I have created a branch to start working on a genericized Proxy API. However, I'm running into an issue that I'm not quite sure how I should deal with. I want to make the methods on ProxyFactory use the varargs language feature. For example:
public Object createDelegatorProxy( ObjectProvider delegateProvider, Class... proxyClasses ); However, I also want to add in a nice helper method like this: public <T> T createDelegatorProxy( ObjectProvider<T> delegateProvider, Class<T> proxyClass ); Now, it's advised to not override varargs methods, but this is a special case. It doesn't really matter which version the compiler picks. The second method's implementation merely delegates to the first method. It's "syntactic sugar." So, if I want to do this in my code: Echo echo = proxyFactory.createDelegatorProxy(echoProvider, Echo.class); Then, I won't have to do any casting at all since I'm using the single-class version of the method. However, if I add another class: Echo echo = proxyFactory.createDelegatorProxy(echoProvider, Echo.class, Serializable.class); This won't compile. You need to cast the return type. It looks like implementing it this way is actually o.k. as far as the compiler is concerned, but I just wanted to get some input from some other folks. Should I just leave the method names the same and folks who use just a single class instance get the type safety for free? Or, do you guys see some ambiguity here that folks might not understand? James --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
