A little bit of brainstorming here. In testing, specifically when comparing expected and actual values, I would like to ignore certain fields located deep in a nested data class structure. I know I can do that with assertj using

assertThat(actual).usingRecursiveComparison().ignoringFields("my.nested.field", MyClass::otherField.name).isEqualTo(expected)

but that is unsatisfactory for two reasons: • The ignore list is either not type-safe (plain strings) or is too verbose (property references) • The ignore list needs to be specified on the call site instead of on the declaration side. The latter point is important in the case of test fixtures which are shared among many tests and which contain some dynamic or computed data, such as dates or hashes. In such cases it would be much more convenient for me if I could ignore the field "on the declaration side, e.g. like this:

my TEST_DATA = MyClass(fixedField = TEST_VALUE, dynamicField = any())

and then use some special assertion function that understands that

any()

. For that to work, I imagine that the return type of

any()

should be both a subclass of the

dynamicField

type as well some marker type for the assertion function. E.g. in pseudo code

interface AnyMarker

inline fun <reified T> any() = object : T, AnyMarker {}

This looks very similar to how Spring creates proxy classes. That relies on the all-open compiler plugin. But I can't use a similar trick here because • It would be quite impractical to mark a class with some annotation for the plugin only for the testing use case • I'm not even sure I can "open" classes such as

String

which would be needed for the important ignore-hash use case. My questions are: • Does this idea make sense? • Is it feasible? If yes, how to effectively implement

any()

(or something similar)?

Trying to understand why the JS vs JVM behaviour is different when async coroutines are launched:

@Test
  fun test() = runTest {
    CoroutineScope(Dispatchers.Default).launch {
      delay(10.seconds)
      println("done")
    }
  }

JVM: completes immediately, without printing "done". JS: takes 10s and prints done. Is there a fundamental reason this is like this? Could JS just ignore the launched coroutine? (Or if it's a problem because it has side effects for other tests, shouldn't the bahaviour be the same for JVM?)

Hi there! Out of curiosity, has anyone found a framework for "web application testing" that they are happy with? Looking for something with a nice DSL to say like "click here", "submit this form with those values" etc.. I haven't found anything active enough, but I'm probably gonna keep looking

I would like to write a generic type-safe TestData class for JUnit parametrized tests. I imagine something like this:

import org.junit.jupiter.params.provider.Arguments

data class TestData<G, E>(val testName: String, val given: G, val expected: E) {
    override fun toString() = testName
}

fun <G, E> List<TestData<G, E>>.asArguments() = map { Arguments.of(it) }

fun myTestMethodTestData() = listOf(
  TestData("test case 1", 1, "1"),
  TestData("test case 2", 2, "2"),
).asArguments()

This works fine but isn't type-safe, i.e. nothing prevents me from writing

fun myTestMethodTestData() = listOf(
  TestData("test case 1", 1, "1"),
  TestData("test case 2", "2", 2),
).asArguments()

Is there any way how that can be achieved? At first I thought I could utilize

@NoInfer

to force the user to specify the type arguments explicitly, e.g.

asArguments<Int, String>()

, but of course that doesn't enforce anything about the types in the

TestData

class instance...

How can I assert that two double arrays are the same within some tolerance? There doesn't seem to be a parameter to specify this in kotlin.test, so it's failing due to float/double precision.

what's the fate of https://github.com/Kotlin/kotlinx-knit? I'd like to use it, but I see the latest release is from 2.5 years ago kodee sad

Hi everyone, Does anyone successfully implement https://github.com/Xray-App/xray-junit-extensions/ in android project for customizing test reports ?

hello, I have a problem with a UI test I want to write... I have a simple app: just a MainActivity that displays a list in a LazyColumn. The data is pulled from the MainViewModel which, in turns, ask it to a repository. I want to make a test for it, but the problem I am having is that I am not being able to configure the FakeRepository before the activity starts and call the viewmodel to fetch the data. I am using Hilt as a DI framework and "createAndroidComposeRule<MainActivity>()" in my test Is there a way to configure the fake before the activity start? or maybe delay the start of the activity until the fake is configured and after that make assertions with a composeRule other than the AndroidComposeTestRule?

I’m running in a weird error while testing android view models from junit tests. If I manually set

Dispatchers.setMain(UnconfinedTestDispatcher(testScheduler))

in my test function, it works. But if I use a Rule instead to do it, it fails. I do create the viewmodel from withing the test function and I do have the

@get:Rule

as the first thing in the test class. I really don’t understand what I’m doing wrong here. Any idea on what I am missing?

Hello! ✋ I would like to share my new tutorial, in which I implemented the Screenshot Testing inside a Compose Multiplatform App. If this is not the best channel to share it, I will delete this. Please, fell free to write me if you find some errors! 🙂 Any kind of feedback is welcome! https://medium.com/@munsra/using-screenshot-tests-in-a-compose-multiplatform-app-8fa171a3975d

hi all, is using the KSP or all-open plugin the only way to mock classes in Multiplatform project tests? currently, we are trying to migrate away from MockK since it doesn't support Multiplatform. we found that most mocking libraries either use KSP or all-open plugins, which require annotating the classes that need to be mocked. we’re not entirely comfortable with this approach, as it introduces test-related annotations/code into

commonMain

. right now, we’re leaning towards using Mokkery since it makes migration easier syntax-wise. however, it still requires us to annotate the classes for mocking. crossposting from #C3PQML5NU since it also make sense to ask here

kodee greetings Hey, I have a very simple question that's been driving me a little potty when using Kotlin Test because I have seen some conflicting opinions

```class FakeRepo: Repo {

private val users = mutableSetOf<String>()

fun addUser(name: String) {

users.add(name)

}

fun getUserCount(): Int {

return users.size

}

}

class SomeService(val repo: Repo) {

fun addUser(name: String) {

userRepository.addUser(name)

}

fun getUserCount(): Int {

return userRepository.getUserCount()

}

}```

If you write tests in this form:

```class SomeServiceTest {

private val repo = FakeRepo()

private val sut = SomeService(repo)

@Test

fun test1() {

service.registerUser("Alice")

val result = sut.getUserCount()

assertEquals(1, result)

}

@Test

fun test2() {

service.registerUser("Bob")

service.registerUser("Charlie")

val result = sut.getUserCount()

assertEquals(2, result)

}

}```

Each individual test will pass if it executes first, but wont pass if it's executed after the other. As the order tests are run isn't guaranteed, this is a common cause of issues with fakes that hold state.

Am I right here? The correct approach would be to init

sut

in a

@BeforeTest

method right? One of my colleagues mentioned

FakeRepo

would be recreated every time without

@BeforeTest

Anybody knows anything about Journeys with Gemini for Android UI Testing? Firebase AI Agent was released recently but I was wondering when Journeys would be released as well.

Hello, I'm looking for an assertion library or any tips to allow me to assert that a data class has all its fields equals to another data class, except for the possible lambdas. I don't care about the lambdas since they can't be tested anyway, a UI test will actually take care of that. But the fields like

label

and other stuff has great value to test. I would like to avoid overriding the equal function of the data class for obvious reasons. Any ideas ?

// Source code
data class Foo(
    val label: String,
    val onClicked: () -> Unit,
    val onValueChanged: (Int) -> Unit,
)

object FooMapper {
    fun map(
        label: String,
        viewModel: FooViewModel,
    ) = Foo(
        label = label,
        onClicked = { viewModel.onFooClicked() },
        onValueChanged = { viewModel.onValueChanged(it.toLong()) },
    )
}

class FooViewModel {
    fun onFooClicked() { /* whatever */ }
    fun onValueChanged(value: Long) { /* whatever */ }
}

// Unit tests
class FooMapperTest {
    @Test
    fun `foo map`() {
        // Given
        val label = "label"
        val fooViewModel = mockk<FooViewModel>()

        // When
        val result = FooMapper.map(label = label, viewModel = fooViewModel)

        // Then
        // !! What should I use there?
        assertEquals(
            Foo(label = label, onClicked = {}, onValueChanged = {}),
            result
        )
    }
}

Does anybody know when we can start using for Android testing?

I’m having drop testing a function that uses a withTimeoutOrNull(). Is there a recommended way of testing such code?

Hello 👋 Is there any good replacement for JUnit

ParameterizedTest

?

hi, is there a way to do deep comparison of two objects(of non data classes) in commonTest (something equivalent of

assertReflectionEquals

in java), i found

shouldBeEqualToComparingFields

in kotest assertions module but that is also jvm only.

kotlin-powerAssert = { id = "org.jetbrains.kotlin.plugin.power-assert", version.ref = "kotlin" }

plugins {
    ...
    alias(libs.plugins.kotlin.powerAssert)
}

import kotlin.test.Test


class ExampleUnitTest {
    @Test
    fun testFunction() {
        val hello = "Hello"
        val world = "world!"
        assert(hello.length == world.length)
    }
}

... No matter what I do, I'm getting just the basic

Assertion failed
java.lang.AssertionError: Assertion failed
	at foo.bar.ExampleUnitTest.testFunction(ExampleUnitTest.kt:11)
	at java.base/java.lang.reflect.Method.invoke(Method.java:569)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)

and not the right messages what am I doing wrong?

Maintainer of kotest here. Can you share anecdotes about flaky tests, what causes your tests to be flaky, how you resolve that, etc.. This will help us improve kotest. TIA!

For JS browser tests, is there a way to hook into the test infrastructure to start/stop a JVM server during the tests? Is anything like this possible?

Are there tools for mutation testing that are compatible with kotlin native? Like for when writing KMP libraries I can not just use Pitest if I got it correct since it’s a JVM thing

Have you found a way for jacoco ignoring compose previews except the annotation with

Generated

word inside?

Could we get the standard testing library to support

assume()

? Granted I'd prefer different naming, but that's less important than having it supported.

on JVM targets (but also in general) why do assertEquals with NaN arguments pass? I'd expect

kotlin.test.assertEquals(Double.NaN, Double.NaN, 1e-6)

to fail because

Double.NaN != Double.NaN

class FooTest {
	private val myDependency = Dependency(StandardTestDispatcher(testScheduler)) <----

	@Test fun foo() = runTest {
		// val dispatcher = StandardTestDispatcher(testScheduler)
		...
	}
}

I like to construct dependencies to the sut in properties, but

testScheduler

is only available from inside the

runTest

lambda, which is "too late" for me. How do I access

testScheduler

from a property?

this goes out to all the test enjoyers https://code.cash.app/burst-test-interceptors

FYI, we’re doing HTML snapshot testing in Redwood. I would love to get this into a standalone library but I haven’t done the work for that yet. Here’s what it looks like!

@Test
  fun happyPath() = runTest {
    val helloWorld = document.createElement("div").apply {
      appendElement("h1") {
        appendText("hello world")
      }
    }
    snapshotter.snapshot(helloWorld, "happyPath", Frame.None)
  }

ℹ️: testballoon TestBalloon 0.6.0-K2.2.0 has been successfully tested to work with Kotlin 2.2.20. With its variants, the latest TestBalloon release 0.6.0 continues to support all Kotlin versions since 2.0.0.

A little point... how could one use Robolectric with test balloon? https://robolectric.org/androidx_test/ seems to have it's own runner...