Inaccurate unnecessary_await_in_return

[ThinkDigitalSoftware]

Describe the issue Linter shows unnecessary_await_in_return if in an expression body, but not when the statement is a block body.

To Reproduce

 onInitialBuild: (store) async {
              await Future.wait(
                  widget.student.guardians.map(
                    (guardian) => guardian.inflate(),
                  ),
                );
            },

No issue here, but when converted to an expression body, it shows a warning.

StoreConnector<AppState, ViewStudentViewModel>(
            onInitialBuild: (store) async => await Future.wait( // warning here.
                  widget.student.guardians.map(
                    (guardian) => guardian.inflate(),
                  ),
                ),

Expected behavior if it's a simple refactor, then it shouldn't change the correctness of the statement. We should either not offer the refactor, or not show a warning here. I know the fat arrow is short for return, but that's not always desirable nor correct, especially when using void functions.

[pq]

Good catch. Thanks for the report!

[ThinkDigitalSoftware]

My pleasure!

[pq]

/cc @a14n

[a14n]

I don't see this as a bug.

In the first snippet there's no return and async/await are needed.

I rather think the refactoring should be changed to have convertions (block/expression) as follow:

Future f() => aFuture;
Future f() async {
  return await aFuture;
}

Future<void> v() => aVoidFuture;
Future<void> v() async {
  await aVoidFuture;
}

[ThinkDigitalSoftware]

In my case, the callback definition isn't async, but it supports it. Just like onPressed() callbacks. so I can't define that it's a Future. Not sure if that makes a difference

typedef OnInitialBuildCallback<ViewModel> = void Function(ViewModel viewModel)

[srawlins]

I agree with @a14n; the code in question doesn't have a return statement, so the lint rule does not apply.