htmlpurifier/maintenance/PH5P.patch

--- C:\Users\Edward\Webs\htmlpurifier\maintenance\PH5P.php	2007-11-05 00:01:51.643585000 -0500
+++ C:\Users\Edward\Webs\htmlpurifier\maintenance/PH5P.new.php	2008-04-05 00:26:39.343160000 -0400
@@ -65,7 +65,7 @@
 
     public function __construct($data) {
         $data = str_replace("\r\n", "\n", $data);
-        $date = str_replace("\r", null, $data);
+        $data = str_replace("\r", null, $data);
 
         $this->data = $data;
         $this->char = -1;
@@ -211,7 +211,10 @@
         // If nothing is returned, emit a U+0026 AMPERSAND character token.
         // Otherwise, emit the character token that was returned.
         $char = (!$entity) ? '&' : $entity;
-        $this->emitToken($char);
+        $this->emitToken(array(
+            'type' => self::CHARACTR,
+            'data' => $char
+        ));
 
         // Finally, switch to the data state.
         $this->state = 'data';
@@ -708,7 +711,7 @@
         } elseif($char === '&') {
             /* U+0026 AMPERSAND (&)
             Switch to the entity in attribute value state. */
-            $this->entityInAttributeValueState('non');
+            $this->entityInAttributeValueState();
 
         } elseif($char === '>') {
             /* U+003E GREATER-THAN SIGN (>)
@@ -738,7 +741,8 @@
             ? '&'
             : $entity;
 
-        $this->emitToken($char);
+        $last = count($this->token['attr']) - 1;
+        $this->token['attr'][$last]['value'] .= $char;
     }
 
     private function bogusCommentState() {
@@ -1066,6 +1070,11 @@
                     $this->char++;
 
                     if(in_array($id, $this->entities)) {
+                        if ($e_name[$c-1] !== ';') {
+                            if ($c < $len && $e_name[$c] == ';') {
+                                $this->char++; // consume extra semicolon
+                            }
+                        }
                         $entity = $id;
                         break;
                     }
@@ -2084,7 +2093,7 @@
                     /* Reconstruct the active formatting elements, if any. */
                     $this->reconstructActiveFormattingElements();
 
-                    $this->insertElement($token);
+                    $this->insertElement($token, true, true);
                 break;
             }
             break;
@@ -3465,7 +3474,18 @@
         }
     }
 
-    private function insertElement($token, $append = true) {
+    private function insertElement($token, $append = true, $check = false) {
+        // Proprietary workaround for libxml2's limitations with tag names
+        if ($check) {
+            // Slightly modified HTML5 tag-name modification,
+            // removing anything that's not an ASCII letter, digit, or hyphen
+            $token['name'] = preg_replace('/[^a-z0-9-]/i', '', $token['name']);
+            // Remove leading hyphens and numbers
+            $token['name'] = ltrim($token['name'], '-0..9');
+            // In theory, this should ever be needed, but just in case
+            if ($token['name'] === '') $token['name'] = 'span'; // arbitrary generic choice
+        }
+        
         $el = $this->dom->createElement($token['name']);
 
         foreach($token['attr'] as $attr) {
@@ -3659,7 +3679,7 @@
         }
     }
 
-    private function generateImpliedEndTags(array $exclude = array()) {
+    private function generateImpliedEndTags($exclude = array()) {
         /* When the steps below require the UA to generate implied end tags,
         then, if the current node is a dd element, a dt element, an li element,
         a p element, a td element, a th  element, or a tr element, the UA must
@@ -3673,7 +3693,8 @@
         }
     }
 
-    private function getElementCategory($name) {
+    private function getElementCategory($node) {
+        $name = $node->tagName;
         if(in_array($name, $this->special))
             return self::SPECIAL;