I'm surprised that https://kotest.io/docs/framework/project-config.html#parallelism does not document how it plays together with https://docs.gradle.org/current/userguide/performance.html#a_run_tests_in_parallel. Any hints on that?

I was thinking of doing something like this:

val result: MyResultType
mockkObject(SomeObject) {
    every { SomeObject.foo() } returns "bar"
    result = callUnitUnderTest()
}

result.a shouldBe 1
result.b shouldBe 2

The idea was to keep the time the object is mocked to a minimum. The object doesn't have to remain mocked while I'm doing the assertions. Unfortunately, the Kotlin compiler gives an error saying

result

is reassigned inside the lambda and uninitialized outside the lambda. This is because it doesn't know that the lambda is guaranteed to be called exactly once. This can be fixed by the use of contracts. Is there any chance it can be done in a future release?

Is there any reason not to bump ktor dependency for kotest-assertions-ktor to 3.x?

Hello. We are in the middle of upgrading to Kotlin 2.1 and we ran into an issue with kotest 5.9 that forced us to try the 6.0 milestones. However, now I am seeing that IsolationMode.InstancePerLeaf has been removed. Is this expected? I couldn’t find anything in the changelogs. If its expected is there a “rough draft” of the guide for migration?

If you have time, please consider upvoting this issue to get Junie to be able to run the Kotest tests it writes. That's apparently too much to ask for 🤦‍♂️ https://youtrack.jetbrains.com/issue/JUNIE-239/Add-support-for-Kotest Thanks @Emre for making.

I have some flaky tests. With JUnit I just use IntelliJ's runnier to repeat until failure. With Spock there's

@RepeatUntilFailure

. What's the Kotest equivalent?

Been using Kotest for some time now and gotten pretty fluent at writing Kotest extensions and Matchers. Just this week I wrote slightly more complex Matcher that I needed to use in both a positive (

shouldBe ….

) and negative

(shouldNotBe …

) which made me realise every Kotest matcher is `invert()`able! To my surprise neither Hamcrest nor AssertJ matcher have such capability. Really nice! 👏

Hi! We're blocked by https://github.com/kotest/kotest/issues/4895, as we really need to have tests running on all targets with the latest Kotlin EAP. Does anyone familiar with the Kotest Codebase have any pointers as to why only the JVM runner ist currently working?

can someone answer this issue https://github.com/kotest/kotest/issues/4905

withData can't cope with a float value in its name function? I have a withData structure like this

val someFloats = listOf(1F, 2F, 4.2F)

withData<Float>({"Testing $someFloat"}, someFloats) { someFloat -> &tc

and the compiler complains about $someFloat in the name function (with our without .toString()). Full(er) disclosure, it's the seventh nested withData, but without the name function it works, and I've tried reducing the number of nests and it still complains. Bug?

I tried to use Kotest with Spring, but I only found the

AnnotationSpec

working well.

@Import(TestcontainersConfiguration::class)
@SpringBootTest(webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT)
class DemoApplicationAnnotationSpec : AnnotationSpec() {
    @LocalServerPort
    private var port: Int = 8080
    lateinit var client: WebClient

    @BeforeEach
    fun setUp() {
        client = WebClient.create("<http://localhost>:$port")
    }

    @Test
    fun `get all posts`() {
        client.get()
            .uri("/posts")
            .accept(MediaType.APPLICATION_JSON)
            .exchangeToFlux {
                assertThat(it.statusCode()).isEqualTo(HttpStatus.OK)
                it.bodyToFlux(Post::class.java)
            }
            .test()
            .expectNextCount(2)
            .verifyComplete()
    }
}

How to convert this to use

StringSpec

FuncSpec

,etc., especially I do not know where to use the Spring special annotation to inject beans, such as

LocalServerPort

,

MockkBean

, etc. I have tried to use LocalServerPort in the test class body, but it can not be accessed in the test block.

xxxStringSpec:StringSpec({
   //1. @LocalServerPort is NOT allowed here
   //2. The port injected in the class body(below) cannot be accessed here.

}){
    @LocalServerPort
    private var port: Int = 8080
}

I'm getting a strange "module-missing" error with Kotest M4. Is this something that still needs to be published?

Could not find io.kotest:kotest-framework-multiplatform-plugin-embeddable-compiler:6.0.0.M4

known issue. use latest 6.0.0-SNAPSHOT

Is there a known issue with M3 on compileTestKotlinIosArm64 in 2.2.0? I'm getting:

e: java.lang.NoSuchMethodError: 'org.jetbrains.kotlin.ir.expressions.IrConstructorCall org.jetbrains.kotlin.ir.builders.ExpressionHelpersKt.irCall(org.jetbrains.kotlin.ir.builders.IrBuilderWithScope, org.jetbrains.kotlin.ir.symbols.IrConstructorSymbol)'

(ah, looks like it is: https://github.com/kotest/kotest/issues/4870)

@Bernd Prünster Now that s01 is gone, how do I actually get a hold of the snapshot version of kotest?

maven { url = uri("<https://central.sonatype.com/repository/maven-snapshots>") }

doesn't seem to do anything.

Kotest Team needs to publish to the new endpoint

That can be a giant headache, make sure you don't make the same mistakes I made: https://kotlinlang.slack.com/archives/C2YH04E3S/p1750990424590899?thread_ts=1745056051.782729&cid=C2YH04E3S

(everything I said there probably also applies to the central.sonatype.com/repository/maven-snapshots repo as well)

@Alexander Ioffe if you have time to create an issue and reference your learnings, I'd appreciate it.

An issue for moving to the new snapshot repo or for migrating away from sonatype-s01 in general?

The former, I think..? Is it relevant to migrate away from s01?

s01.oss.sonatype.org is being decommed as we speak, same with oss.sonatype.org

You need to move to move both snapshots and releases. The former to ossrh-staging-api.central.sonatype.com (or have a look at @mbonnin's nmcp) and the latter to central.sonatype.com/repository/maven-snapshots.

Right. Sounds to me like that warrants a ticket per task, but I'll let you decide. 😉

🧵 please ^^^

So, I got a question regarding the Kotest plugin for IntelliJ! I really, really LOVE to use Kotest over JUnit... but there's a few annoyances where I would like to know if this is an unsolvable problem or a bug... when tests fail, I like to quickly jump to them by clicking on the name in the runner panel... but for Kotest tests that "link" is not there as it is for JUnit tests... because of that, it is really, really tedious to find the source code of the failing test. Any plans to fix this? Thx!

Is it worth considering making

should(matcher: (T) -> Unit)

inline as part of 6.x? Possible the same for

shouldNotBeNull

Kotest 6.0.0.M5 is out. https://github.com/kotest/kotest/releases/tag/v6.0.0.M5 This also marks our first release to the new Maven Central. If you notice any artifacts missing or other oddities related to publications feel free to ping me directly.

FYI, here's a working* KMP MWE using Kotest 6.0.0.M5. *almost: proper test reports are only working on the JVM, but the tests are actually running on native and on the jvm (no JS thouhg, as it seems)