Node::lazyAttach shouldn't lie about being attached

Source/core/html/HTMLPlugInImageElement.cpp

 /*
  * Copyright (C) 2008, 2011, 2012 Apple Inc. All rights reserved.
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Library General Public
  * License as published by the Free Software Foundation; either
  * version 2 of the License, or (at your option) any later version.
  *
  * This library is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ skipping 137 matching lines @@
         RenderImage* image = new (arena) RenderImage(this);
         image->setImageResource(RenderImageResource::create());
         return image;
     }
 
     return new (arena) RenderEmbeddedObject(this);
 }
 
 void HTMLPlugInImageElement::willRecalcStyle(StyleChange)
 {
-    // FIXME: Why is this necessary?  Manual re-attach is almost always wrong.
+    // FIXME: setNeedsWidgetUpdate is used by subclasses to signal that they need
+    // their renderer replaced on the next style recalc. It's not clear why they
+    // don't just do lazyReattach instead. This code looks like a terrible hack.
     if (!useFallbackContent() && needsWidgetUpdate() && renderer() && !isImageType())
         reattach();
 }
 
 void HTMLPlugInImageElement::attach()
 {
     PostAttachCallbackDisabler disabler(this);
 
     bool isImage = isImageType();
     
@@ skipping 58 matching lines @@
 
 void HTMLPlugInImageElement::subframeLoaderWillCreatePlugIn(const KURL& url)
 {
     LOG(Plugins, "%p Plug-in URL: %s", this, m_url.utf8().data());
     LOG(Plugins, "   Loaded URL: %s", url.string().utf8().data());
 
     m_loadedUrl = url;
 }
 
 } // namespace WebCore