[swift-evolution] Why is the status of SE-0110 "implemented"?

[John McCall]

> On Jun 3, 2017, at 3:01 PM, Jens Persson via swift-evolution <swift-evolution at swift.org> wrote:
> I was notified that my SR-2216 and SR-296 had been "fixed as part of implementing SE-0110".
> 
> But AFAICT SE-0110 can't possibly be fully implemented in dev snapshot 2017-06-02, even though the status of SE-0110 is "Implemented (Swift 4)".

Yes, the reflective type system does not yet properly distinguish these cases; it's a bug and will be fixed, but probably not before 4.0.

John.

> 
> 
> The title of SE-0110 is
> "Distinguish between single-tuple and multiple-argument function types"
> 
> Keep that in mind while observing the following demonstration of the (Swift 4) behavior of Developer Snapshot 2017-06-02 (a):
> 
> func f(_ a: Int, _ b: Int) { print("\(a), \(b)") }
> func g(_ tuple: (Int, Int)) { print("\(tuple)") }
> f(1, 2) // 1, 2
> g((1, 2)) // (1, 2)
> // Working as expected.
> 
> // But:
> print(type(of: f) == type(of: g)) // true
> // IMHO this is not how to properly
> // "Distinguish between single-tuple and multiple-argument function types"
> 
> // And as if that wasn't enough (pay close attention):
> var (fCopy, gCopy) = (f, g)
> fCopy(1, 2) // 1, 2
> gCopy((1, 2)) // (1, 2)
> swap(&fCopy, &gCopy)
> fCopy(1, 2) // (1, 2)
> gCopy((1, 2)) // 1, 2
> // Crazy!
> 
> I'm trying to put this behavior and the following pieces together:
> 
> 1. The introdoction from SE-0110:
> "Swift's type system should properly distinguish between functions that take one tuple argument, and functions that take multiple arguments."
> 
> 2. The status of SE-0110:
> Implemented (Swift 4)
> 
> 3. The "fixed as part of implementing SE-0110" notification of my bug reports 
> SR-2216: Confusing behavior related to closure types and tuples
> SR-296: Fix inconsistencies related to tuples, arg/param lists, type params, typealiases
> 
> 
> Perhaps SE-0110 is just not fully implemented in dev snapshot 2017-06-02, but it is implemented somewhere, why else should it get the status "implemented"?
> 
> But then I read the following by Vladimir S (one of the SE-0110 authors?): 
> https://lists.swift.org/pipermail/swift-evolution/Week-of-Mon-20170529/036965.html <https://lists.swift.org/pipermail/swift-evolution/Week-of-Mon-20170529/036965.html>
> 
> It's a very interesting read, and it's nice to see these old inconsistencies are getting some attention (they deserve all the attention they can get), but I'm still confused.
> 
> Is it possible to implement SE-0110 without sorting out the above demonstrated inconsistencies?
> 
> Should status of SE-0110 really be "Implemented"?
> 
> /Jens
> 
> _______________________________________________
> 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/20170603/a95eb4ae/attachment.html>