> In generation time you can have access to a selector, that's why I
> added to options. It was intended to drive the methods that would be
> proxy, even to skip the ones that wouldn't.
>
> When I was planning the implementation of this, I realized how this
> could be problematic for simple windsor scenarios. So I gave up
> implementing the selector at all. I can't remember all the reasons,
> but it should come once you consider all aspects:
> - proxying a type for the first time (scenario 1)
> - re-proxying a type for different scenario (logging + atm)
> - new container on same appdomain, reproxying
> - serialization/deserialization of proxy

So, you'd say selection of methods at code generation time doesn't
make sense for us anyway? I'm not sure. It would make a nice feature,
I think, but if nobody would actually use it...

The problem is that a runtime-based selection won't give us any
benefits over simply using a delegating interceptor, but it does
complicate code generation and ProxyGenerationOptions. I'd vote for
providing an out-of-the-box delegating interceptor rather than adding
this to the codegen _if_ we decide to only support this scenario.
(Even though Kryzstof has already implemented it; maintainance, tests,
etc. would be much easier with a delegating interceptor than
otherwise.)

Fabian

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Castle Project Development List" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/castle-project-devel?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to