From b1d31eea94a8ed9e26ee97ad0e401cb03708d56f Mon Sep 17 00:00:00 2001
From: thomascube <thomas@roundcube.net>
Date: Wed, 09 Feb 2011 11:06:20 -0500
Subject: [PATCH] Fix stripping invalid comments. Changes from r4483 also stripped entire CSS blocks packed in comments
---
program/lib/washtml.php | 50 ++++++++++++++++++++++++++++++++++++--------------
1 files changed, 36 insertions(+), 14 deletions(-)
diff --git a/program/lib/washtml.php b/program/lib/washtml.php
index e19fdf0..0f8dc7e 100644
--- a/program/lib/washtml.php
+++ b/program/lib/washtml.php
@@ -74,6 +74,8 @@
* - added $block_elements
* - changed $ignore_elements behaviour
* - added RFC2397 support
+ * - base URL support
+ * - invalid HTML comments removal before parsing
*/
class washtml
@@ -88,7 +90,7 @@
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');
/* 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');
/* State for linked objects in HTML */
public $extlinks = false;
@@ -144,18 +146,21 @@
'|#[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], $url)) {
if ($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' && $match[0] != 'rbg') //whitelist ?
+ }
+ else if ($match[0] != 'url' && $match[0] != 'rbg') //whitelist ?
$value .= ' ' . $match[0];
$str = substr($str, strlen($match[0]));
}
@@ -180,7 +185,8 @@
else if ($key == 'style' && ($style = $this->wash_style($value)))
$t .= ' style="' . $style . '"';
else if ($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', $value)) {
@@ -215,18 +221,23 @@
switch($node->nodeType) {
case XML_ELEMENT_NODE: //Check element
$tagName = strtolower($node->tagName);
- if($callback = $this->handlers[$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])) {
+ }
+ 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[$tagName])) {
+ // create closing tag for block elements, but also for elements
+ // with content or with some attributes (eg. style, class) (#1486812)
+ ($content != '' || $node->hasAttributes() || isset($this->_block_elements[$tagName]) ? ">$content</$tagName>" : ' />');
+ }
+ else if (isset($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 .= $node->nodeValue;
@@ -249,10 +260,21 @@
/* Main function, give it untrusted HTML, tell it if you allow loading
* remote images and give it a map to convert "cid:" urls. */
- public function wash($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)
+ $html = preg_replace('/<!--[^->]*>/', '', $html);
+
@$node->loadHTML($html);
return $this->dumpHtml($node);
}
--
Gitblit v1.9.1