[Accepted with Modifications] SE-0286: Forward Scan for Trailing Closures

[John_McCall]

The review of SE-0286 has concluded.

Initial feedback on the review was generally positive on the concept, but several reviewers were troubled by the potential for source incompatibilities. The proposal author investigated options for maintaining compatibility better, and when they seemed to work out, the core team elected to amend the proposal during review. The community then expressed strong support for the revised proposal, and the feeling among reviewers was that there was no need for a second review. The core team discussed this and agreed. Accordingly, the revised proposal is accepted with modifications.

The schedule for this proposal and review was compressed quite a bit in order to try to meet the deadline for Swift 5.3. I'd like to thank the community for its patience and forbearance with this sort of unusual process.

As always, thank you for helping to make Swift a better language.

John McCall
Review Manager

[Jumhyn (Frederick Kellison-Linn)]

Thanks for your work managing this review, John. I really appreciated your communication throughout the process regarding the time constraints, and your preemptive solicitation of feedback about whether the community felt that foregoing the second review was appropriate. I think we've ended up in a fantastic place for Swift 5.3!

[Nevin]

Can we clarify the behavior of trailing closures for parameters of type Any, as per @xwu’s comment here?

@xwu

Mmm, this is interesting, and I’m not sure I’d appreciated that. Currently, Any works with trailing closure syntax:

func f(_ x: Any) -> String { "Hello, \(x)!" }
print(f { "World" })
// “Hello, (Function)!”

Unless there’s a compelling reason, I don’t know that we should break that, and I’m not sure it’s touched on in the proposal.

Edit:

The response from @Douglas_Gregor here:

@Douglas_Gregor

From the proposal:

The unlabeled trailing closure will be matched to the next parameter that is either unlabeled or has a declared type that structurally resembles a function type

so this is still well-formed.

Doug

does not seem to address what happens with labeled parameters of type Any:

func foo(f: Any) { print(f) }
foo{ return }

• • •

Also, as a forum usability issue, the fact that the proposal thread has been locked makes it very difficult to quote posts from there for the purpose of responding here. It would be nice if there were a way to quote a post from a locked thread, even without being able to post in that thread.

[John_McCall]

The quoting issue has been noted before, but there isn't anything we can do about it.