On Nov 25, 2008, at 5:27 PM, Gili wrote:
> Paul Sandoz (via Nabble) wrote:
> > In any case you are probably right about there being cases where
> > copying will not work. There is really no way to satisfy all
> > requirements except by actually re-writing Java class byte code at
> > runtime, but that may still have side effects w.r.t. Java
> reflection,
> > depending on what you do.
>
> There is no need to resort to these tricks.
I was referring to the general case.
> All you'd need to do is add
> getInjectableClass and use it in UriInfo. The implementation would be
> very fast. How about we give it a try? If it doesn't work we can
> always
> take it out again...
>
Do you mind if we have a go at after the 1.0.1 release that will occur
on Monday?
I agree we probably can do something to support it in this case in an
IoCComponentProviderFactory is registered. If there are well-defined
and consistent rules for proxying that occur for an IoC framework can
we support a method on IoCComponentProviderFactory instead?
Paul.