Re: On objects, sessions, and class definitions

"Chris A. Henesy" wrote: > I see a few possible solutions to the problem. > The first would be to > store complete object session variables, i.e. store the actuall code contained > in the object's methods as part of the serialized object. > The second would be > to simplify the session_start function so that it returned the serialized > session data instead of unsearializing it back into the environment which > would allow it to be greped first for class names or something before being > unserialized. > The last would be some additional function like > session_class_names() which returns the class names of any serialized object in > the current session. 4) provide a way to register a callback function (or a chain of those) to be called with the class name as argument whenever unserialize encounters a yet unknown classname (which almost immediately will lead us to the necessiy namespaces?) -- Hartmut Holzgraefe hartmut@six.de http://www.six.de +49-711-99091-77 Besuchen Sie uns auf der CeBIT 2001 - in Halle 6 Stand F62/4