There are many other frameworks for Java development that lock you in by forcing you to extend or implement one of their existing classes or interfaces; Struts, Tapestry, and earlier versions of EJB had this approach. The programming model of these frameworks is based on the invasive model. This makes it harder for your code to find bugs in the system, and sometimes it will render your code unintelligible. However, if you are working with Spring Framework, you don't need to implement or extend its existing classes and interfaces, so this is simply POJO-based implementation, following a non-invasive programming model. It makes it easier for your code to find bugs in the system, and keeps the code understandable.

Spring allows you to do programming with very simple non Spring classes, which means there is no need to implement Spring-specific classes or interfaces, so all classes in the Spring-based application are simply POJOs. That means you can compile and run these files without dependency on Spring libraries; you cannot even recognize that these classes are being used by the Spring Framework. In Java-based configuration, you will use Spring annotations, which is the worst case of the Spring-based application.

Let's look at this with the help of the following example:

    package com.packt.chapter1.spring; 
    public class HelloWorld { 
      public String hello() { 
        return "Hello World"; 
      } 
    } 

The preceding class is a simple POJO class with no special indication or implementation related to the framework to make it a Spring component. So this class could function equally well in a Spring application as it could in a non-Spring application. This is the beauty of Spring's non-invasive programming model. Another way that Spring empowers POJO is by collaborating with other POJOs using the DI pattern. Let's see how DI works to help decouple components.

The term dependency injection is not new-it is used by PicoContainer. Dependency injection is a design pattern that promotes loose coupling between the Spring components--that is, between the different collaborating POJOs. So by applying DI to your complex programming, your code will become simpler, easier to understand, and easier to test.

In your application, many objects are working together for a particular functionality as per your requirement. This collaboration between the objects is actually known as dependency injection. Injecting dependency between the working components helps you to unit test every component in your application without tight coupling.

In a working application, what the end user wants is to see the output. To create the output, a few objects in the application work together and are sometimes coupled. So when you are writing these complex application classes, consider the reusability of these classes and make these classes as independent as possible. This is a best practice of coding that will help you in unit testing these classes independently.

How DI works and makes things easy for development and testing

Let's look at DI pattern implementation in your application. It makes things easy to understand, loosely coupled, and testable across the application. Suppose we have a simple application (something more complex than a Hello World example that you might make in your college classes). Every class is working together to perform some business task and help build business needs and expectations. That means that each class in the application has its measure of responsibility for a business task, together with other collaborating objects (its dependencies). Let's look at the following image. This dependency between the objects can create complexity and tight coupling between the dependent objects:

The TransferService component is traditionally dependent on two other components: TransferRepository and AccountRepository

A typical application system consists of several parts working together to carry out a use case. For example, consider the TransferService class, shown next.

TransferService using direct instantiation:

    package com.packt.chapter1.bankapp.transfer; 
    public class TransferService { 
      private AccountRepository accountRepository; 
      public TransferService () { 
        this.accountRepository = new AccountRepository(); 
      } 
      public void transferMoney(Account a, Account b) { 
        accountRepository.transfer(a, b); 
      } 
    } 

The TransferService object needs an AccountRepository object to make money transfer from account a to account b. Hence, it creates an instance of the AccountRepository object directly and uses it. But direct instantiation increases coupling and scatters the object creation code across the application, making it hard to maintain and difficult to write a unit test for TransferService, because, in this case, whenever you want to test the transferMoney() method of the TransferService class by using the assert to unit test, then the transfer() method of the AccountRepository class is also called unlikely by this test. But the developer is not aware about the dependency of AccountRepository on the TransferService class; at least, the developer is not able to test the transferMoney() method of the TransferService class using unit testing.

In enterprise applications, coupling is very dangerous, and it pushes you to a situation where you will not be able to do any enhancement in the application in the future, where any further changes in such an application can create a lot of bugs, and where fixing these bugs can create new bugs. Tightly coupled components are one of the reasons for major problems in these applications. Unnecessary tightly coupled code makes your application non-maintainable, and as time goes by, its code will not be reused, as it cannot be understood by other developers. But sometimes a certain amount of coupling is required for an enterprise application because completely uncoupled components are not possible in real-world cases. Each component in the application has some responsibility for a role and business requirement, to the extent that all components in the application have to be aware of the responsibility of the other components. That means coupling is necessary sometimes, but we have to manage the coupling between required components very carefully.

    package com.packt.chapter1.bankapp.transfer; 
    public class TransferService { 
      private AccountRepository accountRepository; 
      public TransferService() { 
        this.accountRepository = 
          AccountRepositoryFactory.getInstance("jdbc"); 
      } 
      public void transferMoney(Account a, Account b) { 
        accountRepository.transfer(a, b); 
      } 
    } 

    public interface AccountRepository{ 
      void transfer(); 
      //other methods 
    } 

    public class JdbcAccountRepositry implements AccountRepositry{ 
      //...implementation of methods defined in AccountRepositry 
      // ...implementation of other methods 
    } 

According to the DI pattern, dependent objects are given their dependencies at the time of the creation of the objects by some factory or third party. This factory coordinates each object in the system in such a way that each dependent object is not expected to create their dependencies. This means that we have to focus on defining the dependencies instead of resolving the dependencies of collaborating objects in the enterprise application. Let's look at the following image. You will learn that dependencies are injected into the objects that need them:

Dependency injection between the different collaborating components in the application

To illustrate this point, let's look at TransferService in the next section--a TransferService has dependency with AccountRepository and TransferRepository. Here, TransferService is capable of transferring money by any kind implementation of TransferRepository, that is, we can either use JdbcTransferRepository or JpaTransferRepository, depending on whichever comes along according to the deployment environment.

TransferServiceImpl is flexible enough to take on any TransferRepository it's given:

    package com.packt.chapter1.bankapp; 
    public class TransferServiceImpl implements TransferService { 
      private TransferRepository transferRepository; 
      private AccountRepository  accountRepository; 
      public TransferServiceImpl(TransferRepository transferRepository,
       AccountRepository  accountRepository) { 
         this.transferRepository =
          transferRepository;//TransferRepository is injected 
         this.accountRepository  = accountRepository; 
         //AccountRepository is injected 
       } 
       public void transferMoney(Long a, Long b, Amount amount) { 
         Account accountA = accountRepository.findByAccountId(a); 
         Account accountB = accountRepository.findByAccountId(b); 
         transferRepository.transfer(accountA, accountB, amount); 
       } 
    } 

Here you can see that TransferServiceImpl doesn't create its own repositories implementation. Instead, we have given the implementation of repositories at the time of construction as a constructor argument. This is a type of DI known as constructor injection. Here we have passed the repository interface type as an argument of the constructor. Now TransferServiceImpl could use any implementation of repositories, either JDBC, JPA, or mock objects. The point is that TransferServiceImpl isn't coupled to any specific implementation of repositories. It doesn't matter what kind of repository is used to transfer an amount from one account to another account, as long as it implements the repositories interfaces. If you are using the DI pattern of the Spring Framework, loose coupling is one of the key benefits. The DI pattern always promotes P2I, so each object knows about its dependencies by their associated interface rather than associated implementation, so the dependency can easily be swapped out with another implementation of that interface instead of changing to its dependent class implementation.

Spring provides support for assembling such an application system from its parts:

The method for assembling an application system by creating associations between application parts or components is known as wiring. In Spring, there are many ways to wire collaborating components together to make an application system. For instance, we could use either an XML configuration file or a Java configuration file.

Now let's look at how to inject the dependencies of TransferRepository and AccountRepository into a TransferService with Spring:

    <?xml version="1.0" encoding="UTF-8"?> 
    <beans xmlns="http://www.springframework.org/schema/beans" 
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
    xsi:schemaLocation="http://www.springframework.org/schema/beans 
    http://www.springframework.org/schema/beans/spring-beans.xsd"> 
    <bean id="transferService"  
     class="com.packt.chapter1.bankapp.service.TransferServiceImpl"> 
         <constructor-arg ref="accountRepository"/> 
         <constructor-arg ref="transferRepository"/> 
    </bean> 
    <bean id="accountRepository" class="com.
     packt.chapter1.bankapp.repository.JdbcAccountRepository"/> 
    <bean id="transferRepository" class="com.  
     packt.chapter1.bankapp.repository.JdbcTransferRepository"/>     
    
    </beans> 

Here, TransferServiceImpl, JdbcAccountRepository, and JdbcTransferRepository are declared as beans in Spring. In the case of the TransferServiceImpl bean, it's constructed, passing a reference to the AccountRepository and TransferRepository beans as constructor arguments. You might like to know that Spring also allows you to express the same configuration using Java.

Spring offers Java-based configuration as an alternative to XML:

    package com.packt.chapter1.bankapp.config; 
 
    import org.springframework.context.annotation.Bean; 
    import org.springframework.context.annotation.Configuration; 
 
    import com.packt.chapter1.bankapp.repository.AccountRepository; 
    import com.packt.chapter1.bankapp.repository.TransferRepository; 
    import 
     com.packt.chapter1.bankapp.repository.jdbc.JdbcAccountRepository; 
    import 
     com.packt.chapter1.bankapp.repository.jdbc.JdbcTransferRepository; 
    import com.packt.chapter1.bankapp.service.TransferService; 
    import com.packt.chapter1.bankapp.service.TransferServiceImpl; 
 
    @Configuration 
    public class AppConfig { 
    
     @Bean 
     public TransferService transferService(){ 
       return new TransferServiceImpl(accountRepository(),
       transferRepository()); 
     } 
     @Bean 
     public AccountRepository accountRepository() { 
       return new JdbcAccountRepository(); 
     } 
     @Bean 
     public TransferRepository transferRepository() { 
       return new JdbcTransferRepository(); 
     } 
    } 

The benefits of the dependency injection pattern are the same whether you are using an XML-based or a Java-based configuration:

Although TransferService depends on an AccountRepository and TransferRepository, it doesn't care about what type (JDBC or JPA) of implementations of AccountRepository and TransferRepository are used in the application. Only Spring, through its configuration (XML- or Java-based), knows how all the components come together and are instantiated with their required dependencies using the DI pattern. DI makes it possible to change those dependencies with no changes to the dependent classes--that is, we could use either a JDBC implementation or a JPA implementation without changing the implementation of AccountService.

In a Spring application, an implementation of the application context (Spring offers AnnotationConfigApplicationContext for Java-based and ClassPathXmlApplicationContext for XML-based implementations) loads bean definitions and wires them together into a Spring container. The application context in Spring creates and wires the Spring beans for the application at startup. Look into the implementation of the Spring application context with Java-based configuration--It loads the Spring configuration files (AppConfig.java for Java and Sprig.xml for XML) located in the application's classpath. In the following code, the main() method of the TransferMain class uses a AnnotationConfigApplicationContext class to load the configuration class AppConfig.java and get an object of the AccountService class.

Spring offers Java-based configuration as an alternative to XML:

    package com.packt.chapter1.bankapp; 
 
    import org.springframework.context.ConfigurableApplicationContext; 
    import 
     org.springframework.context.annotation
     .AnnotationConfigApplicationContext; 
 
    import com.packt.chapter1.bankapp.config.AppConfig; 
    import com.packt.chapter1.bankapp.model.Amount; 
    import com.packt.chapter1.bankapp.service.TransferService; 
 
    public class TransferMain { 
 
      public static void main(String[] args) { 
        //Load Spring context 
        ConfigurableApplicationContext applicationContext = 
          new AnnotationConfigApplicationContext(AppConfig.class); 
         //Get TransferService bean 
         TransferService transferService = 
          applicationContext.getBean(TransferService.class); 
           //Use transfer method 
         transferService.transferAmmount(100l, 200l,
          new Amount(2000.0)); 
         applicationContext.close(); 
      } 
 
    }    

Here we have a quick introduction to the dependency injection pattern. You'll learn a lot more about the DI pattern in the coming chapters of this book. Now let's look at another way of simplifying Java development using Spring's declarative programming model through aspects and proxy patterns.

In a Spring application, the DI pattern provides us with loose coupling between collaborating software components, but Aspect-Oriented Programming in Spring (Spring AOP) enables you to capture common functionalities that are repetitive throughout your application. So we can say that Spring AOP promotes loose coupling and allows cross-cutting concerns, listed as follows, to be separated in a most elegant fashion. It allows these services to be applied transparently through declaration. With Spring AOP, it is possible to write custom aspects and configure them declaratively.

The generic functionalities that are needed in many places in your application are:

The components listed here are not part of your core application, but these components have some additional responsibilities, commonly referred to as cross-cutting concerns because they tend to cut across multiple components in a system beyond their core responsibilities. If you put these components with your core functionalities, thereby implementing cross-cutting concerns without modularization, it will have two major problems:

The following diagram illustrates this complexity. The business objects are too intimately involved with the cross-cutting concerns. Not only does each object know that it's being logged, secured, and involved in a transactional context, but each object is also responsible for performing those services assigned only to it:

Cross-cutting concerns, such as logging, security and transaction, are often scattered about in modules where those tasks are not their primary concern

Spring AOP enables the modularization of cross-cutting concerns to avoid tangling and scattering. You can apply these modularized concerns to the core business components of the application declaratively without affecting the aforementioned the above components. The aspects ensure that the POJOs remain plain. Spring AOP makes this magic possible by using the Proxy Design Pattern. We will discuss the Proxy Design pattern more in the coming chapters of this book.

How Spring AOP works

The following points describe the work of Spring AOP:

• Implement your mainline application logic: Focusing on the core problem means that, when you are writing the application business logic, you don't need to worry about adding additional functionalities, such as logging, security, and transaction, between the business codes-Spring AOP takes care of it.
• Write aspects to implement your cross-cutting concerns: Spring provides many aspects out of the box, which means you can write additional functionalities in the form of the aspect as independent units in Spring AOP. These aspects have additional responsibilities as cross-cutting concerns beyond the application logic codes.
• Weave the aspects into your application: Adding the cross-cutting behaviors to the right places, that is, after writing the aspects for additional responsibilities, you could declaratively inject them into the right places in the application logic codes.

Let's look at an illustration of AOP in Spring:

AOP-based system evolution--this leaves the application components to focus on their specific business functionalities

In the preceding diagram, Spring AOP separates the cross-cutting concerns, for example, security, transaction, and logging, from the business modules, that is, BankService, CustomerService, and ReportingService. These cross-cutting concerns are applied to predefined points (stripes in the preceding diagram) of the business modules at the running time of the application.

Suppose that you want to log the messages before and after calling the transferAmmount() method of TransferService using the services of a LoggingAspect. The following listing shows the LoggingAspect class you might use.

LoggingAspect call is used for logging the system for TransferService:

    package com.packt.chapter1.bankapp.aspect; 
 
    import org.aspectj.lang.annotation.After; 
    import org.aspectj.lang.annotation.Aspect; 
    import org.aspectj.lang.annotation.Before; 
 
    @Aspect 
    public class LoggingAspect { 
    
     @Before("execution(* *.transferAmount(..))") 
     public void logBeforeTransfer(){ 
       System.out.println("####LoggingAspect.logBeforeTransfer() 
       method called before transfer amount####"); 
     } 
    
     @After("execution(* *.transferAmount(..))") 
     public void logAfterTransfer(){ 
       System.out.println("####LoggingAspect.logAfterTransfer() method
       called after transfer amount####"); 
     } 
    } 

To turn LoggingAspect into an aspect bean, all you need to do is declare it as one in the Spring configuration file. Also, to make it an aspect, you have to add the @Aspect annotation to this class. Here's the updated AppConfig.java file, revised to declare LoggingAspect as an aspect.

Declaring LoggingAspect as an aspect and enabling the Apsect proxy feature of Spring AOP:

    package com.packt.chapter1.bankapp.config; 
 
    import org.springframework.context.annotation.Bean; 
    import org.springframework.context.annotation.Configuration; 
    import
     org.springframework.context.annotation.EnableAspectJAutoProxy; 
 
    import com.packt.chapter1.bankapp.aspect.LoggingAspect; 
    import com.packt.chapter1.bankapp.repository.AccountRepository; 
    import com.packt.chapter1.bankapp.repository.TransferRepository; 
    import
     com.packt.chapter1.bankapp.repository.jdbc.JdbcAccountRepository; 
    import
     com.packt.chapter1.bankapp.repository.jdbc.JdbcTransferRepository; 
    import com.packt.chapter1.bankapp.service.TransferService; 
    import com.packt.chapter1.bankapp.service.TransferServiceImpl; 
 
    @Configuration 
    @EnableAspectJAutoProxy 
    public class AppConfig { 
    
      @Bean 
      public TransferService transferService(){ 
        return new TransferServiceImpl(accountRepository(),
        transferRepository()); 
      } 
      @Bean 
      public AccountRepository accountRepository() { 
        return new JdbcAccountRepository(); 
      } 
      @Bean 
      public TransferRepository transferRepository() { 
        return new JdbcTransferRepository(); 
      } 
      @Bean 
      public LoggingAspect loggingAspect() { 
        return new LoggingAspect(); 
      } 
    } 

Here, we're using Spring's AOP configuration based on Java to declare the LoggingAspect bean as an aspect. First, we declare LoggingAspect as a bean. Then we annotate that bean with the @Aspect annotation.

We annotate logBeforeTransfer() of LoggingAspect with the @Before annotation so that this method is called before the transferAmount() is executed. This is called before advice. Then, we annotate another method of LoggingAspect with the @After annotation to declare that the logAfterTransfer() method should be called after transferAmount() has executed. This is known as after advice.

@EnableAspectJAutoProxy is used to enable Spring AOP features in the application. This annotation actually forces you to apply proxy to some of the components that are defined in the spring configuration file. We'll talk more about Spring AOP later, in Chapter 6, Spring Aspect Oriented Programming with Proxy and Decorator Pattern. For now, it's enough to know that you've asked Spring to call logBeforeTransfer() and logAferTransfer() of LoggingAspect before and after the transferAmount() method of the TransferService class. For now, there are two important points to take away from this example:

• LoggingAspect is still a POJO (if you ignore the @Aspect annotation or are using XML-based configuration)--nothing about it indicates that it's to be used as an aspect.
• It is important to remember that LoggingAspect can be applied to TransferService without TransferService needing to explicitly call it. In fact, TransferService remains completely unaware of the existence of LoggingAspect.

Let's move to another way that Spring simplifies Java development.

At one point in the enterprise application, we saw some code that looked like code we had already written before in the same application. That is actually boilerplate code. It is code that we often have to write again and again in the same application to accomplish common requirements in different parts of the application. Unfortunately, there are a lot of places where Java APIs involve a bunch of boilerplate code. A common example of boilerplate code can be seen when working with JDBC to query data from a database. If you've ever worked with JDBC, you've probably written something in code that deals with the following:

Let's look at the following code, it contains boilerplate code with the JDBC API of the Java:

    public Account getAccountById(long id) { 
      Connection conn = null; 
      PreparedStatement stmt = null; 
      ResultSet rs = null; 
      try { 
        conn = dataSource.getConnection(); 
        stmt = conn.prepareStatement( 
          "select id, name, amount from " + 
          "account where id=?"); 
        stmt.setLong(1, id); 
        rs = stmt.executeQuery(); 
        Account account = null; 
        if (rs.next()) { 
          account = new Account(); 
          account.setId(rs.getLong("id")); 
          account.setName(rs.getString("name")); 
          account.setAmount(rs.getString("amount")); 
        } 
        return account; 
      } catch (SQLException e) { 
      } finally { 
          if(rs != null) { 
            try { 
              rs.close(); 
            } catch(SQLException e) {} 
          } 
          if(stmt != null) { 
            try { 
              stmt.close(); 
            } catch(SQLException e) {} 
          } 
          if(conn != null) { 
            try { 
              conn.close(); 
            } catch(SQLException e) {} 
          } 
        } 
      return null; 
    } 

In the preceding code, we can see that the JDBC code queries the database for an account name and amount. For this simple task, we have to create a connection, then create a statement, and finally query for the results. We also have to catch SQLException, a checked exception, even though there's not a lot you can do if it's thrown. Lastly, we have to clean up the mess, closing down the connection statement and result set. This could also force it to handle JDBC's exception, so you must catch SQLException here as well. This kind of boilerplate code seriously hurts reusability.

Spring JDBC solves the problem of boilerplate code by using the Template Design pattern, and it makes life very easy by removing the common code in templates. This makes the data access code very clean and prevents nagging problems, such as connection leaks, because the Spring Framework ensures that all database resources are released properly.

    public Account getAccountById(long id) { 
      return jdbcTemplate.queryForObject( 
        "select id, name, amoount" + 
        "from account where id=?", 
         new RowMapper<Account>() { 
           public Account mapRow(ResultSet rs, 
            int rowNum) throws SQLException { 
              account = new Account(); 
              account.setId(rs.getLong("id")); 
              account.setName(rs.getString("name")); 
              account.setAmount(rs.getString("amount")); 
              return account; 
            } 
         }, 
      id); 
    }