Deriving collections of enum cases

[jtbandes]

swift-evolution discussion thread (other related threads are linked in the proposal).

Rendered Markdown.

[davidask]

This is a big one, and i gather a lot of people want this. Hoping to see this taken on!

[CodaFi]

I'm concerned that this proposal does not address the issue of differently-versioned cases in enums. For an enum like this

public enum QoSClass {
  @available(OSX 10.10, iOS 8.0, *)
  case background

  @available(OSX 10.10, iOS 8.0, *)
  case utility

  @available(OSX 10.10, iOS 8.0, *)
  case `default`

  @available(OSX 10.10, iOS 8.0, *)
  case userInitiated

  @available(OSX 10.10, iOS 8.0, *)
  case userInteractive

  case unspecified
}

Should the derived conformance look like this?

public static var allCases: [QoSClass] {
    if #available(OSX 10.10, iOS 8.0, *) {
      return [
        QoSClass.background,
        QoSClass.utility,
        QoSClass.`default`,
        QoSClass.userInitiated,
        QoSClass.userInteractive,
        QoSClass.unspecified,
      ]
    } else {
      return [ QoSClass.unspecified ]
    }
}

[CodaFi]

Moreover, do deprecated cases show up in this list?

[DougGregor]

I'm closing out PRs for proposals that are out of scope for Swift 4 stage 1. When we open up the process to the point where this proposal can be in scope, please re-open the pull request.

[kdawgwilk]

Is it possible this might fall into the scope of Swift 4 stage 2 now that priorities have been shuffled?

[jtbandes]

I asked for some clarification about this on Twitter: https://twitter.com/jtbandes/status/834633693138350080

It sounds like it's potentially feasible to do with existing metadata. I've also seen it said that proposals that come with a full or partial implementation are more likely to be accepted 😉 So perhaps if I get some time I'll look at what it would take to add this.

[DougGregor]

I'll bring it up with the full core team, but because this is purely additive, I don't think it meets the Swift 4 stage 2 criteria.

[DougGregor]

Yeah, this is purely additive, and doesn't meet the Swift 4 stage 2 criteria.

[louisdh]

Can this pull request be re-opened for Swift 5? 🙏

[tkremenek]

This can be considered for Swift 5 as long as a potential proposal has an implementation.

[CodaFi]

I have an implementation I would be willing to put up here.

// @jtbandes

[tkremenek]

@CodaFi @jtbandes re-opened

[AttilaTheFun]

Awesome, so glad to see the ball rolling on this!

[jtbandes]

Hi all, sorry for the long delay here. I've pushed a revised proposal with feedback incorporated from recent discussions. @DougGregor / @tkremenek, please advise if you see this as reviewable in its current form. @CodaFi's example implementation still stands, although it will need to be updated to match the naming choice in the final proposal, and to use associatedtype.

[CodaFi]

Excellent. I'll rebase and update my branch this evening.

[DougGregor]

Looks great!