As far as I know spring provides some ways to inject beans into non-managed classes.
It can be done explicitly with AutowireCapableBeanFactory. (How to inject dependencies into a self-instantiated object in Spring?)
But I've faced strange (IMHO) behavior, when spring performs such injection automatically.
Here is an example with spring batch,
Configuration:
#SpringBootConfiguration
public class ProcessorJobConfig {
//.....
#Bean(name = "pullRestTemplate")
public RestTemplate createPullRestTemplate() {
RestTemplate restTemplate = restTemplateBuilder.build();
return restTemplate;
}
#Bean(name = "step")
public Step step(#Autowired ItemReader<Measurement> itemReader,
#Autowired ItemProcessor<Measurement, Event> itemProcessor,
#Autowired ItemWriter<Event> itemWriter) {
return stepBuilderFactory.get("step")
.<Measurement, Event>chunk(Integer.MAX_VALUE)
.reader(itemReader)
.processor(itemProcessor)
.writer(itemWriter)
.build();
}
#Bean(name = "restProcessorJob")
public Job job(#Qualifier("step") Step step) throws Exception {
return jobBuilderFactory.get("restProcessorJob")
.start(step)
.build();
}
#Bean
public ItemReader<Measurement> itemReader() {
RestMeasureReader restMeasureReader = new RestMeasureReader(); // Use new() explicitly
return restMeasureReader;
}
//.....
}
Reader:
public class RestMeasureReader implements ItemReader<Measurement> {
private static final Logger LOGGER = LoggerFactory.getLogger(RestMeasureReader.class);
/**
* NOTE: This field will be injected automatically by spring, even we are using new() to create instance of this class.
*/
#Autowired
#Qualifier("pullRestTemplate")
private RestTemplate restTemplate;
#Override
public Measurement read() throws Exception, UnexpectedInputException, ParseException, NonTransientResourceException {
// do some stuff
}
}
And application itself
#EnableBatchProcessing
#EnableTask
#SpringBootApplication
public class TestAutowiredTaskApplication {
public static void main(String[] args) {
SpringApplication.run(TestAutowiredTaskApplication.class, args);
}
}
Even I use explicit new() to instantiate RestMeasureReader, its RestTemplate field will be injected afterwards.
Is it normal behavior? I do not expect spring to automatically inject fields when creating object with new().
If you are talking about using new inside of your #Configuration class, then yes it is normal behavior. This is you Spring java configs. So it's is Spring managed context. You are not going to call itemReader() in your code explicitly.
So, when you are going to do this:
#Autowired
private ItemReader<Measurement> iterReader;
you will get instance of your RestMeasureReader from Spring's IoC.
But if you will try to do explicitly call new RestMesureReader() inside of your code, you will get a new instance of RestMesureReader not a Spring Proxy with injected #Autowired fields.
Try to remove #Bean from your itemReader() method declaration and won't event be able to autowire RestMesureReader.
So basically #Configuration classes are just a Spring configuration, not a real java code. Even though you call new Spring will still return you a proxy class.
For more information check this guide.
Spring processes beans returned by methods that are annotated with #Bean
This allows you to use autowiring or livecycle callbacks when using Java configuration.
A more minimalistic example:
#Configuration
public class MyConfiguration {
#Bean
public A a() {
return new A();
}
static class A {
#Autowired
private B b;
#PostConstruct
public void onPostConstruct() {
System.out.println("postConstruct: " + b);
}
}
#Component
static class B {
}
}
Here, even if the bean named a is created manually, Spring will inject dependencies (b) and call #PostConstruct callbacks.
Related
I am new to spring framework. I have to use spring boot and have a rest controller as below :-
#RestController
public class StatisticsController {
private TransactionCache transactionCache;
public StatisticsController(TransactionCache transactionCache) {
this.transactionCache = transactionCache;
}
#PostMapping("/tick")
public ResponseEntity<Object> addInstrumentTransaction(#Valid #RequestBody InstrumentTransaction instrumentTransaction) {
transactionCache.addTransaction(instrumentTransaction);
return new ResponseEntity<>(HttpStatus.CREATED);
}
and I have a class which needs to be singleton :-
#Component
public class TransactionStatisticsCacheImpl implements TransactionCache {
private static TransactionStatisticsCacheImpl instance;
public static TransactionStatisticsCacheImpl getInstance(){
if(Objects.isNull(instance)){
synchronized (TransactionStatisticsCacheImpl.class) {
if(Objects.isNull(instance)){
instance = new TransactionStatisticsCacheImpl();
}
}
}
return instance;
}
private TransactionStatisticsCacheImpl() {}
I want to know the correct way to call this singleton class in my rest controller. I know that by default the scope of a bean in spring is singleton. Is this the correct way to call the singleton class in rest controller?
#RestController
public class StatisticsController {
private TransactionCache transactionCache;
public StatisticsController(TransactionCache transactionCache) {
this.transactionCache = transactionCache;
}
#PostMapping("/tick")
public ResponseEntity<Object> addInstrumentTransaction(#Valid #RequestBody InstrumentTransaction instrumentTransaction) {
transactionCache.addTransaction(instrumentTransaction);
return new ResponseEntity<>(HttpStatus.CREATED);
}
or
We need to call it using the getInstance() method? Also do we need to explicitly have the getInstance method in the TransactionStatisticsCacheImpl class?
One of the major advantages of container injection is that you can get the benefits of singleton semantics without all the serious problems of "hard" singletons (such as difficulty testing). Get rid of the getInstance manual business and let Spring take care of ensuring that a single instance is created and used for the context.
Just for clarification: By default, the spring IOC container will create only one instance per bean definition, unless if you specified otherwise using the #Scope stereotype. But if you create an instance using getInstance() the bean pre-processors and post-processors will not work correctly on that bean definition. And also you can use the #Autowired stereotype to inject a bean definition as needed and if you have different implementations for the same definition you can use the #Qualifier stereotype to specify the implementation that you need to inject, alternatively, you can use the constructor injection to inject your bean definition as needed without auto wiring as mentioned here Spring #Autowire on Properties vs Constructor
I would stick to the answers above. However, if you want to preserve further instantiation of the class in your code (or you want to keep your specific implementation of singleton), you can do it with getInstance().
Firstly, get rid of #Component annotation in your class:
// #Component
public class TransactionStatisticsCacheImpl implements TransactionCache {
private static TransactionStatisticsCacheImpl instance;
public static TransactionStatisticsCacheImpl getInstance(){
if(Objects.isNull(instance)){
synchronized (TransactionStatisticsCacheImpl.class) {
if(Objects.isNull(instance)){
instance = new TransactionStatisticsCacheImpl();
}
}
}
return instance;
}
private TransactionStatisticsCacheImpl() {}
}
Then, you may instantiate your singleton #Bean by defining #Configuration class - this way your bean would get managed by spring container.
#Configuration
public class SingletonConfiguration {
#Bean
public TransactionCache transactionCache() {
return TransactionCacheImpl.getInstance();
}
}
Eventually, you can have your singleton injected in your RestController using #Autowired.
#RestController
public class StatisticsController {
private TransactionCache transactionCache;
#Autowired
public StatisticsController(TransactionCache transactionCache) {
this.transactionCache = transactionCache;
}
#PostMapping("/tick")
public ResponseEntity<Object> addInstrumentTransaction(#Valid #RequestBody InstrumentTransaction instrumentTransaction) {
transactionCache.addTransaction(instrumentTransaction);
return new ResponseEntity<>(HttpStatus.CREATED);
}
}
Why can't I use #Autowired in this case?
#SpringBootApplication
public class Application {
#Autowired
BookingService bookingService;
public static void main(String[] args) {
bookingService.book("Alice", "Bob", "Carol");
}
}
but can use #Bean
#SpringBootApplication
public class Application {
#Bean
BookingService bookingService() {
return new BookingService();
}
public static void main(String[] args) {
ApplicationContext ctx = SpringApplication.run(Application.class, args);
BookingService bookingService = ctx.getBean(BookingService.class);
bookingService.book("Alice", "Bob", "Carol");
}
}
Aren't the two ways to generate BookingService the same thing?
#Bean and #Autowired do two very different things. The other answers here explain in a little more detail, but at a simpler level:
#Bean tells Spring 'here is an instance of this class, please keep hold of it and give it back to me when I ask'.
#Autowired says 'please give me an instance of this class, for example, one that I created with an #Bean annotation earlier'.
Does that make sense? In your first example, you're asking Spring to give you an instance of BookingService, but you're never creating one, so Spring has nothing to give you. In your second example, you're creating a new instance of BookingService, telling Spring about it, and then, in the main() method, asking for it back.
If you wanted, you could remove the two additional lines from the second main() method, and combine your two examples as below:
#SpringBootApplication
public class Application {
#Autowired
BookingService bookingService;
#Bean
BookingService bookingService() {
return new BookingService();
}
public static void main(String[] args) {
bookingService.book("Alice", "Bob", "Carol");
}
}
In this case, the #Bean annotation gives Spring the BookingService, and the #Autowired makes use of it.
This would be a slightly pointless example, as you're using it all in the same class, but it becomes useful if you have the #Bean defined in one class, and the #Autowired in a different one.
#Bean
BookingService bookingService() {
return new BookingService();
}
Annotating #Bean only registers the service as a bean(kind of an Object) in the spring application context. In simple words, it is just registration and nothing else.
#Autowired
BookingService bookingService;
Annotating a variable with #Autowired injects a BookingService bean(i.e Object) from Spring Application Context.
(i.e) The registered object with #Bean annotation will be injected to the variable annotated with #Autowired.
Hope this clears your doubt!
great answer by #DaveyDaveDave
In the example instead of
#Bean
BookingService bookingService() {
return new BookingService();
}
You can use #Service annotation on BookingService class
Contrary to what the highest voted answer here claims, they are NOT two very different things. #Bean and #Autowired and interchangeable in most cases.
Suppose you have a #Bean method that returns an instance of a Car. You can literally get rid of that bean method and add #Component on the Car class and then autowire it.
And vice versa. Whatever class you have instantiated using #Autowired, you can instantiate it inside a class with #Configuration annotation using #Bean on the method.
Places where you will use #Bean instead of #Autowired
1>You do not have access to change the class to add #Component annotation, hence you cannot autowire it.
2>You want to customize the instantiation of the class.
For example if you are instantiating a Resilience4J Circuit breaker class, if you do it inside a method with #Bean, you have the option of setting all the config using code like this
#Bean
public CircuitBreaker fooCircuitBreaker() {
CircuitBreakerConfig.Builder builder = CircuitBreakerConfig.custom().
slidingWindowSize(xxx).
failureRateThreshold(xxx).
waitDurationInOpenState(xxx)).
ignoreException(e -> {
if (e instanceof HttpStatusCodeException) {
HttpStatusCodeException httpStatusCodeException = (HttpStatusCodeException) e;
if (httpStatusCodeException.getStatusCode().is4xxClientError()) {
return true;
}
}
return false;
});
circuitBreakerRegistry.addConfiguration(xxx, builder.build());
return circuitBreakerRegistry.circuitBreaker(xxx, xxx);
}
Here's good article about #Autowired annotation: http://www.baeldung.com/spring-autowire
The #Autowired annotation can instantiate your injectables by defining #ComponentScan("namespace.with.your.components.for.inject") on config class
#Configuration
#ComponentScan("com.baeldung.autowire.sample")
public class AppConfig {}
All components must be marked by #Component annotation. It replaces the #Bean annotation.
#Bean is just for the metadata definition to create the bean(equivalent to tag). #Autowired is to inject the dependancy into a bean(equivalent to ref XML tag/attribute).
I have multiple library classes in my project which need to be injected into a service class. This is the error statement for IntegrationFactory class:
Consider defining a bean of type 'com.ignitionone.service.programmanager.integration.IntegrationFactory' in your configuration.
This error is coming on almost every injection where this library class is injected.
I have already added the Library package in #ComponentScan, but, as it is read-only file, I can not annotate the library class. I came to know from some answer here that Spring can not inject classes which it does not manage. This library is not built on spring.
I have tried to create a #Bean method which returns the IntegrationFactory(class in question) in the class where #Inject is used, but this too does not seem to work.
How can this be done, preferably without creating a stub/copy class?
This is EngagementServiceImpl class snippet:
#Inject
public EngagementServiceImpl(EngagementRepository engagementRepository,
#Lazy IntegrationFactory integrationFactory, TokenRepository tokenRepository,
EngagementPartnerRepository engagementPartnerRepository, MetricsService metricsService) {
this.engagementRepository = engagementRepository;
this.integrationFactory = integrationFactory;
this.tokenRepository = tokenRepository;
this.engagementPartnerRepository = engagementPartnerRepository;
this.metricsService = metricsService;
}
This is injection part:
#Autowired
private EngagementService engagementService;
This is ConfigClass:
#Configuration
public class ConfigClass {
#Bean
public IntegrationFactory getIntegrationFactory(){
Map<String, Object> globalConfig = new HashMap<>();
return new IntegrationFactory(globalConfig);
}
#Bean
#Primary
public EntityDataStore getEntityDataStore(){
EntityModel entityModel = Models.ENTITY;
return new EntityDataStore(this.dataSource(), entityModel );
}
#ConfigurationProperties(prefix = "datasource.postgres")
#Bean
#Primary
public DataSource dataSource() {
return DataSourceBuilder
.create()
.build();
}
}
You need to add your bean definitions in a configuration class.
#Configuration
public class ServiceConfig {
#Bean
public IntegrationFactory getIntegrationFactory(){
// return an IntegrationFactory instance
}
}
Then you have to make sure your #Configuration class gets detected by Spring, either by having it within your scanned path or by manually importing it via #Import from somewhere withing you scanned path. An example of #Import, considering you are using Spring Boot.
#Import(ServiceConfig.class)
#SpringBootApplication
public class Application {
public static void main(String[] args) {
SpringApplication.run(Application.class, args);
}
}
Hope this helps!
Your Bean IntegrationFactory can't be found, as it is not annotated with any Spring stereotype and therefore not recognized by the component scan.
As you have multiple options to provide an instance of your class to the application context, read the Spring documentation (which also includes samples) to find out which one fits you the most:
https://docs.spring.io/spring/docs/5.1.0.RELEASE/spring-framework-reference/core.html#beans-java-basic-concepts
One Option would be to create a factory which provides an instance of your class to the application context, like it is stated in the documentation:
#Configuration
public class AppConfig {
#Bean
public IntegrationFactory myIntegrationFactory() {
return new IntegrationFactory();
}
}
Do not forget to add the Configuration to the application context.
I have several classes in a Spring Boot project, some work with #Autowired, some do not. Here my code follows:
Application.java (#Autowired works):
package com.example.myproject;
#ComponentScan(basePackages = {"com.example.myproject"})
#Configuration
#EnableAutoConfiguration
#EnableJpaRepositories(basePackages = "com.example.myproject.repository")
#PropertySource({"classpath:db.properties", "classpath:soap.properties"})
public class Application {
#Autowired
private Environment environment;
public static void main(String[] args) {
SpringApplication.run(Application.class);
}
#Bean
public SOAPConfiguration soapConfiguration() {
SOAPConfiguration SOAPConfiguration = new SOAPConfiguration();
SOAPConfiguration.setUsername(environment.getProperty("SOAP.username"));
SOAPConfiguration.setPassword(environment.getProperty("SOAP.password"));
SOAPConfiguration.setUrl(environment.getProperty("SOAP.root"));
return SOAPConfiguration;
}
HomeController (#Autowired works):
package com.example.myproject.controller;
#Controller
class HomeController {
#Resource
MyRepository myRepository;
MyService (#Autowired does not work):
package com.example.myproject.service;
#Service
public class MyServiceImpl implements MyService {
#Autowired
public SOAPConfiguration soapConfiguration; // is null
private void init() {
log = LogFactory.getLog(MyServiceImpl.class);
log.info("starting init, soapConfiguration: " + soapConfiguration);
url = soapConfiguration.getUrl(); // booom -> NullPointerException
I do not get the SOAPConfiguration but my application breaks with a null pointer exception when I try to access it.
I have already read many Threads here and googled around, but did not find a solution yet. I tried to deliver all necessary information, please let me know if anything misses.
I guess you call init() before the autowiring takes place. Annotate init() with #PostConstruct to make it call automatically after all the spring autowiring.
EDIT: after seeing your comment, I guess you are creating it using new MyServiceImpl(). This takes away the control of the MyServiceImpl from Spring and gives it to you. Autowiring won't work in those case
Did you created a bean for the class SOAPConfiguration in any of your configuration classes? If you want to autowire a class in your project, you need to create a bean for it. For example,
#Configuration
public class SomeConfiguration{
#Bean
public SOAPConfiguration createSOAPConfiguration(){
return new SOAPConfiguration();
}
}
public class SomeOtherClass{
#Autowired
private SOAPConfiguration soapConfiguration;
}
I want configure a component test using spring-test configuration inner class (#Configuration). Tested components has some services which I'd like to mock for the test. These services are classes (no interface used) and have spring annotations (#Autowired) in them. Mockito can easily mock them, however, I found no way of disabling spring autowiring.
Example how I can easily reproduce:
#RunWith(SpringJUnit4ClassRunner.class)
#ContextConfiguration(classes = SomeTest.Beans.class)
public class SomeTest {
// configured in component-config.xml, using ThirdPartyService
#Autowired
private TestedBean entryPoint;
#Test
public void test() {
}
#Configuration
#ImportResource("/spring/component-config.xml")
static class Beans {
#Bean
ThirdPartyService createThirdPartyService() {
return mock(ThirdPartyService.class);
}
}
}
public class ThirdPartyService {
#Autowired
Foo bar;
}
public class TestedBean {
#Autowired
private ThirdPartyService service;
}
In this example "TestBean" represents the service to be mocked. I would NOT like "bar" to be injected by spring! #Bean(autowire = NO) does not help (in fact, that's the default value).
(Please save me from "use interfaces!" comments - the mocked service can be 3rd party which I can't do anything with.)
UPDATE
Springockito partially solves the problem, as long as you don't have to have anything else to configure (so you can't use configuration class with Springockito - it does not support it), but use mocks only.
Still looking for pure spring solution, if there's any...
Here is my solution to your problem:
import static org.mockito.Mockito.mockingDetails;
import org.springframework.beans.BeansException;
import org.springframework.beans.factory.config.InstantiationAwareBeanPostProcessorAdapter;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
#Configuration
public class MockitoSkipAutowireConfiguration {
#Bean MockBeanFactory mockBeanFactory() {
return new MockBeanFactory();
}
private static class MockBeanFactory extends InstantiationAwareBeanPostProcessorAdapter {
#Override
public boolean postProcessAfterInstantiation(Object bean, String beanName) throws BeansException {
return !mockingDetails(bean).isMock();
}
}
}
and then just
#Import(MockitoSkipAutowireConfiguration.class)
in your test #Configuration and you are all set
I solved it by creating FactoryBean for my bean instead of just mocking bean. At this way Spring don't try to autowire fields.
Factory bean helping class:
public class MockitoFactoryBean<T> implements FactoryBean<T> {
private final Class<T> clazz;
public MockitoFactoryBean(Class<T> clazz) {
this.clazz = clazz;
}
#Override public T getObject() throws Exception {
return mock(clazz);
}
#Override public Class<T> getObjectType() {
return clazz;
}
#Override public boolean isSingleton() {
return true;
}
}
Actual test context part:
#Configuration
public class TestContext {
#Bean
public FactoryBean<MockingService> mockingService() {
return new MockitoFactoryBean<>(MockingService.class);
}
}
Check Spring profiles. You don't need to disable auto wiring, you need to inject different beans for different configuration.
You could add the mocked service manually to the spring application context via org.springframework.beans.factory.config.SingletonBeanRegistry#registerSingleton. This way the mock is not post-processed by spring and spring does not attempt to autowire the mock. The mock itself will be injected into your tested bean.
#RunWith(SpringJUnit4ClassRunner.class)
#ContextConfiguration(classes = SomeTest.Beans.class)
public class SomeTest {
// configured in component-config.xml, using ThirdPartyService
#Autowired
private TestedBean entryPoint;
#Autowired
private ThirdPartyService thirdPartyServiceMock;
#Test
public void test() {
}
#Configuration
static class Beans {
#Autowired
private GenericApplicationContext ctx;
#Bean
TestedBean testedBean() {
ctx.getBeanFactory().registerSingleton("thirdPartyService", mock(ThirdPartyService.class));
return new TestedBean();
}
}
public static class ThirdPartyService {
#Autowired
Object bar;
}
public static class TestedBean {
#Autowired
private ThirdPartyService service;
}
}
I am in quite the same situation.
What I found that if you do not set the context loader by #ContextConfiguration annotation on your test class, the default context loader will be used, which derived from AbstractGenericContextLoader. I had a look at its source and turned out it registers all the bean post processors which are responsible for reading annotations such #Autowired. In other words, annotation config is enabled by default.
So the main problem is that there are two configurations which are in conflict: in the java config we said that autowiring is not needed, while the autowired annotation tells the opposite. The real question is how to disable the annotation processing in order to eliminate the undesired configuration.
As far as I know there is no such spring implementation of ContextLoader which would not be derived from AbstractGenericContextLoader so I guess the only we can do is to write our own. It would be something like this:
public static class SimpleContextLoader implements ContextLoader {
#Override
public String[] processLocations(Class<?> type, String... locations) {
return strings;
}
#Override
public ApplicationContext loadContext(String... locations) throws Exception {
// in case of xml configuration
return new ClassPathXmlApplicationContext(strings);
// in case of java configuration (but its name is quite misleading)
// return new AnnotationConfigApplicationContext(TestConfig.class);
}
}
Of course it would be worth to spend more time to find out how to implement ContextLoader properly.
Cheers,
Robert
There are so many ways of doing this, I'm pretty sure that this answer will be incomplete, but here are a few options...
As currently seems to be recommended practice, use constructor injection for your services rather than autowiring the fields directly. This makes testing like this so much easier.
public class SomeTest {
#Mock
private ThirdPartyService mockedBean;
#Before
public void init() {
initMocks(this);
}
#Test
public void test() {
BeanUnderTest bean = new BeanUnderTest(mockedBean);
// ...
}
}
public class BeanUnderTest{
private ThirdPartyService service;
#Autowired
public BeanUnderTest(ThirdPartyService ThirdPartyService) {
this.thirdPartyService = thirdPartyService;
}
}
By doing that, you can also mix up autowired and mocked services by autowiring into the test itself and then constructing the beans under test with the most useful mix of autowired and mocked beans.
A reasonable alternative is to use Spring profiles to define stub services. This is particularly useful when wish to use the same stubbed features in multiple tests:
#Service
#Primary
#Profile("test")
public class MyServiceStub implements MyService {
// ...
}
#RunWith(SpringJUnit4ClassRunner.class)
#ContextConfiguration(classes = SomeTest.Beans.class)
#ActiveProfiles({"test"})
public class SomeTest {
// ...
}
By using the #Primary annotation, it ensures that this stub bean will be used instead of any other bean implementing the MyService interface. I tend to use this approach for things like email services, where by changing profile, I'm able to switch between a real mail server and Wiser.