> On 23 Apr 2017, at 15:55, Ionuț G. Stan via swift-evolution
> <[email protected]> wrote:
>
> Hi all,
>
> I'm unsure if this is the appropriate venue to discuss this, so apologies if
> it's not.
>
> I was wondering if the issue of supporting as-patterns in switch statements
> has been brought up before or not, and what was the outcome of that.
>
> By as-patterns I mean the ability to bind sub-patterns to identifiers. See
> this question of mine on StackOverflow for an example:
> http://stackoverflow.com/questions/43566117/haskell-like-as-patterns-in-swift/43567041#43567041
>
> Maybe an equal sign could be used to denote that, so my example above would
> become:
>
> switch self {
> case .Int: return "int"
> case .Fun(let p = .Fun, let r): return "(\(p)) -> \(r)"
> case .Fun(let p, let r): return "(\(p) -> \(r))"
> }
>
> Note the `let p = .Fun` part.
>
> Alternatively, enhancing the `where` clause of a `case` to support the same
> features as an `if case` expression, would be an acceptable choice, too?
>
> case .Fun(let p, let r) where case .Fun = p:
> return "(\(p)) -> \(r)"
>
> Thanks for reading.
>
> --
> Ionuț G. Stan | http://igstan.ro | http://bucharestfp.ro
> _______________________________________________
> swift-evolution mailing list
> [email protected]
> https://lists.swift.org/mailman/listinfo/swift-evolution
I think the second (using where clauses to provide a more specific constraint)
is probably more consistent with the rest of the language.
switch self {
case .Int:
...
case .Fun(let outer_p, let r) where case .Fun(let inner_p,_) = outer_p, case
.Int = inner_p:
...
case .Fun(let p, let r) where case .Fun(_,_) = p:
...
case .Fun(let p, let r):
...
}
Looks pretty good IMO, although I do wonder if we couldn’t drop the second
“case”.
- Karl
_______________________________________________
swift-evolution mailing list
[email protected]
https://lists.swift.org/mailman/listinfo/swift-evolution
