[Sema] Fix issue 65330 Unhelpful error when missing contextually requ…

[jreference]

…ired as bridging conversion to AnyObject in structural position.

Minor formatting change in previous test case generic-subscript.swift

I'm not sure if these expressions should be considered to be contextual failures as well but it seemed reasonable. So this code change adds some custom logic in the contextual failure diagnosis for TupleElement locators when trying to diagnose AnyObject conversion failures.

Resolves #65330

[xedin]

Thank you for the contribution!

[xedin]

Looks much better now! I think after my inline comments are addressed it would be ready to go.

[jreference]

Thank you very much for the feedback and also for getting back on #57540 . I have made changes accordingly for this PR.

P.S. Do you happen to know of any other good starter bugs in swift-frontend?

[xedin]

Regarding starter bugs I can definitely find you a few, are you interesting in the same general area?

[jreference]

Regarding starter bugs I can definitely find you a few, are you interesting in the same general area?

Yes please, I'm interested in starter bugs within the compiler front-end (Lexing/parsing/semantic analysis etc)

[xedin]

@swift-ci please test

[jreference]

@xedin Can you please merge the PR?

I think I can merge it now that you have approved but https://github.com/apple/swift/blob/main/docs/HowToGuides/FirstPullRequest.md indicates contributors should ask reviewers to merge the PR. Not exactly sure what is the standard practice.

Thanks for all the help!

[xedin]

Feel free to merge yourself next time if the PR is approved and tests are green! I will leave some more starter bug issues in here on Monday.

[jreference]

Got it, thank you!

[xedin]

• Looks like there is some work still that needs to be done for [SR-10988] Suggest return when the last statement of a multi-statement function body would be a valid return value #53378

Example:

protocol Integer {
}
func foo() -> some Integer {
  let x = 0
  x
}

Doesn't produce a fix-it we'd expect.

• [SR-15111] “Type of expression is ambiguous” error when initializing a class #57437
• SwiftUI: Failed to produce diagnostic for expression when using keyword 'open' in some contexts #61039
• [SR-8213] Compiler may indicate "case will never be executed" even if it is unavoidable to execute the case in switch statement. #50745

[jreference]

@xedin

Thanks for the pointers! I'll probably have a look at #57437 first