Writing IOS Unit Tests

iOS Unit Testing (Swift Testing)

Three core principles ensure reliable unit tests:

1. Use Swift Testing framework - Modern @Test and #expect syntax for new tests
2. Keep container alive - Always capture ModelContainer when using SwiftData contexts
3. Safe unwrapping patterns - Never force unwrap; use guard with explicit failures

Scope

If input arguments provided:

Scope / file pattern: $ARGUMENTS

else:

Scope:

• Determine what unit tests need to be created or updated based on either: A - New or modified code in the current phase/PR (analyze changed files and map to testable components); or B - Coverage gaps identified by ./scripts/check-coverage.sh
• Investigate existing tests to avoid duplicative test classes or methods
• Check coverage tier requirements in coverage-config.json

⛔️ MANDATORY: When Tests Fail, Debug First

STOP. Before changing ANY code when a test fails, you MUST analyze the error:

Step 1: Read the Full Error Message

cat logs/latest/raw_output.txt

Step 2: Check Test Isolation

• Is the test failing due to shared state?
• Is ModelContainer being deallocated prematurely?

Step 3: Add Debug Output

print("DEBUG: value = \(value)")
#expect(value != nil, "Value was nil - check initialization path")

Step 4: Analyze BEFORE Changing Code

• Error message shows: What assertion failed and why
• Stack trace shows: Where in the code the failure occurred
• Debug output shows: Actual vs expected values

❌ DO NOT:

• Guess at why tests fail without reading errors
• Change production code to make tests pass
• Remove or weaken assertions

✅ ALWAYS:

• Read the full error output first
• Verify test data setup is correct
• Check ModelContainer lifecycle
• Only then make targeted fixes based on evidence

Framework Overview

• Framework: Swift Testing (modern, preferred) + XCTest (legacy)
• Config: project.yml (XcodeGen) or .xcodeproj
• Output Formatter: xcbeautify
• Unit Test Directory: {ProjectName}Tests/
• Naming Pattern: *Tests.swift

Swift Testing vs XCTest

✅ Swift Testing - Modern, cleaner syntax (USE THIS):

import Testing
@testable import AppName

@Test("User creation with valid data")
func testUserCreation() {
    let user = User(email: "test@example.com", name: "Test User")
    #expect(user.email == "test@example.com")
    #expect(user.name == "Test User")
}

❌ XCTest - Legacy syntax (avoid in new tests):

func testUserCreation() throws {
    let user = User(email: "test@example.com", name: "Test User")
    XCTAssertEqual(user.email, "test@example.com")
    XCTAssertEqual(user.name, "Test User")
}

Basic Test Structure

import Testing
@testable import AppName

@Suite("FeatureName Tests")
struct FeatureNameTests {

    @Test("Calculate next scheduled event time")
    @MainActor
    func testGetNextScheduledTime() async {
        // Given
        let (context, container) = createTestContext()
        _ = container  // Keep alive
        let viewModel = ScheduleViewModel()

        // When
        let nextTime = viewModel.getNextScheduledTime()

        // Then
        #expect(nextTime != nil, "Expected non-nil value")
    }
}

SwiftData Testing Patterns (CRITICAL)

See ./swiftdata-testing-patterns.md for complete patterns including:

• Creating test context with proper container lifecycle
• The critical relationship anti-pattern that causes crashes
• Why array assignment to relationships fails in tests

Safe Unwrapping Patterns

Avoid Force Unwrapping

// ❌ Avoid force unwrapping that can crash tests
let result = viewModel.calculateDose()!

// ✅ Use safe unwrapping with explicit test failures
guard let result = viewModel.calculateDose() else {
    #expect(Bool(false), "calculateDose() returned nil unexpectedly")
    return
}

// ✅ Alternative pattern with nil validation
let result = viewModel.calculateDose()
#expect(result != nil, "Expected calculateDose() to return a value")

Tolerance-Based Assertions

Time/Date Comparisons

// ❌ Exact time comparisons can be flaky
#expect(nextDose == expectedDate)

// ✅ Use tolerance for time-based assertions
let timeDifference = abs(nextDose.timeIntervalSince(expectedDate))
#expect(timeDifference < 60, "Time should be within 1 minute tolerance")

// ✅ For dose scheduling (more generous tolerance)
let timeDifference = abs(nextDoseTime.timeIntervalSinceNow)
#expect(timeDifference < 24 * 60 * 60, "Next dose should be within 24 hours")

Async Testing Best Practices

@MainActor for UI Components

// Always mark async tests with @MainActor when testing UI components
@Test("Verify medication profile selection updates state")
@MainActor
func testMedicationProfileSelection() async {
    let viewModel = OnboardingViewModel()
    let medication = Medication.semaglutide

    // Test async state changes
    await viewModel.selectMedication(medication)

    #expect(viewModel.selectedMedication == medication)
    #expect(viewModel.canProceedToNextStep == true)
}

Common Test Patterns

Testing Error Handling

@Test("Invalid dose amount throws appropriate error")
func testInvalidDoseThrowsError() async {
    let validator = DoseValidator()

    await #expect(throws: MedicationError.invalidDose) {
        try validator.validate(dose: -1.0, medication: .semaglutide)
    }
}

Testing State Changes

@Test("ViewModel updates state correctly on user action")
@MainActor
func testViewModelStateChange() async {
    let viewModel = OnboardingViewModel()

    // Initial state
    #expect(viewModel.currentStep == .welcome)
    #expect(viewModel.canProceedToNextStep == true)

    // Trigger state change
    await viewModel.proceedToNextStep()

    // Verify new state
    #expect(viewModel.currentStep == .medicationSelection)
}

Testing Computed Properties

@Test("MedicationProfile computes next dose time correctly")
func testNextDoseTimeComputation() {
    let profile = MedicationProfile.testProfile()
    profile.frequency = .weekly
    profile.lastDoseDate = Date().addingTimeInterval(-7 * 24 * 60 * 60) // 1 week ago

    let nextDose = profile.nextDoseTime

    // Should be approximately now (within 1 hour tolerance)
    let timeDifference = abs(nextDose.timeIntervalSinceNow)
    #expect(timeDifference < 3600)
}

Mocking Patterns

Protocol-Based Mocking

// Protocol
protocol NotificationCenterProtocol {
    func requestAuthorization(options: UNAuthorizationOptions) async throws -> Bool
    func add(_ request: UNNotificationRequest) async throws
}

// Production
extension UNUserNotificationCenter: NotificationCenterProtocol {}

// Mock
class MockNotificationCenter: NotificationCenterProtocol {
    var authorizationResult = true
    var addedRequests: [UNNotificationRequest] = []

    func requestAuthorization(options: UNAuthorizationOptions) async throws -> Bool {
        return authorizationResult
    }

    func add(_ request: UNNotificationRequest) async throws {
        addedRequests.append(request)
    }
}

What to Mock:

• External services (NotificationCenter, URLSession)
• System frameworks (HealthKit, StoreKit)
• Time-dependent operations

What NOT to Mock:

• Internal services (test with real implementations)
• SwiftData (use in-memory container)
• Pure functions

Test Factories for Consistency

Extension-Based Factories

extension User {
    static func testUser(
        email: String = "test@example.com",
        name: String? = "Test User",
        weight: Double = 70.0
    ) -> User {
        User(email: email, name: name, weight: weight)
    }
}

extension MedicationProfile {
    static func testProfile(
        genericName: String = "semaglutide",
        brandName: String = "Ozempic",
        currentDose: Double = 0.5
    ) -> MedicationProfile {
        MedicationProfile(
            genericName: genericName,
            brandName: brandName,
            currentDose: currentDose
        )
    }
}

Test Data Seeding

Using Preset Configurations

@Test("Test with seeded data")
@MainActor
func myTest() throws {
    let container = try TestDataSeeding.createTestContainer()
    let context = container.mainContext

    // Seed data with preset config
    let result = try TestDataSeeding.seedData(
        into: context,
        config: .thirtyDays  // 30 days, ~4-5 doses, 95% adherence
    )

    // Use seeded data
    #expect(result.doses.count > 0)
    #expect(result.adherenceRate >= 0.90)
}

Available Presets

Coverage Policy (5-Tier System)

See ./coverage-policy-tiers.md for complete coverage requirements including:

• Tier thresholds by category (90% for business logic, 85% for ViewModels, etc.)
• Coverage workflow commands
• Common coverage issues and solutions

Running Unit Tests

Using Project Scripts (Recommended)

# Run all unit tests
./scripts/test.sh unit 1

# Run specific test suite
./scripts/test.sh unit 1 MyServiceTests

# Run with coverage
./scripts/test.sh unit 1 --coverage

# Silent mode (log only, no console output)
./scripts/test.sh unit 1 --coverage --log-only

# View latest test results
cat logs/latest/raw_output.txt

Using xcodebuild Directly

# Run all unit tests
xcodebuild test \
  -scheme AppName \
  -destination 'platform=iOS Simulator,name=iPhone 17 Pro,OS=26.2' \
  -only-testing:AppNameTests \
  | xcbeautify

# Run with coverage enabled
xcodebuild test \
  -scheme AppName \
  -destination 'platform=iOS Simulator,name=iPhone 17 Pro,OS=26.2' \
  -enableCodeCoverage YES \
  -resultBundlePath .coverage/coverage.xcresult \
  -only-testing:AppNameTests \
  | xcbeautify

Swift Testing Framework Limitations

Unit Test Targeting Support:

• ✅ Target Level: -only-testing:AppNameTests (all unit tests)
• ✅ Suite Level: -only-testing:AppNameTests/MyServiceTests (specific test suite)
• ❌ Method Level: Swift Testing doesn't support individual method isolation

Integration Testing Patterns

Testing Component Interactions

@Test("AnalyticsService coordinates User, Dose, and MedicationProfile correctly")
@MainActor
func testAnalyticsServiceIntegration() async throws {
    let (context, container) = createTestContext()
    _ = container

    // Seed realistic test data
    let result = try TestDataSeeding.seedData(
        into: context,
        config: .thirtyDays
    )

    let analyticsService = AnalyticsService(context: context)

    // Test cross-model analytics coordination
    let summary = await analyticsService.generateUserSummary(for: result.user)

    #expect(summary.overallAdherence >= 0.90)
    #expect(summary.medicationEffectiveness.count > 0)
}

Complete Example

import Testing
@testable import AppName

@Suite("PharmacokineticsEngine Tests")
final class PharmacokineticsEngineTests {

    // MARK: - Test Data Setup

    func createTestContext() -> (ModelContext, ModelContainer) {
        let schema = Schema([User.self, MedicationProfile.self, Dose.self])
        let config = ModelConfiguration(
            schema: schema,
            isStoredInMemoryOnly: true,
            cloudKitDatabase: .none
        )
        let container = try! ModelContainer(for: schema, configurations: [config])
        return (ModelContext(container), container)
    }

    // MARK: - Tests

    @Test("Calculate concentration for single dose")
    @MainActor
    func testSingleDoseConcentration() throws {
        let (context, container) = createTestContext()
        _ = container
        let engine = PharmacokineticsEngine()

        // Given: Create test dose
        let dose = Dose(
            amount: 1.0,
            timestamp: Date().addingTimeInterval(-24 * 60 * 60), // 1 day ago
            medication: .semaglutide
        )

        // When: Calculate concentration
        let concentration = engine.calculateConcentration(
            doses: [dose],
            medication: .semaglutide,
            at: Date()
        )

        // Then: Verify calculation
        #expect(concentration > 0.0)
        #expect(concentration < 1.0) // Should have decayed
    }

    @Test("Handle empty dose array gracefully")
    @MainActor
    func testEmptyDoseArray() {
        let engine = PharmacokineticsEngine()

        let concentration = engine.calculateConcentration(
            doses: [],
            medication: .semaglutide,
            at: Date()
        )

        #expect(concentration == 0.0)
    }
}

Common Mistakes and Best Practices

See ./unit-test-anti-patterns.md for complete reference including:

• Common mistakes table with problems and solutions
• Do's and Don'ts for Swift Testing
• SwiftData-specific anti-patterns

Simulator Conflicts

When running tests, ensure no conflicting simulators are active:

xcrun simctl list

Use 1 of the 3 available iPhone 17 Pro simulators to avoid conflicts:

./scripts/test.sh unit 1
./scripts/test.sh unit 2
./scripts/test.sh unit 3