From: Cory H. <ch...@re...> - 2006-11-14 22:01:10
|
Justin Deoliveira wrote: >Hmmmm, well the whole reason for the second method is to make sure that >implementations do not make assumptions about what the return value >should look like. So all the work should really be done in >evaluate(Object), and then just converted on the fly by >evaluate(Object,Class). Making the method final enforces this nicley :). > >Can you give a more concrete example of what you need to do. If need be >we can break out a protected methods call eval( Object, Class ), which >does what evaluate( Object, Class ) does now, and then allow the latter >to be overridden by subclasses. > > Meh.... on second thought it makes more sense to simply have 2 different expressions rather than one that returns completely different objects depending on the context. Ramble, ramble... |