zio-test

ZIO Test is a featherweight testing library for effectful programs.

The library imagines every spec as an ordinary immutable value, providing tremendous potential for composition. Thanks to tight integration with ZIO, specs can use resources (including those requiring disposal), have well- defined linear and parallel semantics, and can benefit from a host of ZIO combinators.

 import zio.test._
 import zio.test.environment.Live
 import zio.clock.nanoTime
 import Assertion.isGreaterThan

 object MyTest extends DefaultRunnableSpec {
   def spec = suite("clock")(
     testM("time is non-zero") {
       assertM(Live.live(nanoTime))(isGreaterThan(0))
     }
   )
 }

The environment package contains testable versions of all the standard ZIO environment types through the TestClock, TestConsole, TestSystem, and TestRandom modules. See the documentation on the individual modules for more detail about using each of them.

If you are using ZIO Test and extending RunnableSpec a TestEnvironment containing all of them will be automatically provided to each of your tests. Otherwise, the easiest way to use the test implementations in ZIO Test is by providing the TestEnvironment to your program.

import zio.test.environment._

myProgram.provideLayer(testEnvironment)

Then all environmental effects, such as printing to the console or generating random numbers, will be implemented by the TestEnvironment and will be fully testable. When you do need to access the "live" environment, for example to print debugging information to the console, just use the live combinator along with the effect as your normally would.

If you are only interested in one of the test implementations for your application, you can also access them a la carte through the make method on each module. Each test module requires some data on initialization. Default data is included for each as DefaultData.

import zio.test.environment._

myProgram.provideM(TestConsole.make(TestConsole.DefaultData))

Finally, you can create a Test object that implements the test interface directly using the makeTest method. This can be useful when you want to access some testing functionality without using the environment type.

import zio.test.environment._

for {
 testRandom <- TestRandom.makeTest(TestRandom.DefaultData)
 n          <- testRandom.nextInt
} yield n

This can also be useful when you are creating a more complex environment to provide the implementation for test services that you mix in.

The laws package provides functionality for describing laws as values. The fundamental abstraction is a set of ZLaws[Caps, R]. These laws model the laws that instances having a capability of type Caps are expected to satisfy. A capability Caps[_] is an abstraction describing some functionality that is common across different data types and obeys certain laws. For example, we can model the capability of two values of a type being compared for equality as follows:

trait Equal[-A] {
 def equal(a1: A, a2: A): Boolean
}

Definitions of equality are expected to obey certain laws:

1. Reflexivity - a1 === a1
2. Symmetry - a1 === a2 ==> a2 === a1
3. Transitivity - (a1 === a2) && (a2 === a3) ==> (a1 === a3)

These laws define what the capabilities mean and ensure that it is safe to abstract across different instances with the same capability.

Using ZIO Test, we can represent these laws as values. To do so, we define each law using one of the ZLaws constructors. For example:

val transitivityLaw = ZLaws.Laws3[Equal]("transitivityLaw") {
 def apply[A: Equal](a1: A, a2: A, a3: A): TestResult =
   ???
}

We can then combine laws using the + operator:

val reflexivityLaw: = ???
val symmetryLaw:    = ???

val equalLaws = reflexivityLaw + symmetryLaw + transitivityLaw

Laws have a run method that takes a generator of values of type A and checks that those values satisfy the laws. In addition, objects can extend ZLawful to provide an even more convenient syntax for users to check that instances satisfy certain laws.

object Equal extends Lawful[Equal]

object Hash extends Lawful[Hash]

object Ord extends Lawful[Ord]

checkAllLaws(Equal + Hash + Ord)(Gen.anyInt)

Note that capabilities compose seamlessly because of contravariance. We can combine laws describing different capabilities to construct a set of laws requiring that instances having all of the capabilities satisfy each of the laws.