Bump to Swift 6 and fix Swift Syntax issues

[davdroman]

I've also made CI runs simpler.

[davdroman]

All tests passing. I know it's an enormous PR but it's mostly deletions (yay! 🎉) and it mostly modernizes things for Swift 6 and Swift Syntax.

[gohanlon]

Wow, thanks for digging into this. And like you said, this is quite the big PR!

• Have you by chance read my post here? Proposal: Addressing gaps in Swift Macro Testing pointfreeco/swift-macro-testing#18. I'm concerned that adopting mainstream swift-macro-testing removes test coverage, especially around fix-its, which is by far the most complex logic in the macro.

• Do you know if the PR's approach to CI will effect the resolution of swift-syntax? My (perhaps incorrect) assumption is that this setup will test against multiple versions of Swift, but will only test against the latest swift-syntax.

• I'd also like to hear your thoughts on dropping vs. keeping pre-Swift 6 support, and also dropping vs. keeping pre swift-syntax 600 support. I lean toward broader support based on the reasoning in this PointFree post: Being a good citizen in the land of SwiftSyntax. But I do agree that few would expect Swift 5.9 compatibility, and maybe same is true for swift-syntax 590.

[davdroman]

• Do you know if the PR's approach to CI will effect the resolution of swift-syntax?

You might want to take a look at this: https://github.com/Matejkob/swift-macro-compatibility-check. I saw it used here https://github.com/pointfreeco/swift-dependencies/blob/main/.github/workflows/ci.yml#L81

But I do agree that few would expect Swift 5.9 compatibility, and maybe same is true for swift-syntax 590.

I always go by the minimum version Apple accepts submissions for in the App Store, which right now is Xcode 26.0 = Swift 6.0. In Apple's eyes Swift 5.9 is "App Store obsolete", and so it's obsolete in my eyes too.

• I'm concerned that adopting mainstream swift-macro-testing removes test coverage, especially around fix-its, which is by far the most complex logic in the macro.

I think this might be a case of perfect getting in the way of good. To me, maintaining a fork of MacroTesting is more agonizing than not getting 100% coverage. But as the author it's of course ultimately up to you.