Swift Testing

Swift Testing Framework

Modern testing with Swift Testing framework. No XCTest.

Overview

Swift Testing replaces XCTest with a modern macro-based approach that's more concise, has better async support, and runs tests in parallel by default. The core principle: if you learned XCTest, unlearn it—Swift Testing works differently.

References

• Apple Documentation
• Migration Guide

Core Concepts

Assertions

Optional Unwrapping

let user = try #require(await fetchUser(id: "123"))
#expect(user.id == "123")

Test Structure

import Testing
@testable import YourModule

@Suite
struct FeatureTests {
    let sut: FeatureType
    
    init() throws {
        sut = FeatureType()
    }
    
    @Test("Description of behavior")
    func testBehavior() {
        #expect(sut.someProperty == expected)
    }
}

Assertion Conversions

Error Testing

#expect(throws: (any Error).self) { try riskyOperation() }
#expect(throws: NetworkError.self) { try fetch() }
#expect(throws: NetworkError.timeout) { try fetch() }
#expect(throws: Never.self) { try safeOperation() }

Parameterized Tests

@Test("Validates inputs", arguments: zip(
    ["a", "b", "c"],
    [1, 2, 3]
))
func testInputs(input: String, expected: Int) {
    #expect(process(input) == expected)
}

Warning: Multiple collections WITHOUT zip creates Cartesian product.

Async Testing

@Test func testAsync() async throws {
    let result = try await fetchData()
    #expect(!result.isEmpty)
}

Confirmations

@Test func testCallback() async {
    await confirmation("callback received") { confirm in
        let sut = SomeType { confirm() }
        sut.triggerCallback()
    }
}

Tags

extension Tag {
    @Tag static var fast: Self
    @Tag static var networking: Self
}

@Test(.tags(.fast, .networking))
func testNetworkCall() { }

Common Pitfalls

1. Overusing #require — Use #expect for most checks
2. Forgetting state isolation — Each test gets a NEW instance
3. Accidental Cartesian product — Always use zip for paired inputs
4. Not using .serialized — Apply for thread-unsafe legacy tests

Common Mistakes

1. Overusing #require — #require is for preconditions only. Using it for normal assertions means the test stops at first failure instead of reporting all failures. Use #expect for assertions, #require only when subsequent assertions depend on the value.

2. Cartesian product bugs — @Test(arguments: [a, b], [c, d]) creates 4 combinations, not 2. Always use zip to pair arguments correctly: arguments: zip([a, b], [c, d]).

3. Forgetting state isolation — Swift Testing creates a new test instance per test method. BUT shared state between tests (static variables, singletons) still leak. Use dependency injection or clean up singletons between tests.

4. Parallel test conflicts — Swift Testing runs tests in parallel by default. Tests touching shared files, databases, or singletons will interfere. Use .serialized or isolation strategies.

5. Not using async naturally — Wrapping async operations in Task { } defeats the purpose. Use async/await directly in test function signature: @Test func testAsync() async throws { }.

6. Confirmation misuse — confirmation is for verifying callbacks were called. Using it for assertions is wrong. Use #expect for assertions, confirmation for callback counts.

Converted and distributed by TomeVault — claim your Tome and manage your conversions.