FunSuite is a Scala testing library with the following goals:
- Reuse JUnit: FunSuite is implemented as a JUnit runner and tries to build on top of existing JUnit functionality where possible. Any tool that knows how to run a JUnit test suite knows how to run FunSuite, including IDEs like IntelliJ.
- Helpful console output: test reports are pretty-printed with colors to help you quickly understand what caused a test failure. FunSuite tries to displays diffs and source locations when possible and it does a best-effort to highlight relevant stack trace elements.
- No Scala dependencies: FunSuite is implemented in ~1k lines of Scala code with no external Scala dependencies. The transitive Java dependencies weigh in total ~500kb, which is mostly just JUnit.
Table of contents
- Getting started
- Features
- Usage
- Tests as values
- Limitations
- Inspirations
- Do we really need another testing library?
- Coming from ScalaTest
- Stability
// Published for 2.11, 2.12 and 2.13. JVM-only.
libraryDependencies += "com.geirsson" %% "funsuite" % "VERSION"
testFrameworks += new TestFramework("funsuite.Framework")
Next, write a test suite.
class MySuite extends funsuite.FunSuite {
test("hello") {
val obtained = 42
val expected = 43
assertEquals(obtained, expected)
}
}
Assertion errors show the source code location where the assertion failed. Use
cmd+click on the location "/path/to/BasicSuite.scala:36
" to open the exact
line number in your editor (may not work in all terminals).
Stack frame elements for classes that are defined in your project sources are highlighted so you can focus on the important parts of the stack trace.
Use assertNoDiff(obtained, expected)
to compare large multi-line strings.
Test failures include the obtained multiline string in a copy-paste friendly format making it easy to update the test as the expected behavior of your program changes.
Override beforeAll()
, beforeEach()
, afterAll()
and afterEach()
to add
custom logic that should run before and after tests run. For example, use this
feature to create temporary files before executing tests or clean up acquired
resources after the test finish.
class MySuite extends funsuite.FunSuite {
// Runs once before all tests start.
override def beforeAll(context: BeforeAll): Unit = ???
// Runs before each individual test.
override def beforeEach(context: BeforeEach): Unit = ???
// Runs after each individual test.
override def afterEach(context: AfterEach): Unit = ???
// Runs once after all tests have completed.
override def afterAll(context: AfterAll): Unit = ???
}
Use assume(condition, explanation)
to skip tests when some conditions do not
hold. For example, use assume
to conditionally run tests based on the
operating system or the Scala compiler version.
import scala.util.Properties
test("paths") {
assume(Properties.isLinux, "this test runs only on Linux")
assume(Properties.versionNumberString.startsWith("2.13"), "this test runs only on Scala 2.13")
}
Use .flaky
to mark a test case that has a tendendency to fail sometimes.
test("requests".flaky) {
// I/O heavy tests that sometimes fail
}
By default, flaky tests fail unless the FUNSUITE_FLAKY_OK
environment variable
is set to true
. Override the isFlakyFailureOk
method to customize when it's
OK for flaky tests to fail.
Use .fail
to mark a test case that is expected to fail.
test("issue-456".fail) {
// Reproduce reported bug
}
A failed test only succeeds if the test body fails. If the test body succeeds, the test fails.
Use .only
to run only a single test.
test("issue-457") {
// will not run
}
test("issue-456".only) {
// only test that runs
}
test("issue-455") {
// will not run
}
Use testOnly -- $GLOB
to filter a fully qualified test name from the command
line.
# sbt shell
> testOnly -- *issue-456
Use testOnly -- --only=$TEST_FILTER
to filter an individual test name from the
command line.
# sbt shell
> testOnly -- --only=issue-456
Use the @Ignore
annotation to skip all tests in a test suite.
@funsuite.Ignore
class MySuite extends funsuite.FunSuite {
test("hello1") {
// will not run
}
test("hello2") {
// will not run
}
// ...
}
Use .ignore
to skip an individual test case in a test suite.
test("issue-456".ignore) {
// will not run
}
Use @Category(...)
to group tests suites together.
package myapp
import org.junit.experimental.categories.Category
class Slow extends funsuite.Tag("Slow")
class Fast extends funsuite.Tag("Fast")
@Category(Array(classOf[Slow]))
class MySlowSuite extends funsuite.FunSuite {
test("slow") {
Thread.sleep(1000)
}
// ...
}
@Category(Array(classOf[Slow], classOf[Fast]))
class MySlowFastSuite extends funsuite.FunSuite {
// ...
}
@Category(Array(classOf[Fast]))
class MyFastSuite extends funsuite.FunSuite {
// ...
}
Next, use --include-category=$CATEGORY
and --exclude-category=$CATEGORY
to
determine what test suites to run from the command line.
# matches: MySlowSuite, MySlowFastSuite
> testOnly -- --include-category=myapp.Slow
# matches: MySlowSuite
> testOnly -- --include-category=myapp.Slow --exclude-category=myapp.Fast
Test results are formatted in a specific way to make it easy to search for them in a large log file.
Test | Prefix |
---|---|
Failed | ==> X |
Ignored | ==> i |
Success | ==> + |
Skipped | ==> s |
Knowing these prefixes may come in handy for example when browsing test logs in a browser. Search for "==> X" to quickly navigate to the failed tests.
FunSuite test suites run in IntelliJ like normal.
It's expected that it's not possible to run individual test cases from IntelliJ
since it does not understand the structure of the test("name") {...}
syntax.
As a workaround, use the .only
marker to run only a single test from IntelliJ.
- test("name") {
+ test("name".only) {
// ...
}
FunSuite test cases are represented with the type funsuite.Test
and can be
manipulated as a normal data structure. Feel free to tweak and extend how you
generate funsuite.Test
to suit your needs.
Extend the base class funsuite.Suite
to customize exactly what Seq[Test]
you
want to run.
class MyCustomSuite extends funsuite.Suite {
// The type returned by bodies of test cases.
// Is defined as `Any` in `funsuite.FunSuite` but it's abstract in `funsuite.Suite`
override type TestValue = Future[String]
override def funsuiteTests() = List(
new Test(
"name",
// compile error if it's not a Future[String]
body = () => Future.successful("Hello world!"),
tags = Set.empty[Tag],
location = Location.generate
)
)
}
The abstract funsuite.Suite
class only includes the before/after APIs and not
other methods like assert
or test()
.
Override funsuiteRunTest()
to extend the default behavior for how test bodies
are evaluated. For example, use this feature to implement a Rerun(N)
modifier
to evaluate the body multiple times.
import scala.util.Properties
case class Rerun(count: Int) extends Tag("Rerun")
class MyWindowsSuite extends funsuite.FunSuite {
override def funsuiteRunTest(options: TestOptions, body: => Any): Any = {
val rerunCount = options.tags.collectFirst {
case Rerun(n) => n
}.getOrElse(1)
1.to(rerunCount).map(_ => super.funsuiteRunTest(options, body))
}
test("files", Rerun(10)) {
println("Hello") // will run 10 times
}
test("files") {
// will run once, like normal
}
}
Override funsuiteTests()
to customize what tests get executed. For example,
use this feature to skip tests based on a dynamic condition.
import scala.util.Properties
case object Windows extends funsuite.Tag("Windows")
class MyWindowsSuite extends funsuite.FunSuite {
override def funsuiteTests(): Any = {
val default = super.funsuiteTests()
if (!Properties.isWin) default
else default.filter(_.tags.contains(Windows))
}
test("files", Windows) {
// will only run in Windows
}
test("files") {
// will run like normal
}
}
JVM-only: FunSuite is currently only published for the JVM. FunSuite uses a JUnit testing interface for sbt that's written in Java so that would need to be changed in order to add Scala.js and Scala Native support. Feel free to open an issue if you would like to contribute cross-platform support.
FunSuite is inspired by several existing testing libraries:
- ScalaTest: the syntax for defining FunSuite test suites is the same as for
org.scalatest.FunSuite
. - JUnit: FunSuite is implemented as a custom JUnit runner and features like
assume
test filters are implemented on top of existing JUnit functionality. - utest: the nicely formatted stack traces and test reports is heavily inspired by the beautifully formatted output in utest.
- ava: the idea for showing the source locations for assertion errors comes from ava, a JavaScript testing library.
FunSuite is built on the idea that >90% of what a JVM testing library needs is already provided by JUnit. However, the default JUnit testing syntax is based on annotations and does not feel idiomatic when used from Scala. FunSuite tries to fill in the gap by providing a small Scala API on top of JUnit.
Add the following settings to run ScalaTest and JUnit suites with the same testing framework as FunSuite.
// build.sbt
testFrameworks := List(
new TestFramework("funsuite.Framework"),
new TestFramework("com.geirsson.junit.PantsFramework")
)
These settings configure all JUnit and ScalaTest suites to run with the same testing interface as FunSuite. This means that you get the same pretty-printing of test reports for JUnit, ScalaTest and FunSuite.
Next, you may want to start migrating your test suites one by one. If you only
use basic ScalaTest features, you should be able to replace usage of
org.scalatest.FunSuite
with minimal changes like below.
- import org.scalatest.funsuite.AnyFunSuite
- import org.scalatest.FunSuite
+ import funsuite.FunSuite
- class MySuite extends FunSuite with BeforeAll with AfterAll {
+ class MySuite extends FunSuite {
test("name") {
// unchanged
}
- ignore("ignored") {
+ test("ignored".ignore) {
// unchanged
}
FunSuite is a new library with no stability guarantees. It's expected that new releases, including patch releases, will have binary and source breaking changes.