test(ui): make ui tests fail when the behaviour they cover breaks

[swolfand]

What

The source/ui half of #985, re-applied on current main. #985 couldn't be rebased. #1038 moved the prebuilt UI onto the Clerk.auth facade, so its AuthStartViewModel tests were written against SignIn.create/SignUp.create mocks that no longer exist. This PR ports the auth tests that were worth keeping onto Clerk.auth (auth.signIn {}, auth.signUp {}, auth.signInWithOAuth/signUpWithOAuth) and carries over the profile and organization changes, which still apply cleanly.

Only test files change. The AuthStartView.kt formatting change from #985 is dropped.

Why each change now fails on a regression

Every behaviour listed below was mutation-checked: the production line was broken, the test was confirmed to fail, and the line was restored.

AuthViewModelTest, ported to Clerk.auth

• startAuthWithSignInOrUpModeSurfacesNonFallbackSignInErrorsWithoutSignUp replaces a test that stopped at Loading with cancelAndIgnoreRemainingEvents(). It now asserts the Error("Invalid email") state, the identifier sent to auth.signIn, and that auth.signUp is never called. It fails if sign-in-or-up falls back to sign-up on any error, not just form_identifier_not_found.
• startAuthWithSignUpChoosesEmailOrUsernameParamsFromIdentifierFormat sends seven identifiers through auth.signUp {} and checks which ones become email and which become username. It fails if the isEmailAddress branch is inverted.
• socialOAuthCanStartWithSignUp now also asserts the final OAuthState.SignUpSuccess. It fails if the redirect path drops the sign-up result.
• Removed authModeEnumShouldHaveCorrectValues and oauthProviderShouldHaveExpectedValues. They only checked enum .name strings, so they couldn't fail on a behaviour change.
• The phone-identifier test (startAuthUsesPhoneNumberAsIdentifierOnlyWhenPhoneFieldIsActive) was already ported by refactor(ui): move the prebuilt UI onto the Clerk.auth facade #1038 and checks the resolved identifier, so it's unchanged.

Sign-in factor code

• SignInAttemptHandlerTest: the failure tests now assert that the "boom" error message reaches onErrorCallback. One of them set receivedError = errorResponse inside the callback and then compared it with itself. All 7 failure tests fail if the handler stops passing errorMessage.
• SignInFactorCodeViewModelTest:
  • Removed the two Turbine Idle→Loading tests. They asserted only the Loading emission, which other tests already assert together with the outcome. Removing Loading from prepare still fails 5 tests, and removing it from attempt still fails 4.
  • "reuse active second-factor verification" now asserts the state stays Idle, matching its first-factor twin. Both fail if Loading is set before the reuse check.
  • The "should throw" names are fixed to "should set NotStarted".
• SignInPrepareHandlerTest: the null-ID tests from main are kept. Main already verifies sendPhoneCode/sendEmailCode there, so test: audit the unit test suite for tests that don't test anything #985's removal no longer applies.

Other auth tests

• AuthStateEffectsTest moves from Paparazzi to the compose test rule. It now asserts that the snackbar shows the error message and that the back stack isn't cleared.
• AuthStateForceMfaTest adds a test that the created session wins over the fallback.
• AuthStartViewHelperTest compares autofill hints instead of asserting what the content type isn't.
• SignInFactorOneViewTest adds a test that resolveFirstFactor switches to the prepared factor when the fallback is no longer supported.

Profile and organization view models

• UserProfileChangePasswordViewModelTest verifies the exact UpdatePasswordParams. It fails if signOutOfOtherSessions is ignored.
• UpdateProfileViewModelTest verifies the update/setProfileImage/deleteProfileImage params and the user.get() refresh. It fails if the refresh is skipped.
• OrganizationVerifiedDomainsViewModelTest adds the domainsEnabled = false case. It fails if canLoadDomains ignores domainsEnabled.
• EmailViewModelTest verifies the primaryEmailAddressId update.
• UserProfileVerifyViewModelTest verifies the code sent to attemptVerification/attemptTotpVerification and that backup codes reach the Verified state. Before, they were stubbed as null.
• UserProfilePasskeyViewModelTest verifies createPasskey() actually ran. Before, its "no events" check passed even if nothing happened.
• AddEmailViewModelTest verifies that the immutable-email path never calls createEmailAddress.
• DeleteAccountViewModelTest verifies that local biometric credentials are kept when deletion fails.
• OrganizationProfileActionConfirmationViewModelTest counts client refreshes on success and failure.
• UserProfileAccountViewModelTest uses MainDispatcherRule, so Dispatchers.Main no longer leaks into other classes.
• PhoneInputUtilsTest checks real flags for upper- and lower-case region codes instead of a made-up ab region.

Removed because they couldn't fail

• CountryInfoTest: tested data-class equals/toString.
• CustomRouteNavKeyTest: duplicated the serialization round-trip.
• LastUsedAuthTest: an exact duplicate of biometricCredentialCountsTowardVisibleMethodCount.
• CustomRowGatingTest: a redundant all { BuiltIn } check.

Left out

• Snapshot goldens. test: audit the unit test suite for tests that don't test anything #985 removed two OrganizationProfile error-state snapshot tests whose goldens are byte-identical to the non-error states (organizationVerifiedDomainsErrorState = …AddDomain, organizationActionDeleteError = …DeleteReady), so the error message isn't actually rendered. ./gradlew :source:ui:verifyPaparazziDebug doesn't pass on this branch with that change applied: 21 failures, all in other tests (sub-0.1% rendering diffs, stale OrganizationProfile goldens, and tests with no golden). So it's not included here. The same goes for test: audit the unit test suite for tests that don't test anything #985's re-recorded goldens and the ClerkTopAppBar/passwordSection_Light snapshot fixes, which need re-recording against a passing baseline.

Results

• :source:ui:testDebugUnitTest: 505 run, 0 failed.
• spotlessApply, :source:api:detekt and :source:ui:detekt are clean.

The API test half is in #1049.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Tests
  • Expanded authentication and sign-in coverage for error messages, factor selection, session resolution, signup identifier handling, and state transitions.
  • Added checks for account and profile actions, including organization deletion, email updates, passkey creation, password changes, verification, and credential retention after deletion failures.
  • Clarified checks for phone-country handling, custom profile rows, and behavior when domain loading is unavailable.
  • Updated UI tests to verify snackbar behavior and email-and-username autofill hints.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository: clerk/clerk-android/.coderabbit.yaml
• Review profile: CHILL
• Plan: Team
• Run ID: 5b9b068c-3a1b-407c-91f7-118c757655a8

📥 Commits

Reviewing files that changed from the base of the PR and between 91ed575 and f46f056.

📒 Files selected for processing (3)

• source/ui/src/test/java/com/clerk/ui/userprofile/account/UserProfileAccountViewModelTest.kt
• source/ui/src/test/java/com/clerk/ui/userprofile/security/delete/DeleteAccountViewModelTest.kt
• source/ui/src/test/java/com/clerk/ui/userprofile/security/passkey/UserProfilePasskeyViewModelTest.kt

Limit details: You’ve used all 6 included reviews currently available. Your 48 included PR review attempts over the past 7 days set your current allowance at 6 reviews per hour.

---

📝 Walkthrough

Walkthrough

This pull request updates UI tests across authentication, sign-in, organization and user profiles, input utilities, and custom rows. The changes revise assertions, add call-count checks, adjust test setup, and remove selected tests.

Changes

UI test coverage

Layer / File(s)	Summary
Authentication behavior tests source/ui/src/test/java/com/clerk/ui/auth/*	Tests check combined autofill hints, snackbar and reset behavior, MFA session selection, authentication parameters, and OAuth signup state. Some enum-value and biometric-credential assertions were removed.
Sign-in factor and code tests source/ui/src/test/java/com/clerk/ui/signin/*	Tests check prepared-factor resolution and error callback messages. View-state test names and assertions were updated, and two state-transition tests were removed.
Organization profile tests source/ui/src/test/java/com/clerk/ui/organizationprofile/*	Tests check client refresh counts after organization deletion and the state when verified domains are disabled.
User profile operation tests source/ui/src/test/java/com/clerk/ui/userprofile/*	Tests add API call-count and argument checks for email, deletion, passkey, password, profile-update, and verification operations. Account tests use a main-dispatcher rule.
Input and custom-row tests source/ui/src/test/java/com/clerk/ui/input/*, source/ui/src/test/java/com/clerk/ui/userprofile/custom/*	Input tests revise flag and country expectations and remove equality and string-format checks. Custom-row tests assert an exact ordered result; a route-key deserialization test was removed.

Priority: ➖ Normal

Estimated code review effort: 3 (Moderate) | ~20 minutes

Change: Other

Suggested reviewers: mikepitre

Merge Risk: ⚪ Minimal · up to f46f0

The reported UI unit tests and static checks are clean, and the snapshot failures are reported in other tests. No actionable merge-blocking risk is established; normal CI checks remain appropriate.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 1.28% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 78 functions across 19 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main change: strengthening UI tests so they detect behavior regressions.
Description check	✅ Passed	The description explains what changed, why the changes were made, which tests to review, and the reported test results. It also covers excluded snapshot changes. The template’s separate “What & why” a…
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

• Fix all pre-merge checks with AI

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

_{Comment @coderabbitai help to get the list of available commands.}