Update to Swift 6

[fpseverino]

• Update package to Swift 6
• Remove all EventLoopFutures in favor of async/await
• Update JWTKit to v5
• Update to Swift Testing
• Adopt swift-format
• Update CI

[codecov]

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

Thanks for integrating Codecov - We've got you covered ☂️

[vamsii777]

Just a quick note to highlight that while many parts of the codebase, such as GoogleJWT and GitHub services, leverage modern async/await constructs, there are still sections, like the Imgur service, that use legacy Future-based approaches.

[fpseverino]

there are still sections, like the Imgur service, that use legacy Future-based approaches.

Those are basically dead code not visible to the user, they're not even part of a target and are stuck in very old versions of Imperial.
Reviving them requires a bit of work, since the protocols have changed, and I think it's best to do it in a separate PR

[ptoffy]

Overall this looks good to me. There's no major changes besides a/a instead of ELFs. Besides the parameters I'd try updating the existential return types to opaque types, that should work e.g for AsyncResponseEncodable. I'd also add formatting

[fpseverino]

I'm afraid that by activating formatting a lot of changes will be made that distract from the important ones. I'll do that in a separate PR

[ptoffy]

Let's also add Swift 6 language mode. I'm fine with the rest, not sure if @0xTim wants to chime in on something. I would also add some tests honestly, but that can be done later.

After thinking again I'm somewhat worried of the closures having to return opaques, as those are publicly passed in methods, but I guess we can see how that works in tests etc

[vamsii777]

Great work!

[fpseverino]

In the last few commits I:

• updated the remaining services from Future APIs to a/a and refactored them in their own target
• changed all classes to structs, since I don't think any of them needed to be a reference type (I could easily be wrong)
• made a lot of public APIs internal or package (I think they were made before the package access level)
• removed FederatedCreatable and OAuthService, as their only purpose IIUC was to get data (like user and email) from the services in a very convoluted way (even the Kodeco book uses a simpler custom solution), without considering that the endpoints they provided could be outdated and require some maintenance that I think we can avoid

Let me know what you think!

[0xTim]

Nothing blocking. I have thoughts about how much we're making package protected in case users want to extend Imperial to support other OAuth providers (especially internal ones that they can't upstream) but it's much easier to go that way than try and lock it down after. Lets get this into people's hands to start trying out

All requested changes have been addressed and resolved and the PR received an approval from Tim