brian.goetz at oracle.com
Sat Dec 19 22:20:48 UTC 2015
> OK. I should have first asked whether there are plans to allow forms
> <any T> boolean equals(T x)
> including as sugar for some superation-like construction.
Yes. That's totally valid (any is a modifier for a type variable
declaration, whether introduced at the class or method level.) Its
worth noting that there's some degree of extra dispatch cost for generic
instance methods (a separate topic, but something to keep in mind) which
is a slight negative for going generic on such a critical method as
equals(). But yes, that's valid.
> Right. I agree that the signature must compatibly change,
> but not necessarily that anything else does.
OK, we're on the same page. I think we're coming at the "superate"
concept from different directions, but both roads lead there.
More information about the valhalla-spec-experts