After upgrading to Kotest 5.0.0 (from 4.6.3) I'm seeing some unit test fail that use Wiremock to start / stop a server in

beforeSpec {}

/

afterSpec {}

, see https://dev.azure.com/oss-review-toolkit/ort/_build/results?buildId=6317&view=ms.vs[…]ld-test-results-tab&runId=38928&resultId=100025&paneView=debug. Is anyone else experiencing problems with Kotest 5 & Wiremock?

I'm running into more troubles after migrating to Kotest 5.0.0. Can it be that the order in which

beforeTest

from a

TestListener

vs. a function from within the test spec are being called? I'm now seeing

beforeTest

from a `TestListener`being caller after the

beforeTest

overridden in a test spec, whereas I believe the order was different before. Is that expected?

Yet one more thing: I just realized that

beforeTest()

in a

WordSpec

also gets called for the outer

WordSpecShouldContainerScope

. What can I use to just have something called for the contained actual tests, i.e. the

WordSpecTerminalScope

?

Do you see any opportunities to speed up running tests via the plugin? I just profiled running a single test in a spec via the green arrow, and it’s taking 5.5 seconds before actually starting the test, mostly in classgraph scanning, which seems unnecessary as the class and test name are being sent to it as

--spec

and

--testpath

args to

io.kotest.engine.launcher.MainKt

in the first place.

Consider editing your IntelliJ Kotest configuration template to add these VM options:

-Dkotest.framework.classpath.scanning.config.disable=true -Dkotest.framework.classpath.scanning.autoscan.disable=true

Knocks 5 seconds off running a simple test for me. (obviously if you’re actually relying on automatically discovered kotest config this is not such a great tip…)

How can I show the exception's message as a clue with

shouldNotThrowAny

?

Kotest 5.0.1 released

Does 5.0.0 change the order in which tests and/or specs are run? We’re seeing some failures after migrating, and its looking like the issue might be due to this. (Not kotests’s fault, of course. Our tests should be able to run in any order. I’m just curious to know if this is a known change.)

Issues with running individual Kotests from IntelliJ in a test class. I setup a new Gradle/Kotlin project in IntelliJ with the Kotest plugin and Kotest 5.0.1 dependencies in the build.gradle. I created a new test class with two tests using StringSpec (I tried FunSpec too). • When I run the test class, both tests run successfully. • When I run the first test, it runs successfully. • When I attempt to run just the 2nd test, I get the message "both tests filtered out" or something like that. I'm kicking off these tests in the IDE using the Kotest plugin panel. Has anyone seen anything like this before? I can provide more details. For now, I've moved on and removed Kotest, and everything is working as expected using pure JUnit5.

Hey folks 🙂 Quick question - is there a specific reason why

eventually

using kotlins

Duration

is still marked with

@ExperimentalTime

? I think since 1.6.0 Duration isn't experimental anymore 🤔

I'm having some issues with the migration, notably with the Koin integration. The KoinExtension registers a

beforeAny

that filters

TestType.Test

, I previously declared my DI mocks in

beforeEach

so that I didn't have to do that extra check. However, with Kotest 5.0,

beforeEach

gets called before

beforeAny

in TestExtensions, which causes my mocks to fail because Koin hasn't started yet. I wonder if the order in TestExtensions should be changed again, as I'd expect the outer scopes to evaluate before inner scopes, thus container → any → each.

Is it possible to write a data driven test using StringSpec, wrapping it in a context? So far I’ve only been able to make it work only like this:

class PriceSpec : StringSpec({
    withData(
        nameFn = { "goods=${it.goods}" },
        // data rows
    ) { (goods, total) ->
        price(goods) shouldBe total
    }

    "test incremental" {
        ...
    }
})

Originally there were 2 tests, then I transformed the first into a ddt, but I lost its name wrapping the generated cases. I would like to give a name to the ddt, but If I put the withData() declaration inside a string block it’ll fail at runtime, and if I wrap it with a context(“name”) as shown in the docs, only the second test will be executed. What am I doing wrong?

I'm using Kotest's data class matchers to compare two data classes that each contain a field with an instance of a sealed class. While the sealed class instances are different, their data is the same. This makes the test fail. Is this expected? I would have assumed sealed class properties should be compared by their values, just like nested data classes.

Hello everyone, we currently have all of our tests written with JUnit and kotest assertions. For our e2e tests we wanted to look into Kotest’s behaviour spec as it would improve readability by a big margin. We have the following scenario:

class SomeE2ETest: E2ETestEnvironment { ... }

where E2ETestEnvironment is a class (annotated with

@SpringBootTest

) that holds common project config, e.g. a preconfigured http client to post something to the api. Converting this test simply to a Kotest test will not work as we can’t do the following:

class SomeE2EKotest: E2ETestEnvironment(), BehaviorSpec({ ... })

because only one class may appear in a super type list. Long story short: I unfortunately couldn’t find any resource about how to do this with Kotest. Our main scenario is that we probably (currently) only translate the E2E Tests, so I’d like to find a way to access the

E2ETestEnvironment

from both JUnit tests and Kotest Tests. Thanks in advance!

after upgrading from 4.6.3 to 5.0.1, integration tests are failing when using a postgres testcontainer + listener 🤔

5.0.2 is out with a bunch of fixes https://kotest.io/docs/changelog.html

Despite potentially looking like a complete buffoon, anyone have ideas about why I can’t seem to get the configuration logging to work correctly? I just bumped to

5.0.2

and am set the system property

kotest.framework.dump.config=true

in my

gradle.properties

but I am not seeing anything in the logs like I used to (prior to the jump to 5+). Is there another way to check the configuration at runtime? * Not a contribution *

I'm having issues with the latest versions of Kotest (5.0.2) with IDEA (2021.3) and plugin (1.1.49-IC-213-EAP-SNAPSHOT), and Kotlin 1.6.0: Some time ago, everything was fine and dandy, but then something changed (it was either IDEA update, me changing the project, or upgrading Gradle or Kotlin version changed), and first I started having problems with running individual tests (like Ross wrote earlier). Then, again something changed, and now I'm able to run the tests via IDEA's Run <test class>: <test> gutter icon of individual tests launching the Kotest runner panel, but the Run <test class> now simply launches a Gradle build, no tests run. IDEA's Gradle > project > verification > test runs no tests, nor does running

./gradlew test

. Currently, I have a root project with multiple subprojects (modules), and I'm applying some conventions (like applying Kotest) to all the subprojects using a Gradle plugin. I'm wondering, if this setup has somehow screwed up the discovery (more in thread)

Did we ever implement those complexity-matchers? I remember we had an idea to have something like

myAlgo shouldRunIn logOfN()

I have a test that fails because the lhs of

shouldBe

is a

List

, and the rhs is a home-grown

Iterable

. This happens in

IterablEq.kt

(kotest 5.0.3) as

fun isValidIterable(it: Any): Boolean {
      return when (it) {
         is List<*>, is Set<*>, is Array<*>, is Collection<*> -> true
         else -> false
      }
   }

Could someone please help me understand why a custom implementation of

Iterable

should not be considered a valid iterable for the purposes of

kotest

?

I’ve noticed some matcher functions handle nullable variables but other’s don’t e.g. this works for a `String?`:

testString shouldHaveMinLength 1

But it doesn’t work for a `Int?`:

testInt shouldBeGreaterThan 0

The error is -

Infix call corresponds to a dot-qualified call 'testInt.shouldBeGreaterThan(0)' which is not allowed on a nullable receiver 'testInt'. Use '?.'-qualified call instead

So I can get around it by doing this:

testInt?.shouldBeGreaterThan(0)

Is this a limitation of Kotest and is there a better way I can handle this?

I'm sure someone has asked this but I can't find it... Just upgraded to 5.0.3 and my tagging strategy to avoid running tests that need to start a docker container seems to no longer work - looks like the tests are skipped, but the listener still starts the container:

internal class AuctionSniperEndToEndContainerTest : StringSpec({

  tags(Docker)

  listener(openfireTestContainer(Paths.get("../docker-openfire")).perSpec())
  ...
}

Hello, basic question here, I usually use Kotest by initializing my mocks with

MockkAnnotations.init(this)

inside the

beforeEach

block and I use

lateinit var

on my mocks. Is this bad or is this the recommended way? I find that using

val

is not always possible as sometimes you have some expectations to setup first.

class MyClassTest : StringSpec() {

	private lateinit var myClass: MyInterface

	@RelaxedMockK
	private lateinit var myMock1: MockableThing1

	@RelaxedMockK
	private lateinit var myMock2: MockableThing2

	init {
		beforeEach {
			MockKAnnotations.init(this)

			myClass = MyClass(
				myMock1,
				myMock2
			)
		}
    }
    ...

I checked that property testing is really slow in KMP compared to JVM. Not sure if this is caused by Kotest or just because KMP tests are slow and property testing is running a lot iterations (probably this last one). Any way to improve it? I tried 100k iterations and it was running under 30 secs in jvm, but in other native target is taking over 2 or 3 minutes

How do you assert matchers on nested object contents? Something like capturing slots in MockK. E.g. I have an object like this:

Success(Either.Left(IOError(cause=com.google.gson.stream.MalformedJsonException)))

I want to assert certain things: • that the contents of

Success

is of type

Left

• that the contents of

Left

is of type

IOError

• that the

cause

is of type

MalformedJsonException

I have a test class that looks like this:

class NetworkEitherCallAdapterTest : StringSpec({

  private lateinit var server: MockWebServer
  private lateinit var service: CallErrorTestClient

  beforeAny {
      server = MockWebServer()
      server.start()
      service = Retrofit.Builder()
      .baseUrl(server.url("/"))
      .addConverterFactory(GsonConverterFactory.create())
      .addCallAdapterFactory(EitherCallAdapterFactory.create())
      .build()
      .create(CallErrorTestClient::class.java)
  }
  afterAny { server.shutdown() }

  "should return ResponseMock for 200 with valid JSON" {
      server.enqueue(MockResponse().setBody("""{"response":"Arrow rocks"}"""))
      val body = service.getEither()
      body shouldBe ResponseMock("Arrow rocks").right()
  }

  // More tests...

I want to run the test class 3 times with different

service

instances (which will use different

ConverterFactory

each). What is the most readable way to achieve it? I don't want to parameterize every test, I want to parameterize the whole class and repeat it.

Hello, I am new on Kotest and i read it's intro which says it is testing framework for multi-platform test. So can i use it for android project only? i mean its not multi-platform project.

One thing I noticed while creating an

Arb

automagically using

Arb.bind

is that I’m not getting any null support. For example, if I have the following data class:

data class Foo(
  val myInt: Int?
)

and I generate an arb like so:

val fooArb = Arb.bind<Foo>()

I have never encountered a generated value with

myInt == null

. Is this expected? Is there a way to get this behavior? * Not a contribution*