Updating HTMLPurifier to 4.9.3

Source: https://htmlpurifier.org/download
Release date: 2017-06-19
This commit is contained in:
Mikael Nordfeldth 2017-07-10 13:46:07 +02:00
parent 3158f9c33a
commit a4a6a8469e
54 changed files with 919 additions and 212 deletions

View File

@ -7,7 +7,7 @@
* primary concern and you are using an opcode cache. PLEASE DO NOT EDIT THIS
* FILE, changes will be overwritten the next time the script is run.
*
* @version 4.7.0
* @version 4.9.3
*
* @warning
* You must *not* include any other HTML Purifier files before this file,
@ -137,6 +137,8 @@ require 'HTMLPurifier/AttrTransform/SafeObject.php';
require 'HTMLPurifier/AttrTransform/SafeParam.php';
require 'HTMLPurifier/AttrTransform/ScriptRequired.php';
require 'HTMLPurifier/AttrTransform/TargetBlank.php';
require 'HTMLPurifier/AttrTransform/TargetNoopener.php';
require 'HTMLPurifier/AttrTransform/TargetNoreferrer.php';
require 'HTMLPurifier/AttrTransform/Textarea.php';
require 'HTMLPurifier/ChildDef/Chameleon.php';
require 'HTMLPurifier/ChildDef/Custom.php';
@ -175,6 +177,8 @@ require 'HTMLPurifier/HTMLModule/StyleAttribute.php';
require 'HTMLPurifier/HTMLModule/Tables.php';
require 'HTMLPurifier/HTMLModule/Target.php';
require 'HTMLPurifier/HTMLModule/TargetBlank.php';
require 'HTMLPurifier/HTMLModule/TargetNoopener.php';
require 'HTMLPurifier/HTMLModule/TargetNoreferrer.php';
require 'HTMLPurifier/HTMLModule/Text.php';
require 'HTMLPurifier/HTMLModule/Tidy.php';
require 'HTMLPurifier/HTMLModule/XMLCommonAttributes.php';
@ -225,5 +229,6 @@ require 'HTMLPurifier/URIScheme/https.php';
require 'HTMLPurifier/URIScheme/mailto.php';
require 'HTMLPurifier/URIScheme/news.php';
require 'HTMLPurifier/URIScheme/nntp.php';
require 'HTMLPurifier/URIScheme/tel.php';
require 'HTMLPurifier/VarParser/Flexible.php';
require 'HTMLPurifier/VarParser/Native.php';

View File

@ -19,7 +19,7 @@
*/
/*
HTML Purifier 4.7.0 - Standards Compliant HTML Filtering
HTML Purifier 4.9.3 - Standards Compliant HTML Filtering
Copyright (C) 2006-2008 Edward Z. Yang
This library is free software; you can redistribute it and/or
@ -58,12 +58,12 @@ class HTMLPurifier
* Version of HTML Purifier.
* @type string
*/
public $version = '4.7.0';
public $version = '4.9.3';
/**
* Constant with version of HTML Purifier.
*/
const VERSION = '4.7.0';
const VERSION = '4.9.3';
/**
* Global configuration object.
@ -104,7 +104,7 @@ class HTMLPurifier
/**
* Initializes the purifier.
*
* @param HTMLPurifier_Config $config Optional HTMLPurifier_Config object
* @param HTMLPurifier_Config|mixed $config Optional HTMLPurifier_Config object
* for all instances of the purifier, if omitted, a default
* configuration is supplied (which can be overridden on a
* per-use basis).

View File

@ -131,6 +131,8 @@ require_once $__dir . '/HTMLPurifier/AttrTransform/SafeObject.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/SafeParam.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/ScriptRequired.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/TargetBlank.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/TargetNoopener.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/TargetNoreferrer.php';
require_once $__dir . '/HTMLPurifier/AttrTransform/Textarea.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Chameleon.php';
require_once $__dir . '/HTMLPurifier/ChildDef/Custom.php';
@ -169,6 +171,8 @@ 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/TargetBlank.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/TargetNoopener.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/TargetNoreferrer.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Text.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/Tidy.php';
require_once $__dir . '/HTMLPurifier/HTMLModule/XMLCommonAttributes.php';
@ -219,5 +223,6 @@ 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/URIScheme/tel.php';
require_once $__dir . '/HTMLPurifier/VarParser/Flexible.php';
require_once $__dir . '/HTMLPurifier/VarParser/Native.php';

View File

@ -19,8 +19,8 @@ class HTMLPurifier_Arborize
if ($token instanceof HTMLPurifier_Token_End) {
$token->start = null; // [MUT]
$r = array_pop($stack);
assert($r->name === $token->name);
assert(empty($token->attr));
//assert($r->name === $token->name);
//assert(empty($token->attr));
$r->endCol = $token->col;
$r->endLine = $token->line;
$r->endArmor = $token->armor;
@ -32,7 +32,7 @@ class HTMLPurifier_Arborize
$stack[] = $node;
}
}
assert(count($stack) == 1);
//assert(count($stack) == 1);
return $stack[0];
}

View File

@ -21,6 +21,11 @@ class HTMLPurifier_AttrCollections
* @param HTMLPurifier_HTMLModule[] $modules Hash array of HTMLPurifier_HTMLModule members
*/
public function __construct($attr_types, $modules)
{
$this->doConstruct($attr_types, $modules);
}
public function doConstruct($attr_types, $modules)
{
// load extensions from the modules
foreach ($modules as $module) {

View File

@ -86,7 +86,13 @@ abstract class HTMLPurifier_AttrDef
*/
protected function mungeRgb($string)
{
return preg_replace('/rgb\((\d+)\s*,\s*(\d+)\s*,\s*(\d+)\)/', 'rgb(\1,\2,\3)', $string);
$p = '\s*(\d+(\.\d+)?([%]?))\s*';
if (preg_match('/(rgba|hsla)\(/', $string)) {
return preg_replace('/(rgba|hsla)\('.$p.','.$p.','.$p.','.$p.'\)/', '\1(\2,\5,\8,\11)', $string);
}
return preg_replace('/(rgb|hsl)\('.$p.','.$p.','.$p.'\)/', '\1(\2,\5,\8)', $string);
}
/**

View File

@ -25,15 +25,42 @@ class HTMLPurifier_AttrDef_CSS extends HTMLPurifier_AttrDef
$css = $this->parseCDATA($css);
$definition = $config->getCSSDefinition();
$allow_duplicates = $config->get("CSS.AllowDuplicates");
// we're going to break the spec and explode by semicolons.
// This is because semicolon rarely appears in escaped form
// Doing this is generally flaky but fast
// IT MIGHT APPEAR IN URIs, see HTMLPurifier_AttrDef_CSSURI
// for details
$declarations = explode(';', $css);
// According to the CSS2.1 spec, the places where a
// non-delimiting semicolon can appear are in strings
// escape sequences. So here is some dumb hack to
// handle quotes.
$len = strlen($css);
$accum = "";
$declarations = array();
$quoted = false;
for ($i = 0; $i < $len; $i++) {
$c = strcspn($css, ";'\"", $i);
$accum .= substr($css, $i, $c);
$i += $c;
if ($i == $len) break;
$d = $css[$i];
if ($quoted) {
$accum .= $d;
if ($d == $quoted) {
$quoted = false;
}
} else {
if ($d == ";") {
$declarations[] = $accum;
$accum = "";
} else {
$accum .= $d;
$quoted = $d;
}
}
}
if ($accum != "") $declarations[] = $accum;
$propvalues = array();
$new_declarations = '';
/**
* Name of the current CSS property being validated.
@ -83,8 +110,12 @@ class HTMLPurifier_AttrDef_CSS extends HTMLPurifier_AttrDef
if ($result === false) {
continue;
}
if ($allow_duplicates) {
$new_declarations .= "$property:$result;";
} else {
$propvalues[$property] = $result;
}
}
$context->destroy('CurrentCSSProperty');
@ -92,7 +123,6 @@ class HTMLPurifier_AttrDef_CSS extends HTMLPurifier_AttrDef
// slightly inefficient, but it's the only way of getting rid of
// duplicates. Perhaps config to optimize it, but not now.
$new_declarations = '';
foreach ($propvalues as $prop => $value) {
$new_declarations .= "$prop:$value;";
}

View File

@ -6,6 +6,16 @@
class HTMLPurifier_AttrDef_CSS_Color extends HTMLPurifier_AttrDef
{
/**
* @type HTMLPurifier_AttrDef_CSS_AlphaValue
*/
protected $alpha;
public function __construct()
{
$this->alpha = new HTMLPurifier_AttrDef_CSS_AlphaValue();
}
/**
* @param string $color
* @param HTMLPurifier_Config $config
@ -29,59 +39,104 @@ class HTMLPurifier_AttrDef_CSS_Color extends HTMLPurifier_AttrDef
return $colors[$lower];
}
if (strpos($color, 'rgb(') !== false) {
// rgb literal handling
if (preg_match('#(rgb|rgba|hsl|hsla)\(#', $color, $matches) === 1) {
$length = strlen($color);
if (strpos($color, ')') !== $length - 1) {
return false;
}
$triad = substr($color, 4, $length - 4 - 1);
$parts = explode(',', $triad);
if (count($parts) !== 3) {
// get used function : rgb, rgba, hsl or hsla
$function = $matches[1];
$parameters_size = 3;
$alpha_channel = false;
if (substr($function, -1) === 'a') {
$parameters_size = 4;
$alpha_channel = true;
}
/*
* Allowed types for values :
* parameter_position => [type => max_value]
*/
$allowed_types = array(
1 => array('percentage' => 100, 'integer' => 255),
2 => array('percentage' => 100, 'integer' => 255),
3 => array('percentage' => 100, 'integer' => 255),
);
$allow_different_types = false;
if (strpos($function, 'hsl') !== false) {
$allowed_types = array(
1 => array('integer' => 360),
2 => array('percentage' => 100),
3 => array('percentage' => 100),
);
$allow_different_types = true;
}
$values = trim(str_replace($function, '', $color), ' ()');
$parts = explode(',', $values);
if (count($parts) !== $parameters_size) {
return false;
}
$type = false; // to ensure that they're all the same type
$type = false;
$new_parts = array();
$i = 0;
foreach ($parts as $part) {
$i++;
$part = trim($part);
if ($part === '') {
return false;
}
$length = strlen($part);
if ($part[$length - 1] === '%') {
// handle percents
if (!$type) {
$type = 'percentage';
} elseif ($type !== 'percentage') {
// different check for alpha channel
if ($alpha_channel === true && $i === count($parts)) {
$result = $this->alpha->validate($part, $config, $context);
if ($result === false) {
return false;
}
$num = (float)substr($part, 0, $length - 1);
if ($num < 0) {
$num = 0;
$new_parts[] = (string)$result;
continue;
}
if ($num > 100) {
$num = 100;
}
$new_parts[] = "$num%";
if (substr($part, -1) === '%') {
$current_type = 'percentage';
} else {
// handle integers
if (!$type) {
$type = 'integer';
} elseif ($type !== 'integer') {
$current_type = 'integer';
}
if (!array_key_exists($current_type, $allowed_types[$i])) {
return false;
}
$num = (int)$part;
if ($num < 0) {
$num = 0;
if (!$type) {
$type = $current_type;
}
if ($num > 255) {
$num = 255;
if ($allow_different_types === false && $type != $current_type) {
return false;
}
$new_parts[] = (string)$num;
$max_value = $allowed_types[$i][$current_type];
if ($current_type == 'integer') {
// Return value between range 0 -> $max_value
$new_parts[] = (int)max(min($part, $max_value), 0);
} elseif ($current_type == 'percentage') {
$new_parts[] = (float)max(min(rtrim($part, '%'), $max_value), 0) . '%';
}
}
$new_triad = implode(',', $new_parts);
$color = "rgb($new_triad)";
$new_values = implode(',', $new_parts);
$color = $function . '(' . $new_values . ')';
} else {
// hexadecimal handling
if ($color[0] === '#') {
@ -100,6 +155,7 @@ class HTMLPurifier_AttrDef_CSS_Color extends HTMLPurifier_AttrDef
}
return $color;
}
}
// vim: et sw=4 sts=4

View File

@ -130,6 +130,8 @@ class HTMLPurifier_AttrDef_CSS_FontFamily extends HTMLPurifier_AttrDef
// <http://ja.wikipedia.org/wiki/MS_明朝>. See
// the CSS3 spec for more examples:
// <http://www.w3.org/TR/2011/WD-css3-fonts-20110324/localizedfamilynames.png>
// You can see live samples of these on the Internet:
// <http://www.google.co.jp/search?q=font-family++明朝|ゴシック>
// However, most of these fonts have ASCII equivalents:
// for example, 'MS Mincho', and it's considered
// professional to use ASCII font names instead of

View File

@ -33,6 +33,9 @@ class HTMLPurifier_AttrDef_CSS_URI extends HTMLPurifier_AttrDef_URI
return false;
}
$uri_string = substr($uri_string, 4);
if (strlen($uri_string) == 0) {
return false;
}
$new_length = strlen($uri_string) - 1;
if ($uri_string[$new_length] != ')') {
return false;

View File

@ -72,8 +72,13 @@ class HTMLPurifier_AttrDef_HTML_ID extends HTMLPurifier_AttrDef
// we purposely avoid using regex, hopefully this is faster
if ($config->get('Attr.ID.HTML5') === true) {
if (preg_match('/[\t\n\x0b\x0c ]/', $id)) {
return false;
}
} else {
if (ctype_alpha($id)) {
$result = true;
// OK
} else {
if (!ctype_alpha(@$id[0])) {
return false;
@ -83,7 +88,10 @@ class HTMLPurifier_AttrDef_HTML_ID extends HTMLPurifier_AttrDef
$id,
'A..Za..z0..9:-._'
);
$result = ($trim === '');
if ($trim !== '') {
return false;
}
}
}
$regexp = $config->get('Attr.IDBlacklistRegexp');
@ -91,14 +99,14 @@ class HTMLPurifier_AttrDef_HTML_ID extends HTMLPurifier_AttrDef
return false;
}
if (!$this->selector && $result) {
if (!$this->selector) {
$id_accumulator->add($id);
}
// if no change was made to the ID, return the result
// else, return the new id if stripping whitespace made it
// valid, or return false.
return $result ? $id : false;
return $id;
}
}

View File

@ -76,24 +76,33 @@ class HTMLPurifier_AttrDef_URI_Host extends HTMLPurifier_AttrDef
// fairly well supported.
$underscore = $config->get('Core.AllowHostnameUnderscore') ? '_' : '';
// Based off of RFC 1738, but amended so that
// as per RFC 3696, the top label need only not be all numeric.
// The productions describing this are:
$a = '[a-z]'; // alpha
$an = '[a-z0-9]'; // alphanum
$and = "[a-z0-9-$underscore]"; // alphanum | "-"
// domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum
$domainlabel = "$an($and*$an)?";
// toplabel = alpha | alpha *( alphanum | "-" ) alphanum
$toplabel = "$a($and*$an)?";
$domainlabel = "$an(?:$and*$an)?";
// AMENDED as per RFC 3696
// toplabel = alphanum | alphanum *( alphanum | "-" ) alphanum
// side condition: not all numeric
$toplabel = "$an(?:$and*$an)?";
// hostname = *( domainlabel "." ) toplabel [ "." ]
if (preg_match("/^($domainlabel\.)*$toplabel\.?$/i", $string)) {
if (preg_match("/^(?:$domainlabel\.)*($toplabel)\.?$/i", $string, $matches)) {
if (!ctype_digit($matches[1])) {
return $string;
}
}
// PHP 5.3 and later support this functionality natively
if (function_exists('idn_to_ascii')) {
$string = idn_to_ascii($string);
// If we have Net_IDNA2 support, we can support IRIs by
// punycoding them. (This is the most portable thing to do,
// since otherwise we have to assume browsers support
if ($config->get('Core.EnableIDNA')) {
} elseif ($config->get('Core.EnableIDNA')) {
$idna = new Net_IDNA2(array('encoding' => 'utf8', 'overlong' => false, 'strict' => true));
// we need to encode each period separately
$parts = explode('.', $string);
@ -114,13 +123,14 @@ class HTMLPurifier_AttrDef_URI_Host extends HTMLPurifier_AttrDef
}
}
$string = implode('.', $new_parts);
if (preg_match("/^($domainlabel\.)*$toplabel\.?$/i", $string)) {
return $string;
}
} catch (Exception $e) {
// XXX error reporting
}
}
// Try again
if (preg_match("/^($domainlabel\.)*$toplabel\.?$/i", $string)) {
return $string;
}
return false;
}
}

View File

@ -32,8 +32,7 @@ class HTMLPurifier_AttrTransform_ImgRequired extends HTMLPurifier_AttrTransform
if ($src) {
$alt = $config->get('Attr.DefaultImageAlt');
if ($alt === null) {
// truncate if the alt is too long
$attr['alt'] = substr(basename($attr['src']), 0, 40);
$attr['alt'] = basename($attr['src']);
} else {
$attr['alt'] = $alt;
}

View File

@ -0,0 +1,37 @@
<?php
// must be called POST validation
/**
* Adds rel="noopener" to any links which target a different window
* than the current one. This is used to prevent malicious websites
* from silently replacing the original window, which could be used
* to do phishing.
* This transform is controlled by %HTML.TargetNoopener.
*/
class HTMLPurifier_AttrTransform_TargetNoopener extends HTMLPurifier_AttrTransform
{
/**
* @param array $attr
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return array
*/
public function transform($attr, $config, $context)
{
if (isset($attr['rel'])) {
$rels = explode(' ', $attr['rel']);
} else {
$rels = array();
}
if (isset($attr['target']) && !in_array('noopener', $rels)) {
$rels[] = 'noopener';
}
if (!empty($rels) || isset($attr['rel'])) {
$attr['rel'] = implode(' ', $rels);
}
return $attr;
}
}

View File

@ -0,0 +1,37 @@
<?php
// must be called POST validation
/**
* Adds rel="noreferrer" to any links which target a different window
* than the current one. This is used to prevent malicious websites
* from silently replacing the original window, which could be used
* to do phishing.
* This transform is controlled by %HTML.TargetNoreferrer.
*/
class HTMLPurifier_AttrTransform_TargetNoreferrer extends HTMLPurifier_AttrTransform
{
/**
* @param array $attr
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return array
*/
public function transform($attr, $config, $context)
{
if (isset($attr['rel'])) {
$rels = explode(' ', $attr['rel']);
} else {
$rels = array();
}
if (isset($attr['target']) && !in_array('noreferrer', $rels)) {
$rels[] = 'noreferrer';
}
if (!empty($rels) || isset($attr['rel'])) {
$attr['rel'] = implode(' ', $rels);
}
return $attr;
}
}

View File

@ -225,6 +225,10 @@ class HTMLPurifier_CSSDefinition extends HTMLPurifier_Definition
);
$max = $config->get('CSS.MaxImgLength');
$this->info['min-width'] =
$this->info['max-width'] =
$this->info['min-height'] =
$this->info['max-height'] =
$this->info['width'] =
$this->info['height'] =
$max === null ?
@ -370,6 +374,19 @@ class HTMLPurifier_CSSDefinition extends HTMLPurifier_Definition
);
$this->info['page-break-inside'] = new HTMLPurifier_AttrDef_Enum(array('auto', 'avoid'));
$border_radius = new HTMLPurifier_AttrDef_CSS_Composite(
array(
new HTMLPurifier_AttrDef_CSS_Percentage(true), // disallow negative
new HTMLPurifier_AttrDef_CSS_Length('0') // disallow negative
));
$this->info['border-top-left-radius'] =
$this->info['border-top-right-radius'] =
$this->info['border-bottom-right-radius'] =
$this->info['border-bottom-left-radius'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_radius, 2);
// TODO: support SLASH syntax
$this->info['border-radius'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_radius, 4);
}
/**

View File

@ -38,13 +38,19 @@ class HTMLPurifier_ChildDef_List extends HTMLPurifier_ChildDef
return false;
}
// if li is not allowed, delete parent node
if (!isset($config->getHTMLDefinition()->info['li'])) {
trigger_error("Cannot allow ul/ol without allowing li", E_USER_WARNING);
return false;
}
// the new set of children
$result = array();
// a little sanity check to make sure it's not ALL whitespace
$all_whitespace = true;
$current_li = false;
$current_li = null;
foreach ($children as $node) {
if (!empty($node->is_whitespace)) {
@ -65,7 +71,7 @@ class HTMLPurifier_ChildDef_List extends HTMLPurifier_ChildDef
// to handle non-list elements; non-list elements should
// not be appended to an existing li; only li created
// for non-list. This distinction is not currently made.
if ($current_li === false) {
if ($current_li === null) {
$current_li = new HTMLPurifier_Node_Element('li');
$result[] = $current_li;
}

View File

@ -203,7 +203,7 @@ class HTMLPurifier_ChildDef_Table extends HTMLPurifier_ChildDef
$current_tr_tbody->children[] = $node;
break;
case '#PCDATA':
assert($node->is_whitespace);
//assert($node->is_whitespace);
if ($current_tr_tbody === null) {
$ret[] = $node;
} else {

View File

@ -21,7 +21,7 @@ class HTMLPurifier_Config
* HTML Purifier's version
* @type string
*/
public $version = '4.7.0';
public $version = '4.9.3';
/**
* Whether or not to automatically finalize
@ -333,7 +333,7 @@ class HTMLPurifier_Config
}
// Raw type might be negative when using the fully optimized form
// of stdclass, which indicates allow_null == true
// of stdClass, which indicates allow_null == true
$rtype = is_int($def) ? $def : $def->type;
if ($rtype < 0) {
$type = -$rtype;

View File

@ -24,11 +24,11 @@ class HTMLPurifier_ConfigSchema
*
* array(
* 'Namespace' => array(
* 'Directive' => new stdclass(),
* 'Directive' => new stdClass(),
* )
* )
*
* The stdclass may have the following properties:
* The stdClass may have the following properties:
*
* - If isAlias isn't set:
* - type: Integer type of directive, see HTMLPurifier_VarParser for definitions
@ -39,8 +39,8 @@ class HTMLPurifier_ConfigSchema
* - namespace: Namespace this directive aliases to
* - name: Directive name this directive aliases to
*
* In certain degenerate cases, stdclass will actually be an integer. In
* that case, the value is equivalent to an stdclass with the type
* In certain degenerate cases, stdClass will actually be an integer. In
* that case, the value is equivalent to an stdClass with the type
* property set to the integer. If the integer is negative, type is
* equal to the absolute value of integer, and allow_null is true.
*
@ -105,7 +105,7 @@ class HTMLPurifier_ConfigSchema
*/
public function add($key, $default, $type, $allow_null)
{
$obj = new stdclass();
$obj = new stdClass();
$obj->type = is_int($type) ? $type : HTMLPurifier_VarParser::$types[$type];
if ($allow_null) {
$obj->allow_null = true;
@ -152,14 +152,14 @@ class HTMLPurifier_ConfigSchema
*/
public function addAlias($key, $new_key)
{
$obj = new stdclass;
$obj = new stdClass;
$obj->key = $new_key;
$obj->isAlias = true;
$this->info[$key] = $obj;
}
/**
* Replaces any stdclass that only has the type property with type integer.
* Replaces any stdClass that only has the type property with type integer.
*/
public function postProcess()
{

View File

@ -0,0 +1,10 @@
Attr.ID.HTML5
TYPE: bool/null
DEFAULT: null
VERSION: 4.8.0
--DESCRIPTION--
In HTML5, restrictions on the format of the id attribute have been significantly
relaxed, such that any string is valid so long as it contains no spaces and
is at least one character. In lieu of a general HTML5 compatibility flag,
set this configuration directive to true to use the relaxed rules.
--# vim: et sw=4 sts=4

View File

@ -0,0 +1,11 @@
CSS.AllowDuplicates
TYPE: bool
DEFAULT: false
VERSION: 4.8.0
--DESCRIPTION--
<p>
By default, HTML Purifier removes duplicate CSS properties,
like <code>color:red; color:blue</code>. If this is set to
true, duplicate properties are allowed.
</p>
--# vim: et sw=4 sts=4

View File

@ -1,5 +1,5 @@
Cache.SerializerPermissions
TYPE: int
TYPE: int/null
VERSION: 4.3.0
DEFAULT: 0755
--DESCRIPTION--
@ -8,4 +8,9 @@ DEFAULT: 0755
Directory permissions of the files and directories created inside
the DefinitionCache/Serializer or other custom serializer path.
</p>
<p>
In HTML Purifier 4.8.0, this also supports <code>NULL</code>,
which means that no chmod'ing or directory creation shall
occur.
</p>
--# vim: et sw=4 sts=4

View File

@ -0,0 +1,16 @@
Core.AggressivelyRemoveScript
TYPE: bool
VERSION: 4.9.0
DEFAULT: true
--DESCRIPTION--
<p>
This directive enables aggressive pre-filter removal of
script tags. This is not necessary for security,
but it can help work around a bug in libxml where embedded
HTML elements inside script sections cause the parser to
choke. To revert to pre-4.9.0 behavior, set this to false.
This directive has no effect if %Core.Trusted is true,
%Core.RemoveScriptContents is false, or %Core.HiddenElements
does not contain script.
</p>
--# vim: et sw=4 sts=4

View File

@ -0,0 +1,36 @@
Core.LegacyEntityDecoder
TYPE: bool
VERSION: 4.9.0
DEFAULT: false
--DESCRIPTION--
<p>
Prior to HTML Purifier 4.9.0, entities were decoded by performing
a global search replace for all entities whose decoded versions
did not have special meanings under HTML, and replaced them with
their decoded versions. We would match all entities, even if they did
not have a trailing semicolon, but only if there weren't any trailing
alphanumeric characters.
</p>
<table>
<tr><th>Original</th><th>Text</th><th>Attribute</th></tr>
<tr><td>&amp;yen;</td><td>&yen;</td><td>&yen;</td></tr>
<tr><td>&amp;yen</td><td>&yen;</td><td>&yen;</td></tr>
<tr><td>&amp;yena</td><td>&amp;yena</td><td>&amp;yena</td></tr>
<tr><td>&amp;yen=</td><td>&yen;=</td><td>&yen;=</td></tr>
</table>
<p>
In HTML Purifier 4.9.0, we changed the behavior of entity parsing
to match entities that had missing trailing semicolons in less
cases, to more closely match HTML5 parsing behavior:
</p>
<table>
<tr><th>Original</th><th>Text</th><th>Attribute</th></tr>
<tr><td>&amp;yen;</td><td>&yen;</td><td>&yen;</td></tr>
<tr><td>&amp;yen</td><td>&yen;</td><td>&yen;</td></tr>
<tr><td>&amp;yena</td><td>&yen;a</td><td>&amp;yena</td></tr>
<tr><td>&amp;yen=</td><td>&yen;=</td><td>&amp;yen=</td></tr>
</table>
<p>
This flag reverts back to pre-HTML Purifier 4.9.0 behavior.
</p>
--# vim: et sw=4 sts=4

View File

@ -0,0 +1,10 @@
--# vim: et sw=4 sts=4
HTML.TargetNoopener
TYPE: bool
VERSION: 4.8.0
DEFAULT: TRUE
--DESCRIPTION--
If enabled, noopener rel attributes are added to links which have
a target attribute associated with them. This prevents malicious
destinations from overwriting the original window.
--# vim: et sw=4 sts=4

View File

@ -0,0 +1,9 @@
HTML.TargetNoreferrer
TYPE: bool
VERSION: 4.8.0
DEFAULT: TRUE
--DESCRIPTION--
If enabled, noreferrer rel attributes are added to links which have
a target attribute associated with them. This prevents malicious
destinations from overwriting the original window.
--# vim: et sw=4 sts=4

View File

@ -8,6 +8,7 @@ array (
'ftp' => true,
'nntp' => true,
'news' => true,
'tel' => true,
)
--DESCRIPTION--
Whitelist that defines the schemes that a URI is allowed to have. This

View File

@ -1,5 +1,5 @@
URI.DefaultScheme
TYPE: string
TYPE: string/null
DEFAULT: 'http'
--DESCRIPTION--
@ -7,4 +7,9 @@ DEFAULT: 'http'
Defines through what scheme the output will be served, in order to
select the proper object validator when no scheme information is present.
</p>
<p>
Starting with HTML Purifier 4.9.0, the default scheme can be null, in
which case we reject all URIs which do not have explicit schemes.
</p>
--# vim: et sw=4 sts=4

View File

@ -9,7 +9,7 @@ DEFAULT: NULL
absolute URIs 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>https://searx.laquadrature.net/?q=%s</code>).
<code>http://www.google.com/url?q=%s</code>).
</p>
<p>
Uses for this directive:

View File

@ -118,7 +118,7 @@ abstract class HTMLPurifier_DefinitionCache
/**
* Clears all expired (older version or revision) objects from cache
* @note Be carefuly implementing this method as flush. Flush must
* @note Be careful implementing this method as flush. Flush must
* not interfere with other Definition types, and cleanup()
* should not be repeatedly called by userland code.
* @param HTMLPurifier_Config $config

View File

@ -97,6 +97,12 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
}
$dir = $this->generateDirectoryPath($config);
$dh = opendir($dir);
// Apparently, on some versions of PHP, readdir will return
// an empty string if you pass an invalid argument to readdir.
// So you need this test. See #49.
if (false === $dh) {
return false;
}
while (false !== ($filename = readdir($dh))) {
if (empty($filename)) {
continue;
@ -106,6 +112,8 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
}
unlink($dir . '/' . $filename);
}
closedir($dh);
return true;
}
/**
@ -119,6 +127,10 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
}
$dir = $this->generateDirectoryPath($config);
$dh = opendir($dir);
// See #49 (and above).
if (false === $dh) {
return false;
}
while (false !== ($filename = readdir($dh))) {
if (empty($filename)) {
continue;
@ -131,6 +143,8 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
unlink($dir . '/' . $filename);
}
}
closedir($dh);
return true;
}
/**
@ -186,11 +200,9 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
if ($result !== false) {
// set permissions of the new file (no execute)
$chmod = $config->get('Cache.SerializerPermissions');
if (!$chmod) {
$chmod = 0644; // invalid config or simpletest
if ($chmod !== null) {
chmod($file, $chmod & 0666);
}
$chmod = $chmod & 0666;
chmod($file, $chmod);
}
return $result;
}
@ -204,8 +216,10 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
{
$directory = $this->generateDirectoryPath($config);
$chmod = $config->get('Cache.SerializerPermissions');
if (!$chmod) {
$chmod = 0755; // invalid config or simpletest
if ($chmod === null) {
// TODO: This races
if (is_dir($directory)) return true;
return mkdir($directory);
}
if (!is_dir($directory)) {
$base = $this->generateBaseDirectoryPath($config);
@ -219,15 +233,16 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
} elseif (!$this->_testPermissions($base, $chmod)) {
return false;
}
mkdir($directory, $chmod);
if (!$this->_testPermissions($directory, $chmod)) {
if (!mkdir($directory, $chmod)) {
trigger_error(
'Base directory ' . $base . ' does not exist,
please create or change using %Cache.SerializerPath',
'Could not create directory ' . $directory . '',
E_USER_WARNING
);
return false;
}
if (!$this->_testPermissions($directory, $chmod)) {
return false;
}
} elseif (!$this->_testPermissions($directory, $chmod)) {
return false;
}
@ -256,7 +271,7 @@ class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCac
);
return false;
}
if (function_exists('posix_getuid')) {
if (function_exists('posix_getuid') && $chmod !== null) {
// POSIX system, we can give more specific advice
if (fileowner($dir) === posix_getuid()) {
// we can chmod it ourselves

View File

@ -101,6 +101,14 @@ class HTMLPurifier_Encoder
* It will parse according to UTF-8 and return a valid UTF8 string, with
* non-SGML codepoints excluded.
*
* Specifically, it will permit:
* \x{9}\x{A}\x{D}\x{20}-\x{7E}\x{A0}-\x{D7FF}\x{E000}-\x{FFFD}\x{10000}-\x{10FFFF}
* Source: https://www.w3.org/TR/REC-xml/#NT-Char
* Arguably this function should be modernized to the HTML5 set
* of allowed characters:
* https://www.w3.org/TR/html5/syntax.html#preprocessing-the-input-stream
* which simultaneously expand and restrict the set of allowed characters.
*
* @param string $str The string to clean
* @param bool $force_php
* @return string
@ -122,15 +130,12 @@ class HTMLPurifier_Encoder
* function that needs to be able to understand UTF-8 characters.
* As of right now, only smart lossless character encoding converters
* would need that, and I'm probably not going to implement them.
* Once again, PHP 6 should solve all our problems.
*/
public static function cleanUTF8($str, $force_php = false)
{
// UTF-8 validity is checked since PHP 4.3.5
// This is an optimization: if the string is already valid UTF-8, no
// need to do PHP stuff. 99% of the time, this will be the case.
// The regexp matches the XML char production, as well as well as excluding
// non-SGML codepoints U+007F to U+009F
if (preg_match(
'/^[\x{9}\x{A}\x{D}\x{20}-\x{7E}\x{A0}-\x{D7FF}\x{E000}-\x{FFFD}\x{10000}-\x{10FFFF}]*$/Du',
$str
@ -255,6 +260,7 @@ class HTMLPurifier_Encoder
// 7F-9F is not strictly prohibited by XML,
// but it is non-SGML, and thus we don't allow it
(0xA0 <= $mUcs4 && 0xD7FF >= $mUcs4) ||
(0xE000 <= $mUcs4 && 0xFFFD >= $mUcs4) ||
(0x10000 <= $mUcs4 && 0x10FFFF >= $mUcs4)
)
) {

View File

@ -16,6 +16,138 @@ class HTMLPurifier_EntityParser
*/
protected $_entity_lookup;
/**
* Callback regex string for entities in text.
* @type string
*/
protected $_textEntitiesRegex;
/**
* Callback regex string for entities in attributes.
* @type string
*/
protected $_attrEntitiesRegex;
/**
* Tests if the beginning of a string is a semi-optional regex
*/
protected $_semiOptionalPrefixRegex;
public function __construct() {
// From
// http://stackoverflow.com/questions/15532252/why-is-reg-being-rendered-as-without-the-bounding-semicolon
$semi_optional = "quot|QUOT|lt|LT|gt|GT|amp|AMP|AElig|Aacute|Acirc|Agrave|Aring|Atilde|Auml|COPY|Ccedil|ETH|Eacute|Ecirc|Egrave|Euml|Iacute|Icirc|Igrave|Iuml|Ntilde|Oacute|Ocirc|Ograve|Oslash|Otilde|Ouml|REG|THORN|Uacute|Ucirc|Ugrave|Uuml|Yacute|aacute|acirc|acute|aelig|agrave|aring|atilde|auml|brvbar|ccedil|cedil|cent|copy|curren|deg|divide|eacute|ecirc|egrave|eth|euml|frac12|frac14|frac34|iacute|icirc|iexcl|igrave|iquest|iuml|laquo|macr|micro|middot|nbsp|not|ntilde|oacute|ocirc|ograve|ordf|ordm|oslash|otilde|ouml|para|plusmn|pound|raquo|reg|sect|shy|sup1|sup2|sup3|szlig|thorn|times|uacute|ucirc|ugrave|uml|uuml|yacute|yen|yuml";
// NB: three empty captures to put the fourth match in the right
// place
$this->_semiOptionalPrefixRegex = "/&()()()($semi_optional)/";
$this->_textEntitiesRegex =
'/&(?:'.
// hex
'[#]x([a-fA-F0-9]+);?|'.
// dec
'[#]0*(\d+);?|'.
// string (mandatory semicolon)
// NB: order matters: match semicolon preferentially
'([A-Za-z_:][A-Za-z0-9.\-_:]*);|'.
// string (optional semicolon)
"($semi_optional)".
')/';
$this->_attrEntitiesRegex =
'/&(?:'.
// hex
'[#]x([a-fA-F0-9]+);?|'.
// dec
'[#]0*(\d+);?|'.
// string (mandatory semicolon)
// NB: order matters: match semicolon preferentially
'([A-Za-z_:][A-Za-z0-9.\-_:]*);|'.
// string (optional semicolon)
// don't match if trailing is equals or alphanumeric (URL
// like)
"($semi_optional)(?![=;A-Za-z0-9])".
')/';
}
/**
* Substitute entities with the parsed equivalents. Use this on
* textual data in an HTML document (as opposed to attributes.)
*
* @param string $string String to have entities parsed.
* @return string Parsed string.
*/
public function substituteTextEntities($string)
{
return preg_replace_callback(
$this->_textEntitiesRegex,
array($this, 'entityCallback'),
$string
);
}
/**
* Substitute entities with the parsed equivalents. Use this on
* attribute contents in documents.
*
* @param string $string String to have entities parsed.
* @return string Parsed string.
*/
public function substituteAttrEntities($string)
{
return preg_replace_callback(
$this->_attrEntitiesRegex,
array($this, 'entityCallback'),
$string
);
}
/**
* Callback function for substituteNonSpecialEntities() that does the work.
*
* @param array $matches PCRE matches array, with 0 the entire match, and
* either index 1, 2 or 3 set with a hex value, dec value,
* or string (respectively).
* @return string Replacement string.
*/
protected function entityCallback($matches)
{
$entity = $matches[0];
$hex_part = @$matches[1];
$dec_part = @$matches[2];
$named_part = empty($matches[3]) ? @$matches[4] : $matches[3];
if ($hex_part !== NULL && $hex_part !== "") {
return HTMLPurifier_Encoder::unichr(hexdec($hex_part));
} elseif ($dec_part !== NULL && $dec_part !== "") {
return HTMLPurifier_Encoder::unichr((int) $dec_part);
} else {
if (!$this->_entity_lookup) {
$this->_entity_lookup = HTMLPurifier_EntityLookup::instance();
}
if (isset($this->_entity_lookup->table[$named_part])) {
return $this->_entity_lookup->table[$named_part];
} else {
// exact match didn't match anything, so test if
// any of the semicolon optional match the prefix.
// Test that this is an EXACT match is important to
// prevent infinite loop
if (!empty($matches[3])) {
return preg_replace_callback(
$this->_semiOptionalPrefixRegex,
array($this, 'entityCallback'),
$entity
);
}
return $entity;
}
}
}
// LEGACY CODE BELOW
/**
* Callback regex string for parsing entities.
* @type string
@ -144,7 +276,7 @@ class HTMLPurifier_EntityParser
$entity;
} else {
return isset($this->_special_ent2dec[$matches[3]]) ?
$this->_special_ent2dec[$matches[3]] :
$this->_special_dec2str[$this->_special_ent2dec[$matches[3]]] :
$entity;
}
}

View File

@ -95,7 +95,10 @@ class HTMLPurifier_Filter_ExtractStyleBlocks extends HTMLPurifier_Filter
if ($tidy !== null) {
$this->_tidy = $tidy;
}
$html = preg_replace_callback('#<style(?:\s.*)?>(.+)</style>#isU', array($this, 'styleCallback'), $html);
// NB: this must be NON-greedy because if we have
// <style>foo</style> <style>bar</style>
// we must not grab foo</style> <style>bar
$html = preg_replace_callback('#<style(?:\s.*)?>(.*)<\/style>#isU', array($this, 'styleCallback'), $html);
$style_blocks = $this->_styleMatches;
$this->_styleMatches = array(); // reset
$context->register('StyleBlocks', $style_blocks); // $context must not be reused

View File

@ -146,7 +146,7 @@ class HTMLPurifier_Generator
$attr = $this->generateAttributes($token->attr, $token->name);
if ($this->_flashCompat) {
if ($token->name == "object") {
$flash = new stdclass();
$flash = new stdClass();
$flash->attr = $token->attr;
$flash->param = array();
$this->_flashStack[] = $flash;

View File

@ -0,0 +1,21 @@
<?php
/**
* Module adds the target-based noopener attribute transformation to a tags. It
* is enabled by HTML.TargetNoopener
*/
class HTMLPurifier_HTMLModule_TargetNoopener extends HTMLPurifier_HTMLModule
{
/**
* @type string
*/
public $name = 'TargetNoopener';
/**
* @param HTMLPurifier_Config $config
*/
public function setup($config) {
$a = $this->addBlankElement('a');
$a->attr_transform_post[] = new HTMLPurifier_AttrTransform_TargetNoopener();
}
}

View File

@ -0,0 +1,21 @@
<?php
/**
* Module adds the target-based noreferrer attribute transformation to a tags. It
* is enabled by HTML.TargetNoreferrer
*/
class HTMLPurifier_HTMLModule_TargetNoreferrer extends HTMLPurifier_HTMLModule
{
/**
* @type string
*/
public $name = 'TargetNoreferrer';
/**
* @param HTMLPurifier_Config $config
*/
public function setup($config) {
$a = $this->addBlankElement('a');
$a->attr_transform_post[] = new HTMLPurifier_AttrTransform_TargetNoreferrer();
}
}

View File

@ -271,6 +271,14 @@ class HTMLPurifier_HTMLModuleManager
if ($config->get('HTML.TargetBlank')) {
$modules[] = 'TargetBlank';
}
// NB: HTML.TargetNoreferrer and HTML.TargetNoopener must be AFTER HTML.TargetBlank
// so that its post-attr-transform gets run afterwards.
if ($config->get('HTML.TargetNoreferrer')) {
$modules[] = 'TargetNoreferrer';
}
if ($config->get('HTML.TargetNoopener')) {
$modules[] = 'TargetNoopener';
}
// merge in custom modules
$modules = array_merge($modules, $this->userModules);

View File

@ -27,13 +27,18 @@ class HTMLPurifier_Injector_Linkify extends HTMLPurifier_Injector
if (strpos($token->data, '://') === false) {
// our really quick heuristic failed, abort
// this may not work so well if we want to match things like
// "domainname.com", but then again, most people don't
// "google.com", but then again, most people don't
return;
}
// there is/are URL(s). Let's split the string:
// Note: this regex is extremely permissive
$bits = preg_split('#((?:https?|ftp)://[^\s\'",<>()]+)#Su', $token->data, -1, PREG_SPLIT_DELIM_CAPTURE);
// there is/are URL(s). Let's split the string.
// We use this regex:
// https://gist.github.com/gruber/249502
// but with @cscott's backtracking fix and also
// the Unicode characters un-Unicodified.
$bits = preg_split(
'/\\b((?:[a-z][\\w\\-]+:(?:\\/{1,3}|[a-z0-9%])|www\\d{0,3}[.]|[a-z0-9.\\-]+[.][a-z]{2,4}\\/)(?:[^\\s()<>]|\\((?:[^\\s()<>]|(?:\\([^\\s()<>]+\\)))*\\))+(?:\\((?:[^\\s()<>]|(?:\\([^\\s()<>]+\\)))*\\)|[^\\s`!()\\[\\]{};:\'".,<>?\x{00ab}\x{00bb}\x{201c}\x{201d}\x{2018}\x{2019}]))/iu',
$token->data, -1, PREG_SPLIT_DELIM_CAPTURE);
$token = array();

View File

@ -46,6 +46,12 @@ class HTMLPurifier_Injector_RemoveEmpty extends HTMLPurifier_Injector
$this->removeNbsp = $config->get('AutoFormat.RemoveEmpty.RemoveNbsp');
$this->removeNbspExceptions = $config->get('AutoFormat.RemoveEmpty.RemoveNbsp.Exceptions');
$this->exclude = $config->get('AutoFormat.RemoveEmpty.Predicate');
foreach ($this->exclude as $key => $attrs) {
if (!is_array($attrs)) {
// HACK, see HTMLPurifier/Printer/ConfigForm.php
$this->exclude[$key] = explode(';', $attrs);
}
}
$this->attrValidator = new HTMLPurifier_AttrValidator();
}

View File

@ -36,6 +36,7 @@ class HTMLPurifier_Injector_SafeObject extends HTMLPurifier_Injector
);
/**
* These are all lower-case keys.
* @type array
*/
protected $allowedParam = array(
@ -43,7 +44,7 @@ class HTMLPurifier_Injector_SafeObject extends HTMLPurifier_Injector
'movie' => true,
'flashvars' => true,
'src' => true,
'allowFullScreen' => true, // if omitted, assume to be 'false'
'allowfullscreen' => true, // if omitted, assume to be 'false'
);
/**
@ -93,9 +94,11 @@ class HTMLPurifier_Injector_SafeObject extends HTMLPurifier_Injector
$token->attr['name'] === $this->addParam[$n]) {
// keep token, and add to param stack
$this->paramStack[$i][$n] = true;
} elseif (isset($this->allowedParam[$n])) {
} elseif (isset($this->allowedParam[strtolower($n)])) {
// keep token, don't do anything to it
// (could possibly check for duplicates here)
// Note: In principle, parameters should be case sensitive.
// But it seems they are not really; so accept any case.
} else {
$token = false;
}

View File

@ -96,7 +96,7 @@ class HTMLPurifier_Lexer
break;
}
if (class_exists('DOMDocument') &&
if (class_exists('DOMDocument', false) &&
method_exists('DOMDocument', 'loadHTML') &&
!extension_loaded('domxml')
) {
@ -169,21 +169,24 @@ class HTMLPurifier_Lexer
'&#x27;' => "'"
);
public function parseText($string, $config) {
return $this->parseData($string, false, $config);
}
public function parseAttr($string, $config) {
return $this->parseData($string, true, $config);
}
/**
* Parses special entities into the proper characters.
*
* This string will translate escaped versions of the special characters
* into the correct ones.
*
* @warning
* You should be able to treat the output of this function as
* completely parsed, but that's only because all other entities should
* have been handled previously in substituteNonSpecialEntities()
*
* @param string $string String character data to be parsed.
* @return string Parsed character data.
*/
public function parseData($string)
public function parseData($string, $is_attr, $config)
{
// following functions require at least one character
if ($string === '') {
@ -209,7 +212,15 @@ class HTMLPurifier_Lexer
}
// hmm... now we have some uncommon entities. Use the callback.
if ($config->get('Core.LegacyEntityDecoder')) {
$string = $this->_entity_parser->substituteSpecialEntities($string);
} else {
if ($is_attr) {
$string = $this->_entity_parser->substituteAttrEntities($string);
} else {
$string = $this->_entity_parser->substituteTextEntities($string);
}
}
return $string;
}
@ -323,7 +334,9 @@ class HTMLPurifier_Lexer
}
// expand entities that aren't the big five
if ($config->get('Core.LegacyEntityDecoder')) {
$html = $this->_entity_parser->substituteNonSpecialEntities($html);
}
// clean into wellformed UTF-8 string for an SGML context: this has
// to be done after entity expansion because the entities sometimes
@ -335,6 +348,13 @@ class HTMLPurifier_Lexer
$html = preg_replace('#<\?.+?\?>#s', '', $html);
}
$hidden_elements = $config->get('Core.HiddenElements');
if ($config->get('Core.AggressivelyRemoveScript') &&
!($config->get('HTML.Trusted') || !$config->get('Core.RemoveScriptContents')
|| empty($hidden_elements["script"]))) {
$html = preg_replace('#<script[^>]*>.*?</script>#i', '', $html);
}
return $html;
}
@ -345,13 +365,18 @@ class HTMLPurifier_Lexer
public function extractBody($html)
{
$matches = array();
$result = preg_match('!<body[^>]*>(.*)</body>!is', $html, $matches);
$result = preg_match('|(.*?)<body[^>]*>(.*)</body>|is', $html, $matches);
if ($result) {
return $matches[1];
} else {
return $html;
// Make sure it's not in a comment
$comment_start = strrpos($matches[1], '<!--');
$comment_end = strrpos($matches[1], '-->');
if ($comment_start === false ||
($comment_end !== false && $comment_end > $comment_start)) {
return $matches[2];
}
}
return $html;
}
}
// vim: et sw=4 sts=4

View File

@ -72,12 +72,20 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
$doc->loadHTML($html);
restore_error_handler();
$body = $doc->getElementsByTagName('html')->item(0)-> // <html>
getElementsByTagName('body')->item(0); // <body>
$div = $body->getElementsByTagName('div')->item(0); // <div>
$tokens = array();
$this->tokenizeDOM(
$doc->getElementsByTagName('html')->item(0)-> // <html>
getElementsByTagName('body')->item(0), // <body>
$tokens
);
$this->tokenizeDOM($div, $tokens, $config);
// If the div has a sibling, that means we tripped across
// a premature </div> tag. So remove the div we parsed,
// and then tokenize the rest of body. We can't tokenize
// the sibling directly as we'll lose the tags in that case.
if ($div->nextSibling) {
$body->removeChild($div);
$this->tokenizeDOM($body, $tokens, $config);
}
return $tokens;
}
@ -88,7 +96,7 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
* @param HTMLPurifier_Token[] $tokens Array-list of already tokenized tokens.
* @return HTMLPurifier_Token of node appended to previously passed tokens.
*/
protected function tokenizeDOM($node, &$tokens)
protected function tokenizeDOM($node, &$tokens, $config)
{
$level = 0;
$nodes = array($level => new HTMLPurifier_Queue(array($node)));
@ -97,7 +105,7 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
while (!$nodes[$level]->isEmpty()) {
$node = $nodes[$level]->shift(); // FIFO
$collect = $level > 0 ? true : false;
$needEndingTag = $this->createStartNode($node, $tokens, $collect);
$needEndingTag = $this->createStartNode($node, $tokens, $collect, $config);
if ($needEndingTag) {
$closingNodes[$level][] = $node;
}
@ -127,7 +135,7 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
* @return bool if the token needs an endtoken
* @todo data and tagName properties don't seem to exist in DOMNode?
*/
protected function createStartNode($node, &$tokens, $collect)
protected function createStartNode($node, &$tokens, $collect, $config)
{
// intercept non element nodes. WE MUST catch all of them,
// but we're not getting the character reference nodes because
@ -151,7 +159,7 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
}
}
}
$tokens[] = $this->factory->createText($this->parseData($data));
$tokens[] = $this->factory->createText($this->parseText($data, $config));
return false;
} elseif ($node->nodeType === XML_COMMENT_NODE) {
// this is code is only invoked for comments in script/style in versions
@ -252,7 +260,7 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
* @param HTMLPurifier_Context $context
* @return string
*/
protected function wrapHTML($html, $config, $context)
protected function wrapHTML($html, $config, $context, $use_div = true)
{
$def = $config->getDefinition('HTML');
$ret = '';
@ -271,7 +279,11 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
$ret .= '<html><head>';
$ret .= '<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />';
// No protection if $html contains a stray </div>!
$ret .= '</head><body>' . $html . '</body></html>';
$ret .= '</head><body>';
if ($use_div) $ret .= '<div>';
$ret .= $html;
if ($use_div) $ret .= '</div>';
$ret .= '</body></html>';
return $ret;
}
}

View File

@ -129,12 +129,12 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
// We are not inside tag and there still is another tag to parse
$token = new
HTMLPurifier_Token_Text(
$this->parseData(
$this->parseText(
substr(
$html,
$cursor,
$position_next_lt - $cursor
)
), $config
)
);
if ($maintain_line_numbers) {
@ -154,11 +154,11 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
// Create Text of rest of string
$token = new
HTMLPurifier_Token_Text(
$this->parseData(
$this->parseText(
substr(
$html,
$cursor
)
), $config
)
);
if ($maintain_line_numbers) {
@ -324,8 +324,8 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
$token = new
HTMLPurifier_Token_Text(
'<' .
$this->parseData(
substr($html, $cursor)
$this->parseText(
substr($html, $cursor), $config
)
);
if ($maintain_line_numbers) {
@ -429,7 +429,7 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
if ($value === false) {
$value = '';
}
return array($key => $this->parseData($value));
return array($key => $this->parseAttr($value, $config));
}
// setup loop environment
@ -518,7 +518,7 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
if ($value === false) {
$value = '';
}
$array[$key] = $this->parseData($value);
$array[$key] = $this->parseAttr($value, $config);
$cursor++;
} else {
// boolattr

View File

@ -21,7 +21,7 @@ class HTMLPurifier_Lexer_PH5P extends HTMLPurifier_Lexer_DOMLex
public function tokenizeHTML($html, $config, $context)
{
$new_html = $this->normalize($html, $config, $context);
$new_html = $this->wrapHTML($new_html, $config, $context);
$new_html = $this->wrapHTML($new_html, $config, $context, false /* no div */);
try {
$parser = new HTML5($new_html);
$doc = $parser->save();
@ -36,7 +36,7 @@ class HTMLPurifier_Lexer_PH5P extends HTMLPurifier_Lexer_DOMLex
$doc->getElementsByTagName('html')->item(0)-> // <html>
getElementsByTagName('body')->item(0) // <body>
,
$tokens
$tokens, $config
);
return $tokens;
}
@ -1515,6 +1515,7 @@ class HTML5
// Consume the maximum number of characters possible, with the
// consumed characters case-sensitively matching one of the
// identifiers in the first column of the entities table.
$e_name = $this->characters('0-9A-Za-z;', $this->char + 1);
$len = strlen($e_name);
@ -1547,7 +1548,7 @@ class HTML5
// Return a character token for the character corresponding to the
// entity name (as given by the second column of the entities table).
return html_entity_decode('&' . $entity . ';', ENT_QUOTES, 'UTF-8');
return html_entity_decode('&' . rtrim($entity, ';') . ';', ENT_QUOTES, 'UTF-8');
}
private function emitToken($token)

View File

@ -327,6 +327,10 @@ class HTMLPurifier_Printer_ConfigForm_default extends HTMLPurifier_Printer
case HTMLPurifier_VarParser::HASH:
$nvalue = '';
foreach ($value as $i => $v) {
if (is_array($v)) {
// HACK
$v = implode(";", $v);
}
$nvalue .= "$i:$v" . PHP_EOL;
}
$value = $nvalue;

View File

@ -165,7 +165,7 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
if (empty($zipper->front)) break;
$token = $zipper->prev($token);
// indicate that other injectors should not process this token,
// but we need to reprocess it
// but we need to reprocess it. See Note [Injector skips]
unset($token->skip[$i]);
$token->rewind = $i;
if ($token instanceof HTMLPurifier_Token_Start) {
@ -210,6 +210,7 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
if ($token instanceof HTMLPurifier_Token_Text) {
foreach ($this->injectors as $i => $injector) {
if (isset($token->skip[$i])) {
// See Note [Injector skips]
continue;
}
if ($token->rewind !== null && $token->rewind !== $i) {
@ -367,6 +368,7 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
if ($ok) {
foreach ($this->injectors as $i => $injector) {
if (isset($token->skip[$i])) {
// See Note [Injector skips]
continue;
}
if ($token->rewind !== null && $token->rewind !== $i) {
@ -422,6 +424,7 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
$token->start = $current_parent;
foreach ($this->injectors as $i => $injector) {
if (isset($token->skip[$i])) {
// See Note [Injector skips]
continue;
}
if ($token->rewind !== null && $token->rewind !== $i) {
@ -534,12 +537,17 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
*/
protected function processToken($token, $injector = -1)
{
// Zend OpCache miscompiles $token = array($token), so
// avoid this pattern. See: https://github.com/ezyang/htmlpurifier/issues/108
// normalize forms of token
if (is_object($token)) {
$token = array(1, $token);
$tmp = $token;
$token = array(1, $tmp);
}
if (is_int($token)) {
$token = array($token);
$tmp = $token;
$token = array($tmp);
}
if ($token === false) {
$token = array(1);
@ -561,7 +569,12 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
list($old, $r) = $this->zipper->splice($this->token, $delete, $token);
if ($injector > -1) {
// determine appropriate skips
// See Note [Injector skips]
// Determine appropriate skips. Here's what the code does:
// *If* we deleted one or more tokens, copy the skips
// of those tokens into the skips of the new tokens (in $token).
// Also, mark the newly inserted tokens as having come from
// $injector.
$oldskip = isset($old[0]) ? $old[0]->skip : array();
foreach ($token as $object) {
$object->skip = $oldskip;
@ -597,4 +610,50 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
}
}
// Note [Injector skips]
// ~~~~~~~~~~~~~~~~~~~~~
// When I originally designed this class, the idea behind the 'skip'
// property of HTMLPurifier_Token was to help avoid infinite loops
// in injector processing. For example, suppose you wrote an injector
// that bolded swear words. Naively, you might write it so that
// whenever you saw ****, you replaced it with <strong>****</strong>.
//
// When this happens, we will reprocess all of the tokens with the
// other injectors. Now there is an opportunity for infinite loop:
// if we rerun the swear-word injector on these tokens, we might
// see **** and then reprocess again to get
// <strong><strong>****</strong></strong> ad infinitum.
//
// Thus, the idea of a skip is that once we process a token with
// an injector, we mark all of those tokens as having "come from"
// the injector, and we never run the injector again on these
// tokens.
//
// There were two more complications, however:
//
// - With HTMLPurifier_Injector_RemoveEmpty, we noticed that if
// you had <b><i></i></b>, after you removed the <i></i>, you
// really would like this injector to go back and reprocess
// the <b> tag, discovering that it is now empty and can be
// removed. So we reintroduced the possibility of infinite looping
// by adding a "rewind" function, which let you go back to an
// earlier point in the token stream and reprocess it with injectors.
// Needless to say, we need to UN-skip the token so it gets
// reprocessed.
//
// - Suppose that you successfuly process a token, replace it with
// one with your skip mark, but now another injector wants to
// process the skipped token with another token. Should you continue
// to skip that new token, or reprocess it? If you reprocess,
// you can end up with an infinite loop where one injector converts
// <a> to <b>, and then another injector converts it back. So
// we inherit the skips, but for some reason, I thought that we
// should inherit the skip from the first token of the token
// that we deleted. Why? Well, it seems to work OK.
//
// If I were to redesign this functionality, I would absolutely not
// go about doing it this way: the semantics are just not very well
// defined, and in any case you probably wanted to operate on trees,
// not token streams.
// vim: et sw=4 sts=4

View File

@ -26,7 +26,7 @@ abstract class HTMLPurifier_Token
public $armor = array();
/**
* Used during MakeWellFormed.
* Used during MakeWellFormed. See Note [Injector skips]
* @type
*/
public $skip;

View File

@ -85,11 +85,13 @@ class HTMLPurifier_URI
$def = $config->getDefinition('URI');
$scheme_obj = $def->getDefaultScheme($config, $context);
if (!$scheme_obj) {
if ($def->defaultScheme !== null) {
// something funky happened to the default scheme object
trigger_error(
'Default scheme object "' . $def->defaultScheme . '" was not readable',
E_USER_WARNING
);
} // suppress error if it's null
return false;
}
}

View File

@ -79,9 +79,18 @@ class HTMLPurifier_URIScheme_data extends HTMLPurifier_URIScheme
} else {
$raw_data = $data;
}
if ( strlen($raw_data) < 12 ) {
// error; exif_imagetype throws exception with small files,
// and this likely indicates a corrupt URI/failed parse anyway
return false;
}
// XXX probably want to refactor this into a general mechanism
// for filtering arbitrary content types
if (function_exists('sys_get_temp_dir')) {
$file = tempnam(sys_get_temp_dir(), "");
} else {
$file = tempnam("/tmp", "");
}
file_put_contents($file, $raw_data);
if (function_exists('exif_imagetype')) {
$image_code = exif_imagetype($file);

View File

@ -0,0 +1,46 @@
<?php
/**
* Validates tel (for phone numbers).
*
* The relevant specifications for this protocol are RFC 3966 and RFC 5341,
* but this class takes a much simpler approach: we normalize phone
* numbers so that they only include (possibly) a leading plus,
* and then any number of digits and x'es.
*/
class HTMLPurifier_URIScheme_tel extends HTMLPurifier_URIScheme
{
/**
* @type bool
*/
public $browsable = false;
/**
* @type bool
*/
public $may_omit_host = true;
/**
* @param HTMLPurifier_URI $uri
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return bool
*/
public function doValidate(&$uri, $config, $context)
{
$uri->userinfo = null;
$uri->host = null;
$uri->port = null;
// Delete all non-numeric characters, non-x characters
// from phone number, EXCEPT for a leading plus sign.
$uri->path = preg_replace('/(?!^\+)[^\dx]/', '',
// Normalize e(x)tension to lower-case
str_replace('X', 'x', $uri->path));
return true;
}
}
// vim: et sw=4 sts=4

View File

@ -1 +1 @@
4.7.0
4.9.3