TestStore runs not being atomic

[luismachado-sumer]

Description

Stumbling into a bit of an weird issue.

final class AppReducerTests: XCTestCase {
    func testAppReducer_Onboarding() async {
        let store = await TestStore(
            initialState: AppReducer.State()
        ) {
            AppReducer()
        } withDependencies: {
            $0.manager = .noop
        }

        await store.send(\.onLoad)

        await store.receive(\.onSetOnboardingVisible) {
            $0.isOnboardingVisible = true
        }

        await store.send(.onboarding(.dismiss))

        await store.receive(\.onSetOnboardingVisible) {
            $0.isOnboardingVisible = false
        }
    }

    func testAppReducer_User() async {
        let store = await TestStore(
            initialState: AppReducer.State()
        ) {
            AppReducer()
        } withDependencies: {
            $0.manager = .noop
        }

        await store.send(\.onLoad)
        await store.send(\.onAppear)
    }
}

If I run these two test separately they both work. However, if I run the full TestCase the 2nd one fails (seemingly picking up an event from the first test).

Unhandled actions: [
  [0]: .onSetOnboardingVisible(true)
]

Shouldn't the TestStore runs be atomic? Any suggestions on what to try?

Checklist

• I have determined whether this bug is also reproducible in a vanilla SwiftUI project.
• If possible, I've reproduced the issue using the main branch of this package.
• This issue hasn't been addressed in an existing GitHub issue or discussion.

Expected behavior

No response

Actual behavior

No response

Reproducing project

No response

The Composable Architecture version information

No response

Destination operating system

No response

Xcode version information

Xcode 16.4

Swift Compiler version information

[stephencelis]

@luismachado-sumer You'll need to provide a project reproducing the issue for us to determine if this is a library issue and help you debug the problem. In the code sample above the two test stores should be fully independent, and as far as XCTest is concerned, the tests are run in entirely different subprocesses, isolated from one another, so there should definitely be no influence between the tests, and whatever you're seeing is probably an entirely different problem.

[stephencelis]

@luismachado-sumer I'm going to convert this to a discussion since at the moment it's not clear whether this is a bug or not.