[google_sign_in] Convert iOS tests to Swift

[stuartmorgan-g]

Converts the existing native unit tests from Objective-C to Swift.

This was done as a few steps:

• Direct conversion of the existing code (first pass by Gemini, but I checked it over and did some fixup).
• Converted from XCTest to Swift Testing, for Migrate iOS plugin tests from XCTest to Swift Testing flutter#180787
• Replaced global state with new loadGoogleServiceInfo and createTestPlugin helpers, because there was no reason the test needed to be so stateful, and removing the state makes the test-specific setup easier to follow (e.g., a subset of the tests didn't actually use the test-global plugin instance, but instead replaced it with a local copy, which made things confusing when moving between tests). It also allows adding suites.
• Added suites for each method.
• Combined some tests into parameterized tests.

Part of flutter/flutter#119103
Part of flutter/flutter#180787

Pre-Review Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
• I read the Tree Hygiene page, which explains my responsibilities.
• I read and followed the relevant style guides and ran the auto-formatter.
• I signed the CLA.
• The title of the PR starts with the name of the package surrounded by square brackets, e.g. [shared_preferences]
• I linked to at least one issue that this PR fixes in the description above.
• I updated pubspec.yaml with an appropriate new version according to the pub versioning philosophy, or I have commented below to indicate which version change exemption this PR falls under¹.
• I updated CHANGELOG.md to add a description of the change, following repository CHANGELOG style, or I have commented below to indicate which CHANGELOG exemption this PR falls under¹.
• I updated/added any relevant documentation (doc comments with ///).
• I added new tests to check the change I am making, or I have commented below to indicate which test exemption this PR falls under¹.
• All existing and new tests are passing.

Footnotes

1. Regular contributors who have demonstrated familiarity with the repository guidelines only need to comment if the PR is not auto-exempted by repo tooling. ↩ ↩² ↩³

[stuartmorgan-g]

@jmagman Do you want to review the Swift Testing aspect here since you've been looking at it? (This PR was prompted by wanting to get some hands-on experience with Swift Testing, and I'd been wanting to do the Swift conversion here already, so it seemed like a good opportunity to do both.)

[gemini-code-assist]

Code Review

This pull request is a great step forward, successfully converting the iOS unit tests from Objective-C to Swift and adopting the new Swift Testing framework. The code is well-structured, and the refactoring to reduce global state by introducing helper functions significantly improves test isolation and readability. I have a few minor suggestions to further enhance the clarity and robustness of the new Swift tests.

[jmagman]

I see warnings in Xcode:

Implicit capture of 'fakeUserProfile' requires that 'TestProfileData' conforms to 'Sendable'
Class 'TestProfileData' does not conform to the 'Sendable' protocol

Do the test fake helper classes inheriting from Objective-C classes need @unchecked Sendable?

[stuartmorgan-g]

I adjusted the test to only pass primitive types into the callback, rather than the test types.

[jmagman]

You can get pretty cute with test parameterization (if you want to in this review, or the mechanical change is fine too).

For example, you can collapse these two tests into one, with parameters:

Suggested change

	@Test func refreshTokensNoAuthKeychainError() async {
	let plugin = createTestPlugin()
	let fakeUser = addSignedInUser(to: plugin)
	let sdkError = NSError(
	domain: kGIDSignInErrorDomain, code: GIDSignInError.hasNoAuthInKeychain.rawValue,
	userInfo: nil)
	fakeUser.error = sdkError
	await confirmation("completion called") { confirmed in
	plugin.refreshedAuthorizationTokens(forUser: fakeUser.userID!) { result, error in
	#expect(error == nil)
	#expect(result?.success == nil)
	#expect(result?.error?.type == .noAuthInKeychain)
	confirmed()
	}
	}
	}
	@Test func refreshTokensCancelledError() async {
	let plugin = createTestPlugin()
	let fakeUser = addSignedInUser(to: plugin)
	let sdkError = NSError(
	domain: kGIDSignInErrorDomain, code: GIDSignInError.canceled.rawValue, userInfo: nil)
	fakeUser.error = sdkError
	await confirmation("completion called") { confirmed in
	plugin.refreshedAuthorizationTokens(forUser: fakeUser.userID!) { result, error in
	#expect(error == nil)
	#expect(result?.success == nil)
	#expect(result?.error?.type == .canceled)
	confirmed()
	}
	}
	}
	@Test(arguments: [
	(GIDSignInError.hasNoAuthInKeychain.rawValue, FSIGoogleSignInErrorCode.noAuthInKeychain),
	(GIDSignInError.canceled.rawValue, .canceled),
	])
	func refreshTokensKnownErrors(
	errorCode: Int,
	expectedErrorCode: FSIGoogleSignInErrorCode
	) async {
	let plugin = createTestPlugin()
	let fakeUser = addSignedInUser(to: plugin)
	fakeUser.error = NSError(domain: kGIDSignInErrorDomain, code: errorCode, userInfo: nil)
	await confirmation("completion called") { confirmed in
	plugin.refreshedAuthorizationTokens(forUser: fakeUser.userID!) { result, error in
	#expect(error == nil)
	#expect(result?.success == nil)
	#expect(result?.error?.type == expectedErrorCode)
	confirmed()
	}
	}
	}

[stuartmorgan-g]

I've parameterized several sets of tests. Unfortunately we lose the helpful test names (e.g. refreshTokensNoAuthKeychainError is now refreshTokensGIDSignInErrorDomainErrors > -4, FSIGoogleSignInErrorCode(rawValue: 2)), but I don't think that's probably a big deal. If it becomes annoying we can always add an extra fake initial parameter that's the test "name".

I also did a bit more refactoring to make the test fixture fully stateless so that I could adopt suites, and group each method's tests into a suite.

[LongCatIsLooong]

I thought FSIGIDGoogleUser is now a protocol so it is supposed to be an existential type here (any FSIGIDGoogleUser?)?

[stuartmorgan-g]

Thanks, my Swift experience is still thin, and I'd missed the memo on the switch to explicit existentials in more recent versions of Swift. I think I've fixed all the cases here.

[LongCatIsLooong]

LGTM. On a side note I didn't know they were planning to enforce that, created flutter/flutter#181005

[stuartmorgan-g]

I forgot Swift imports don't like Obj-C submodules. I'm not sure why it was failing in CI but working locally (Xcode version?) but to resolve it I just removed the modulemap, so there's no longer a distinction between test and non-test headers; this is the same thing we've had to do with some other plugins.

It shouldn't actually affect anything, since nobody should be importing the native framework anyway, but also it'll be temporary since I plan to convert the plugin to Swift shortly anyway.

[stuartmorgan-g]

(I did add a version bump though, since it's no longer a test-only change, even though I don't expect any client impact.)