The Cucumber Guice module allows you to use Google Guice dependency injection in your Cucumber tests. Guice comes as standard with singleton scope and 'no scope'. This module adds Cucumber scenario scope to the scopes available for use in your test code. The rest of this documentation assumes you have at least a basic understanding of Guice. Please refer to the Guice wiki if necessary, see Google Guice - Motivation
Add the cucumber-guice
dependency to your pom.xml
:
<dependencies>
[...]
<dependency>
<groupId>io.cucumber</groupId>
<artifactId>cucumber-guice</artifactId>
<version>${cucumber.version}</version>
<scope>test</scope>
</dependency>
[...]
</dependencies>
It's important to realise the differences in how this module functions when compared with earlier versions. The changes are as follows.
A Guice injector is created at the start of each test scenario and is destroyed at the end of each test scenario. There is no scenario scope, just singleton and 'no scope'.
A Guice injector is created once before any tests are run and is destroyed after the last test has run. Before each test scenario, a new scenario scope is created. At the end of the test scenario the scenario scope is destroyed. Singleton scope exists throughout all test scenarios.
Users wishing to migrate should replace @Singleton
annotations
with @ScenarioScope
annotations. Guice modules should also have
their singleton bindings updated. All bindings in
Scopes.SINGLETON
should be replaced with bindings in
CucumberScopes.SCENARIO
.
By including the cucumber-guice
jar on your
CLASSPATH
your Step Definitions will be instantiated by Guice.
There are two main modes of using the module: with scope annotations and with
module bindings. The two modes can also be mixed. When mixing modes, it is
important to realise that binding a class in a scope in a module takes
precedence if the same class is also bound using a scope annotation.
An implementation of this interface is used to obtain an
com.google.inject.Injector
that is used to provide instances of all the classes that are used to run the Cucumber
tests. The injector should be configured with a binding for ScenarioScope
.
Usually you will want to bind your step definition classes in either scenario scope or in singleton scope. It is not recommended to leave your step definition classes with no scope as it means that Cucumber will instantiate a new instance of the class for each step within a scenario that uses that step definition.
Cucumber will create exactly one instance of a class bound in scenario scope for each scenario in which it is used. You should use scenario scope when you want to store state during a scenario but do not want the state to interfere with subsequent scenarios.
Cucumber will create just one instance of a class bound in singleton scope that will last for the lifetime of all test scenarios in the test run. You should use singleton scope if your classes are stateless. You can also use singleton scope when your classes contain state but with caution. You should be absolutely sure that a state change in one scenario could not possibly influence the success or failure of a subsequent scenario. As an example of when you might use a singleton, imagine you have an http client that is expensive to create. By holding a reference to the client in a class bound in singleton scope, you can reuse the client in multiple scenarios.
This is the easy route if you're new to Guice. To bind a class in scenario
scope add the io.cucumber.guice.ScenarioScoped
annotation to the
class definition. The class should have a no-args constructor or one
constructor that is annotated with javax.inject.Inject
. For
example:
import cucumber.runtime.java.guice.ScenarioScoped;
import javax.inject.Inject;
@ScenarioScoped
public class ScenarioScopedSteps {
private final Object someInjectedDependency;
@Inject
public ScenarioScopedSteps(Object someInjectedDependency) {
this.someInjectedDependency = someInjectedDependency;
}
}
To bind a class in singleton scope add the
javax.inject.Singleton
annotation to the class definition. One
strategy for using stateless step definitions is to use providers to share
stateful scenario-scoped instances between stateless singleton step
definition instances. For example:
import javax.inject.Inject;
import javax.inject.Singleton;
@Singleton
public class MyStatelessSteps {
private final Provider<MyStatefulObject> providerMyStatefulObject;
@Inject
public MyStatelessSteps(Provider<MyStatefulObject> providerMyStatefulObject) {
this.providerMyStatefulObject = providerMyStatefulObject;
}
@Given("^I have (\\d+) cukes in my belly$")
public void I_have_cukes_in_my_belly(int n) {
providerMyStatefulObject.get().iHaveCukesInMyBelly(n);
}
}
There is an alternative explanation of using providers for mixing scopes on the Guice wiki.
As an alternative to using annotations you may prefer to declare Guice
bindings in a class that implements com.google.inject.Module
. To
do this, you should create a class that implements
io.cucumber.guice.api.InjectorSource
. This gives you complete
control over how you obtain a Guice injector and it's Guice modules. The
injector must provide a binding for
io.cucumber.guice.ScenarioScope
. It should also provide a
binding for the io.cucumber.guice.ScenarioScoped
annotation if
your classes are using the annotation. The easiest way to do this it to use
CucumberModules.createScenarioModule()
. For example:
import com.google.inject.Guice;
import com.google.inject.Injector;
import com.google.inject.Stage;
import io.cucumber.guice.CucumberModules;
import io.cucumber.guice.InjectorSource;
public class YourInjectorSource implements InjectorSource {
@Override
public Injector getInjector() {
return Guice.createInjector(Stage.PRODUCTION, CucumberModules.createScenarioModule(), new YourModule());
}
}