- Fix displaying of HTML messages with unknown/malformed tags (#1486003)

- Some other changes for styled HTML display

2 files changed, 38 insertions, 25 deletions

diff --git a/program/lib/washtml.php b/program/lib/washtml.php
index afc5f47ff..efd525df3 100644
--- a/program/lib/washtml.php
+++ b/program/lib/washtml.php
@@ -69,6 +69,10 @@
  * Dont be a fool:
  *  - Dont alter data on a GET: '<img src="http://yourhost/mail?action=delete&uid=3267" />'
  *  - ...
+ *
+ * Roundcube Changes:
+ * - added $block_elements
+ * - changed $ignore_elements behaviour
  */
 
 class washtml
@@ -76,8 +80,8 @@ 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');
   
-  /* Ignore these HTML tags but process their content */
-  static $ignore_elements = array('html', 'head', 'body');
+  /* Ignore these HTML tags and their content */
+  static $ignore_elements = array('script', 'applet', 'embed', 'object', 'style');
   
   /* Allowed HTML attributes */
   static $html_attribs = array('name', 'class', 'title', 'alt', 'width', 'height', 'align', 'nowrap', 'col', 'row', 'id', 'rowspan', 'colspan', 'cellspacing', 'cellpadding', 'valign', 'bgcolor', 'color', 'border', 'bordercolorlight', 'bordercolordark', 'face', 'marginwidth', 'marginheight', 'axis', 'border', 'abbr', 'char', 'charoff', 'clear', 'compact', 'coords', 'vspace', 'hspace', 'cellborder', 'size', 'lang', 'dir');  
@@ -209,14 +213,13 @@ class washtml
         } else if(isset($this->_html_elements[$tagName])) {
           $content = $this->dumpHtml($node);
           $dump .= '<' . $tagName . $this->wash_attribs($node) .
-//            ($content?">$content</$tagName>":' />');
-// Roundcube Trac: #1485974
             ($content || isset($this->_block_elements[$tagName]) ? ">$content</$tagName>" : ' />');
         } else if(isset($this->_ignore_elements[$tagName])) {
-          $dump .= '<!-- ' . htmlspecialchars($tagName, ENT_QUOTES) . ' ignored -->';
-          $dump .= $this->dumpHtml($node); //Just ignored
-        } else
           $dump .= '<!-- ' . htmlspecialchars($tagName, ENT_QUOTES) . ' not allowed -->';
+        } 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;
@@ -249,4 +252,4 @@ class washtml
 
 }
 
-?>
\ No newline at end of file
+?>
diff --git a/program/steps/mail/func.inc b/program/steps/mail/func.inc
index b969c04e9..5cf0f7664 100644
--- a/program/steps/mail/func.inc
+++ b/program/steps/mail/func.inc
@@ -428,7 +428,10 @@ function rcmail_js_message_list($a_headers, $insert_top=FALSE, $replace=TRUE)
         {
         $action = $mbox==$CONFIG['drafts_mbox'] ? 'compose' : 'show';
         $uid_param = $mbox==$CONFIG['drafts_mbox'] ? '_draft_uid' : '_uid';
-        $cont = abbreviate_string(trim($IMAP->decode_header($header->$col)), 160);
+	$cont = trim($IMAP->decode_header($header->$col));
+	if ($browser->ie)
+	  $cont = rc_utf8_clean($cont);
+        $cont = abbreviate_string($cont, 160);
         if (!$cont) $cont = rcube_label('nosubject');
         $cont = $browser->ie ? Q($cont) : sprintf('<a href="%s" onclick="return rcube_event.cancel(event)">%s</a>', Q(rcmail_url($action, array($uid_param=>$header->uid, '_mbox'=>$mbox))), Q($cont));
         }
@@ -668,36 +671,28 @@ function rcmail_wash_html($html, $p = array(), $cid_replaces)
   global $REMOTE_OBJECTS;
   
   $p += array('safe' => false, 'inline_html' => true);
-  
+
   // special replacements (not properly handled by washtml class)
   $html_search = array(
     '/(<\/nobr>)(\s+)(<nobr>)/i',	// space(s) between <NOBR>
-    '/(<[\/]*st1:[^>]+>)/i',		// Microsoft's Smart Tags <ST1>
-    '/<\/?rte_text>/i',			// Rich Text Editor tags (#1485647)
-    '/<\/?broadcast[^>]*>/i',		// invoices from the Apple Store contains <broadcast> tags (#1485962)
     '/<title>.*<\/title>/i',		// PHP bug #32547 workaround: remove title tag
-    '/<html[^>]*>/im',			// malformed html: remove html tags (#1485139)
-    '/<\/html>/i',			// malformed html: remove html tags (#1485139)
     '/^(\0\0\xFE\xFF|\xFF\xFE\0\0|\xFE\xFF|\xFF\xFE|\xEF\xBB\xBF)/',	// byte-order mark (only outlook?)
   );
   $html_replace = array(
     '\\1'.' &nbsp; '.'\\3',
     '',
     '',
-    '',
-    '',
-    '',
-    '',
-    '',
   );
   $html = preg_replace($html_search, $html_replace, $html);
 
+  // fix (unknown/malformed) HTML tags before "wash"
+  $html = preg_replace_callback('/(<[\/!]*)([^ >]+)/', 'rcmail_html_tag_callback', $html);
+
   // charset was converted to UTF-8 in rcube_imap::get_message_part() -> change charset specification in HTML accordingly
   $charset_pattern = '/(\s+content=[\'"]?\w+\/\w+;\s*charset)=([a-z0-9-_]+)/i';
   if (preg_match($charset_pattern, $html)) {
     $html = preg_replace($charset_pattern, '\\1='.RCMAIL_CHARSET, $html);
-  }
-  else {
+  } else {
     // add head for malformed messages, washtml cannot work without that
     if (!preg_match('/<head[^>]*>(.*)<\/head>/Uims', $html))
       $html = '<head></head>'. $html;
@@ -728,9 +723,8 @@ function rcmail_wash_html($html, $p = array(), $cid_replaces)
   $washer = new washtml($wash_opts);
   $washer->add_callback('form', 'rcmail_washtml_callback');
 
-  if ($p['safe']) {  // allow CSS styles, will be sanitized by rcmail_washtml_callback()
-    $washer->add_callback('style', 'rcmail_washtml_callback');
-  }
+  // allow CSS styles, will be sanitized by rcmail_washtml_callback()
+  $washer->add_callback('style', 'rcmail_washtml_callback');
     
   $html = $washer->wash($html);
   $REMOTE_OBJECTS = $washer->extlinks;
@@ -883,6 +877,22 @@ function rcmail_washtml_callback($tagname, $attrib, $content)
 
 
 /**
+ * Callback function for HTML tags fixing
+ */
+function rcmail_html_tag_callback($matches)
+{
+  $tagname = $matches[2];
+
+  $tagname = preg_replace(array(
+    '/:.*$/',		// Microsoft's Smart Tags <st1:xxxx>
+    '/[^a-z0-9_-]/i',	// forbidden characters
+    ), '', $tagname);
+
+  return $matches[1].$tagname;
+}
+
+
+/**
  * return table with message headers
  */
 function rcmail_message_headers($attrib, $headers=NULL)