I'm being hit with the issue that spock doesn't allow Mocks to be created outside of the specification - How to create Spock mocks outside of a specification class?
This seems to be still outstanding so am asking is that giving that i've got a complex and nested DI graph what is the most efficient way to 'inject' a mock representation deep in the graph?
Ideally, I have one bean definition set for normal deployment and another when running unit tests and it is this definition set being the applicable Mocks
e.g.
#Configuration
#Profile("deployment")
public class MyBeansForDeployment {
#Bean
public MyInterface myBean() {
return new MyConcreateImplmentation();
}
}
&&
#Configuration
#Profile("test")
public class MyBeansForUnitTests {
#Bean
public MyInterface myBean() {
return new MyMockImplementation();
}
}
Since Spock 1.1, you can create mocks outside of a specification class (detached mocks). One of the options is DetachedMockFactory. Take a look at the documentation or my answer to the question you linked.
You could try to implement a BeanPostProcessor that will replace the beans that you want with test doubles, such as shown below:
public class TestDoubleInjector implements BeanPostProcessor {
...
private static Map<String, Object[]> testDoubleBeanReplacements = new HashMap<>();
public void replaceBeanWithTestDouble(String beanName, Object testDouble, Class testDoubleType) {
testDoubleBeanReplacements.put(beanName, new Object[]{testDouble, testDoubleType});
}
#Override
public Object postProcessAfterInitialization(Object bean, String beanName) throws BeansException {
if (testDoubleBeanReplacements.containsKey(beanName)) {
return testDoubleBeanReplacements.get(beanName)[TEST_DOUBLE_OBJ];
}
return bean;
}
In your test, setup your mocks like shown below before initializing the application context. Make sure to include the TestDoubleInjector as a bean in your test context.
TestDoubleInjector testDoubleInjector = new TestDoubleInjector()
testDoubleInjector.replaceBeanWithTestDouble('beanToReplace', mock(MyBean.class), MyBean.class)
It could be done using HotSwappableTargetSource
#WebAppConfiguration
#SpringApplicationConfiguration(TestApp)
#IntegrationTest('server.port:0')
class HelloSpec extends Specification {
#Autowired
#Qualifier('swappableHelloService')
HotSwappableTargetSource swappableHelloService
def "test mocked"() {
given: 'hello service is mocked'
def mockedHelloService = Mock(HelloService)
and:
swappableHelloService.swap(mockedHelloService)
when:
//hit endpoint
then:
//asserts
and: 'check interactions'
interaction {
1 * mockedHelloService.hello(postfix) >> { ""Mocked, $postfix"" as String }
}
where:
postfix | _
randomAlphabetic(10) | _
}
}
And this is TestApp (override the bean you want to mock with proxy)
class TestApp extends App {
//override hello service bean
#Bean(name = HelloService.HELLO_SERVICE_BEAN_NAME)
public ProxyFactoryBean helloService(#Qualifier("swappableHelloService") HotSwappableTargetSource targetSource) {
def proxyFactoryBean = new ProxyFactoryBean()
proxyFactoryBean.setTargetSource(targetSource)
proxyFactoryBean
}
#Bean
public HotSwappableTargetSource swappableHelloService() {
new HotSwappableTargetSource(new HelloService());
}
}
Have a look at this example https://github.com/sf-git/spock-spring
Related
Spring 6, Quartz, and a SimpleTrigger based scheduled task.
#Component
#Slf4j
public class Greeting {
public void sayHello() {
log.debug("Hello at {}:", LocalDateTime.now());
}
}
Quartz config:
#Configuration
class QuartzConfig{
#Bean
MethodInvokingJobDetailFactoryBean greetingJobDetailFactoryBean() {
var jobFactory = new MethodInvokingJobDetailFactoryBean();
jobFactory.setTargetBeanName("greeting");
jobFactory.setTargetMethod("sayHello");
return jobFactory;
}
#Bean
public SimpleTriggerFactoryBean simpleTriggerFactoryBean() {
SimpleTriggerFactoryBean simpleTrigger = new SimpleTriggerFactoryBean();
simpleTrigger.setJobDetail(greetingJobDetailFactoryBean().getObject());
simpleTrigger.setStartDelay(1_000);
simpleTrigger.setRepeatInterval(5_000);
return simpleTrigger;
}
#Bean
public SchedulerFactoryBean schedulerFactoryBean() {
var factory = new SchedulerFactoryBean();
factory.setTriggers(
simpleTriggerFactoryBean().getObject(),
cronTriggerFactoryBean().getObject()
);
return factory;
}
And I tried to use awaitility to check the invocations.
#SpringJUnitConfig(value = {
QuartzConfig.class,
Greeting.class
})
public class GreetingTest {
#Autowired
Greeting greeting;
Greeting greetingSpy;
#BeforeEach
public void setUp() {
this.greetingSpy = spy(greeting);
}
#Test
public void whenWaitTenSecond_thenScheduledIsCalledAtLeastTenTimes() {
await()
.atMost(Duration.ofSeconds(10))
.untilAsserted(() -> verify(greetingSpy, atLeast(1)).sayHello());
}
}
Running the tests, it is failed.
org.awaitility.core.ConditionTimeoutException: Assertion condition defined as a com.example.demo.GreetingTest
Wanted but not invoked:
greeting.sayHello();
-> at com.example.demo.GreetingTest.lambda$whenWaitTenSecond_thenScheduledIsCalledAtLeastTenTimes$0(GreetingTest.java:36)
Actually, there were zero interactions with this mock.
within 10 seconds.
In the jobDetailFactorBean, I used jobFactory.setTargetBeanName("greeting"); to setup the target beans here, it should pass the Greeting bean directly.
Updated: resolved myself, check here.
You're creating a spy that in no way interacts with the actual code:
#BeforeEach
public void setUp() {
this.greetingSpy = spy(greeting);
}
This would have to be injected into the Spring context as a bean and used everywhere, where greeting is used. Spring actually provides such functionality: #SpyBean.
Instead of autowiring a greeting and wrapping it with a spy that does not interact with anything in the context, replace the #Autowired with #SpyBean annotation. Thanks to that a spy bean will be created and injected within the Spring context:
#SpyBean
Greeting greeting;
I created a commit in GitHub repository, where you can see the whole code - the test passes. I had to add the cronTriggerFactoryBean() method to the configuration as it is omitted in your question.
If you cannot use Spring Boot, you can create the spy within Spring context yourself using configuration:
static class Config {
#Bean
#Primary
Greeting greeting() {
return spy(new Greeting());
}
}
Thanks to that when you inject the bean, it will be possible to act on it with Mockito (remember to include the Config class in the #SpringJUnitConfig annotation).
I created another commit in the GitHub repository - the test passes. You can see the whole code there.
I am using Spring, Junit and Mockito. I need to override beans defined in the main spring configuration using another mockito test configuration (injecting mock beans only as needed). Nested beans have been #Autowired in the application.
Update:
Based on alfcope's answer below, it is important to add the name attribute so that spring can allow the primary bean (mock) to override the original one. Otherwise you get this exception:
org.mockito.exceptions.misusing.MissingMethodInvocationException:
when() requires an argument which has to be 'a method call on a mock'.
For example:
when(mock.getArticles()).thenReturn(articles);
The info message in the spring log shows:
Skipping bean definition for [BeanMethod:name=bar,declaringClass=test.package.MockitoTestConfiguration]: a definition for bean 'bar' already exists. This top-level bean definition is considered as an override.
Example:
I have a simplified example below that works. Here, Bar is the nested inside Foo, and I need to mock Bar for testing:
#Component
public class Foo
{
#Autowired
private Bar bar;
public String getResponseFromBar(String request)
{
String response = bar.someMethod(String request);
//do something else with this reponse
return response;
}
}
#Component
public class Bar {
public String someMethod(String request) {
String response = null;
//do something
return response;
}
}
Now for testing, let's say I want to inject a mockbar instead of the real bar. How can I achieve this in my test class below?
#Profile("test")
#Configuration
public class MockitoTestConfiguration {
//adding the name attribute is important.
#Bean(name="mockBar")
#Primary
public Bar bar() {
logger.debug("injecting mock bar");
return Mockito.mock(Bar.class);
}
}
Actual test case:
#RunWith(SpringJUnit4ClassRunner.class)
#ContextConfiguration(locations = "classpath*:test-context.xml")
public class FooTest {
#Autowired
Foo foo;
#Autowired
Bar mockBar; //need this to set up the mock response in the test case.
#Test
public void testAMethodInFoo_WithBarInjectedByMockito() {
//set up the mockBar response
Mockito.when(mockBar.someMethod("1")).thenReturn("1-response");
String response = foo.getResponseFromBar();
assertEquals("1-response", response);
}
}
Based on the ConfigurationClassBeanDefinitionReader code I guess you are using xml configuration to define your main bean. If so, just add a name when creating your mockito bean.
#Bean(name="mockbar")
#Primary
public Bar bar() {
logger.debug("injecting mock bar");
return Mockito.mock(Bar.class);
}
This way Spring will allow you to have both beans, and as you are using #Primary it will be the one used by your tests.
Spring overriding primary bean with non-primary bean
ConfigurationClassBeanDefinitionReader Code
Alternatively, if you use Mockito, you can do this and completely do away with the extra MockitoTestConfiguration class and named primary mock beans in the test profile. Just simply do this:
#RunWith(SpringJUnit4ClassRunner.class)
#ContextConfiguration(locations = "classpath*:test-context.xml")
public class FooTest {
#Autowired
#InjectMocks
Foo foo;
#Mock
Bar mockBar; //need this to set up the mock response in the test case.
#Before
public void setUp() throws Exception {
MockitoAnnotations.initMocks(this);
}
#Test
public void testAMethodInFoo_WithBarInjectedByMockito() {
//set up the mockBar response
Mockito.when(mockBar.someMethod("1")).thenReturn("1-response");
String response = foo.getResponseFromBar();
assertEquals("1-response", response);
}
}
There have been several arguments around not using ApplicationContext.getBean() to get a bean reference, of which most are based on logic that it violates the principles of Inversion of control.
Is there a way to get reference to prototype scoped bean without calling context.getBean() ?
Consider to use Spring Boot!
Than you can do something like this...
Runner:
#SpringBootApplication
public class Runner{
public static void main(String[] args) {
SpringApplication.run(Runner.class, args);
}
}
Some Controller:
#Controller
public class MyController {
// Spring Boot injecting beans through #Autowired annotation
#Autowired
#Qualifier("CoolFeature") // Use Qualifier annotation to mark a class, if for example
// you have more than one concreate class with differant implementations of some interface.
private CoolFeature myFeature;
public void testFeature(){
myFeature.doStuff();
}
}
Some cool feature:
#Component("CoolFeature") // To identify with Qualifier
public class CoolFeature{
#Autowired
private SomeOtherBean utilityBean;
public void doStuff(){
// use utilityBean in some way
}
}
No XML files to handle.
We can still access context for manual configurations if needed.
Suggested reading:
Spring Boot Reference
Pro Spring Boot
This type of problem can be solved using method injection, which is described in more detail here: https://docs.spring.io/spring/docs/current/spring-framework-reference/html/beans.html#beans-factory-method-injection
This is the most common approach to create prototype bean:
abstract class MyService {
void doSome() {
OtherService otherService = getOtherService();
}
abstract OtherService getOtherService();
}
#Configuration
class Config {
#Bean
public MyService myService() {
return new MyService() {
OtherService getOtherService() {
return otherService();
}
}
}
#Bean
#Scope("prototype")
public OtherService otherService() {
return new OtherService();
}
}
I'm a Spring rookie and trying to benefit from the advantages of the easy 'profile' handling of Spring. I already worked through this tutorial: https://spring.io/blog/2011/02/14/spring-3-1-m1-introducing-profile and now I'd like to adapt that concept to an easy example.
I've got two profiles: dev and prod. I imagine a #Configuration class for each profile where I can instantiate different beans (implementing a common interface respectively) depending on the set profile.
My currently used classes look like this:
StatusController.java
#RestController
#RequestMapping("/status")
public class StatusController {
private final EnvironmentAwareBean environmentBean;
#Autowired
public StatusController(EnvironmentAwareBean environmentBean) {
this.environmentBean = environmentBean;
}
#RequestMapping(method = RequestMethod.GET)
Status getStatus() {
Status status = new Status();
status.setExtra("environmentBean=" + environmentBean.getString());
return status;
}
}
EnvironmentAwareBean.java
public interface EnvironmentAwareBean {
String getString();
}
EnvironmentAwareBean.java
#Service
public class DevBean implements EnvironmentAwareBean {
#Override
public String getString() {
return "development";
}
}
EnvironmentAwareBean.java
#Service
public class ProdBean implements EnvironmentAwareBean {
#Override
public String getString() {
return "production";
}
}
DevConfig.java
#Configuration
#Profile("dev")
public class DevConfig {
#Bean
public EnvironmentAwareBean getDevBean() {
return new DevBean();
}
}
ProdConfig.java
#Configuration
#Profile("prod")
public class ProdConfig {
#Bean
public EnvironmentAwareBean getProdBean() {
return new ProdBean();
}
}
Running the example throws this exception during startup (SPRING_PROFILES_DEFAULT is set to dev):
(...) UnsatisfiedDependencyException: (...) nested exception is org.springframework.beans.factory.NoUniqueBeanDefinitionException: No qualifying bean of type [EnvironmentAwareBean] is defined: expected single matching bean but found 3: prodBean,devBean,getDevBean
Is my approach far from a recommended configuration? In my opinion it would make more sense to annotate each Configuration with the #Profile annotation instead of doing it for each and every bean and possibly forgetting some variants when new classes are added later on.
Your implementations of EnvironmentAwareBean are all annotated with #Service.
This means they will all be picked up by component scanning and hence you get more than one matching bean. Do they need to be annotated with #Service?
Annotating each #Configuration with the #Profile annotation is fine. Another way as an educational exercise would be to not use #Profile and instead annotate the #Bean or Config classes with your own implementation of #Conditional.
The method DefaultProduitGeneriqueService.valider is not catched by the method traceWhenReturnedValueDoesntExistOrNotNecessary and I don't understand why?
package fr.generali.nova.atp.service.metier.impl;
public class DefaultProduitGeneriqueService extends DefaultService implements IProduitGeneriqueService, IBacAware {
...
#Override
#Traceable(value = ETraceableMessages.VALIDATION_PRODUIT_GENERIQUE, hasReturnedValue=Traceable.HAS_NOT_RETURNS_VALUE)
public void valider(ElementNiveauUn element) {
...
}
...
}
package fr.generali.nova.atp.logging.advisor;
#Aspect
public class TraceableAdvisor {
#Before(value = "execution(* fr.generali.nova.atp.service.metier.impl.*.*(..)) && #annotation(traceable) && args(element)", argNames = "element,traceable")
public void traceWhenReturnedValueDoesntExistOrNotNecessary(ElementNiveauUn element, Traceable traceable) {
...
}
}
Assuming that the service interfaces are in package fr.generali.nova.atp.service.metier.api:
package fr.generali.nova.atp.service.metier.api;
public interface IProduitGeneriqueService {
void valider(ElementNiveauUn element);
}
And the service implementations are in package fr.generali.nova.atp.service.metier.impl:
package fr.generali.nova.atp.service.metier.impl;
public class DefaultProduitGeneriqueServiceImpl implements IProduitGeneriqueService {
#Override
#Traceable(value = ETraceableMessages.VALIDATION_PRODUIT_GENERIQUE, hasReturnedValue=Traceable.HAS_NOT_RETURNS_VALUE)
public void valider(ElementNiveauUn element) {
// TODO: implement
}
}
Your aspect should look like this:
package fr.generali.nova.atp.logging.advisor;
#Aspect
public class TraceableAdvisor {
#Before(value = "execution(* fr.generali.nova.atp.service.metier.api.*.*(..)) && #annotation(traceable) && args(element)", argNames = "element,traceable")
public void traceWhenReturnedValueDoesntExistOrNotNecessary(ElementNiveauUn element, Traceable traceable) {
// TODO: implement
System.err.println("traced...");
}
}
The default proxy strategy for Spring AOP is JDK interface-based proxies, so Your pointcut expression should match the interface method execution, not the implementation method execution, and Your poincut expression may match either interface mothod execution or implementation method execution.
And remember to include an AspectJAutoProxyCreator in your configuration using for example <aspectj-autoproxy /> tag.
And here is a simple test to prove everyting is working:
public class TraceableAdvisorTest {
#Configuration
public static class TestConfiguration {
#Bean
public IProduitGeneriqueService produitGeneriqueService() {
return new DefaultProduitGeneriqueServiceImpl();
}
#Bean
public TraceableAdvisor traceableAdvisor() {
return new TraceableAdvisor();
}
#Bean
public AnnotationAwareAspectJAutoProxyCreator autoProxyCreator() {
return new AnnotationAwareAspectJAutoProxyCreator();
}
}
private AnnotationConfigApplicationContext testApplicationContext;
#Test
public void testTraceWhenReturnedValueDoesntExistOrNotNecessary() {
this.testApplicationContext = new AnnotationConfigApplicationContext();
this.testApplicationContext.register(TestConfiguration.class);
this.testApplicationContext.refresh();
IProduitGeneriqueService service = BeanFactoryUtils.beanOfType(this.testApplicationContext, IProduitGeneriqueService.class);
System.err.println("BEFORE");
service.valider(null);
System.err.println("AFTER");
}
}
The err output is:
BEFORE
traced...
AFTER
For all combinations:
fr.generali.nova.atp.service.metier.api.*.*(..)
fr.generali.nova.atp.service.metier.impl.*.*(..)
fr.generali.nova.atp.service.metier..*.*(..)
Make sure both beans are properly configured, either through annotations or in your appCtx.
It looks like your Aspect is definitely right, but how about the other class? Is it Spring enabled?
Also, if both classes are indeed configured correctly, are you sure that the instance being passed is a Spring bean and not a "new" instance from a constructor?