Re: incomplete-object-patch I

Sounds like a good idea to me. Zeev On Fri, 23 Jun 2000, Stanislav Malyshev wrote: > SS>> attached is a first version of the incomplete object patch > SS>> which implements the discussed semantics. Suggestions (i.e. a > SS>> better wording for reserved_class_message()) and comments are > SS>> welcome. > > I'd make one change in this - that "reserved" class won't be in serializer > but passed from outside (from sessions, for example). This would allow: > > 1. Pass NULL and revert to giving ERROR on wrong class. > 2. Use different behaviours on resered class for different modules using > serializer. Now, even if you deserialize from string, it mentions sessions > (though call to deserializer could bear no realtion to sessions). > > This would require adding a parameter to serializing functions, but I > think this would pay itself. > > Also, I'd propose to compare to reserved class by poiner, not by name - > that's the same on meaning and should be much faster. > > -- Zeev Suraski <zeev@zend.com> http://www.zend.com/