does kotest support mapping like strikt ```val subject = Pan kotlinlang #kotest

val subject = Pantheon.NORSE
expectThat(subject)
  .get(Pantheon::ruler) // reference to a property
  .get { toString() }   // return type of a method call
  .isEqualTo("Odin")

sam

10/14/2020, 2:32 PM

subject.ruler shouldBe "Odin"

That would be the equivalent for testing a single property

MiSikora

10/14/2020, 2:52 PM

Will it convert

ruler

to a String or is

toString()

missing? Seems kind of spooky if it converts

sam

10/14/2020, 3:19 PM

I assumed panthon.ruler was a string, so if it's not, then

subject.ruler.toString() shouldBe "Odin"

👍 1

christophsturm

10/14/2020, 7:48 PM

The difference is in the failure message.

With the strikt syntax you get something like expected pantheon.Norse to have ruler Odin.

I Guess I can just use strikt with kotest

sam

10/14/2020, 8:11 PM

Yeah I see what you mean. The word "ruler" won't be present in the kotest error.

I think the closest right now would be

withClue("greek ruler") {
   Pantheon.Greek.ruler() shouldBe "Zeus"
}

Then your error would be something like

greek ruler
Expected :"Odin"
Actual   :"Zeus"

https://kotest.io/clues/

Or as you say, you don't have to use kotest assertions (or you can use them and strickt)

christophsturm

10/15/2020, 9:58 AM

I’m right now using strikt and minutest, but I’m looking at kotest for the multiplatform support, and because its not junit dependent

sam

10/15/2020, 10:35 AM

Right. The JVM support is currently on top of JUnit5, unless you use the experimental gradle plugin, then you can ditch junit completely. The kotest intellij plugin is junit free. Hopefully in 4.4 / 4.5 we can make the JUnit support 2nd class in favour of our native gradle support (the output is much better). The JS support uses the mocha/karma runners (whichever one the JS plugin pulls down now).

2 Views

#kotest Join Slack