool indicating whether or not zero is allowed
* @param $positive Bool indicating whether or not positive values are allowed
*/
function HTMLPurifier_AttrDef_Integer(
public function __construct(
$negative = true, $zero = true, $positive = true
) {
$this->negative = $negative;
@@ -40,7 +38,7 @@ class HTMLPurifier_AttrDef_Integer extends HTMLPurifier_AttrDef
$this->positive = $positive;
}
function validate($integer, $config, &$context) {
public function validate($integer, $config, $context) {
$integer = $this->parseCDATA($integer);
if ($integer === '') return false;
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates the HTML attribute lang, effectively a language code.
* @note Built according to RFC 3066, which obsoleted RFC 1766
@@ -9,7 +7,7 @@ require_once 'HTMLPurifier/AttrDef.php';
class HTMLPurifier_AttrDef_Lang extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
// moodle change - we use special lang strings unfortunatelly
return ereg_replace('[^0-9a-zA-Z_-]', '', $string);
@@ -1,14 +1,12 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates arbitrary text according to the HTML spec.
*/
class HTMLPurifier_AttrDef_Text extends HTMLPurifier_AttrDef
{
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
return $this->parseCDATA($string);
}
+5 -66
View File
@@ -1,66 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/URIParser.php';
require_once 'HTMLPurifier/URIScheme.php';
require_once 'HTMLPurifier/URISchemeRegistry.php';
require_once 'HTMLPurifier/AttrDef/URI/Host.php';
require_once 'HTMLPurifier/PercentEncoder.php';
require_once 'HTMLPurifier/AttrDef/URI/Email.php';
// special case filtering directives
HTMLPurifier_ConfigSchema::define(
'URI', 'Munge', null, 'string/null', '
<p>
Munges all browsable (usually http, https and ftp)
absolute URI\'s into another URI, usually a URI redirection service.
This directive accepts a URI, formatted with a <code>%s</code> where
the url-encoded original URI should be inserted (sample:
<code>http://www.google.com/url?q=%s</code>).
</p>
<p>
Uses for this directive:
</p>
<ul>
<li>
Prevent PageRank leaks, while being fairly transparent
to users (you may also want to add some client side JavaScript to
override the text in the statusbar). <strong>Notice</strong>:
Many security experts believe that this form of protection does not deter spam-bots.
</li>
<li>
Redirect users to a splash page telling them they are leaving your
website. While this is poor usability practice, it is often mandated
in corporate environments.
</li>
</ul>
<p>
This directive has been available since 1.3.0.
</p>
');
// disabling directives
HTMLPurifier_ConfigSchema::define(
'URI', 'Disable', false, 'bool', '
<p>
Disables all URIs in all forms. Not sure why you\'d want to do that
(after all, the Internet\'s founded on the notion of a hyperlink).
This directive has been available since 1.3.0.
</p>
');
HTMLPurifier_ConfigSchema::defineAlias('Attr', 'DisableURI', 'URI', 'Disable');
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableResources', false, 'bool', '
<p>
Disables embedding resources, essentially meaning no pictures. You can
still link to them though. See %URI.DisableExternalResources for why
this might be a good idea. This directive has been available since 1.3.0.
</p>
');
/**
* Validates a URI as defined by RFC 3986.
* @note Scheme-specific mechanics deferred to HTMLPurifier_URIScheme
@@ -68,18 +7,18 @@ HTMLPurifier_ConfigSchema::define(
class HTMLPurifier_AttrDef_URI extends HTMLPurifier_AttrDef
{
var $parser;
var $embedsResource;
protected $parser;
protected $embedsResource;
/**
* @param $embeds_resource_resource Does the URI here result in an extra HTTP request?
*/
function HTMLPurifier_AttrDef_URI($embeds_resource = false) {
public function __construct($embeds_resource = false) {
$this->parser = new HTMLPurifier_URIParser();
$this->embedsResource = (bool) $embeds_resource;
}
function validate($uri, $config, &$context) {
public function validate($uri, $config, $context) {
if ($config->get('URI', 'Disable')) return false;
@@ -100,7 +39,7 @@ class HTMLPurifier_AttrDef_URI extends HTMLPurifier_AttrDef
if (!$result) break;
// chained filtering
$uri_def =& $config->getDefinition('URI');
$uri_def = $config->getDefinition('URI');
$result = $uri_def->filter($uri, $config, $context);
if (!$result) break;
@@ -1,8 +1,6 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
class HTMLPurifier_AttrDef_URI_Email extends HTMLPurifier_AttrDef
abstract class HTMLPurifier_AttrDef_URI_Email extends HTMLPurifier_AttrDef
{
/**
@@ -15,4 +13,3 @@ class HTMLPurifier_AttrDef_URI_Email extends HTMLPurifier_AttrDef
}
// sub-implementations
//moodlefix require_once 'HTMLPurifier/AttrDef/URI/Email/SimpleCheck.php';
@@ -0,0 +1,20 @@
<?php
/**
* Primitive email validation class based on the regexp found at
* http://www.regular-expressions.info/email.html
*/
class HTMLPurifier_AttrDef_URI_Email_SimpleCheck extends HTMLPurifier_AttrDef_URI_Email
{
public function validate($string, $config, $context) {
// no support for named mailboxes i.e. "Bob <[email protected]>"
// that needs more percent encoding to be done
if ($string == '') return false;
$string = trim($string);
$result = preg_match('/^[A-Z0-9._%-]+@[A-Z0-9.-]+\.[A-Z]{2,4}$/i', $string);
return $result ? $string : false;
}
}
@@ -1,9 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/AttrDef/URI/IPv4.php';
require_once 'HTMLPurifier/AttrDef/URI/IPv6.php';
/**
* Validates a host according to the IPv4, IPv6 and DNS (future) specifications.
*/
@@ -13,19 +9,19 @@ class HTMLPurifier_AttrDef_URI_Host extends HTMLPurifier_AttrDef
/**
* Instance of HTMLPurifier_AttrDef_URI_IPv4 sub-validator
*/
var $ipv4;
protected $ipv4;
/**
* Instance of HTMLPurifier_AttrDef_URI_IPv6 sub-validator
*/
var $ipv6;
protected $ipv6;
function HTMLPurifier_AttrDef_URI_Host() {
public function __construct() {
$this->ipv4 = new HTMLPurifier_AttrDef_URI_IPv4();
$this->ipv6 = new HTMLPurifier_AttrDef_URI_IPv6();
}
function validate($string, $config, &$context) {
public function validate($string, $config, $context) {
$length = strlen($string);
if ($string === '') return '';
if ($length > 1 && $string[0] === '[' && $string[$length-1] === ']') {
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
/**
* Validates an IPv4 address
* @author Feyd @ forums.devnetwork.net (public domain)
@@ -11,11 +9,10 @@ class HTMLPurifier_AttrDef_URI_IPv4 extends HTMLPurifier_AttrDef
/**
* IPv4 regex, protected so that IPv6 can reuse it
* @protected
*/
var $ip4;
protected $ip4;
function validate($aIP, $config, &$context) {
public function validate($aIP, $config, $context) {
if (!$this->ip4) $this->_loadRegex();
@@ -32,7 +29,7 @@ class HTMLPurifier_AttrDef_URI_IPv4 extends HTMLPurifier_AttrDef
* Lazy load function to prevent regex from being stuffed in
* cache.
*/
function _loadRegex() {
protected function _loadRegex() {
$oct = '(?:25[0-5]|2[0-4][0-9]|1[0-9]{2}|[1-9][0-9]|[0-9])'; // 0-255
$this->ip4 = "(?:{$oct}\\.{$oct}\\.{$oct}\\.{$oct})";
}
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef/URI/IPv4.php';
/**
* Validates an IPv6 address.
* @author Feyd @ forums.devnetwork.net (public domain)
@@ -11,7 +9,7 @@ require_once 'HTMLPurifier/AttrDef/URI/IPv4.php';
class HTMLPurifier_AttrDef_URI_IPv6 extends HTMLPurifier_AttrDef_URI_IPv4
{
function validate($aIP, $config, &$context) {
public function validate($aIP, $config, $context) {
if (!$this->ip4) $this->_loadRegex();
@@ -14,7 +14,7 @@
* more details.
*/
class HTMLPurifier_AttrTransform
abstract class HTMLPurifier_AttrTransform
{
/**
@@ -26,9 +26,7 @@ class HTMLPurifier_AttrTransform
* @param $context Mandatory HTMLPurifier_Context object
* @returns Processed attribute array.
*/
function transform($attr, $config, &$context) {
trigger_error('Cannot call abstract function', E_USER_ERROR);
}
abstract public function transform($attr, $config, $context);
/**
* Prepends CSS properties to the style attribute, creating the
@@ -36,7 +34,7 @@ class HTMLPurifier_AttrTransform
* @param $attr Attribute array to process (passed by reference)
* @param $css CSS to prepend
*/
function prependCSS(&$attr, $css) {
public function prependCSS(&$attr, $css) {
$attr['style'] = isset($attr['style']) ? $attr['style'] : '';
$attr['style'] = $css . $attr['style'];
}
@@ -46,7 +44,7 @@ class HTMLPurifier_AttrTransform
* @param $attr Attribute array to process (passed by reference)
* @param $key Key of attribute to confiscate
*/
function confiscateAttr(&$attr, $key) {
public function confiscateAttr(&$attr, $key) {
if (!isset($attr[$key])) return null;
$value = $attr[$key];
unset($attr[$key]);
@@ -1,26 +1,14 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
// this MUST be placed in post, as it assumes that any value in dir is valid
HTMLPurifier_ConfigSchema::define(
'Attr', 'DefaultTextDir', 'ltr', 'string',
'Defines the default text direction (ltr or rtl) of the document '.
'being parsed. This generally is the same as the value of the dir '.
'attribute in HTML, or ltr if that is not specified.'
);
HTMLPurifier_ConfigSchema::defineAllowedValues(
'Attr', 'DefaultTextDir', array( 'ltr', 'rtl' )
);
/**
* Post-trasnform that ensures that bdo tags have the dir attribute set.
*/
class HTMLPurifier_AttrTransform_BdoDir extends HTMLPurifier_AttrTransform
{
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (isset($attr['dir'])) return $attr;
$attr['dir'] = $config->get('Attr', 'DefaultTextDir');
return $attr;
@@ -1,14 +1,11 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Pre-transform that changes deprecated bgcolor attribute to CSS.
*/
class HTMLPurifier_AttrTransform_BgColor
extends HTMLPurifier_AttrTransform {
class HTMLPurifier_AttrTransform_BgColor extends HTMLPurifier_AttrTransform {
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr['bgcolor'])) return $attr;
@@ -1,33 +1,30 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Pre-transform that changes converts a boolean attribute to fixed CSS
*/
class HTMLPurifier_AttrTransform_BoolToCSS
extends HTMLPurifier_AttrTransform {
class HTMLPurifier_AttrTransform_BoolToCSS extends HTMLPurifier_AttrTransform {
/**
* Name of boolean attribute that is trigger
*/
var $attr;
protected $attr;
/**
* CSS declarations to add to style, needs trailing semicolon
*/
var $css;
protected $css;
/**
* @param $attr string attribute name to convert from
* @param $css string CSS declarations to add to style (needs semicolon)
*/
function HTMLPurifier_AttrTransform_BoolToCSS($attr, $css) {
public function __construct($attr, $css) {
$this->attr = $attr;
$this->css = $css;
}
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr[$this->attr])) return $attr;
unset($attr[$this->attr]);
$this->prependCSS($attr, $this->css);
@@ -1,13 +1,11 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Pre-transform that changes deprecated border attribute to CSS.
*/
class HTMLPurifier_AttrTransform_Border extends HTMLPurifier_AttrTransform {
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr['border'])) return $attr;
$border_width = $this->confiscateAttr($attr, 'border');
// some validation should happen here
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Generic pre-transform that converts an attribute with a fixed number of
* values (enumerated) to CSS.
@@ -11,32 +9,32 @@ class HTMLPurifier_AttrTransform_EnumToCSS extends HTMLPurifier_AttrTransform {
/**
* Name of attribute to transform from
*/
var $attr;
protected $attr;
/**
* Lookup array of attribute values to CSS
*/
var $enumToCSS = array();
protected $enumToCSS = array();
/**
* Case sensitivity of the matching
* @warning Currently can only be guaranteed to work with ASCII
* values.
*/
var $caseSensitive = false;
protected $caseSensitive = false;
/**
* @param $attr String attribute name to transform from
* @param $enumToCSS Lookup array of attribute values to CSS
* @param $case_sensitive Boolean case sensitivity indicator, default false
*/
function HTMLPurifier_AttrTransform_EnumToCSS($attr, $enum_to_css, $case_sensitive = false) {
public function __construct($attr, $enum_to_css, $case_sensitive = false) {
$this->attr = $attr;
$this->enumToCSS = $enum_to_css;
$this->caseSensitive = (bool) $case_sensitive;
}
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr[$this->attr])) return $attr;
@@ -1,24 +1,7 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
// must be called POST validation
HTMLPurifier_ConfigSchema::define(
'Attr', 'DefaultInvalidImage', '', 'string',
'This is the default image an img tag will be pointed to if it does '.
'not have a valid src attribute. In future versions, we may allow the '.
'image tag to be removed completely, but due to design issues, this is '.
'not possible right now.'
);
HTMLPurifier_ConfigSchema::define(
'Attr', 'DefaultInvalidImageAlt', 'Invalid image', 'string',
'This is the content of the alt tag of an invalid image if the user '.
'had not previously specified an alt attribute. It has no effect when the '.
'image is valid but there was no alt attribute present.'
);
/**
* Transform that supplies default values for the src and alt attributes
* in img tags, as well as prevents the img tag from being removed
@@ -28,7 +11,7 @@ HTMLPurifier_ConfigSchema::define(
class HTMLPurifier_AttrTransform_ImgRequired extends HTMLPurifier_AttrTransform
{
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
$src = true;
if (!isset($attr['src'])) {
@@ -1,27 +1,24 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Pre-transform that changes deprecated hspace and vspace attributes to CSS
*/
class HTMLPurifier_AttrTransform_ImgSpace
extends HTMLPurifier_AttrTransform {
class HTMLPurifier_AttrTransform_ImgSpace extends HTMLPurifier_AttrTransform {
var $attr;
var $css = array(
protected $attr;
protected $css = array(
'hspace' => array('left', 'right'),
'vspace' => array('top', 'bottom')
);
function HTMLPurifier_AttrTransform_ImgSpace($attr) {
public function __construct($attr) {
$this->attr = $attr;
if (!isset($this->css[$attr])) {
trigger_error(htmlspecialchars($attr) . ' is not valid space attribute');
}
}
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr[$this->attr])) return $attr;
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Post-transform that copies lang's value to xml:lang (and vice-versa)
* @note Theoretically speaking, this could be a pre-transform, but putting
@@ -10,7 +8,7 @@ require_once 'HTMLPurifier/AttrTransform.php';
class HTMLPurifier_AttrTransform_Lang extends HTMLPurifier_AttrTransform
{
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
$lang = isset($attr['lang']) ? $attr['lang'] : false;
$xml_lang = isset($attr['xml:lang']) ? $attr['xml:lang'] : false;
@@ -1,22 +1,20 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Class for handling width/height length attribute transformations to CSS
*/
class HTMLPurifier_AttrTransform_Length extends HTMLPurifier_AttrTransform
{
var $name;
var $cssName;
protected $name;
protected $cssName;
function HTMLPurifier_AttrTransform_Length($name, $css_name = null) {
public function __construct($name, $css_name = null) {
$this->name = $name;
$this->cssName = $css_name ? $css_name : $name;
}
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr[$this->name])) return $attr;
$length = $this->confiscateAttr($attr, $this->name);
if(ctype_digit($length)) $length .= 'px';
@@ -1,14 +1,12 @@
<?php
require_once 'HTMLPurifier/AttrTransform.php';
/**
* Pre-transform that changes deprecated name attribute to ID if necessary
*/
class HTMLPurifier_AttrTransform_Name extends HTMLPurifier_AttrTransform
{
function transform($attr, $config, &$context) {
public function transform($attr, $config, $context) {
if (!isset($attr['name'])) return $attr;
$id = $this->confiscateAttr($attr, 'name');
if ( isset($attr['id'])) return $attr;
@@ -0,0 +1,14 @@
<?php
/**
* Implements required attribute stipulation for <script>
*/
class HTMLPurifier_AttrTransform_ScriptRequired extends HTMLPurifier_AttrTransform
{
public function transform($attr, $config, $context) {
if (!isset($attr['type'])) {
$attr['type'] = 'text/javascript';
}
return $attr;
}
}
+4 -18
View File
@@ -1,18 +1,5 @@
<?php
require_once 'HTMLPurifier/AttrDef/Lang.php';
require_once 'HTMLPurifier/AttrDef/Enum.php';
require_once 'HTMLPurifier/AttrDef/HTML/Bool.php';
require_once 'HTMLPurifier/AttrDef/HTML/ID.php';
require_once 'HTMLPurifier/AttrDef/HTML/Length.php';
require_once 'HTMLPurifier/AttrDef/HTML/MultiLength.php';
require_once 'HTMLPurifier/AttrDef/HTML/Nmtokens.php';
require_once 'HTMLPurifier/AttrDef/HTML/Pixels.php';
require_once 'HTMLPurifier/AttrDef/HTML/Color.php';
require_once 'HTMLPurifier/AttrDef/Integer.php';
require_once 'HTMLPurifier/AttrDef/Text.php';
require_once 'HTMLPurifier/AttrDef/URI.php';
/**
* Provides lookup array of attribute types to HTMLPurifier_AttrDef objects
*/
@@ -20,15 +7,14 @@ class HTMLPurifier_AttrTypes
{
/**
* Lookup array of attribute string identifiers to concrete implementations
* @protected
*/
var $info = array();
protected $info = array();
/**
* Constructs the info array, supplying default implementations for attribute
* types.
*/
function HTMLPurifier_AttrTypes() {
public function __construct() {
// pseudo-types, must be instantiated via shorthand
$this->info['Enum'] = new HTMLPurifier_AttrDef_Enum();
$this->info['Bool'] = new HTMLPurifier_AttrDef_HTML_Bool();
@@ -57,7 +43,7 @@ class HTMLPurifier_AttrTypes
* @param $type String type name
* @return Object AttrDef for type
*/
function get($type) {
public function get($type) {
// determine if there is any extra info tacked on
if (strpos($type, '#') !== false) list($type, $string) = explode('#', $type, 2);
@@ -77,7 +63,7 @@ class HTMLPurifier_AttrTypes
* @param $type String type name
* @param $impl Object AttrDef for type
*/
function set($type, $impl) {
public function set($type, $impl) {
$this->info[$type] = $impl;
}
}
@@ -18,7 +18,7 @@ class HTMLPurifier_AttrValidator
* @param $config Instance of HTMLPurifier_Config
* @param $context Instance of HTMLPurifier_Context
*/
function validateToken(&$token, &$config, &$context) {
public function validateToken(&$token, &$config, $context) {
$definition = $config->getHTMLDefinition();
$e =& $context->get('ErrorCollector', true);
@@ -34,7 +34,10 @@ class HTMLPurifier_AttrValidator
$current_token =& $context->get('CurrentToken', true);
if (!$current_token) $context->register('CurrentToken', $token);
if ($token->type !== 'start' && $token->type !== 'empty') return $token;
if (
!$token instanceof HTMLPurifier_Token_Start &&
!$token instanceof HTMLPurifier_Token_Empty
) return $token;
// create alias to global definition array, see also $defs
// DEFINITION CALL
@@ -0,0 +1,96 @@
<?php
// constants are slow, so we use as few as possible
if (!defined('HTMLPURIFIER_PREFIX')) {
define('HTMLPURIFIER_PREFIX', realpath(dirname(__FILE__) . '/..'));
}
// accomodations for versions earlier than 5.0.2
// borrowed from PHP_Compat, LGPL licensed, by Aidan Lister <[email protected]>
if (!defined('PHP_EOL')) {
switch (strtoupper(substr(PHP_OS, 0, 3))) {
case 'WIN':
define('PHP_EOL', "\r\n");
break;
case 'DAR':
define('PHP_EOL', "\r");
break;
default:
define('PHP_EOL', "\n");
}
}
/**
* Bootstrap class that contains meta-functionality for HTML Purifier such as
* the autoload function.
*
* @note
* This class may be used without any other files from HTML Purifier.
*/
class HTMLPurifier_Bootstrap
{
/**
* Autoload function for HTML Purifier
* @param $class Class to load
*/
public static function autoload($class) {
$file = HTMLPurifier_Bootstrap::getPath($class);
if (!$file) return false;
require HTMLPURIFIER_PREFIX . '/' . $file;
return true;
}
/**
* Returns the path for a specific class.
*/
public static function getPath($class) {
if (strncmp('HTMLPurifier', $class, 12) !== 0) return false;
// Custom implementations
if (strncmp('HTMLPurifier_Language_', $class, 22) === 0) {
$code = str_replace('_', '-', substr($class, 22));
$file = 'HTMLPurifier/Language/classes/' . $code . '.php';
} else {
$file = str_replace('_', '/', $class) . '.php';
}
if (!file_exists(HTMLPURIFIER_PREFIX . '/' . $file)) return false;
return $file;
}
/**
* "Pre-registers" our autoloader on the SPL stack.
*/
public static function registerAutoload() {
$autoload = array('HTMLPurifier_Bootstrap', 'autoload');
if ( ($funcs = spl_autoload_functions()) === false ) {
spl_autoload_register($autoload);
} elseif (function_exists('spl_autoload_unregister')) {
$compat = version_compare(PHP_VERSION, '5.1.2', '<=') &&
version_compare(PHP_VERSION, '5.1.0', '>=');
foreach ($funcs as $func) {
if (is_array($func)) {
// :TRICKY: There are some compatibility issues and some
// places where we need to error out
$reflector = new ReflectionMethod($func[0], $func[1]);
if (!$reflector->isStatic()) {
throw new Exception('
HTML Purifier autoloader registrar is not compatible
with non-static object methods due to PHP Bug #44144;
Please do not use HTMLPurifier.autoload.php (or any
file that includes this file); instead, place the code:
spl_autoload_register(array(\'HTMLPurifier_Bootstrap\', \'autoload\'))
after your own autoloaders.
');
}
// Suprisingly, spl_autoload_register supports the
// Class::staticMethod callback format, although call_user_func doesn't
if ($compat) $func = implode('::', $func);
}
spl_autoload_unregister($func);
}
spl_autoload_register($autoload);
foreach ($funcs as $func) spl_autoload_register($func);
}
}
}
+76 -30
View File
@@ -1,32 +1,5 @@
<?php
require_once 'HTMLPurifier/Definition.php';
require_once 'HTMLPurifier/AttrDef/CSS/Background.php';
require_once 'HTMLPurifier/AttrDef/CSS/BackgroundPosition.php';
require_once 'HTMLPurifier/AttrDef/CSS/Border.php';
require_once 'HTMLPurifier/AttrDef/CSS/Color.php';
require_once 'HTMLPurifier/AttrDef/CSS/Composite.php';
require_once 'HTMLPurifier/AttrDef/CSS/DenyElementDecorator.php';
require_once 'HTMLPurifier/AttrDef/CSS/Font.php';
require_once 'HTMLPurifier/AttrDef/CSS/FontFamily.php';
require_once 'HTMLPurifier/AttrDef/CSS/Length.php';
require_once 'HTMLPurifier/AttrDef/CSS/ListStyle.php';
require_once 'HTMLPurifier/AttrDef/CSS/Multiple.php';
require_once 'HTMLPurifier/AttrDef/CSS/Percentage.php';
require_once 'HTMLPurifier/AttrDef/CSS/TextDecoration.php';
require_once 'HTMLPurifier/AttrDef/CSS/URI.php';
require_once 'HTMLPurifier/AttrDef/Enum.php';
HTMLPurifier_ConfigSchema::define(
'CSS', 'DefinitionRev', 1, 'int', '
<p>
Revision identifier for your custom definition. See
%HTML.DefinitionRev for details. This directive has been available
since 2.0.0.
</p>
');
/**
* Defines allowed CSS attributes and what their values are.
* @see HTMLPurifier_HTMLDefinition
@@ -34,17 +7,17 @@ HTMLPurifier_ConfigSchema::define(
class HTMLPurifier_CSSDefinition extends HTMLPurifier_Definition
{
var $type = 'CSS';
public $type = 'CSS';
/**
* Assoc array of attribute name to definition object.
*/
var $info = array();
public $info = array();
/**
* Constructs the info array. The meat of this class.
*/
function doSetup($config) {
protected function doSetup($config) {
$this->info['text-align'] = new HTMLPurifier_AttrDef_Enum(
array('left', 'right', 'center', 'justify'), false);
@@ -226,7 +199,80 @@ class HTMLPurifier_CSSDefinition extends HTMLPurifier_Definition
// partial support
$this->info['white-space'] = new HTMLPurifier_AttrDef_Enum(array('nowrap'));
if ($config->get('CSS', 'Proprietary')) {
$this->doSetupProprietary($config);
}
if ($config->get('CSS', 'AllowTricky')) {
$this->doSetupTricky($config);
}
$allow_important = $config->get('CSS', 'AllowImportant');
// wrap all attr-defs with decorator that handles !important
foreach ($this->info as $k => $v) {
$this->info[$k] = new HTMLPurifier_AttrDef_CSS_ImportantDecorator($v, $allow_important);
}
$this->setupConfigStuff($config);
}
protected function doSetupProprietary($config) {
// Internet Explorer only scrollbar colors
$this->info['scrollbar-arrow-color'] = new HTMLPurifier_AttrDef_CSS_Color();
$this->info['scrollbar-base-color'] = new HTMLPurifier_AttrDef_CSS_Color();
$this->info['scrollbar-darkshadow-color'] = new HTMLPurifier_AttrDef_CSS_Color();
$this->info['scrollbar-face-color'] = new HTMLPurifier_AttrDef_CSS_Color();
$this->info['scrollbar-highlight-color'] = new HTMLPurifier_AttrDef_CSS_Color();
$this->info['scrollbar-shadow-color'] = new HTMLPurifier_AttrDef_CSS_Color();
// technically not proprietary, but CSS3, and no one supports it
$this->info['opacity'] = new HTMLPurifier_AttrDef_CSS_AlphaValue();
$this->info['-moz-opacity'] = new HTMLPurifier_AttrDef_CSS_AlphaValue();
$this->info['-khtml-opacity'] = new HTMLPurifier_AttrDef_CSS_AlphaValue();
// only opacity, for now
$this->info['filter'] = new HTMLPurifier_AttrDef_CSS_Filter();
}
protected function doSetupTricky($config) {
$this->info['display'] = new HTMLPurifier_AttrDef_Enum(array(
'inline', 'block', 'list-item', 'run-in', 'compact',
'marker', 'table', 'inline-table', 'table-row-group',
'table-header-group', 'table-footer-group', 'table-row',
'table-column-group', 'table-column', 'table-cell', 'table-caption', 'none'
));
$this->info['visibility'] = new HTMLPurifier_AttrDef_Enum(array(
'visible', 'hidden', 'collapse'
));
}
/**
* Performs extra config-based processing. Based off of
* HTMLPurifier_HTMLDefinition.
* @todo Refactor duplicate elements into common class (probably using
* composition, not inheritance).
*/
protected function setupConfigStuff($config) {
// setup allowed elements
$support = "(for information on implementing this, see the ".
"support forums) ";
$allowed_attributes = $config->get('CSS', 'AllowedProperties');
if ($allowed_attributes !== null) {
foreach ($this->info as $name => $d) {
if(!isset($allowed_attributes[$name])) unset($this->info[$name]);
unset($allowed_attributes[$name]);
}
// emit errors
foreach ($allowed_attributes as $name => $d) {
// :TODO: Is this htmlspecialchars() call really necessary?
$name = htmlspecialchars($name);
trigger_error("Style attribute '$name' is not supported $support", E_USER_WARNING);
}
}
}
}
+5 -25
View File
@@ -1,50 +1,32 @@
<?php
// HTMLPurifier_ChildDef and inheritance have three types of output:
// true = leave nodes as is
// false = delete parent node and all children
// array(...) = replace children nodes with these
HTMLPurifier_ConfigSchema::define(
'Core', 'EscapeInvalidChildren', false, 'bool',
'When true, a child is found that is not allowed in the context of the '.
'parent element will be transformed into text as if it were ASCII. When '.
'false, that element and all internal tags will be dropped, though text '.
'will be preserved. There is no option for dropping the element but '.
'preserving child nodes.'
);
/**
* Defines allowed child nodes and validates tokens against it.
*/
class HTMLPurifier_ChildDef
abstract class HTMLPurifier_ChildDef
{
/**
* Type of child definition, usually right-most part of class name lowercase.
* Used occasionally in terms of context.
* @public
*/
var $type;
public $type;
/**
* Bool that indicates whether or not an empty array of children is okay
*
* This is necessary for redundant checking when changes affecting
* a child node may cause a parent node to now be disallowed.
*
* @public
*/
var $allow_empty;
public $allow_empty;
/**
* Lookup array of all elements that this definition could possibly allow
*/
var $elements = array();
public $elements = array();
/**
* Validates nodes according to definition and returns modification.
*
* @public
* @param $tokens_of_children Array of HTMLPurifier_Token
* @param $config HTMLPurifier_Config object
* @param $context HTMLPurifier_Context object
@@ -52,9 +34,7 @@ class HTMLPurifier_ChildDef
* @return bool false to remove parent node
* @return array of replacement child tokens
*/
function validateChildren($tokens_of_children, $config, &$context) {
trigger_error('Call to abstract function', E_USER_ERROR);
}
abstract public function validateChildren($tokens_of_children, $config, $context);
}
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/ChildDef.php';
/**
* Definition that uses different definitions depending on context.
*
@@ -16,29 +14,27 @@ class HTMLPurifier_ChildDef_Chameleon extends HTMLPurifier_ChildDef
/**
* Instance of the definition object to use when inline. Usually stricter.
* @public
*/
var $inline;
public $inline;
/**
* Instance of the definition object to use when block.
* @public
*/
var $block;
public $block;
var $type = 'chameleon';
public $type = 'chameleon';
/**
* @param $inline List of elements to allow when inline.
* @param $block List of elements to allow when block.
*/
function HTMLPurifier_ChildDef_Chameleon($inline, $block) {
public function __construct($inline, $block) {
$this->inline = new HTMLPurifier_ChildDef_Optional($inline);
$this->block = new HTMLPurifier_ChildDef_Optional($block);
$this->elements = $this->block->elements;
}
function validateChildren($tokens_of_children, $config, &$context) {
public function validateChildren($tokens_of_children, $config, $context) {
if ($context->get('IsInline') === false) {
return $this->block->validateChildren(
$tokens_of_children, $config, $context);
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/ChildDef.php';
/**
* Custom validation class, accepts DTD child definitions
*
@@ -12,28 +10,28 @@ require_once 'HTMLPurifier/ChildDef.php';
*/
class HTMLPurifier_ChildDef_Custom extends HTMLPurifier_ChildDef
{
var $type = 'custom';
var $allow_empty = false;
public $type = 'custom';
public $allow_empty = false;
/**
* Allowed child pattern as defined by the DTD
*/
var $dtd_regex;
public $dtd_regex;
/**
* PCRE regex derived from $dtd_regex
* @private
*/
var $_pcre_regex;
private $_pcre_regex;
/**
* @param $dtd_regex Allowed child pattern from the DTD
*/
function HTMLPurifier_ChildDef_Custom($dtd_regex) {
public function __construct($dtd_regex) {
$this->dtd_regex = $dtd_regex;
$this->_compileRegex();
}
/**
* Compiles the PCRE regex from a DTD regex ($dtd_regex to $_pcre_regex)
*/
function _compileRegex() {
protected function _compileRegex() {
$raw = str_replace(' ', '', $this->dtd_regex);
if ($raw{0} != '(') {
$raw = "($raw)";
@@ -61,7 +59,7 @@ class HTMLPurifier_ChildDef_Custom extends HTMLPurifier_ChildDef
$this->_pcre_regex = $reg;
}
function validateChildren($tokens_of_children, $config, &$context) {
public function validateChildren($tokens_of_children, $config, $context) {
$list_of_children = '';
$nesting = 0; // depth into the nest
foreach ($tokens_of_children as $token) {
@@ -69,9 +67,9 @@ class HTMLPurifier_ChildDef_Custom extends HTMLPurifier_ChildDef
$is_child = ($nesting == 0); // direct
if ($token->type == 'start') {
if ($token instanceof HTMLPurifier_Token_Start) {
$nesting++;
} elseif ($token->type == 'end') {
} elseif ($token instanceof HTMLPurifier_Token_End) {
$nesting--;
}
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/ChildDef.php';
/**
* Definition that disallows all elements.
* @warning validateChildren() in this class is actually never called, because
@@ -11,10 +9,10 @@ require_once 'HTMLPurifier/ChildDef.php';
*/
class HTMLPurifier_ChildDef_Empty extends HTMLPurifier_ChildDef
{
var $allow_empty = true;
var $type = 'empty';
function HTMLPurifier_ChildDef_Empty() {}
function validateChildren($tokens_of_children, $config, &$context) {
public $allow_empty = true;
public $type = 'empty';
public function __construct() {}
public function validateChildren($tokens_of_children, $config, $context) {
return array();
}
}
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/ChildDef/Required.php';
/**
* Definition that allows a set of elements, and allows no children.
* @note This is a hack to reuse code from HTMLPurifier_ChildDef_Required,
@@ -11,9 +9,9 @@ require_once 'HTMLPurifier/ChildDef/Required.php';
*/
class HTMLPurifier_ChildDef_Optional extends HTMLPurifier_ChildDef_Required
{
var $allow_empty = true;
var $type = 'optional';
function validateChildren($tokens_of_children, $config, &$context) {
public $allow_empty = true;
public $type = 'optional';
public function validateChildren($tokens_of_children, $config, $context) {
$result = parent::validateChildren($tokens_of_children, $config, $context);
if ($result === false) {
if (empty($tokens_of_children)) return true;
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/ChildDef.php';
/**
* Definition that allows a set of elements, but disallows empty children.
*/
@@ -11,11 +9,11 @@ class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef
* Lookup table of allowed elements.
* @public
*/
var $elements = array();
public $elements = array();
/**
* @param $elements List of allowed element names (lowercase).
*/
function HTMLPurifier_ChildDef_Required($elements) {
public function __construct($elements) {
if (is_string($elements)) {
$elements = str_replace(' ', '', $elements);
$elements = explode('|', $elements);
@@ -30,9 +28,9 @@ class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef
}
$this->elements = $elements;
}
var $allow_empty = false;
var $type = 'required';
function validateChildren($tokens_of_children, $config, &$context) {
public $allow_empty = false;
public $type = 'required';
public function validateChildren($tokens_of_children, $config, $context) {
// if there are no tokens, delete parent node
if (empty($tokens_of_children)) return false;
@@ -59,7 +57,7 @@ class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef
// generator
static $gen = null;
if ($gen === null) {
$gen = new HTMLPurifier_Generator();
$gen = new HTMLPurifier_Generator($config, $context);
}
foreach ($tokens_of_children as $token) {
@@ -71,9 +69,9 @@ class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef
$is_child = ($nesting == 0);
if ($token->type == 'start') {
if ($token instanceof HTMLPurifier_Token_Start) {
$nesting++;
} elseif ($token->type == 'end') {
} elseif ($token instanceof HTMLPurifier_Token_End) {
$nesting--;
}
@@ -81,7 +79,7 @@ class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef
$is_deleting = false;
if (!isset($this->elements[$token->name])) {
$is_deleting = true;
if ($pcdata_allowed && $token->type == 'text') {
if ($pcdata_allowed && $token instanceof HTMLPurifier_Token_Text) {
$result[] = $token;
} elseif ($pcdata_allowed && $escape_invalid_children) {
$result[] = new HTMLPurifier_Token_Text(
@@ -91,7 +89,7 @@ class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef
continue;
}
}
if (!$is_deleting || ($pcdata_allowed && $token->type == 'text')) {
if (!$is_deleting || ($pcdata_allowed && $token instanceof HTMLPurifier_Token_Text)) {
$result[] = $token;
} elseif ($pcdata_allowed && $escape_invalid_children) {
$result[] =
@@ -1,19 +1,16 @@
<?php
require_once 'HTMLPurifier/ChildDef/Required.php';
/**
* Takes the contents of blockquote when in strict and reformats for validation.
*/
class HTMLPurifier_ChildDef_StrictBlockquote
extends HTMLPurifier_ChildDef_Required
class HTMLPurifier_ChildDef_StrictBlockquote extends HTMLPurifier_ChildDef_Required
{
var $real_elements;
var $fake_elements;
var $allow_empty = true;
var $type = 'strictblockquote';
var $init = false;
function validateChildren($tokens_of_children, $config, &$context) {
protected $real_elements;
protected $fake_elements;
public $allow_empty = true;
public $type = 'strictblockquote';
protected $init = false;
public function validateChildren($tokens_of_children, $config, $context) {
$def = $config->getHTMLDefinition();
if (!$this->init) {
@@ -45,8 +42,8 @@ extends HTMLPurifier_ChildDef_Required
if (!$is_inline) {
if (!$depth) {
if (
($token->type == 'text' && !$token->is_whitespace) ||
($token->type != 'text' && !isset($this->elements[$token->name]))
($token instanceof HTMLPurifier_Token_Text && !$token->is_whitespace) ||
(!$token instanceof HTMLPurifier_Token_Text && !isset($this->elements[$token->name]))
) {
$is_inline = true;
$ret[] = $block_wrap_start;
@@ -55,7 +52,7 @@ extends HTMLPurifier_ChildDef_Required
} else {
if (!$depth) {
// starting tokens have been inline text / empty
if ($token->type == 'start' || $token->type == 'empty') {
if ($token instanceof HTMLPurifier_Token_Start || $token instanceof HTMLPurifier_Token_Empty) {
if (isset($this->elements[$token->name])) {
// ended
$ret[] = $block_wrap_end;
@@ -65,8 +62,8 @@ extends HTMLPurifier_ChildDef_Required
}
}
$ret[] = $token;
if ($token->type == 'start') $depth++;
if ($token->type == 'end') $depth--;
if ($token instanceof HTMLPurifier_Token_Start) $depth++;
if ($token instanceof HTMLPurifier_Token_End) $depth--;
}
if ($is_inline) $ret[] = $block_wrap_end;
return $ret;
@@ -1,18 +1,16 @@
<?php
require_once 'HTMLPurifier/ChildDef.php';
/**
* Definition for tables
*/
class HTMLPurifier_ChildDef_Table extends HTMLPurifier_ChildDef
{
var $allow_empty = false;
var $type = 'table';
var $elements = array('tr' => true, 'tbody' => true, 'thead' => true,
public $allow_empty = false;
public $type = 'table';
public $elements = array('tr' => true, 'tbody' => true, 'thead' => true,
'tfoot' => true, 'caption' => true, 'colgroup' => true, 'col' => true);
function HTMLPurifier_ChildDef_Table() {}
function validateChildren($tokens_of_children, $config, &$context) {
public function __construct() {}
public function validateChildren($tokens_of_children, $config, $context) {
if (empty($tokens_of_children)) return false;
// this ensures that the loop gets run one last time before closing
@@ -41,9 +39,9 @@ class HTMLPurifier_ChildDef_Table extends HTMLPurifier_ChildDef
if ($token === false) {
// terminating sequence started
} elseif ($token->type == 'start') {
} elseif ($token instanceof HTMLPurifier_Token_Start) {
$nesting++;
} elseif ($token->type == 'end') {
} elseif ($token instanceof HTMLPurifier_Token_End) {
$nesting--;
}
@@ -112,7 +110,7 @@ class HTMLPurifier_ChildDef_Table extends HTMLPurifier_ChildDef
$collection[] = $token;
continue;
default:
if ($token->type == 'text' && $token->is_whitespace) {
if ($token instanceof HTMLPurifier_Token_Text && $token->is_whitespace) {
$collection[] = $token;
$tag_index++;
}
+101 -111
View File
@@ -1,29 +1,5 @@
<?php
require_once 'HTMLPurifier/ConfigSchema.php';
// member variables
require_once 'HTMLPurifier/HTMLDefinition.php';
require_once 'HTMLPurifier/CSSDefinition.php';
require_once 'HTMLPurifier/URIDefinition.php';
require_once 'HTMLPurifier/Doctype.php';
require_once 'HTMLPurifier/DefinitionCacheFactory.php';
// accomodations for versions earlier than 4.3.10 and 5.0.2
// borrowed from PHP_Compat, LGPL licensed, by Aidan Lister <[email protected]>
if (!defined('PHP_EOL')) {
switch (strtoupper(substr(PHP_OS, 0, 3))) {
case 'WIN':
define('PHP_EOL', "\r\n");
break;
case 'DAR':
define('PHP_EOL', "\r");
break;
default:
define('PHP_EOL', "\n");
}
}
/**
* Configuration object that triggers customizable behavior.
*
@@ -35,6 +11,8 @@ if (!defined('PHP_EOL')) {
* because a configuration object should always be forwarded,
* otherwise, you run the risk of missing a parameter and then
* being stumped when a configuration directive doesn't work.
*
* @todo Reconsider some of the public member variables
*/
class HTMLPurifier_Config
{
@@ -42,69 +20,83 @@ class HTMLPurifier_Config
/**
* HTML Purifier's version
*/
var $version = '2.1.4';
/**
* Two-level associative array of configuration directives
*/
var $conf;
/**
* Reference HTMLPurifier_ConfigSchema for value checking
*/
var $def;
/**
* Indexed array of definitions
*/
var $definitions;
/**
* Bool indicator whether or not config is finalized
*/
var $finalized = false;
public $version = '3.1.0';
/**
* Bool indicator whether or not to automatically finalize
* the object if a read operation is done
*/
var $autoFinalize = true;
public $autoFinalize = true;
// protected member variables
/**
* Namespace indexed array of serials for specific namespaces (see
* getSerial for more info).
* getSerial() for more info).
*/
var $serials = array();
protected $serials = array();
/**
* Serial for entire configuration object
*/
var $serial;
protected $serial;
/**
* Two-level associative array of configuration directives
*/
protected $conf;
/**
* Parser for variables
*/
protected $parser;
/**
* Reference HTMLPurifier_ConfigSchema for value checking
* @note This is public for introspective purposes. Please don't
* abuse!
*/
public $def;
/**
* Indexed array of definitions
*/
protected $definitions;
/**
* Bool indicator whether or not config is finalized
*/
protected $finalized = false;
/**
* @param $definition HTMLPurifier_ConfigSchema that defines what directives
* are allowed.
*/
function HTMLPurifier_Config(&$definition) {
public function __construct($definition) {
$this->conf = $definition->defaults; // set up, copy in defaults
$this->def = $definition; // keep a copy around for checking
$this->parser = new HTMLPurifier_VarParser_Flexible();
}
/**
* Convenience constructor that creates a config object based on a mixed var
* @static
* @param mixed $config Variable that defines the state of the config
* object. Can be: a HTMLPurifier_Config() object,
* an array of directives based on loadArray(),
* or a string filename of an ini file.
* @param HTMLPurifier_ConfigSchema Schema object
* @return Configured HTMLPurifier_Config object
*/
function create($config) {
if (is_a($config, 'HTMLPurifier_Config')) {
public static function create($config, $schema = null) {
if ($config instanceof HTMLPurifier_Config) {
// pass-through
return $config;
}
$ret = HTMLPurifier_Config::createDefault();
if (!$schema) {
$ret = HTMLPurifier_Config::createDefault();
} else {
$ret = new HTMLPurifier_Config($schema);
}
if (is_string($config)) $ret->loadIni($config);
elseif (is_array($config)) $ret->loadArray($config);
return $ret;
@@ -112,11 +104,10 @@ class HTMLPurifier_Config
/**
* Convenience constructor that creates a default configuration object.
* @static
* @return Default HTMLPurifier_Config object.
*/
function createDefault() {
$definition =& HTMLPurifier_ConfigSchema::instance();
public static function createDefault() {
$definition = HTMLPurifier_ConfigSchema::instance();
$config = new HTMLPurifier_Config($definition);
return $config;
}
@@ -126,7 +117,7 @@ class HTMLPurifier_Config
* @param $namespace String namespace
* @param $key String key
*/
function get($namespace, $key, $from_alias = false) {
public function get($namespace, $key) {
if (!$this->finalized && $this->autoFinalize) $this->finalize();
if (!isset($this->def->info[$namespace][$key])) {
// can't add % due to SimpleTest bug
@@ -147,7 +138,7 @@ class HTMLPurifier_Config
* Retreives an array of directives to values from a given namespace
* @param $namespace String namespace
*/
function getBatch($namespace) {
public function getBatch($namespace) {
if (!$this->finalized && $this->autoFinalize) $this->finalize();
if (!isset($this->def->info[$namespace])) {
trigger_error('Cannot retrieve undefined namespace ' . htmlspecialchars($namespace),
@@ -164,7 +155,7 @@ class HTMLPurifier_Config
* before processing!
* @param $namespace Namespace to get serial for
*/
function getBatchSerial($namespace) {
public function getBatchSerial($namespace) {
if (empty($this->serials[$namespace])) {
$batch = $this->getBatch($namespace);
unset($batch['DefinitionRev']);
@@ -177,7 +168,7 @@ class HTMLPurifier_Config
* Returns a md5 signature for the entire configuration object
* that uniquely identifies that particular configuration
*/
function getSerial() {
public function getSerial() {
if (empty($this->serial)) {
$this->serial = md5(serialize($this->getAll()));
}
@@ -187,7 +178,7 @@ class HTMLPurifier_Config
/**
* Retrieves all directives, organized by namespace
*/
function getAll() {
public function getAll() {
if (!$this->finalized && $this->autoFinalize) $this->finalize();
return $this->conf;
}
@@ -198,7 +189,7 @@ class HTMLPurifier_Config
* @param $key String key
* @param $value Mixed value
*/
function set($namespace, $key, $value, $from_alias = false) {
public function set($namespace, $key, $value, $from_alias = false) {
if ($this->isFinalized('Cannot set directive after finalization')) return;
if (!isset($this->def->info[$namespace][$key])) {
trigger_error('Cannot set undefined directive ' . htmlspecialchars("$namespace.$key") . ' to value',
@@ -208,18 +199,25 @@ class HTMLPurifier_Config
if ($this->def->info[$namespace][$key]->class == 'alias') {
if ($from_alias) {
trigger_error('Double-aliases not allowed, please fix '.
'ConfigSchema bug with' . "$namespace.$key");
'ConfigSchema bug with' . "$namespace.$key", E_USER_ERROR);
return;
}
$this->set($this->def->info[$namespace][$key]->namespace,
$this->def->info[$namespace][$key]->name,
$this->set($new_ns = $this->def->info[$namespace][$key]->namespace,
$new_dir = $this->def->info[$namespace][$key]->name,
$value, true);
trigger_error("$namespace.$key is an alias, preferred directive name is $new_ns.$new_dir", E_USER_NOTICE);
return;
}
try {
$value = $this->parser->parse(
$value,
$type = $this->def->info[$namespace][$key]->type,
$this->def->info[$namespace][$key]->allow_null
);
} catch (HTMLPurifier_VarParserException $e) {
trigger_error('Value for ' . "$namespace.$key" . ' is of invalid type, should be ' . $type, E_USER_WARNING);
return;
}
$value = $this->def->validate(
$value,
$type = $this->def->info[$namespace][$key]->type,
$this->def->info[$namespace][$key]->allow_null
);
if (is_string($value)) {
// resolve value alias if defined
if (isset($this->def->info[$namespace][$key]->aliases[$value])) {
@@ -234,10 +232,6 @@ class HTMLPurifier_Config
}
}
}
if ($this->def->isError($value)) {
trigger_error('Value for ' . "$namespace.$key" . ' is of invalid type, should be ' . $type, E_USER_WARNING);
return;
}
$this->conf[$namespace][$key] = $value;
// reset definitions if the directives they depend on changed
@@ -252,30 +246,29 @@ class HTMLPurifier_Config
/**
* Convenience function for error reporting
* @private
*/
function _listify($lookup) {
private function _listify($lookup) {
$list = array();
foreach ($lookup as $name => $b) $list[] = $name;
return implode(', ', $list);
}
/**
* Retrieves reference to the HTML definition.
* Retrieves object reference to the HTML definition.
* @param $raw Return a copy that has not been setup yet. Must be
* called before it's been setup, otherwise won't work.
*/
function &getHTMLDefinition($raw = false) {
$def =& $this->getDefinition('HTML', $raw);
return $def; // prevent PHP 4.4.0 from complaining
public function getHTMLDefinition($raw = false) {
return $this->getDefinition('HTML', $raw);
}
/**
* Retrieves reference to the CSS definition
* Retrieves object reference to the CSS definition
* @param $raw Return a copy that has not been setup yet. Must be
* called before it's been setup, otherwise won't work.
*/
function &getCSSDefinition($raw = false) {
$def =& $this->getDefinition('CSS', $raw);
return $def;
public function getCSSDefinition($raw = false) {
return $this->getDefinition('CSS', $raw);
}
/**
@@ -283,7 +276,7 @@ class HTMLPurifier_Config
* @param $type Type of definition: HTML, CSS, etc
* @param $raw Whether or not definition should be returned raw
*/
function &getDefinition($type, $raw = false) {
public function getDefinition($type, $raw = false) {
if (!$this->finalized && $this->autoFinalize) $this->finalize();
$factory = HTMLPurifier_DefinitionCacheFactory::instance();
$cache = $factory->create($type, $this);
@@ -317,17 +310,13 @@ class HTMLPurifier_Config
} elseif ($type == 'URI') {
$this->definitions[$type] = new HTMLPurifier_URIDefinition();
} else {
trigger_error("Definition of $type type not supported");
$false = false;
return $false;
throw new HTMLPurifier_Exception("Definition of $type type not supported");
}
// quick abort if raw
if ($raw) {
if (is_null($this->get($type, 'DefinitionID'))) {
// fatally error out if definition ID not set
trigger_error("Cannot retrieve raw version without specifying %$type.DefinitionID", E_USER_ERROR);
$false = new HTMLPurifier_Error();
return $false;
throw new HTMLPurifier_Exception("Cannot retrieve raw version without specifying %$type.DefinitionID");
}
return $this->definitions[$type];
}
@@ -343,7 +332,7 @@ class HTMLPurifier_Config
* Namespace.Directive => Value
* @param $config_array Configuration associative array
*/
function loadArray($config_array) {
public function loadArray($config_array) {
if ($this->isFinalized('Cannot load directives after finalization')) return;
foreach ($config_array as $key => $value) {
$key = str_replace('_', '.', $key);
@@ -366,10 +355,11 @@ class HTMLPurifier_Config
* that are allowed in a web-form context as per an allowed
* namespaces/directives list.
* @param $allowed List of allowed namespaces/directives
* @static
*/
function getAllowedDirectivesForForm($allowed) {
$schema = HTMLPurifier_ConfigSchema::instance();
public static function getAllowedDirectivesForForm($allowed, $schema = null) {
if (!$schema) {
$schema = HTMLPurifier_ConfigSchema::instance();
}
if ($allowed !== true) {
if (is_string($allowed)) $allowed = array($allowed);
$allowed_ns = array();
@@ -411,11 +401,11 @@ class HTMLPurifier_Config
* @param $index Index/name that the config variables are in
* @param $allowed List of allowed namespaces/directives
* @param $mq_fix Boolean whether or not to enable magic quotes fix
* @static
* @param $schema Instance of HTMLPurifier_ConfigSchema to use, if not global copy
*/
function loadArrayFromForm($array, $index, $allowed = true, $mq_fix = true) {
$ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix);
$config = HTMLPurifier_Config::create($ret);
public static function loadArrayFromForm($array, $index = false, $allowed = true, $mq_fix = true, $schema = null) {
$ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix, $schema);
$config = HTMLPurifier_Config::create($ret, $schema);
return $config;
}
@@ -423,21 +413,20 @@ class HTMLPurifier_Config
* Merges in configuration values from $_GET/$_POST to object. NOT STATIC.
* @note Same parameters as loadArrayFromForm
*/
function mergeArrayFromForm($array, $index, $allowed = true, $mq_fix = true) {
$ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix);
public function mergeArrayFromForm($array, $index = false, $allowed = true, $mq_fix = true) {
$ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix, $this->def);
$this->loadArray($ret);
}
/**
* Prepares an array from a form into something usable for the more
* strict parts of HTMLPurifier_Config
* @static
*/
function prepareArrayFromForm($array, $index, $allowed = true, $mq_fix = true) {
$array = (isset($array[$index]) && is_array($array[$index])) ? $array[$index] : array();
$mq = get_magic_quotes_gpc() && $mq_fix;
public static function prepareArrayFromForm($array, $index = false, $allowed = true, $mq_fix = true, $schema = null) {
if ($index !== false) $array = (isset($array[$index]) && is_array($array[$index])) ? $array[$index] : array();
$mq = $mq_fix && function_exists('get_magic_quotes_gpc') && get_magic_quotes_gpc();
$allowed = HTMLPurifier_Config::getAllowedDirectivesForForm($allowed);
$allowed = HTMLPurifier_Config::getAllowedDirectivesForForm($allowed, $schema);
$ret = array();
foreach ($allowed as $key) {
list($ns, $directive) = $key;
@@ -457,7 +446,7 @@ class HTMLPurifier_Config
* Loads configuration values from an ini file
* @param $filename Name of ini file
*/
function loadIni($filename) {
public function loadIni($filename) {
if ($this->isFinalized('Cannot load directives after finalization')) return;
$array = parse_ini_file($filename, true);
$this->loadArray($array);
@@ -467,7 +456,7 @@ class HTMLPurifier_Config
* Checks whether or not the configuration object is finalized.
* @param $error String error message, or false for no error
*/
function isFinalized($error = false) {
public function isFinalized($error = false) {
if ($this->finalized && $error) {
trigger_error($error, E_USER_ERROR);
}
@@ -478,17 +467,18 @@ class HTMLPurifier_Config
* Finalizes configuration only if auto finalize is on and not
* already finalized
*/
function autoFinalize() {
public function autoFinalize() {
if (!$this->finalized && $this->autoFinalize) $this->finalize();
}
/**
* Finalizes a configuration object, prohibiting further change
*/
function finalize() {
public function finalize() {
$this->finalized = true;
}
}
+2 -2
View File
@@ -3,7 +3,7 @@
/**
* Base class for configuration entity
*/
class HTMLPurifier_ConfigDef {
var $class = false;
abstract class HTMLPurifier_ConfigDef {
public $class = false;
}
@@ -1,7 +1,5 @@
<?php
require_once 'HTMLPurifier/ConfigDef.php';
/**
* Structure object containing definition of a directive.
* @note This structure does not contain default values
@@ -9,20 +7,18 @@ require_once 'HTMLPurifier/ConfigDef.php';
class HTMLPurifier_ConfigDef_Directive extends HTMLPurifier_ConfigDef
{
var $class = 'directive';
public $class = 'directive';
function HTMLPurifier_ConfigDef_Directive(
public function __construct(
$type = null,
$descriptions = null,
$allow_null = null,
$allowed = null,
$aliases = null
) {
if ( $type !== null) $this->type = $type;
if ($descriptions !== null) $this->descriptions = $descriptions;
if ( $allow_null !== null) $this->allow_null = $allow_null;
if ( $allowed !== null) $this->allowed = $allowed;
if ( $aliases !== null) $this->aliases = $aliases;
if ( $type !== null) $this->type = $type;
if ( $allow_null !== null) $this->allow_null = $allow_null;
if ( $allowed !== null) $this->allowed = $allowed;
if ( $aliases !== null) $this->aliases = $aliases;
}
/**
@@ -37,43 +33,23 @@ class HTMLPurifier_ConfigDef_Directive extends HTMLPurifier_ConfigDef
* - hash (array of key => value)
* - mixed (anything goes)
*/
var $type = 'mixed';
/**
* Plaintext descriptions of the configuration entity is. Organized by
* file and line number, so multiple descriptions are allowed.
*/
var $descriptions = array();
public $type = 'mixed';
/**
* Is null allowed? Has no effect for mixed type.
* @bool
*/
var $allow_null = false;
public $allow_null = false;
/**
* Lookup table of allowed values of the element, bool true if all allowed.
*/
var $allowed = true;
public $allowed = true;
/**
* Hash of value aliases, i.e. values that are equivalent.
*/
var $aliases = array();
/**
* Advisory list of directive aliases, i.e. other directives that
* redirect here
*/
var $directiveAliases = array();
/**
* Adds a description to the array
*/
function addDescription($file, $line, $description) {
if (!isset($this->descriptions[$file])) $this->descriptions[$file] = array();
$this->descriptions[$file][$line] = $description;
}
public $aliases = array();
}
@@ -1,24 +1,22 @@
<?php
require_once 'HTMLPurifier/ConfigDef.php';
/**
* Structure object describing a directive alias
*/
class HTMLPurifier_ConfigDef_DirectiveAlias extends HTMLPurifier_ConfigDef
{
var $class = 'alias';
public $class = 'alias';
/**
* Namespace being aliased to
*/
var $namespace;
public $namespace;
/**
* Directive being aliased to
*/
var $name;
public $name;
function HTMLPurifier_ConfigDef_DirectiveAlias($namespace, $name) {
public function __construct($namespace, $name) {
$this->namespace = $namespace;
$this->name = $name;
}
@@ -1,22 +1,10 @@
<?php
require_once 'HTMLPurifier/ConfigDef.php';
/**
* Structure object describing of a namespace
*/
class HTMLPurifier_ConfigDef_Namespace extends HTMLPurifier_ConfigDef {
function HTMLPurifier_ConfigDef_Namespace($description = null) {
$this->description = $description;
}
var $class = 'namespace';
/**
* String description of what kinds of directives go in this namespace.
*/
var $description;
class HTMLPurifier_ConfigDef_Namespace extends HTMLPurifier_ConfigDef
{
public $class = 'namespace';
}
+98 -351
View File
@@ -1,31 +1,7 @@
<?php
require_once 'HTMLPurifier/Error.php';
require_once 'HTMLPurifier/ConfigDef.php';
require_once 'HTMLPurifier/ConfigDef/Namespace.php';
require_once 'HTMLPurifier/ConfigDef/Directive.php';
require_once 'HTMLPurifier/ConfigDef/DirectiveAlias.php';
if (!defined('HTMLPURIFIER_SCHEMA_STRICT')) define('HTMLPURIFIER_SCHEMA_STRICT', false);
/**
* Configuration definition, defines directives and their defaults.
* @note If you update this, please update Printer_ConfigForm
* @todo The ability to define things multiple times is confusing and should
* be factored out to its own function named registerDependency() or
* addNote(), where only the namespace.name and an extra descriptions
* documenting the nature of the dependency are needed. Since it's
* possible that the dependency is registered before the configuration
* is defined, deferring it to some sort of cache until it actually
* gets defined would be wise, keeping it opaque until it does get
* defined. We could add a finalize() method which would cause it to
* error out if we get a dangling dependency. It's difficult, however,
* to know whether or not it's a dependency, or a codependency, that is
* neither of them fully depends on it. Where does the configuration go
* then? This could be partially resolved by allowing blanket definitions
* and then splitting them up into finer-grained versions, however, there
* might be implementation difficulties in ini files regarding order of
* execution.
*/
class HTMLPurifier_ConfigSchema {
@@ -33,172 +9,75 @@ class HTMLPurifier_ConfigSchema {
* Defaults of the directives and namespaces.
* @note This shares the exact same structure as HTMLPurifier_Config::$conf
*/
var $defaults = array();
public $defaults = array();
/**
* Definition of the directives.
*/
var $info = array();
public $info = array();
/**
* Definition of namespaces.
* Application-wide singleton
*/
var $info_namespace = array();
static protected $singleton;
/**
* Lookup table of allowed types.
* Variable parser.
*/
var $types = array(
'string' => 'String',
'istring' => 'Case-insensitive string',
'text' => 'Text',
'itext' => 'Case-insensitive text',
'int' => 'Integer',
'float' => 'Float',
'bool' => 'Boolean',
'lookup' => 'Lookup array',
'list' => 'Array list',
'hash' => 'Associative array',
'mixed' => 'Mixed'
);
protected $parser;
public function __construct() {
$this->parser = new HTMLPurifier_VarParser_Flexible();
}
/**
* Initializes the default namespaces.
* Unserializes the default ConfigSchema.
*/
function initialize() {
$this->defineNamespace('Core', 'Core features that are always available.');
$this->defineNamespace('Attr', 'Features regarding attribute validation.');
$this->defineNamespace('URI', 'Features regarding Uniform Resource Identifiers.');
$this->defineNamespace('HTML', 'Configuration regarding allowed HTML.');
$this->defineNamespace('CSS', 'Configuration regarding allowed CSS.');
$this->defineNamespace('AutoFormat', 'Configuration for activating auto-formatting functionality (also known as <code>Injector</code>s)');
$this->defineNamespace('AutoFormatParam', 'Configuration for customizing auto-formatting functionality');
$this->defineNamespace('Output', 'Configuration relating to the generation of (X)HTML.');
$this->defineNamespace('Cache', 'Configuration for DefinitionCache and related subclasses.');
$this->defineNamespace('Test', 'Developer testing configuration for our unit tests.');
public static function makeFromSerial() {
return unserialize(file_get_contents(HTMLPURIFIER_PREFIX . '/HTMLPurifier/ConfigSchema/schema.ser'));
}
/**
* Retrieves an instance of the application-wide configuration definition.
* @static
*/
function &instance($prototype = null) {
static $instance;
public static function instance($prototype = null) {
if ($prototype !== null) {
$instance = $prototype;
} elseif ($instance === null || $prototype === true) {
$instance = new HTMLPurifier_ConfigSchema();
$instance->initialize();
HTMLPurifier_ConfigSchema::$singleton = $prototype;
} elseif (HTMLPurifier_ConfigSchema::$singleton === null || $prototype === true) {
HTMLPurifier_ConfigSchema::$singleton = HTMLPurifier_ConfigSchema::makeFromSerial();
}
return $instance;
return HTMLPurifier_ConfigSchema::$singleton;
}
/**
* Defines a directive for configuration
* @static
* @warning Will fail of directive's namespace is defined
* @warning Will fail of directive's namespace is defined.
* @warning This method's signature is slightly different from the legacy
* define() static method! Beware!
* @param $namespace Namespace the directive is in
* @param $name Key of directive
* @param $default Default value of directive
* @param $type Allowed type of the directive. See
* HTMLPurifier_DirectiveDef::$type for allowed values
* @param $description Description of directive for documentation
* @param $allow_null Whether or not to allow null values
*/
function define($namespace, $name, $default, $type, $description) {
$def =& HTMLPurifier_ConfigSchema::instance();
// basic sanity checks
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!isset($def->info[$namespace])) {
trigger_error('Cannot define directive for undefined namespace',
E_USER_ERROR);
return;
}
if (!ctype_alnum($name)) {
trigger_error('Directive name must be alphanumeric',
E_USER_ERROR);
return;
}
if (empty($description)) {
trigger_error('Description must be non-empty',
E_USER_ERROR);
return;
}
}
if (isset($def->info[$namespace][$name])) {
// already defined
if (
$def->info[$namespace][$name]->type !== $type ||
$def->defaults[$namespace][$name] !== $default
) {
trigger_error('Inconsistent default or type, cannot redefine');
return;
}
} else {
// needs defining
// process modifiers (OPTIMIZE!)
$type_values = explode('/', $type, 2);
$type = $type_values[0];
$modifier = isset($type_values[1]) ? $type_values[1] : false;
$allow_null = ($modifier === 'null');
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!isset($def->types[$type])) {
trigger_error('Invalid type for configuration directive',
E_USER_ERROR);
return;
}
$default = $def->validate($default, $type, $allow_null);
if ($def->isError($default)) {
trigger_error('Default value does not match directive type',
E_USER_ERROR);
return;
}
}
$def->info[$namespace][$name] =
new HTMLPurifier_ConfigDef_Directive();
$def->info[$namespace][$name]->type = $type;
$def->info[$namespace][$name]->allow_null = $allow_null;
$def->defaults[$namespace][$name] = $default;
}
if (!HTMLPURIFIER_SCHEMA_STRICT) return;
$backtrace = debug_backtrace();
$file = $def->mungeFilename($backtrace[0]['file']);
$line = $backtrace[0]['line'];
$def->info[$namespace][$name]->addDescription($file,$line,$description);
public function add($namespace, $name, $default, $type, $allow_null) {
$default = $this->parser->parse($default, $type, $allow_null);
$this->info[$namespace][$name] = new HTMLPurifier_ConfigDef_Directive();
$this->info[$namespace][$name]->type = $type;
$this->info[$namespace][$name]->allow_null = $allow_null;
$this->defaults[$namespace][$name] = $default;
}
/**
* Defines a namespace for directives to be put into.
* @static
* @warning This is slightly different from the corresponding static
* method.
* @param $namespace Namespace's name
* @param $description Description of the namespace
*/
function defineNamespace($namespace, $description) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (isset($def->info[$namespace])) {
trigger_error('Cannot redefine namespace', E_USER_ERROR);
return;
}
if (!ctype_alnum($namespace)) {
trigger_error('Namespace name must be alphanumeric',
E_USER_ERROR);
return;
}
if (empty($description)) {
trigger_error('Description must be non-empty',
E_USER_ERROR);
return;
}
}
$def->info[$namespace] = array();
$def->info_namespace[$namespace] = new HTMLPurifier_ConfigDef_Namespace();
$def->info_namespace[$namespace]->description = $description;
$def->defaults[$namespace] = array();
public function addNamespace($namespace) {
$this->info[$namespace] = array();
$this->defaults[$namespace] = array();
}
/**
@@ -206,231 +85,99 @@ class HTMLPurifier_ConfigSchema {
*
* Directive value aliases are convenient for developers because it lets
* them set a directive to several values and get the same result.
* @static
* @param $namespace Directive's namespace
* @param $name Name of Directive
* @param $alias Name of aliased value
* @param $real Value aliased value will be converted into
* @param $aliases Hash of aliased values to the real alias
*/
function defineValueAliases($namespace, $name, $aliases) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (HTMLPURIFIER_SCHEMA_STRICT && !isset($def->info[$namespace][$name])) {
trigger_error('Cannot set value alias for non-existant directive',
E_USER_ERROR);
return;
}
public function addValueAliases($namespace, $name, $aliases) {
foreach ($aliases as $alias => $real) {
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!$def->info[$namespace][$name] !== true &&
!isset($def->info[$namespace][$name]->allowed[$real])
) {
trigger_error('Cannot define alias to value that is not allowed',
E_USER_ERROR);
return;
}
if (isset($def->info[$namespace][$name]->allowed[$alias])) {
trigger_error('Cannot define alias over allowed value',
E_USER_ERROR);
return;
}
}
$def->info[$namespace][$name]->aliases[$alias] = $real;
$this->info[$namespace][$name]->aliases[$alias] = $real;
}
}
/**
* Defines a set of allowed values for a directive.
* @static
* @warning This is slightly different from the corresponding static
* method definition.
* @param $namespace Namespace of directive
* @param $name Name of directive
* @param $allowed_values Arraylist of allowed values
* @param $allowed Lookup array of allowed values
*/
function defineAllowedValues($namespace, $name, $allowed_values) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (HTMLPURIFIER_SCHEMA_STRICT && !isset($def->info[$namespace][$name])) {
trigger_error('Cannot define allowed values for undefined directive',
E_USER_ERROR);
return;
}
$directive =& $def->info[$namespace][$name];
$type = $directive->type;
if (HTMLPURIFIER_SCHEMA_STRICT && $type != 'string' && $type != 'istring') {
trigger_error('Cannot define allowed values for directive whose type is not string',
E_USER_ERROR);
return;
}
if ($directive->allowed === true) {
$directive->allowed = array();
}
foreach ($allowed_values as $value) {
$directive->allowed[$value] = true;
}
if (
HTMLPURIFIER_SCHEMA_STRICT &&
$def->defaults[$namespace][$name] !== null &&
!isset($directive->allowed[$def->defaults[$namespace][$name]])
) {
trigger_error('Default value must be in allowed range of variables',
E_USER_ERROR);
$directive->allowed = true; // undo undo!
return;
}
public function addAllowedValues($namespace, $name, $allowed) {
$type = $this->info[$namespace][$name]->type;
$this->info[$namespace][$name]->allowed = $allowed;
}
/**
* Defines a directive alias for backwards compatibility
* @static
* @param $namespace
* @param $name Directive that will be aliased
* @param $new_namespace
* @param $new_name Directive that the alias will be to
*/
function defineAlias($namespace, $name, $new_namespace, $new_name) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!isset($def->info[$namespace])) {
trigger_error('Cannot define directive alias in undefined namespace',
E_USER_ERROR);
return;
}
if (!ctype_alnum($name)) {
trigger_error('Directive name must be alphanumeric',
E_USER_ERROR);
return;
}
if (isset($def->info[$namespace][$name])) {
trigger_error('Cannot define alias over directive',
E_USER_ERROR);
return;
}
if (!isset($def->info[$new_namespace][$new_name])) {
trigger_error('Cannot define alias to undefined directive',
E_USER_ERROR);
return;
}
if ($def->info[$new_namespace][$new_name]->class == 'alias') {
trigger_error('Cannot define alias to alias',
E_USER_ERROR);
return;
}
public function addAlias($namespace, $name, $new_namespace, $new_name) {
$this->info[$namespace][$name] = new HTMLPurifier_ConfigDef_DirectiveAlias($new_namespace, $new_name);
}
// DEPRECATED METHODS
/** @see HTMLPurifier_ConfigSchema->set() */
public static function define($namespace, $name, $default, $type, $description) {
HTMLPurifier_ConfigSchema::deprecated(__METHOD__);
// process modifiers (OPTIMIZE!)
$type_values = explode('/', $type, 2);
$type = $type_values[0];
$modifier = isset($type_values[1]) ? $type_values[1] : false;
$allow_null = ($modifier === 'null');
$def = HTMLPurifier_ConfigSchema::instance();
$def->add($namespace, $name, $default, $type, $allow_null);
}
/** @see HTMLPurifier_ConfigSchema->addNamespace() */
public static function defineNamespace($namespace, $description) {
HTMLPurifier_ConfigSchema::deprecated(__METHOD__);
$def = HTMLPurifier_ConfigSchema::instance();
$def->addNamespace($namespace);
}
/** @see HTMLPurifier_ConfigSchema->addValueAliases() */
public static function defineValueAliases($namespace, $name, $aliases) {
HTMLPurifier_ConfigSchema::deprecated(__METHOD__);
$def = HTMLPurifier_ConfigSchema::instance();
$def->addValueAliases($namespace, $name, $aliases);
}
/** @see HTMLPurifier_ConfigSchema->addAllowedValues() */
public static function defineAllowedValues($namespace, $name, $allowed_values) {
HTMLPurifier_ConfigSchema::deprecated(__METHOD__);
$allowed = array();
foreach ($allowed_values as $value) {
$allowed[$value] = true;
}
$def->info[$namespace][$name] =
new HTMLPurifier_ConfigDef_DirectiveAlias(
$new_namespace, $new_name);
$def->info[$new_namespace][$new_name]->directiveAliases[] = "$namespace.$name";
$def = HTMLPurifier_ConfigSchema::instance();
$def->addAllowedValues($namespace, $name, $allowed);
}
/** @see HTMLPurifier_ConfigSchema->addAlias() */
public static function defineAlias($namespace, $name, $new_namespace, $new_name) {
HTMLPurifier_ConfigSchema::deprecated(__METHOD__);
$def = HTMLPurifier_ConfigSchema::instance();
$def->addAlias($namespace, $name, $new_namespace, $new_name);
}
/** @deprecated, use HTMLPurifier_VarParser->parse() */
public function validate($a, $b, $c = false) {
trigger_error("HTMLPurifier_ConfigSchema->validate deprecated, use HTMLPurifier_VarParser->parse instead", E_USER_NOTICE);
return $this->parser->parse($a, $b, $c);
}
/**
* Validate a variable according to type. Return null if invalid.
* Throws an E_USER_NOTICE stating that a method is deprecated.
*/
function validate($var, $type, $allow_null = false) {
if (!isset($this->types[$type])) {
trigger_error('Invalid type', E_USER_ERROR);
return;
}
if ($allow_null && $var === null) return null;
switch ($type) {
case 'mixed':
//if (is_string($var)) $var = unserialize($var);
return $var;
case 'istring':
case 'string':
case 'text': // no difference, just is longer/multiple line string
case 'itext':
if (!is_string($var)) break;
if ($type === 'istring' || $type === 'itext') $var = strtolower($var);
return $var;
case 'int':
if (is_string($var) && ctype_digit($var)) $var = (int) $var;
elseif (!is_int($var)) break;
return $var;
case 'float':
if (is_string($var) && is_numeric($var)) $var = (float) $var;
elseif (!is_float($var)) break;
return $var;
case 'bool':
if (is_int($var) && ($var === 0 || $var === 1)) {
$var = (bool) $var;
} elseif (is_string($var)) {
if ($var == 'on' || $var == 'true' || $var == '1') {
$var = true;
} elseif ($var == 'off' || $var == 'false' || $var == '0') {
$var = false;
} else {
break;
}
} elseif (!is_bool($var)) break;
return $var;
case 'list':
case 'hash':
case 'lookup':
if (is_string($var)) {
// special case: technically, this is an array with
// a single empty string item, but having an empty
// array is more intuitive
if ($var == '') return array();
if (strpos($var, "\n") === false && strpos($var, "\r") === false) {
// simplistic string to array method that only works
// for simple lists of tag names or alphanumeric characters
$var = explode(',',$var);
} else {
$var = preg_split('/(,|[\n\r]+)/', $var);
}
// remove spaces
foreach ($var as $i => $j) $var[$i] = trim($j);
if ($type === 'hash') {
// key:value,key2:value2
$nvar = array();
foreach ($var as $keypair) {
$c = explode(':', $keypair, 2);
if (!isset($c[1])) continue;
$nvar[$c[0]] = $c[1];
}
$var = $nvar;
}
}
if (!is_array($var)) break;
$keys = array_keys($var);
if ($keys === array_keys($keys)) {
if ($type == 'list') return $var;
elseif ($type == 'lookup') {
$new = array();
foreach ($var as $key) {
$new[$key] = true;
}
return $new;
} else break;
}
if ($type === 'lookup') {
foreach ($var as $key => $value) {
$var[$key] = true;
}
}
return $var;
}
$error = new HTMLPurifier_Error();
return $error;
private static function deprecated($method) {
trigger_error("Static HTMLPurifier_ConfigSchema::$method deprecated, use add*() method instead", E_USER_NOTICE);
}
/**
* Takes an absolute path and munges it into a more manageable relative path
*/
function mungeFilename($filename) {
if (!HTMLPURIFIER_SCHEMA_STRICT) return $filename;
$offset = strrpos($filename, 'HTMLPurifier');
$filename = substr($filename, $offset);
$filename = str_replace('\\', '/', $filename);
return $filename;
}
/**
* Checks if var is an HTMLPurifier_Error object
*/
function isError($var) {
if (!is_object($var)) return false;
if (!is_a($var, 'HTMLPurifier_Error')) return false;
return true;
}
}
@@ -0,0 +1,49 @@
<?php
/**
* Converts HTMLPurifier_ConfigSchema_Interchange to our runtime
* representation used to perform checks on user configuration.
*/
class HTMLPurifier_ConfigSchema_Builder_ConfigSchema
{
public function build($interchange) {
$schema = new HTMLPurifier_ConfigSchema();
foreach ($interchange->namespaces as $n) {
$schema->addNamespace($n->namespace);
}
foreach ($interchange->directives as $d) {
$schema->add(
$d->id->namespace,
$d->id->directive,
$d->default,
$d->type,
$d->typeAllowsNull
);
if ($d->allowed !== null) {
$schema->addAllowedValues(
$d->id->namespace,
$d->id->directive,
$d->allowed
);
}
foreach ($d->aliases as $alias) {
$schema->addAlias(
$alias->namespace,
$alias->directive,
$d->id->namespace,
$d->id->directive
);
}
if ($d->valueAliases !== null) {
$schema->addValueAliases(
$d->id->namespace,
$d->id->directive,
$d->valueAliases
);
}
}
return $schema;
}
}
@@ -0,0 +1,106 @@
<?php
/**
* Converts HTMLPurifier_ConfigSchema_Interchange to an XML format,
* which can be further processed to generate documentation.
*/
class HTMLPurifier_ConfigSchema_Builder_Xml extends XMLWriter
{
protected $interchange;
protected function writeHTMLDiv($html) {
$this->startElement('div');
$purifier = HTMLPurifier::getInstance();
$html = $purifier->purify($html);
$this->writeAttribute('xmlns', 'http://www.w3.org/1999/xhtml');
$this->writeRaw($html);
$this->endElement(); // div
}
protected function export($var) {
if ($var === array()) return 'array()';
return var_export($var, true);
}
public function build($interchange) {
// global access, only use as last resort
$this->interchange = $interchange;
$this->setIndent(true);
$this->startDocument('1.0', 'UTF-8');
$this->startElement('configdoc');
$this->writeElement('title', $interchange->name);
foreach ($interchange->namespaces as $namespace) {
$this->buildNamespace($namespace);
}
$this->endElement(); // configdoc
$this->flush();
}
public function buildNamespace($namespace) {
$this->startElement('namespace');
$this->writeAttribute('id', $namespace->namespace);
$this->writeElement('name', $namespace->namespace);
$this->startElement('description');
$this->writeHTMLDiv($namespace->description);
$this->endElement(); // description
foreach ($this->interchange->directives as $directive) {
if ($directive->id->namespace !== $namespace->namespace) continue;
$this->buildDirective($directive);
}
$this->endElement(); // namespace
}
public function buildDirective($directive) {
$this->startElement('directive');
$this->writeAttribute('id', $directive->id->toString());
$this->writeElement('name', $directive->id->directive);
$this->startElement('aliases');
foreach ($directive->aliases as $alias) $this->writeElement('alias', $alias->toString());
$this->endElement(); // aliases
$this->startElement('constraints');
if ($directive->version) $this->writeElement('version', $directive->version);
$this->startElement('type');
if ($directive->typeAllowsNull) $this->writeAttribute('allow-null', 'yes');
$this->text($directive->type);
$this->endElement(); // type
if ($directive->allowed) {
$this->startElement('allowed');
foreach ($directive->allowed as $value => $x) $this->writeElement('value', $value);
$this->endElement(); // allowed
}
$this->writeElement('default', $this->export($directive->default));
$this->writeAttribute('xml:space', 'preserve');
if ($directive->external) {
$this->startElement('external');
foreach ($directive->external as $project) $this->writeElement('project', $project);
$this->endElement();
}
$this->endElement(); // constraints
if ($directive->deprecatedVersion) {
$this->startElement('deprecated');
$this->writeElement('version', $directive->deprecatedVersion);
$this->writeElement('use', $directive->deprecatedUse->toString());
$this->endElement(); // deprecated
}
$this->startElement('description');
$this->writeHTMLDiv($directive->description);
$this->endElement(); // description
$this->endElement(); // directive
}
}
@@ -0,0 +1,9 @@
<?php
/**
* Exceptions related to configuration schema
*/
class HTMLPurifier_ConfigSchema_Exception extends HTMLPurifier_Exception
{
}
@@ -0,0 +1,55 @@
<?php
/**
* Generic schema interchange format that can be converted to a runtime
* representation (HTMLPurifier_ConfigSchema) or HTML documentation. Members
* are completely validated.
*/
class HTMLPurifier_ConfigSchema_Interchange
{
/**
* Name of the application this schema is describing.
*/
public $name;
/**
* Array of Namespace ID => array(namespace info)
*/
public $namespaces = array();
/**
* Array of Directive ID => array(directive info)
*/
public $directives = array();
/**
* Adds a namespace array to $namespaces
*/
public function addNamespace($namespace) {
if (isset($this->namespaces[$i = $namespace->namespace])) {
throw new HTMLPurifier_ConfigSchema_Exception("Cannot redefine namespace '$i'");
}
$this->namespaces[$i] = $namespace;
}
/**
* Adds a directive array to $directives
*/
public function addDirective($directive) {
if (isset($this->directives[$i = $directive->id->toString()])) {
throw new HTMLPurifier_ConfigSchema_Exception("Cannot redefine directive '$i'");
}
$this->directives[$i] = $directive;
}
/**
* Convenience function to perform standard validation. Throws exception
* on failed validation.
*/
public function validate() {
$validator = new HTMLPurifier_ConfigSchema_Validator();
return $validator->validate($this);
}
}
@@ -0,0 +1,75 @@
<?php
/**
* Interchange component class describing configuration directives.
*/
class HTMLPurifier_ConfigSchema_Interchange_Directive
{
/**
* ID of directive, instance of HTMLPurifier_ConfigSchema_Interchange_Id.
*/
public $id;
/**
* String type, e.g. 'integer' or 'istring'.
*/
public $type;
/**
* Default value, e.g. 3 or 'DefaultVal'.
*/
public $default;
/**
* HTML description.
*/
public $description;
/**
* Boolean whether or not null is allowed as a value.
*/
public $typeAllowsNull = false;
/**
* Lookup table of allowed scalar values, e.g. array('allowed' => true).
* Null if all values are allowed.
*/
public $allowed;
/**
* List of aliases for the directive,
* e.g. array(new HTMLPurifier_ConfigSchema_Interchange_Id('Ns', 'Dir'))).
*/
public $aliases = array();
/**
* Hash of value aliases, e.g. array('alt' => 'real'). Null if value
* aliasing is disabled (necessary for non-scalar types).
*/
public $valueAliases;
/**
* Version of HTML Purifier the directive was introduced, e.g. '1.3.1'.
* Null if the directive has always existed.
*/
public $version;
/**
* ID of directive that supercedes this old directive, is an instance
* of HTMLPurifier_ConfigSchema_Interchange_Id. Null if not deprecated.
*/
public $deprecatedUse;
/**
* Version of HTML Purifier this directive was deprecated. Null if not
* deprecated.
*/
public $deprecatedVersion;
/**
* List of external projects this directive depends on, e.g. array('CSSTidy').
*/
public $external = array();
}
@@ -0,0 +1,29 @@
<?php
/**
* Represents a directive ID in the interchange format.
*/
class HTMLPurifier_ConfigSchema_Interchange_Id
{
public $namespace, $directive;
public function __construct($namespace, $directive) {
$this->namespace = $namespace;
$this->directive = $directive;
}
/**
* @warning This is NOT magic, to ensure that people don't abuse SPL and
* cause problems for PHP 5.0 support.
*/
public function toString() {
return $this->namespace . '.' . $this->directive;
}
public static function make($id) {
list($namespace, $directive) = explode('.', $id);
return new HTMLPurifier_ConfigSchema_Interchange_Id($namespace, $directive);
}
}
@@ -0,0 +1,19 @@
<?php
/**
* Interchange component class describing namespaces.
*/
class HTMLPurifier_ConfigSchema_Interchange_Namespace
{
/**
* Name of namespace defined.
*/
public $namespace;
/**
* HTML description.
*/
public $description;
}
@@ -0,0 +1,175 @@
<?php
class HTMLPurifier_ConfigSchema_InterchangeBuilder
{
/**
* Used for processing DEFAULT, nothing else.
*/
protected $varParser;
public function __construct($varParser = null) {
$this->varParser = $varParser ? $varParser : new HTMLPurifier_VarParser_Native();
}
public static function buildFromDirectory($dir = null) {
$parser = new HTMLPurifier_StringHashParser();
$builder = new HTMLPurifier_ConfigSchema_InterchangeBuilder();
$interchange = new HTMLPurifier_ConfigSchema_Interchange();
if (!$dir) $dir = HTMLPURIFIER_PREFIX . '/HTMLPurifier/ConfigSchema/schema/';
$info = parse_ini_file($dir . 'info.ini');
$interchange->name = $info['name'];
$files = array();
$dh = opendir($dir);
while (false !== ($file = readdir($dh))) {
if (!$file || $file[0] == '.' || strrchr($file, '.') !== '.txt') {
continue;
}
$files[] = $file;
}
closedir($dh);
sort($files);
foreach ($files as $file) {
$builder->build(
$interchange,
new HTMLPurifier_StringHash( $parser->parseFile($dir . $file) )
);
}
return $interchange;
}
/**
* Builds an interchange object based on a hash.
* @param $interchange HTMLPurifier_ConfigSchema_Interchange object to build
* @param $hash HTMLPurifier_ConfigSchema_StringHash source data
*/
public function build($interchange, $hash) {
if (!$hash instanceof HTMLPurifier_StringHash) {
$hash = new HTMLPurifier_StringHash($hash);
}
if (!isset($hash['ID'])) {
throw new HTMLPurifier_ConfigSchema_Exception('Hash does not have any ID');
}
if (strpos($hash['ID'], '.') === false) {
$this->buildNamespace($interchange, $hash);
} else {
$this->buildDirective($interchange, $hash);
}
$this->_findUnused($hash);
}
public function buildNamespace($interchange, $hash) {
$namespace = new HTMLPurifier_ConfigSchema_Interchange_Namespace();
$namespace->namespace = $hash->offsetGet('ID');
if (isset($hash['DESCRIPTION'])) {
$namespace->description = $hash->offsetGet('DESCRIPTION');
}
$interchange->addNamespace($namespace);
}
public function buildDirective($interchange, $hash) {
$directive = new HTMLPurifier_ConfigSchema_Interchange_Directive();
// These are required elements:
$directive->id = $this->id($hash->offsetGet('ID'));
$id = $directive->id->toString(); // convenience
if (isset($hash['TYPE'])) {
$type = explode('/', $hash->offsetGet('TYPE'));
if (isset($type[1])) $directive->typeAllowsNull = true;
$directive->type = $type[0];
} else {
throw new HTMLPurifier_ConfigSchema_Exception("TYPE in directive hash '$id' not defined");
}
if (isset($hash['DEFAULT'])) {
try {
$directive->default = $this->varParser->parse($hash->offsetGet('DEFAULT'), $directive->type, $directive->typeAllowsNull);
} catch (HTMLPurifier_VarParserException $e) {
throw new HTMLPurifier_ConfigSchema_Exception($e->getMessage() . " in DEFAULT in directive hash '$id'");
}
}
if (isset($hash['DESCRIPTION'])) {
$directive->description = $hash->offsetGet('DESCRIPTION');
}
if (isset($hash['ALLOWED'])) {
$directive->allowed = $this->lookup($this->evalArray($hash->offsetGet('ALLOWED')));
}
if (isset($hash['VALUE-ALIASES'])) {
$directive->valueAliases = $this->evalArray($hash->offsetGet('VALUE-ALIASES'));
}
if (isset($hash['ALIASES'])) {
$raw_aliases = trim($hash->offsetGet('ALIASES'));
$aliases = preg_split('/\s*,\s*/', $raw_aliases);
foreach ($aliases as $alias) {
$directive->aliases[] = $this->id($alias);
}
}
if (isset($hash['VERSION'])) {
$directive->version = $hash->offsetGet('VERSION');
}
if (isset($hash['DEPRECATED-USE'])) {
$directive->deprecatedUse = $this->id($hash->offsetGet('DEPRECATED-USE'));
}
if (isset($hash['DEPRECATED-VERSION'])) {
$directive->deprecatedVersion = $hash->offsetGet('DEPRECATED-VERSION');
}
if (isset($hash['EXTERNAL'])) {
$directive->external = preg_split('/\s*,\s*/', trim($hash->offsetGet('EXTERNAL')));
}
$interchange->addDirective($directive);
}
/**
* Evaluates an array PHP code string without array() wrapper
*/
protected function evalArray($contents) {
return eval('return array('. $contents .');');
}
/**
* Converts an array list into a lookup array.
*/
protected function lookup($array) {
$ret = array();
foreach ($array as $val) $ret[$val] = true;
return $ret;
}
/**
* Convenience function that creates an HTMLPurifier_ConfigSchema_Interchange_Id
* object based on a string Id.
*/
protected function id($id) {
return HTMLPurifier_ConfigSchema_Interchange_Id::make($id);
}
/**
* Triggers errors for any unused keys passed in the hash; such keys
* may indicate typos, missing values, etc.
* @param $hash Instance of ConfigSchema_StringHash to check.
*/
protected function _findUnused($hash) {
$accessed = $hash->getAccessed();
foreach ($hash as $k => $v) {
if (!isset($accessed[$k])) {
trigger_error("String hash key '$k' not used by builder", E_USER_NOTICE);
}
}
}
}
@@ -0,0 +1,222 @@
<?php
/**
* Performs validations on HTMLPurifier_ConfigSchema_Interchange
*
* @note If you see '// handled by InterchangeBuilder', that means a
* design decision in that class would prevent this validation from
* ever being necessary. We have them anyway, however, for
* redundancy.
*/
class HTMLPurifier_ConfigSchema_Validator
{
/**
* Easy to access global objects.
*/
protected $interchange, $aliases;
/**
* Context-stack to provide easy to read error messages.
*/
protected $context = array();
/**
* HTMLPurifier_VarParser to test default's type.
*/
protected $parser;
public function __construct() {
$this->parser = new HTMLPurifier_VarParser();
}
/**
* Validates a fully-formed interchange object. Throws an
* HTMLPurifier_ConfigSchema_Exception if there's a problem.
*/
public function validate($interchange) {
$this->interchange = $interchange;
$this->aliases = array();
// PHP is a bit lax with integer <=> string conversions in
// arrays, so we don't use the identical !== comparison
foreach ($interchange->namespaces as $i => $namespace) {
if ($i != $namespace->namespace) $this->error(false, "Integrity violation: key '$i' does not match internal id '{$namespace->namespace}'");
$this->validateNamespace($namespace);
}
foreach ($interchange->directives as $i => $directive) {
$id = $directive->id->toString();
if ($i != $id) $this->error(false, "Integrity violation: key '$i' does not match internal id '$id'");
$this->validateDirective($directive);
}
return true;
}
/**
* Validates a HTMLPurifier_ConfigSchema_Interchange_Namespace object.
*/
public function validateNamespace($n) {
$this->context[] = "namespace '{$n->namespace}'";
$this->with($n, 'namespace')
->assertNotEmpty()
->assertAlnum(); // implicit assertIsString handled by InterchangeBuilder
$this->with($n, 'description')
->assertNotEmpty()
->assertIsString(); // handled by InterchangeBuilder
array_pop($this->context);
}
/**
* Validates a HTMLPurifier_ConfigSchema_Interchange_Id object.
*/
public function validateId($id) {
$id_string = $id->toString();
$this->context[] = "id '$id_string'";
if (!$id instanceof HTMLPurifier_ConfigSchema_Interchange_Id) {
// handled by InterchangeBuilder
$this->error(false, 'is not an instance of HTMLPurifier_ConfigSchema_Interchange_Id');
}
if (!isset($this->interchange->namespaces[$id->namespace])) {
$this->error('namespace', 'does not exist'); // assumes that the namespace was validated already
}
$this->with($id, 'directive')
->assertNotEmpty()
->assertAlnum(); // implicit assertIsString handled by InterchangeBuilder
array_pop($this->context);
}
/**
* Validates a HTMLPurifier_ConfigSchema_Interchange_Directive object.
*/
public function validateDirective($d) {
$id = $d->id->toString();
$this->context[] = "directive '$id'";
$this->validateId($d->id);
$this->with($d, 'description')
->assertNotEmpty();
// BEGIN - handled by InterchangeBuilder
$this->with($d, 'type')
->assertNotEmpty();
$this->with($d, 'typeAllowsNull')
->assertIsBool();
try {
// This also tests validity of $d->type
$this->parser->parse($d->default, $d->type, $d->typeAllowsNull);
} catch (HTMLPurifier_VarParserException $e) {
$this->error('default', 'had error: ' . $e->getMessage());
}
// END - handled by InterchangeBuilder
if (!is_null($d->allowed) || !empty($d->valueAliases)) {
// allowed and valueAliases require that we be dealing with
// strings, so check for that early.
if (!isset(HTMLPurifier_VarParser::$stringTypes[$d->type])) {
$this->error('type', 'must be a string type when used with allowed or value aliases');
}
}
$this->validateDirectiveAllowed($d);
$this->validateDirectiveValueAliases($d);
$this->validateDirectiveAliases($d);
array_pop($this->context);
}
/**
* Extra validation if $allowed member variable of
* HTMLPurifier_ConfigSchema_Interchange_Directive is defined.
*/
public function validateDirectiveAllowed($d) {
if (is_null($d->allowed)) return;
$this->with($d, 'allowed')
->assertNotEmpty()
->assertIsLookup(); // handled by InterchangeBuilder
if (is_string($d->default) && !isset($d->allowed[$d->default])) {
$this->error('default', 'must be an allowed value');
}
$this->context[] = 'allowed';
foreach ($d->allowed as $val => $x) {
if (!is_string($val)) $this->error("value $val", 'must be a string');
}
array_pop($this->context);
}
/**
* Extra validation if $valueAliases member variable of
* HTMLPurifier_ConfigSchema_Interchange_Directive is defined.
*/
public function validateDirectiveValueAliases($d) {
if (is_null($d->valueAliases)) return;
$this->with($d, 'valueAliases')
->assertIsArray(); // handled by InterchangeBuilder
$this->context[] = 'valueAliases';
foreach ($d->valueAliases as $alias => $real) {
if (!is_string($alias)) $this->error("alias $alias", 'must be a string');
if (!is_string($real)) $this->error("alias target $real from alias '$alias'", 'must be a string');
if ($alias === $real) {
$this->error("alias '$alias'", "must not be an alias to itself");
}
}
if (!is_null($d->allowed)) {
foreach ($d->valueAliases as $alias => $real) {
if (isset($d->allowed[$alias])) {
$this->error("alias '$alias'", 'must not be an allowed value');
} elseif (!isset($d->allowed[$real])) {
$this->error("alias '$alias'", 'must be an alias to an allowed value');
}
}
}
array_pop($this->context);
}
/**
* Extra validation if $aliases member variable of
* HTMLPurifier_ConfigSchema_Interchange_Directive is defined.
*/
public function validateDirectiveAliases($d) {
$this->with($d, 'aliases')
->assertIsArray(); // handled by InterchangeBuilder
$this->context[] = 'aliases';
foreach ($d->aliases as $alias) {
$this->validateId($alias);
$s = $alias->toString();
if (isset($this->interchange->directives[$s])) {
$this->error("alias '$s'", 'collides with another directive');
}
if (isset($this->aliases[$s])) {
$other_directive = $this->aliases[$s];
$this->error("alias '$s'", "collides with alias for directive '$other_directive'");
}
$this->aliases[$s] = $d->id->toString();
}
array_pop($this->context);
}
// protected helper functions
/**
* Convenience function for generating HTMLPurifier_ConfigSchema_ValidatorAtom
* for validating simple member variables of objects.
*/
protected function with($obj, $member) {
return new HTMLPurifier_ConfigSchema_ValidatorAtom($this->getFormattedContext(), $obj, $member);
}
/**
* Emits an error, providing helpful context.
*/
protected function error($target, $msg) {
if ($target !== false) $prefix = ucfirst($target) . ' in ' . $this->getFormattedContext();
else $prefix = ucfirst($this->getFormattedContext());
throw new HTMLPurifier_ConfigSchema_Exception(trim($prefix . ' ' . $msg));
}
/**
* Returns a formatted context string.
*/
protected function getFormattedContext() {
return implode(' in ', array_reverse($this->context));
}
}
@@ -0,0 +1,66 @@
<?php
/**
* Fluent interface for validating the contents of member variables.
* This should be immutable. See HTMLPurifier_ConfigSchema_Validator for
* use-cases. We name this an 'atom' because it's ONLY for validations that
* are independent and usually scalar.
*/
class HTMLPurifier_ConfigSchema_ValidatorAtom
{
protected $context, $obj, $member, $contents;
public function __construct($context, $obj, $member) {
$this->context = $context;
$this->obj = $obj;
$this->member = $member;
$this->contents =& $obj->$member;
}
public function assertIsString() {
if (!is_string($this->contents)) $this->error('must be a string');
return $this;
}
public function assertIsBool() {
if (!is_bool($this->contents)) $this->error('must be a boolean');
return $this;
}
public function assertIsArray() {
if (!is_array($this->contents)) $this->error('must be an array');
return $this;
}
public function assertNotNull() {
if ($this->contents === null) $this->error('must not be null');
return $this;
}
public function assertAlnum() {
$this->assertIsString();
if (!ctype_alnum($this->contents)) $this->error('must be alphanumeric');
return $this;
}
public function assertNotEmpty() {
if (empty($this->contents)) $this->error('must not be empty');
return $this;
}
public function assertIsLookup() {
$this->assertIsArray();
foreach ($this->contents as $v) {
if ($v !== true) $this->error('must be a lookup array');
}
return $this;
}
protected function error($msg) {
throw new HTMLPurifier_ConfigSchema_Exception(ucfirst($this->member) . ' in ' . $this->context . ' ' . $msg);
}
}
File diff suppressed because one or more lines are too long
@@ -0,0 +1,11 @@
Attr.AllowedFrameTargets
TYPE: lookup
DEFAULT: array()
--DESCRIPTION--
Lookup table of all allowed link frame targets. Some commonly used link
targets include _blank, _self, _parent and _top. Values should be
lowercase, as validation will be done in a case-sensitive manner despite
W3C's recommendation. XHTML 1.0 Strict does not permit the target attribute
so this directive will have no effect in that doctype. XHTML 1.1 does not
enable the Target module by default, you will have to manually enable it
(see the module documentation for more details.)
@@ -0,0 +1,8 @@
Attr.AllowedRel
TYPE: lookup
VERSION: 1.6.0
DEFAULT: array()
--DESCRIPTION--
List of allowed forward document relationships in the rel attribute. Common
values may be nofollow or print. By default, this is empty, meaning that no
document relationships are allowed.
@@ -0,0 +1,8 @@
Attr.AllowedRev
TYPE: lookup
VERSION: 1.6.0
DEFAULT: array()
--DESCRIPTION--
List of allowed reverse document relationships in the rev attribute. This
attribute is a bit of an edge-case; if you don't know what it is for, stay
away.
@@ -0,0 +1,8 @@
Attr.DefaultInvalidImage
TYPE: string
DEFAULT: ''
--DESCRIPTION--
This is the default image an img tag will be pointed to if it does not have
a valid src attribute. In future versions, we may allow the image tag to
be removed completely, but due to design issues, this is not possible right
now.
@@ -0,0 +1,7 @@
Attr.DefaultInvalidImageAlt
TYPE: string
DEFAULT: 'Invalid image'
--DESCRIPTION--
This is the content of the alt tag of an invalid image if the user had not
previously specified an alt attribute. It has no effect when the image is
valid but there was no alt attribute present.
@@ -0,0 +1,9 @@
Attr.DefaultTextDir
TYPE: string
DEFAULT: 'ltr'
--DESCRIPTION--
Defines the default text direction (ltr or rtl) of the document being
parsed. This generally is the same as the value of the dir attribute in
HTML, or ltr if that is not specified.
--ALLOWED--
'ltr', 'rtl'
@@ -0,0 +1,15 @@
Attr.EnableID
TYPE: bool
DEFAULT: false
VERSION: 1.2.0
--DESCRIPTION--
Allows the ID attribute in HTML. This is disabled by default due to the
fact that without proper configuration user input can easily break the
validation of a webpage by specifying an ID that is already on the
surrounding HTML. If you don't mind throwing caution to the wind, enable
this directive, but I strongly recommend you also consider blacklisting IDs
you use (%Attr.IDBlacklist) or prefixing all user supplied IDs
(%Attr.IDPrefix). When set to true HTML Purifier reverts to the behavior of
pre-1.2.0 versions.
--ALIASES--
HTML.EnableAttrID
@@ -0,0 +1,4 @@
Attr.IDBlacklist
TYPE: list
DEFAULT: array()
DESCRIPTION: Array of IDs not allowed in the document.
@@ -0,0 +1,8 @@
Attr.IDBlacklistRegexp
TYPE: string/null
VERSION: 1.6.0
DEFAULT: NULL
--DESCRIPTION--
PCRE regular expression to be matched against all IDs. If the expression is
matches, the ID is rejected. Use this with care: may cause significant
degradation. ID matching is done after all other validation.
@@ -0,0 +1,11 @@
Attr.IDPrefix
TYPE: string
VERSION: 1.2.0
DEFAULT: ''
--DESCRIPTION--
String to prefix to IDs. If you have no idea what IDs your pages may use,
you may opt to simply add a prefix to all user-submitted ID attributes so
that they are still usable, but will not conflict with core page IDs.
Example: setting the directive to 'user_' will result in a user submitted
'foo' to become 'user_foo' Be sure to set %HTML.EnableAttrID to true
before using this.
@@ -0,0 +1,13 @@
Attr.IDPrefixLocal
TYPE: string
VERSION: 1.2.0
DEFAULT: ''
--DESCRIPTION--
Temporary prefix for IDs used in conjunction with %Attr.IDPrefix. If you
need to allow multiple sets of user content on web page, you may need to
have a seperate prefix that changes with each iteration. This way,
seperately submitted user content displayed on the same page doesn't
clobber each other. Ideal values are unique identifiers for the content it
represents (i.e. the id of the row in the database). Be sure to add a
seperator (like an underscore) at the end. Warning: this directive will
not work unless %Attr.IDPrefix is set to a non-empty value!
@@ -0,0 +1,2 @@
Attr
DESCRIPTION: Features regarding attribute validation.
@@ -0,0 +1,30 @@
AutoFormat.AutoParagraph
TYPE: bool
VERSION: 2.0.1
DEFAULT: false
--DESCRIPTION--
<p>
This directive turns on auto-paragraphing, where double newlines are
converted in to paragraphs whenever possible. Auto-paragraphing:
</p>
<ul>
<li>Always applies to inline elements or text in the root node,</li>
<li>Applies to inline elements or text with double newlines in nodes
that allow paragraph tags,</li>
<li>Applies to double newlines in paragraph tags</li>
</ul>
<p>
<code>p</code> tags must be allowed for this directive to take effect.
We do not use <code>br</code> tags for paragraphing, as that is
semantically incorrect.
</p>
<p>
To prevent auto-paragraphing as a content-producer, refrain from using
double-newlines except to specify a new paragraph or in contexts where
it has special meaning (whitespace usually has no meaning except in
tags like <code>pre</code>, so this should not be difficult.) To prevent
the paragraphing of inline text adjacent to block elements, wrap them
in <code>div</code> tags (the behavior is slightly different outside of
the root node.)
</p>
@@ -0,0 +1,12 @@
AutoFormat.Custom
TYPE: list
VERSION: 2.0.1
DEFAULT: array()
--DESCRIPTION--
<p>
This directive can be used to add custom auto-format injectors.
Specify an array of injector names (class name minus the prefix)
or concrete implementations. Injector class must exist.
</p>
@@ -0,0 +1,12 @@
AutoFormat.Linkify
TYPE: bool
VERSION: 2.0.1
DEFAULT: false
--DESCRIPTION--
<p>
This directive turns on linkification, auto-linking http, ftp and
https URLs. <code>a</code> tags with the <code>href</code> attribute
must be allowed.
</p>

Some files were not shown because too many files have changed in this diff Show More