From b3206b4b2822b8c9d18c4730aa1afdf72a758f8c Mon Sep 17 00:00:00 2001
From: Aleksander Machniak <alec@alec.pl>
Date: Wed, 15 Aug 2012 04:13:17 -0400
Subject: [PATCH] Fix XSS issue with href="javascript:" not being removed (#1488613)
---
program/lib/washtml.php | 144 +++++++++++++++++++++++++++++++++++------------
1 files changed, 106 insertions(+), 38 deletions(-)
diff --git a/program/lib/washtml.php b/program/lib/washtml.php
index e19fdf0..98ae5ed 100644
--- a/program/lib/washtml.php
+++ b/program/lib/washtml.php
@@ -74,22 +74,43 @@
* - added $block_elements
* - changed $ignore_elements behaviour
* - added RFC2397 support
+ * - base URL support
+ * - invalid HTML comments removal before parsing
+ * - "fixing" unitless CSS values for XHTML output
*/
class washtml
{
/* Allowed HTML elements (default) */
- static $html_elements = array('a', 'abbr', 'acronym', 'address', 'area', 'b', 'basefont', 'bdo', 'big', 'blockquote', 'br', 'caption', 'center', 'cite', 'code', 'col', 'colgroup', 'dd', 'del', 'dfn', 'dir', 'div', 'dl', 'dt', 'em', 'fieldset', 'font', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'i', 'ins', 'label', 'legend', 'li', 'map', 'menu', 'nobr', 'ol', 'p', 'pre', 'q', 's', 'samp', 'small', 'span', 'strike', 'strong', 'sub', 'sup', 'table', 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr', 'tt', 'u', 'ul', 'var', 'wbr', 'img');
-
+ static $html_elements = array('a', 'abbr', 'acronym', 'address', 'area', 'b',
+ 'basefont', 'bdo', 'big', 'blockquote', 'br', 'caption', 'center',
+ 'cite', 'code', 'col', 'colgroup', 'dd', 'del', 'dfn', 'dir', 'div', 'dl',
+ 'dt', 'em', 'fieldset', 'font', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'i',
+ 'ins', 'label', 'legend', 'li', 'map', 'menu', 'nobr', 'ol', 'p', 'pre', 'q',
+ 's', 'samp', 'small', 'span', 'strike', 'strong', 'sub', 'sup', 'table',
+ 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr', 'tt', 'u', 'ul', 'var', 'wbr', 'img',
+ // form elements
+ 'button', 'input', 'textarea', 'select', 'option', 'optgroup'
+ );
+
/* Ignore these HTML tags and their content */
static $ignore_elements = array('script', 'applet', 'embed', 'object', 'style');
-
+
/* Allowed HTML attributes */
- static $html_attribs = array('name', 'class', 'title', 'alt', 'width', 'height', 'align', 'nowrap', 'col', 'row', 'id', 'rowspan', 'colspan', 'cellspacing', 'cellpadding', 'valign', 'bgcolor', 'color', 'border', 'bordercolorlight', 'bordercolordark', 'face', 'marginwidth', 'marginheight', 'axis', 'border', 'abbr', 'char', 'charoff', 'clear', 'compact', 'coords', 'vspace', 'hspace', 'cellborder', 'size', 'lang', 'dir');
+ static $html_attribs = array('name', 'class', 'title', 'alt', 'width', 'height',
+ 'align', 'nowrap', 'col', 'row', 'id', 'rowspan', 'colspan', 'cellspacing',
+ 'cellpadding', 'valign', 'bgcolor', 'color', 'border', 'bordercolorlight',
+ 'bordercolordark', 'face', 'marginwidth', 'marginheight', 'axis', 'border',
+ 'abbr', 'char', 'charoff', 'clear', 'compact', 'coords', 'vspace', 'hspace',
+ 'cellborder', 'size', 'lang', 'dir', 'usemap',
+ // attributes of form elements
+ 'type', 'rows', 'cols', 'disabled', 'readonly', 'checked', 'multiple', 'value'
+ );
/* Block elements which could be empty but cannot be returned in short form (<tag />) */
- static $block_elements = array('div', 'p', 'pre', 'blockquote', 'a', 'font');
-
+ static $block_elements = array('div', 'p', 'pre', 'blockquote', 'a', 'font', 'center',
+ 'table', 'ul', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'ol', 'dl', 'strong', 'i', 'b', 'u', 'span');
+
/* State for linked objects in HTML */
public $extlinks = false;
@@ -98,7 +119,7 @@
/* Registered callback functions for tags */
private $handlers = array();
-
+
/* Allowed HTML elements */
private $_html_elements = array();
@@ -110,10 +131,11 @@
/* Allowed HTML attributes */
private $_html_attribs = array();
-
+
/* Constructor */
- public function __construct($p = array()) {
+ public function __construct($p = array())
+ {
$this->_html_elements = array_flip((array)$p['html_elements']) + array_flip(self::$html_elements) ;
$this->_html_attribs = array_flip((array)$p['html_attribs']) + array_flip(self::$html_attribs);
$this->_ignore_elements = array_flip((array)$p['ignore_elements']) + array_flip(self::$ignore_elements);
@@ -121,53 +143,70 @@
unset($p['html_elements'], $p['html_attribs'], $p['ignore_elements'], $p['block_elements']);
$this->config = $p + array('show_washed'=>true, 'allow_remote'=>false, 'cid_map'=>array());
}
-
+
/* Register a callback function for a certain tag */
public function add_callback($tagName, $callback)
{
$this->handlers[$tagName] = $callback;
}
-
+
/* Check CSS style */
- private function wash_style($style) {
+ private function wash_style($style)
+ {
$s = '';
foreach (explode(';', $style) as $declaration) {
if (preg_match('/^\s*([a-z\-]+)\s*:\s*(.*)\s*$/i', $declaration, $match)) {
$cssid = $match[1];
- $str = $match[2];
+ $str = $match[2];
$value = '';
+
while (sizeof($str) > 0 &&
preg_match('/^(url\(\s*[\'"]?([^\'"\)]*)[\'"]?\s*\)'./*1,2*/
'|rgb\(\s*[0-9]+\s*,\s*[0-9]+\s*,\s*[0-9]+\s*\)'.
'|-?[0-9.]+\s*(em|ex|px|cm|mm|in|pt|pc|deg|rad|grad|ms|s|hz|khz|%)?'.
- '|#[0-9a-f]{3,6}|[a-z0-9\-]+'.
- ')\s*/i', $str, $match)) {
+ '|#[0-9a-f]{3,6}'.
+ '|[a-z0-9", -]+'.
+ ')\s*/i', $str, $match)
+ ) {
if ($match[2]) {
- if ($src = $this->config['cid_map'][$match[2]])
+ if (($src = $this->config['cid_map'][$match[2]])
+ || ($src = $this->config['cid_map'][$this->config['base_url'].$match[2]])) {
$value .= ' url('.htmlspecialchars($src, ENT_QUOTES) . ')';
- else if (preg_match('/^(http|https|ftp):.*$/i', $match[2], $url)) {
+ }
+ else if (preg_match('!^(https?:)?//[a-z0-9/._+-]+$!i', $match[2], $url)) {
if ($this->config['allow_remote'])
$value .= ' url('.htmlspecialchars($url[0], ENT_QUOTES).')';
else
$this->extlinks = true;
}
- else if (preg_match('/^data:.+/i', $url)) { // RFC2397
- $value .= ' url('.htmlspecialchars($url, ENT_QUOTES).')';
+ else if (preg_match('/^data:.+/i', $match[2])) { // RFC2397
+ $value .= ' url('.htmlspecialchars($match[2], ENT_QUOTES).')';
}
- } else if ($match[0] != 'url' && $match[0] != 'rbg') //whitelist ?
+ }
+ else { //whitelist ?
$value .= ' ' . $match[0];
+
+ // #1488535: Fix size units, so width:800 would be changed to width:800px
+ if (preg_match('/(left|right|top|bottom|width|height)/i', $cssid) && preg_match('/^[0-9]+$/', $match[0])) {
+ $value .= 'px';
+ }
+ }
+
$str = substr($str, strlen($match[0]));
}
- if ($value)
+
+ if (isset($value[0])) {
$s .= ($s?' ':'') . $cssid . ':' . $value . ';';
+ }
}
}
return $s;
}
/* Take a node and return allowed attributes and check values */
- private function wash_attribs($node) {
+ private function wash_attribs($node)
+ {
$t = '';
$washed;
@@ -175,12 +214,18 @@
$key = strtolower($key);
$value = $node->getAttribute($key);
if (isset($this->_html_attribs[$key]) ||
- ($key == 'href' && preg_match('/^(http:|https:|ftp:|mailto:|#).+/i', $value)))
+ ($key == 'href' && !preg_match('!^javascript!i', $value)
+ && preg_match('!^([a-z][a-z0-9.+-]+:|//|#).+!i', $value))
+ ) {
$t .= ' ' . $key . '="' . htmlspecialchars($value, ENT_QUOTES) . '"';
- else if ($key == 'style' && ($style = $this->wash_style($value)))
- $t .= ' style="' . $style . '"';
+ }
+ else if ($key == 'style' && ($style = $this->wash_style($value))) {
+ $quot = strpos($style, '"') !== false ? "'" : '"';
+ $t .= ' style=' . $quot . $style . $quot;
+ }
else if ($key == 'background' || ($key == 'src' && strtolower($node->tagName) == 'img')) { //check tagName anyway
- if ($src = $this->config['cid_map'][$value]) {
+ if (($src = $this->config['cid_map'][$value])
+ || ($src = $this->config['cid_map'][$this->config['base_url'].$value])) {
$t .= ' ' . $key . '="' . htmlspecialchars($src, ENT_QUOTES) . '"';
}
else if (preg_match('/^(http|https|ftp):.+/i', $value)) {
@@ -195,7 +240,8 @@
else if (preg_match('/^data:.+/i', $value)) { // RFC2397
$t .= ' ' . $key . '="' . htmlspecialchars($value, ENT_QUOTES) . '"';
}
- } else
+ }
+ else
$washed .= ($washed?' ':'') . $key;
}
return $t . ($washed && $this->config['show_washed']?' x-washed="'.$washed.'"':'');
@@ -204,7 +250,8 @@
/* The main loop that recurse on a node tree.
* It output only allowed tags with allowed attributes
* and allowed inline styles */
- private function dumpHtml($node) {
+ private function dumpHtml($node)
+ {
if(!$node->hasChildNodes())
return '';
@@ -215,18 +262,21 @@
switch($node->nodeType) {
case XML_ELEMENT_NODE: //Check element
$tagName = strtolower($node->tagName);
- if($callback = $this->handlers[$tagName]) {
- $dump .= call_user_func($callback, $tagName, $this->wash_attribs($node), $this->dumpHtml($node));
- } else if(isset($this->_html_elements[$tagName])) {
+ if ($callback = $this->handlers[$tagName]) {
+ $dump .= call_user_func($callback, $tagName, $this->wash_attribs($node), $this->dumpHtml($node), $this);
+ }
+ else if (isset($this->_html_elements[$tagName])) {
$content = $this->dumpHtml($node);
$dump .= '<' . $tagName . $this->wash_attribs($node) .
($content != '' || isset($this->_block_elements[$tagName]) ? ">$content</$tagName>" : ' />');
- } else if(isset($this->_ignore_elements[$tagName])) {
+ }
+ else if (isset($this->_ignore_elements[$tagName])) {
$dump .= '<!-- ' . htmlspecialchars($tagName, ENT_QUOTES) . ' not allowed -->';
- } else {
+ }
+ else {
$dump .= '<!-- ' . htmlspecialchars($tagName, ENT_QUOTES) . ' ignored -->';
$dump .= $this->dumpHtml($node); // ignore tags not its content
- }
+ }
break;
case XML_CDATA_SECTION_NODE:
$dump .= $node->nodeValue;
@@ -249,14 +299,32 @@
/* Main function, give it untrusted HTML, tell it if you allow loading
* remote images and give it a map to convert "cid:" urls. */
- public function wash($html) {
- //Charset seems to be ignored (probably if defined in the HTML document)
+ public function wash($html)
+ {
+ // Charset seems to be ignored (probably if defined in the HTML document)
$node = new DOMDocument('1.0', $this->config['charset']);
$this->extlinks = false;
+
+ // Find base URL for images
+ if (preg_match('/<base\s+href=[\'"]*([^\'"]+)/is', $html, $matches))
+ $this->config['base_url'] = $matches[1];
+ else
+ $this->config['base_url'] = '';
+
+ // Remove invalid HTML comments (#1487759)
+ // Don't remove valid conditional comments
+ $html = preg_replace('/<!--[^->[\n]*>/', '', $html);
+
@$node->loadHTML($html);
return $this->dumpHtml($node);
}
-}
+ /**
+ * Getter for config parameters
+ */
+ public function get_config($prop)
+ {
+ return $this->config[$prop];
+ }
-?>
+}
--
Gitblit v1.9.1