ScalaCheck

A clever way of creating expectations in specs2 is to use the ScalaCheck library.

To declare ScalaCheck properties you first need to extend the org.specs2.ScalaCheck trait. Then you can pass functions returning any kind of Result (Boolean, Result, MatchResult or a ScalaCheck Prop) to the prop method and use the resulting Prop as your example body:

s2"addition and multiplication are related ${ prop { (a: Int) => a + a == 2 * a } }"

The function that is checked can either return:

// a Boolean
s2"addition and multiplication are related ${ prop { (a: Int) => a + a == 2 * a } }"

// a MatchResult
s2"addition and multiplication are related ${ prop { (a: Int) => a + a must_== 2 * a } }"

// a Prop
s2"addition and multiplication are related ${ prop { (a: Int) => (a > 0) ==> (a + a must_== 2 * a) } }"

Note that if you pass functions using MatchResults you will get better failure messages than just using boolean expressions.

By default the properties created with prop will be shrinking counter-examples. But as you will see below there lots of different ways to parameterize ScalaCheck properties in specs2, including declaring if shrinking must be done.

Arbitrary instances

ScalaCheck requires an implicit Arbitrary[T] instance for each parameter of type T used in a property. If you rather want to pick up a specific Arbitrary[T] for a given property argument you can modify the prop with to use another Arbitrary instance:

s2"""
  a simple property       $ex1
  a more complex property $ex2
"""

def abStringGen = (Gen.oneOf("a", "b") |@| Gen.oneOf("a", "b"))(_+_)

implicit def abStrings: Arbitrary[String] =
  Arbitrary(abStringGen)

def ex1 = prop((s: String) => s must contain("a") or contain("b")).setArbitrary(abStrings)

// use the setArbitrary<n> method for the nth argument
def ex2 = prop((s1: String, s2: String) => (s1+s2) must contain("a") or contain("b")).
            setArbitrary1(abStrings).setArbitrary2(abStrings)

It is also possible to pass a Gen[T] instance instead of an Arbitrary[T]:

val abStringGen = (Gen.oneOf("a", "b") |@| Gen.oneOf("a", "b"))(_+_)

def ex1 = prop((s: String) => s must contain("a") or contain("b")).setGen(abStringGen)

With Shrink / Pretty

Specific Shrink and Pretty instances can also be specified at the property level:

val shrinkString: Shrink[String] = ???

// set a specific shrink instance on the second parameter
prop((s1: String, s2: String) => s1.nonEmpty or s2.nonEmpty).setShrink2(shrinkString)

// set a specific pretty instance
prop((s: String) => s must contain("a") or contain("b")).setPretty((s: String) =>
  Pretty((prms: Pretty.Params) => if (prms.verbosity >= 1) s.toUpperCase else s))

// or simply if you don't use the Pretty parameters
prop((s: String) => s must contain("a") or contain("b")).pretty((_: String).toUpperCase)

Contexts

ScalaCheck properties are sometimes used to test stateful applications rather than pure functions. For example you want to test that a function is writing files somewhere and you would like those files to be deleted after each property execution:

def createFile(f: File): Unit = ???
def deleteTmpDir: Unit = ???

prop { f: File =>
  createFile(f)
  f.exists
}.after(deleteTmpDir) // before and beforeAfter can also be used there

You can also “prepare” the property to be tested based on the generated arguments:

def createFile(directory: File, f: File): Unit = ???
// this method will keep the arguments intact but can
// have a side-effect to prepare the system
def setupDirectoryAndFile = (directory: File, file: File) => (directory, file)

prop { (directory: File, f: File) =>
  createFile(directory, f)
  f.exists
}.prepare(setupDirectoryAndFile)

Note that there is a way to model stateful systems with ScalaCheck which goes beyond the simple setup/teardown testing done here.

Test properties

Default values

ScalaCheck test generation can be tuned with a few properties. If you want to change the default settings, you have to use implicit values:

implicit val params = Parameters(minTestsOk = 20) // add ".verbose" to get additional console printing

The parameters you can modify are:

Parameter Default Description
minTestsOk 100 minimum of tests which must be ok before the property is ok
maxDiscardRatio 5.0f if the data generation discards too many values, then the property can’t be proven
minSize 0 minimum size for the “sized” data generators, like list generators
maxSize 100 maximum size for the “sized” data generators
workers 1 number of threads checking the property
rng new java.util.Random the random number generator
callback a ScalaCheck TestCallback (see the ScalaCheck documentation)
loader a custom classloader (see the ScalaCheck documentation)
prettyParams a Pretty.Params instance to set the verbosity level when displaying Pretty instances

Property level

It is also possible to specifically set the execution parameters on a given property:

class ScalaCheckSpec extends mutable.Specification with ScalaCheck {
  "this is a specific property" >> prop { (a: Int, b: Int) =>
    (a + b) must_== (b + a)
  }.set(minTestsOk = 200, workers = 3) // use "display" instead of "set" for additional console printing
}

You can also set the random generator that is used in all the ScalaCheck generators:

case class MyRandomGenerator() extends java.util.Random {
  // implement a deterministic generator for example
}

"this is a specific property" ! prop { (a: Int, b: Int) =>
  (a + b) must_== (b + a)
}.set(rng = MyRandomGenerator(), minTestsOk = 200, workers = 3)

Command-line

Some properties can be overridden from the command line

Parameter Command line
minTestsOk scalacheck.mintestsok
maxDiscardRatio scalacheck.maxdiscardratio
minSize scalacheck.minsize
maxSize scalacheck.maxsize
workers scalacheck.workers
verbose scalacheck.verbose

Expectations

By default, a successful example using a Prop will be reported as 1 success and 100 (or minTestsOk) expectations. If you don’t want the number of expectations to appear in the specification statistics just mix-in your specification the org.specs2.scalacheck.OneExpectationPerProp trait.

Collect values

It is important to validate that generated values are meaningful. In order to do this you can use collect to collect values:

// for a property with just one argument
prop((i: Int) => i % 2 == 0).collect
// for a property with just 2 arguments
// collect the second value only
prop((i: Int, j: Int) => i > 0 && j % 2 == 0).collect2
// collect the second value but map it to something else
prop((i: Int, j: Int) => i > 0 && j % 2 == 0).collectArg2((n: Int) => "the value "+n)
// collect all values and display
prop((i: Int, j: Int) => i > 0 && j % 2 == 0).collectAll.verbose

Note that, by default, nothing will be printed on screen unless you set the reporting to verbose by either:

Equivalence

The ==> operator in ScalaCheck helps you specify under which conditions a given property is applicable. However it only works one way, you cannot declare that a property must be true “if and only if” some conditions are respected.

With specs2 and the org.specs2.execute.ResultImplicits trait you can use the <==> operator to declare the equivalence of 2 Results, whether they are properties or booleans or MatchResults. So you can write:

// replace 55 with whatever you think "old" is...
prop((i: Int) => (i >= 18 && i <= 55) <==> isYoung(i))