It works if you remove "htmlspecialchars". Except I don't quite understand what you want to do with that class. Do you want to have a generic class where you can pass start and end text and extract everything between them and in addition choose if you want to encode all html entities?