[swift-evolution] I think we need more to our type calculus

[Slava Pestov]

> On Feb 8, 2017, at 12:09 PM, Daryle Walker via swift-evolution <swift-evolution at swift.org> wrote:
> 
> I’ve been trying to get the maximum of a list of counts. I started by using “myCollection.map { $0.myCountProperty }.reduce(0, max)”. Then I thought I should really use the value closest to negative infinity as the base. The problem is how to get that value. The current hard-coded “0” works because the count type ultimately aliases “Int”, but shouldn’t have to count on that. I put in “MyCountType.min”, where “MyCountType” is hard coded from the docs of “myCountProperty”, but I shouldn’t have to do that either.
> 
> There is a “type(of:)” global function, which I did use. But not only did I have to make up an expression to go in there, its return value, some sort of meta-type stuff I don’t understand, can’t be used on the right side of a “typealias” construct. I ultimately used “let lowestCount = type(of: anElement.myCountProperty).min” to get what I needed.
> 
> Some queries/requests:
> 
> 1. Is the expression within “type(of:)” evaluated?

Yes. Eg,

class Base {}
class Derived : Base {}

func foo() -> Base { return Derived() }

type(of: foo()) // this is Derived.self, not Base.self

> 2. I think we need the equivalent of "std::declval” from C++. Especially if the answer to (1) is yes.
> 3. Why isn’t the return type(?) of “type(of:)” compile-time enough to be used as a type-alias?

type(of:) is not a real function because it’s return type depends on the input type in a funny way. If the input type T is an existential, then type(of:) returns an existential metatype; otherwise, it returns a concrete metatype. So it’s almost as if the signature is (T) -> (T.Type), but not quite.

Slava

> 4. Should we add something like the C++ type-traits collection of generic structures to give use type information as type-aliases or type-level instances?
> 
> — 
> Daryle Walker
> Mac, Internet, and Video Game Junkie
> darylew AT mac DOT com 
> 
> _______________________________________________
> swift-evolution mailing list
> swift-evolution at swift.org
> https://lists.swift.org/mailman/listinfo/swift-evolution

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20170208/aeb99c1d/attachment.html>