Testing Java Applications Cheat Sheet 1.2
Testing Java Applications Cheat Sheet 1.2
Maven. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
How can I verify the mock was invoked during test execution? . . . . . . . . . . . . 34
Introduction
This cheat sheet follows a question & answer style to provide quickstart
solutions for testing Java applications:
Both JUnit 5 and Mockito are your main building blocks when it comes to testing
Java applications. Therefore, mastering them is essential for your testing success.
The following code examples can be copied and pasted to your project as-is. In
addition, you’ll find instructions for both Maven & Gradle.
While this document might not cover all your questions around JUnit 5 and
Mockito, it follows the Pareto principle and targets 20% of the features you use
80% of the time.
One important note before we jump into the Q&A style, always remember
Introduction | 1
Testing Java Applications Cheat Sheet
The Java testing ecosystem is huge. Over the years, a lot of libraries have
emerged, especially for niche areas. This book aims to cover all testing tools &
libraries that should be part of your testing toolbox as a Java Developer. Each
tool/library is present with a standardized cookbook-style approach.
The central goal of this book is to enrich your testing toolbox with new tools
& libraries that you might not have even heard about (yet). Furthermore, you’ll
also learn about new features of testing tools that you are already using.
The book will be your go-to resources for an overview of Java’s excellent testing
ecosystem with a right-sized introduction for tools & libraries of the following
categories:
• Test Frameworks
• Assertion Libraries
• Mocking Frameworks
• Infrastructure Libraries
• Utility Libraries
• Performance Testing
2 | Introduction
Testing Java Applications Cheat Sheet
This free 14 days email course introduces you to the Java Testing landscape
with tips, tricks, and recipes.
Regardless of whether you use Spring Boot, Jakarta EE, Quarkus, etc. for your
application, the underlying testing libraries & strategies remain the same:
All course lessons are sent right to your inbox for your convenience.
Introduction | 3
Testing Java Applications Cheat Sheet
If you are coming from a different programming language, you’ll love the user-
friendly API of Mockito.
Having worked with JavaScript for almost three years, I still have to go to the Jest
documentation every time I want to mock stuff.
The knowledge you gain with this course applies to testing all your Java
applications, independent of the framework (Spring Boot, Jakarta EE, Micronaut,
Helidon, Quarkus, etc.).
• Creating mocks
• Stubbing methods calls
• Verify the interaction of our mocks
4 | Introduction
Testing Java Applications Cheat Sheet
Not only will your technical testing skills improve but also you’ll deploy to
production with more confidence and sleep better at night thanks to a
sophisticated test suite.
Introduction | 5
Testing Java Applications Cheat Sheet
Maven
When writing applications with Java, we can’t just pass our .java files to the JVM
(Java Virtual Machine) to run our program. We first have to compile our Java
source code to bytecode (.class files) using the Java Compiler (javac). Next, we
pass this bytecode to the JVM (java binary on our machines) which then
interprets our program and/or compiles parts of it even further to native
machine code.
Given this two-step process, someone has to compile our Java classes and
package our application accordingly. Manually calling javac and passing the
correct classpath is a cumbersome task. A build tool automates this process. As
developers, we then only have to execute one command, and everything get’s
build automatically.
The two most adopted build tools for the Java ecosystem are Maven and
Gradle. Ancient devs might still prefer Ant, while latest-greatest devs might
advocate for Bazel as a build tool for their Java applications.
To build and test our Java applications, we need a JDK (Java Development Kit)
installed on our machine and Maven. We can either install Maven as a
command-line tool (i.e., place the Maven binary on our system’s PATH) or use the
portable Maven Wrapper.
The Maven Wrapper is a convenient way to work with Maven without having to
install it locally. It allows us to conveniently build Java projects with Maven
without having to install and configure Maven as a CLI tool on our machine
When creating a new Spring Boot project, for example, you might have already
wondered what the mvnw and mvnw.cmd files inside the root of the project are
used for. That’s the Maven Wrapper (the idea is borrowed from Gradle).
Maven provides a set of default Archetypes artifacts for several purposes like
a new web app, a new Maven plugin project, or a simple quickstart project.
We bootstrap a new Java project from one of these Archetypes using the mvn
command-line tool:
mvn archetype:generate \
-DarchetypeGroupId=org.apache.maven.archetypes \
-DarchetypeArtifactId=maven-archetype-quickstart \
-DarchetypeVersion=1.4 \
-DgroupId=com.mycompany \
-DartifactId=order-service
The skeleton projects we create with the official Maven Archetypes are a good
place to start.
We can create a new testing playground project using this custom Maven
Archetype with the following Maven command (for Linux & Mac):
mvn archetype:generate \
-DarchetypeGroupId=de.rieckpil.archetypes \
-DarchetypeArtifactId=java-testing-toolkit \
-DarchetypeVersion=1.0.0 \
-DgroupId=com.mycompany \
-DartifactId=order-service
For Windows (both PowerShell and CMD), we can use the following command to
bootstrap a new project from this template:
The generated project comes with a basic set of the most central Java testing
libraries. We can use it as a blueprint for our next project or explore testing Java
applications with this playground.
In summary, the following default configuration and libraries are part of this
project shell:
• Java 11
• JUnit Jupiter, Mockito, and Testcontainers dependencies
• A basic unit test
Next, we have to ensure a JDK 11 (or higher) is on our PATH and also JAVA_HOME
points to the installation folder:
$ java -version
openjdk version "11.0.10" 2021-01-19
OpenJDK Runtime Environment AdoptOpenJDK (build 11.0.10+9)
OpenJDK 64-Bit Server VM AdoptOpenJDK (build 11.0.10+9, mixed mode)
# Windows
$ echo %JAVA_HOME%
C:\Program Files\AdoptOpenJDK\jdk-11.0.10.9-hotspot
As a final verification step, we can now build and test this project with Maven:
....
We can now open and import the project to our editor or IDE (IntelliJ IDEA,
Eclipse, NetBeans, Visual Code, etc.) to inspect the generated project in more
detail.
• src/test/java
This folder is the main place to add our Java test classes (.java files). As a
general recommendation, we should try to mirror the package structure of our
production code (src/main/java). Especially if there’s a direct relationship
between the test and the source class.
Most of the IDEs and editors provide further support to jump to a test class.
IntelliJ IDEA, for example, provides a shortcut (Ctrl + Shift + T) to navigate from a
source file to its test classes(s) and vice-versa.
• src/test/resources
As part of this folder, we store static files that are only relevant for our test. This
might be a CSV file to import test customers for an integration test, a dummy
JSON response for testing our HTTP clients, or a configuration file.
• target/test-classes
At this location, Maven places our compiled test classes (.class files) and test
resources whenever the Maven compiler compiles our test sources. We can
explicitly trigger this with mvn test-compile and add a clean if we want to
remove the existing content of the entire target folder first.
Usually, there’s no need to perform any manual operations inside this folder as it
contains build artifacts. Nevertheless, it’s helpful to investigate the content for
this folder whenever we face test failures because we, e.g., can’t read a file from
the classpath. Taking a look at this folder (after the Maven compiler did its work),
can help to understand where a resources file ended up on the classpath.
• pom.xml
This is the heart of our Maven project. The abbreviation stands for Project Object
Model. Within this file, we define metadata about our project (e.g., description,
artifactId, developers, etc.), which dependencies we require, and the
configuration of our plugins.
The Maven Surefire Plugin, more about this plugin later, is designed to run our
unit tests. The following patterns are the defaults so that the plugin will detect a
class as a test:
• **/Test*.java
• **/*Test.java
• **/*Tests.java
• **/*TestCase.java
Several smart people came up with a definition for this term. One of such smart
people is Michael Feathers. He’s turning the definition around and defines
Kevlin Henney is also a great source of inspiration for a definition of the term
unit test.
The Maven Failsafe Plugin, designed to run our integration tests, detects our
integration tests by the following default patterns:
• **/IT*.java
• **/*IT.java
• **/*ITCase.java
We can also override the default patterns for both plugins and come up with a
different naming convention. However, sticking to the defaults is recommended.
Each of the three built-in lifecycles has a list of build phases. For our testing
example, the default lifecycle is important.
• validate: validate that our project setup is correct (e.g., we have the
correct Maven folder structure)
• compile: compile our source code with javac
• test: run our unit tests
• package: build our project in its distributable format (e.g., JAR or WAR)
• verify: run our integration tests and further checks (e.g., the OWASP
dependency check)
• install: install the distributable format into our local repository (~/.m2
folder)
• deploy: deploy the project to a remote repository (e.g., Maven Central or a
company hosted Artifactory
These build phases represent the central phases of the default lifecycle. There
are actually more phases. For a complete list, please refer to the Lifecycle
Reference of the official Maven documentation.
Whenever we execute a build phase, our project will go through all build phases
and sequentially until the build phase we specified. To phrase it differently, when
we run mvn package, for example, Maven will execute the default lifecycle
phases up to package in order:
If one of the build phases in the chain fails, the entire build process will
terminate. Imagine our Java source code has a missing semicolon, the compile
phase would detect this and terminate the process. As with a corrupt source file,
there’ll be no compiled .class file to test.
When it comes to testing our Java project, both the test and verify build
phases are of importance. As part of the test phase, we’re running our unit tests
with the Maven Surefire Plugin, and with verify our integration tests are
executed by the Maven Failsafe Plugin.
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
class MainTest {
@BeforeEach
void setUp() {
this.cut = new Main();
}
@Test
void shouldReturnFormattedUppercase() {
String input = "duke";
assertEquals("DUKE", result);
}
}
Depending on the Maven version and distribution format of our application (e.g.,
JAR or WAR), Maven defines default versions for the core plugins. Besides the
Maven Compiler Plugin, the Maven Resource Plugin, and other plugins, the
Maven Surefire Plugin is such a core plugin.
When packaging our application as a JAR file and using Maven 3.8.1, for example,
Maven picks the Maven Surefire Plugin with version 2.12.4 by default unless we
override it. As the default versions are sometimes a little behind the latest plugin
versions, it’s worth updating the plugin versions and manually specifying the
plugin version inside our pom.xml:
<project>
<!-- dependencies -->
<build>
<plugins>
<plugin>
<artifactId>maven-surefire-plugin</artifactId>
<version>3.0.0-M5</version>
</plugin>
</plugins>
</build>
</project>
As part of the test phase of the default lifecycle, we’ll now see the Maven
Surefire Plugin executing our tests:
$ mvn test
For the example above, we’re running one unit test with JUnit 5 (testing provider).
There’s no need to configure the testing provider anywhere, as with recent
Surefire versions, the plugin will pick up the correct test provider by itself. The
Maven Surefire Plugin integrates both JUnit and TestNG as testing providers
out-of-the-box.
If we don’t want to execute all build phases before running our tests, we can also
explicitly execute the test goal of the Surefire plugin:
mvn surefire:test
But keep in mind that we have to ensure that the test classes have been
compiled first (e.g., by a previous build).
We can further tweak and configure the Maven Surefire Plugin to, e.g., parallelize
the execution of our unit tests. This is only relevant for JUnit 4, as JUnit 5 (JUnit
Jupiter to be precise) supports parallelization on the test framework level.
Whenever we want to skip our unit tests when building our project, we can use
an additional parameter:
In contrast to our unit tests, the integration tests usually take more time, more
setup effort (e.g., start Docker containers for external infrastructure with
Testcontainers), and test multiple components of our application together.
We integrate the Maven Failsafe Plugin by adding it to the build section of our
pom.xml:
<project>
<build>
<!-- further plugins -->
<plugin>
<artifactId>maven-failsafe-plugin</artifactId>
<version>3.0.0-M5</version>
<executions>
<execution>
<goals>
<goal>integration-test</goal>
<goal>verify</goal>
</goals>
</execution>
</executions>
</plugin>
</plugins>
</build>
</project>
The Maven Failsafe Plugin is invoked as part of the verify build phase of the
default lifecycle. That’s right after the package build phase where we build our
distributable artifact (e.g., JAR):
If we want to run our integration tests manually, we can do so with the following
command:
For scenarios where we don’t want to run our integration test (but still our unit
tests), we can add -DskipITs to our Maven execution:
Similar to the Maven Surefire Plugin, we can also run a subset of our integration
tests:
When using the command above, make sure the test classes have been
compiled previously, as otherwise, there won’t be any test execution.
There’s also a property available to entirely skip the compilation of test classes
and avoid running any tests when building our project (not recommended):
of the Maven build lifecycle and how and when different Maven Plugins interact
is something to understand first.
With the help of Maven Archetypes or using a framework initializer, we can easily
bootstrap new Maven projects. There’s no need to install Maven as a CLI tool for
our machine as we can instead use the portable Maven Wrapper.
Furthermore, keep this in mind when testing your Java applications and use
Maven as the build tool:
• With Maven, we can separate the unit and integration test execution
• The Maven Surefire Plugin runs our unit tests
• The Maven Failsafe Plugin runs our integration tests
• By following the default naming conventions for both plugins, we can easily
separate our tests
• The Maven default lifecycle consists of several build phases that are
executed in order and sequentially
• Use the Java Testing Toolkit Maven archetype for your next testing
adventure
For a long time, JUnit 4.12 was the primary framework version. In 2017 JUnit 5
was launched and is now composed of several modules:
The JUnit team invested a lot in this refactoring to have a more platform-based
approach with a comprehensive extension model.
Nevertheless, migrating from JUnit 4.X to 5.X requires effort. Like @Test, all
annotations now reside in the package org.junit.jupiter.api and some
annotations were renamed or dropped and have to be replaced.
More information about the migration process from JUnit 4 to JUnit 5 in one of
the next sections.
Throughout this cheat sheet, I’ll use the term JUnit 5 to refer to
JUnit Jupiter (which is formally incorrect).
If you only use JUnit Jupiter for your tests and have no JUnit 3 or JUnit 4 left-overs,
your Maven project requires the following dependency:
<dependency>
<groupId>org.junit.jupiter</groupId>
<artifactId>junit-jupiter</artifactId>
<version>5.7.1</version>
<scope>test</scope>
</dependency>
For Gradle:
dependencies {
testImplementation('org.junit.jupiter:junit-jupiter:5.7.1')
}
Gradle supports native support for the JUnit Platform since version 4.6. To
enable it, we have to adjust the test task declaration as to the following:
test {
useJUnitPlatform()
}
For those of you that still have tests written in JUnit 4, make sure to include the
following dependencies:
<dependency>
<groupId>org.junit.jupiter</groupId>
<artifactId>junit-jupiter</artifactId>
<version>5.7.1</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.junit.vintage</groupId>
<artifactId>junit-vintage-engine</artifactId>
<version>5.7.1</version>
<scope>test</scope>
</dependency>
For Gradle:
dependencies {
testImplementation('org.junit.jupiter:junit-jupiter:5.7.1')
testImplementation('org.junit.vintage:junit-vintage-engine:5.7.1')
}
If your project uses Spring Boot, the Spring Boot Starter Test dependency
includes everything you need. You only have to decide whether or not you want
to include the Vintage Engine.
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-test</artifactId>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-test</artifactId>
<scope>test</scope>
<exclusions>
<exclusion>
<groupId>org.junit.vintage</groupId>
<artifactId>junit-vintage-engine</artifactId>
</exclusion>
</exclusions>
</dependency>
UPDATE: As of Spring Boot 2.4.0, the Spring Boot Starter Test no longer includes
the JUnit Vintage Engine.
For further application setup support, take a look at the different JUnit 5 project
setup examples on GitHub.
The following list is a short overview of the differences between both framework
versions:
If your codebase is using JUnit 4, changing the annotations to the JUnit 5 ones is
the first step. The most effort is required for migrating custom JUnit 4 rules to
JUnit 5 extensions.
You can automate the trivial parts of the migration, with e.g., Sam Brannen’s
script. The JUnit documentation also includes an own section with hints &
pitfalls for the migration
import org.junit.jupiter.api.Test;
class FirstTest {
@Test // ①
void firstTest() {
System.out.println("Running my first test with JUnit 5");
}
All you need is to annotate a non-private method (the Java class can be package-
private) with @Test.
Your tests must reside inside the src/test/java folder to follow default
Maven/Java conventions.
Running mvn test or gradle test will then execute your test:
class LifecycleTest {
public LifecycleTest() {
System.out.println("-- Constructor called");
}
@BeforeAll
static void beforeAllTests() {
System.out.println("- Before all tests");
}
@BeforeEach
void beforeEachTest() {
System.out.println("--- Before each test");
}
@AfterAll
static void afterAllTests() {
System.out.println("- After all tests");
}
@AfterEach
void afterEachTest() {
System.out.println("--- After each test");
}
@Test
void shouldAddIntegers() {
System.out.println("Test: Adding integers with Java");
assertEquals(4, 2 + 2);
}
@Test
void shouldSubtractIntegers() {
System.out.println("Test: Subtracting integers with Java");
assertEquals(4, 6 - 2);
}
}
-- Constructor called
However, if you want to write a more readable sentence, you can use
@DisplayName for your test:
class DisplayNameTest {
@Test
@DisplayName("This test verifies that Java can add integers")
void shouldAddIntegers() {
System.out.println("Test: Adding integers with Java");
assertEquals(4, 2 + 2);
}
@Test
@DisplayName("This test verifies that Java can subtract integers")
void shouldSubtractIntegers() {
System.out.println("Test: Subtracting integers with Java");
assertEquals(4, 6 - 2);
}
}
If you project uses Kotlin, you can achieve the same with:
@Test
fun `should throw an exception when dividing by zero`() {
//
}
Once your test suite grows, you might want to structure them. You can place
tests that verify a similar outcome together.
If you want to structure tests within a test class, you can use the @Nested
annotation:
class NestedTest {
@Nested
class Addition {
@Test
void shouldAddIntegers() {
System.out.println("Test: Adding integers with Java");
assertEquals(4, 2 + 2);
}
}
@Nested
class Subtraction {
@Test
void shouldSubtractIntegers() {
System.out.println("Test: Subtracting integers with Java");
assertEquals(4, 6 - 2);
}
}
}
Or you can tag them (either the whole test class or specific tests) with @Tag:
class TaggingTest {
@Test
@Tag("slow")
void shouldAddIntegers() {
System.out.println("Test: Adding integers with Java");
assertEquals(4, 2 + 2);
}
@Test
@Tag("fast")
void shouldSubtractIntegers() {
System.out.println("Test: Subtracting integers with Java");
assertEquals(4, 6 - 2);
}
}
JUnit 5 allows the following sources to feed input for a parameterized test:
• inline values
• method sources
• enums
• CSV files
class ParameterizedExampleTest {
@ParameterizedTest
@ValueSource(ints = {2, 4, 6, 8, 10})
void shouldAddIntegersValueSource(Integer input) {
assertEquals(input * 2, input + input);
}
@ParameterizedTest
@MethodSource("integerProvider")
void shouldAddIntegersMethodSource(Integer input) {
assertEquals(input * 2, input + input);
}
@ParameterizedTest
@CsvFileSource(resources = "/integers.csv")
void shouldAddIntegersCsvSource(Integer firstOperand,
Integer secondOperand,
Integer expectedResult) {
assertEquals(expectedResult, firstOperand + secondOperand);
}
}
import org.junit.jupiter.api.Test;
class AssertionExampleTest {
@Test
void demoJUnit5Assertions() {
assertEquals(4, 2 + 2);
assertNotEquals(4, 2 + 1);
class AssertThrowsExampleTest {
@Test
void invocationShouldThrowException() {
assertThrows(IllegalArgumentException.class, () -> {
divide(5, 0);
});
}
The assertThrows method also returns the exception. You can store it inside a
variable to make further inspections and, e.g., check the message or the cause of
the exception:
junit.jupiter.execution.parallel.enabled = true
junit.jupiter.execution.parallel.mode.default = concurrent
This enables parallel execution for all your tests and sets the execution mode to
concurrent. Compared to same_thread, concurrent does not enforce to
execute the test in the same thread of the parent. For a per test class or method
mode configuration, you can use the @Execution annotation.
There are multiple ways to set these configuration values. One solution is to
use the Maven Surefire plugin for the configuration:
<plugin>
<artifactId>maven-surefire-plugin</artifactId>
<version>3.0.0-M5</version>
<configuration>
<properties>
<configurationParameters>
junit.jupiter.execution.parallel.enabled = true
junit.jupiter.execution.parallel.mode.default = concurrent
</configurationParameters>
</properties>
</configuration>
</plugin>
You should benefit the most when using this feature for unit tests. Enabling
parallelization for integration tests might not be possible or easy to achieve,
depending on your setup.
Therefore, I recommend executing your unit tests with the Maven Surefire Plugin
and configure parallelization for them. All your integration tests can then be run
with the Maven Failsafe Plugin where you don’t specify these JUnit 5
configuration parameters.
For a more fine-grain parallelism configuration, take a look at the official JUnit
5 documentation.
If you have used Spring Boot or Mockito in the past, you probably already saw
@ExtendWith on top of your test class, e.g.
@ExtendWith(SpringExtension.class).
This is the declarative approach to register an extension. You also use the
programmatic approach using @RegisterExtension or use Java’s
ServiceLoader to automatically register them.
• BeforeAllCallback
• BeforeEachCallback
• BeforeTestExecutionCallback
• AfterTestExecutionCallback
• AfterEachCallback
• AfterAllCallback
• ParameterResolver
• TestExecutionExceptionHandler
• InvocationInterceptor
• etc.
For example, injecting Spring Beans (using @Autowired) during your test is
achieved with a ParameterResolver.
Next, whenever you use the @Testcontainers annotation for your test,
Testcontainers registers an extension in the background to start and stop your
container definitions (using @Container) accordingly.
@Retention(RetentionPolicy.RUNTIME)
@Target(ElementType.PARAMETER)
public @interface RandomUUID {
}
@Override
public boolean supportsParameter(ParameterContext parameterContext, ExtensionContext extensionContext) {
return parameterContext.isAnnotated(RandomUUID.class);
}
@Override
public Object resolveParameter(ParameterContext parameterContext, ExtensionContext extensionContext) {
return UUID.randomUUID().toString();
}
}
Let’s use this parameter resolver and register it for one of our tests:
import org.junit.jupiter.api.RepeatedTest;
import org.junit.jupiter.api.extension.ExtendWith;
@ExtendWith(RandomUUIDParameterResolver.class)
public class ParameterInjectionTest {
@RepeatedTest(5)
public void testUUIDInjection(@RandomUUIDParameterResolver.RandomUUID String uuid) {
System.out.println("Random UUID: " + uuid);
}
}
The main reason to use Mockito is to stub methods calls and verify
interaction of objects. The first is essential if you write unit tests, and your test
class requires other objects to work (so-called collaborators). As your unit test
should focus on just testing your class under test, you mock the behavior of the
dependent objects of this class.
A visual overview might explain this even better. Let’s take the following
example. Our application has a PricingService that makes use of public
methods of the InventoryService.
When we now want to write a unit test for the PricingService we don’t want
to interact with the real InventoryService, but rather a mock of it.
This way, we can control the behavior of the collaborator and test different
scenarios:
<dependency>
<groupId>org.mockito</groupId>
<artifactId>mockito-core</artifactId>
<version>3.9.0</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.mockito</groupId>
<artifactId>mockito-junit-jupiter</artifactId>
<version>3.9.0</version>
<scope>test</scope>
</dependency>
dependencies {
testImplementation('org.mockito:mockito-core:3.9.0')
testImplementation('org.mockito:mockito-junit-jupiter:3.9.0')
}
You can’t create mocks from final classes and methods with
Mockito. Everything else is mockable. Be aware of this if you are
using Kotlin, as you have to declare your classes as open.
If we now want to create a mock for the InventoryService while unit testing
the PricingService, we can do the following:
import org.junit.jupiter.api.Test;
class PricingServiceTest {
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
// Test comes here
}
}
or
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
import org.mockito.Mock;
import org.mockito.MockitoAnnotations;
class PricingServiceTest {
@Mock
private InventoryService inventoryService;
@BeforeEach
void setup() {
MockitoAnnotations.openMocks(this); // ①
this.pricingService = new PricingService(inventoryService);
}
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
// Test comes here
}
}
There is a third way to create mocks that I prefer for my tests. More about this
approach on the next page.
import org.junit.jupiter.api.Test;
import org.junit.jupiter.api.extension.ExtendWith;
import org.mockito.InjectMocks;
import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension;
@ExtendWith(MockitoExtension.class) // ①
class PricingServiceTest {
@Mock // ②
private InventoryService inventoryService;
@InjectMocks // ③
private PricingService pricingService;
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
// Test comes here
}
}
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
when(inventoryService.isCurrentlyInStockOfCompetitor("MacBook")).thenReturn(false);
For this stubbing to work your setup has to match the exact method
arguments your stub is called with during test execution.
This might be sometimes hard to achieve, and you can use the following
workaround:
@Test
void mockWithAnyInputParameter() {
when(inventoryService.isCurrentlyInStockOfCompetitor(anyString())).thenReturn(true);
@Test
void mockWithAnyClassInputParameter() {
when(inventoryService.isCurrentlyInStockOfCompetitor(any(String.class))).thenReturn(true);
Apart from returning a value, you can also instruct your mock to throw an
exception:
@Test
void shouldThrowExceptionWhenCheckingForAvailability() {
when(inventoryService.isCurrentlyInStockOfCompetitor("MacBook"))
.thenThrow(new RuntimeException("Network error"));
If you want to return a value based on the method argument of your mock,
thenAnswer fits perfectly:
@Test
void mockWithThenAnswer() {
when(inventoryService.isCurrentlyInStockOfCompetitor(any(String.class))).thenAnswer(invocation -> {
String productName = invocation.getArgument(0);
return productName.contains("Mac");
});
I’ve a recorded a video that explains where thenAnswer can reduce your
boilerplate setup.
• for primitive types (e.g. int, boolean): the default value, e.g. 0 for int and
false for boolean
• for reference types (e.g. String or Customer): null
• for collections: []
• for Java’s Optional: Optional.empty()
@ExtendWith(MockitoExtension.class)
class PricingServiceTest {
@Mock
private InventoryService inventoryService;
@InjectMocks
private PricingService pricingService;
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
// the inventoryService will return false
BigDecimal result = pricingService.calculatePrice("MacBook");
}
}
On the other side, you might also want to test that no interaction happened with
a mock.
@ExtendWith(MockitoExtension.class)
class PricingServiceVerificationTest {
@Mock
private InventoryService inventoryService;
@InjectMocks
private PricingService pricingService;
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
when(inventoryService.isCurrentlyInStockOfCompetitor(anyString())).thenReturn(false);
verify(inventoryService).isCurrentlyInStockOfCompetitor("MacBook");
// verifyNoInteractions(otherMock);
import org.junit.jupiter.api.Test;
import org.junit.jupiter.api.extension.ExtendWith;
import org.mockito.ArgumentCaptor;
import org.mockito.Captor;
import org.mockito.InjectMocks;
import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension;
import java.math.BigDecimal;
@ExtendWith(MockitoExtension.class)
class PricingServiceArgumentCaptureTest {
@Mock
private InventoryService inventoryService;
@InjectMocks
private PricingService pricingService;
@Captor
private ArgumentCaptor<String> stringArgumentCaptor;
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
when(inventoryService.isCurrentlyInStockOfCompetitor(anyString())).thenReturn(false);
verify(inventoryService).isCurrentlyInStockOfCompetitor(stringArgumentCaptor.capture());
After capturing the arguments, you can then inspect them and write assertions
for them.
If you stub a method during your test setup that is not used during your test
execution, Mockito will fail the test:
org.mockito.exceptions.misusing.UnnecessaryStubbingException:
Unnecessary stubbings detected.
Clean & maintainable test code requires zero unnecessary code.
Following stubbings are unnecessary (click to navigate to relevant line of code):
1. -> at
de.rieckpil.products.mockito.PricingServiceMockitoSettingsTest.shouldReturnHigherPriceIfProductIsInStockOfCo
mpetitor(PricingServiceMockitoSettingsTest.java:27)
Please remove unnecessary stubbings or use 'lenient' strictness. More info: javadoc for
UnnecessaryStubbingException class.
import org.junit.jupiter.api.Test;
import org.junit.jupiter.api.extension.ExtendWith;
import org.mockito.InjectMocks;
import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension;
import org.mockito.junit.jupiter.MockitoSettings;
import org.mockito.quality.Strictness;
@ExtendWith(MockitoExtension.class)
@MockitoSettings(strictness = Strictness.WARN)
class PricingServiceMockitoSettingsTest {
@Mock
private InventoryService inventoryService;
@InjectMocks
private PricingService pricingService;
@Test
void shouldReturnHigherPriceIfProductIsInStockOfCompetitor() {
when(inventoryService.isAvailable("MacBook")).thenReturn(false);
when(inventoryService.isCurrentlyInStockOfCompetitor("MacBook")).thenReturn(false);
pricingService.calculatePrice("MacBook");
}
}
You can override this strictness setting to be either lenient, produce warn logs or
fail on unnecessary stubs (default):
• Strictness.LENIENT
• Strictness.WARN
• Strictness.STRICT_STUBS
Let’s take a look at the following OrderService that makes use of two static
methods UUID.randomUUID() and LocalDateTime.now():
return order;
}
}
Starting with version 3.4.0 can now mock these static calls with Mockito
import org.junit.jupiter.api.Test;
import org.mockito.MockedStatic;
import org.mockito.Mockito;
import java.time.LocalDateTime;
import java.util.UUID;
class OrderServiceTest {
@Test
void shouldIncludeRandomOrderIdWhenNoParentOrderExists() {
try (MockedStatic<UUID> mockedUuid = Mockito.mockStatic(UUID.class)) {
mockedUuid.when(UUID::randomUUID).thenReturn(defaultUuid);
assertEquals("8d8b30e3-de52-4f1c-a71c-9905a8043dac", result.getId());
}
}
@Test
void shouldIncludeCurrentTimeWhenCreatingANewOrder() {
try (MockedStatic<LocalDateTime> mockedLocalDateTime = Mockito.mockStatic(LocalDateTime.class)) {
mockedLocalDateTime.when(LocalDateTime::now).thenReturn(defaultLocalDateTime);
assertEquals(defaultLocalDateTime, result.getCreationDate());
}
}
}
A good is example is testing your web layer. Therefore, you can use
@WebMvcTest and MockMvc to start a mocked Servlet environment to access
your controller endpoints.
For such test, you usually don’t want any interaction with your service layer and
hence mock any collaborator of your controller with @MockBean.