Spring AOP 4 官网文档

11. Aspect Oriented Programming with Spring

11.1 Introduction

Aspect-Oriented Programming (AOP) complements Object-Oriented Programming (OOP) by providing another way of thinking about program structure. The key unit of modularity in OOP is the class, whereas in AOP the unit of modularity is the aspect. Aspects enable the modularization of concerns such as transaction management that cut across multiple types and objects. (Such concerns are often termed crosscutting concerns in AOP literature.)

One of the key components of Spring is the AOP framework. While the Spring IoC container does not depend on AOP, meaning you do not need to use AOP if you don’t want to, AOP complements Spring IoC to provide a very capable middleware solution.

Spring 2.0 AOP

Spring 2.0 introduces a simpler and more powerful way of writing custom aspects using either a schema-based approach or the @AspectJ annotation style. Both of these styles offer fully typed advice and use of the AspectJ pointcut language, while still using Spring AOP for weaving.

The Spring 2.0 schema- and @AspectJ-based AOP support is discussed in this chapter. Spring 2.0 AOP remains fully backwards compatible with Spring 1.2 AOP, and the lower-level AOP support offered by the Spring 1.2 APIs is discussed in the following chapter.

AOP is used in the Spring Framework to…​

• …​ provide declarative enterprise services, especially as a replacement for EJB declarative services. The most important such service is declarative transaction management.
• …​ allow users to implement custom aspects, complementing their use of OOP with AOP.

	If you are interested only in generic declarative services or other pre-packaged declarative middleware services such as pooling, you do not need to work directly with Spring AOP, and can skip most of this chapter.

11.1.1 AOP concepts

Let us begin by defining some central AOP concepts and terminology. These terms are not Spring-specific…​ unfortunately, AOP terminology is not particularly intuitive; however, it would be even more confusing if Spring used its own terminology.

• Aspect: a modularization of a concern that cuts across multiple classes. Transaction management is a good example of a crosscutting concern in enterprise Java applications. In Spring AOP, aspects are implemented using regular classes (the schema-based approach) or regular classes annotated with the @Aspect annotation (the @AspectJ style).
• Join point: a point during the execution of a program, such as the execution of a method or the handling of an exception. In Spring AOP, a join point always represents a method execution.
• Advice: action taken by an aspect at a particular join point. Different types of advice include "around," "before" and "after" advice. (Advice types are discussed below.) Many AOP frameworks, including Spring, model an advice as an interceptor, maintaining a chain of interceptors around the join point.
• Pointcut: a predicate that matches join points. Advice is associated with a pointcut expression and runs at any join point matched by the pointcut (for example, the execution of a method with a certain name). The concept of join points as matched by pointcut expressions is central to AOP, and Spring uses the AspectJ pointcut expression language by default.
• Introduction: declaring additional methods or fields on behalf of a type. Spring AOP allows you to introduce new interfaces (and a corresponding implementation) to any advised object. For example, you could use an introduction to make a bean implement an IsModified interface, to simplify caching. (An introduction is known as an inter-type declaration in the AspectJ community.)
• Target object: object being advised by one or more aspects. Also referred to as the advised object. Since Spring AOP is implemented using runtime proxies, this object will always be a proxied object.
• AOP proxy: an object created by the AOP framework in order to implement the aspect contracts (advise method executions and so on). In the Spring Framework, an AOP proxy will be a JDK dynamic proxy or a CGLIB proxy.
• Weaving: linking aspects with other application types or objects to create an advised object. This can be done at compile time (using the AspectJ compiler, for example), load time, or at runtime. Spring AOP, like other pure Java AOP frameworks, performs weaving at runtime.

Types of advice:

• Before advice: Advice that executes before a join point, but which does not have the ability to prevent execution flow proceeding to the join point (unless it throws an exception).
• After returning advice: Advice to be executed after a join point completes normally: for example, if a method returns without throwing an exception.
• After throwing advice: Advice to be executed if a method exits by throwing an exception.
• After (finally) advice: Advice to be executed regardless of the means by which a join point exits (normal or exceptional return).
• Around advice: Advice that surrounds a join point such as a method invocation. This is the most powerful kind of advice. Around advice can perform custom behavior before and after the method invocation. It is also responsible for choosing whether to proceed to the join point or to shortcut the advised method execution by returning its own return value or throwing an exception.

Around advice is the most general kind of advice. Since Spring AOP, like AspectJ, provides a full range of advice types, we recommend that you use the least powerful advice type that can implement the required behavior. For example, if you need only to update a cache with the return value of a method, you are better off implementing an after returning advice than an around advice, although an around advice can accomplish the same thing. Using the most specific advice type provides a simpler programming model with less potential for errors. For example, you do not need to invoke the proceed() method on the JoinPoint used for around advice, and hence cannot fail to invoke it.

In Spring 2.0, all advice parameters are statically typed, so that you work with advice parameters of the appropriate type (the type of the return value from a method execution for example) rather than Object arrays.

The concept of join points, matched by pointcuts, is the key to AOP which distinguishes it from older technologies offering only interception. Pointcuts enable advice to be targeted independently of the Object-Oriented hierarchy. For example, an around advice providing declarative transaction management can be applied to a set of methods spanning multiple objects (such as all business operations in the service layer).

11.1.2 Spring AOP capabilities and goals

Spring AOP is implemented in pure Java. There is no need for a special compilation process. Spring AOP does not need to control the class loader hierarchy, and is thus suitable for use in a Servlet container or application server.

Spring AOP currently supports only method execution join points (advising the execution of methods on Spring beans). Field interception is not implemented, although support for field interception could be added without breaking the core Spring AOP APIs. If you need to advise field access and update join points, consider a language such as AspectJ.

Spring AOP’s approach to AOP differs from that of most other AOP frameworks. The aim is not to provide the most complete AOP implementation (although Spring AOP is quite capable); it is rather to provide a close integration between AOP implementation and Spring IoC to help solve common problems in enterprise applications.

Thus, for example, the Spring Framework’s AOP functionality is normally used in conjunction with the Spring IoC container. Aspects are configured using normal bean definition syntax (although this allows powerful "autoproxying" capabilities): this is a crucial difference from other AOP implementations. There are some things you cannot do easily or efficiently with Spring AOP, such as advise very fine-grained objects (such as domain objects typically): AspectJ is the best choice in such cases. However, our experience is that Spring AOP provides an excellent solution to most problems in enterprise Java applications that are amenable to AOP.

Spring AOP will never strive to compete with AspectJ to provide a comprehensive AOP solution. We believe that both proxy-based frameworks like Spring AOP and full-blown frameworks such as AspectJ are valuable, and that they are complementary, rather than in competition. Spring seamlessly integrates Spring AOP and IoC with AspectJ, to enable all uses of AOP to be catered for within a consistent Spring-based application architecture. This integration does not affect the Spring AOP API or the AOP Alliance API: Spring AOP remains backward-compatible. See the following chapter for a discussion of the Spring AOP APIs.

One of the central tenets of the Spring Framework is that of non-invasiveness; this is the idea that you should not be forced to introduce framework-specific classes and interfaces into your business/domain model. However, in some places the Spring Framework does give you the option to introduce Spring Framework-specific dependencies into your codebase: the rationale in giving you such options is because in certain scenarios it might be just plain easier to read or code some specific piece of functionality in such a way. The Spring Framework (almost) always offers you the choice though: you have the freedom to make an informed decision as to which option best suits your particular use case or scenario. One such choice that is relevant to this chapter is that of which AOP framework (and which AOP style) to choose. You have the choice of AspectJ and/or Spring AOP, and you also have the choice of either the @AspectJ annotation-style approach or the Spring XML configuration-style approach. The fact that this chapter chooses to introduce the @AspectJ-style approach first should not be taken as an indication that the Spring team favors the @AspectJ annotation-style approach over the Spring XML configuration-style. See Section 11.4, “Choosing which AOP declaration style to use” for a more complete discussion of the whys and wherefores of each style.

11.1.3 AOP Proxies

Spring AOP defaults to using standard JDK dynamic proxies for AOP proxies. This enables any interface (or set of interfaces) to be proxied.

Spring AOP can also use CGLIB proxies. This is necessary to proxy classes rather than interfaces. CGLIB is used by default if a business object does not implement an interface. As it is good practice to program to interfaces rather than classes; business classes normally will implement one or more business interfaces. It is possible to force the use of CGLIB, in those (hopefully rare) cases where you need to advise a method that is not declared on an interface, or where you need to pass a proxied object to a method as a concrete type.

It is important to grasp the fact that Spring AOP is proxy-based. See Section 11.6.1, “Understanding AOP proxies” for a thorough examination of exactly what this implementation detail actually means.

11.2 @AspectJ support

@AspectJ refers to a style of declaring aspects as regular Java classes annotated with annotations. The @AspectJ style was introduced by the AspectJ project as part of the AspectJ 5 release. Spring interprets the same annotations as AspectJ 5, using a library supplied by AspectJ for pointcut parsing and matching. The AOP runtime is still pure Spring AOP though, and there is no dependency on the AspectJ compiler or weaver.

	Using the AspectJ compiler and weaver enables use of the full AspectJ language, and is discussed in Section 11.8, “Using AspectJ with Spring applications”.

11.2.1 Enabling @AspectJ Support

To use @AspectJ aspects in a Spring configuration you need to enable Spring support for configuring Spring AOP based on @AspectJ aspects, and autoproxying beans based on whether or not they are advised by those aspects. By autoproxying we mean that if Spring determines that a bean is advised by one or more aspects, it will automatically generate a proxy for that bean to intercept method invocations and ensure that advice is executed as needed.

The @AspectJ support can be enabled with XML or Java style configuration. In either case you will also need to ensure that AspectJ’s aspectjweaver.jar library is on the classpath of your application (version 1.6.8 or later). This library is available in the 'lib' directory of an AspectJ distribution or via the Maven Central repository.

Enabling @AspectJ Support with Java configuration

To enable @AspectJ support with Java @Configuration add the @EnableAspectJAutoProxy annotation:

@Configuration
@EnableAspectJAutoProxy
public class AppConfig {

}

Enabling @AspectJ Support with XML configuration

To enable @AspectJ support with XML based configuration use the aop:aspectj-autoproxy element:

This assumes that you are using schema support as described in Chapter 41, XML Schema-based configuration. See Section 41.2.7, “the aop schema” for how to import the tags in the aop namespace.

11.2.2 Declaring an aspect

With the @AspectJ support enabled, any bean defined in your application context with a class that is an @AspectJ aspect (has the @Aspect annotation) will be automatically detected by Spring and used to configure Spring AOP. The following example shows the minimal definition required for a not-very-useful aspect:

A regular bean definition in the application context, pointing to a bean class that has the @Aspect annotation:

 id="myAspect" class="org.xyz.NotVeryUsefulAspect">
    

And the NotVeryUsefulAspect class definition, annotated with org.aspectj.lang.annotation.Aspect annotation;

package org.xyz;
import org.aspectj.lang.annotation.Aspect;

@Aspect
public class NotVeryUsefulAspect {

}

Aspects (classes annotated with @Aspect) may have methods and fields just like any other class. They may also contain pointcut, advice, and introduction (inter-type) declarations.

You may register aspect classes as regular beans in your Spring XML configuration, or autodetect them through classpath scanning - just like any other Spring-managed bean. However, note that the @Aspect annotation is not sufficient for autodetection in the classpath: For that purpose, you need to add a separate @Component annotation (or alternatively a custom stereotype annotation that qualifies, as per the rules of Spring’s component scanner).

	In Spring AOP, it is not possible to have aspects themselves be the target of advice from other aspects. The @Aspect annotation on a class marks it as an aspect, and hence excludes it from auto-proxying.

11.2.3 Declaring a pointcut

Recall that pointcuts determine join points of interest, and thus enable us to control when advice executes. Spring AOP only supports method execution join points for Spring beans, so you can think of a pointcut as matching the execution of methods on Spring beans. A pointcut declaration has two parts: a signature comprising a name and any parameters, and a pointcut expression that determines exactly which method executions we are interested in. In the @AspectJ annotation-style of AOP, a pointcut signature is provided by a regular method definition, and the pointcut expression is indicated using the @Pointcut annotation (the method serving as the pointcut signature must have a void return type).

An example will help make this distinction between a pointcut signature and a pointcut expression clear. The following example defines a pointcut named 'anyOldTransfer' that will match the execution of any method named 'transfer':

@Pointcut("execution(* transfer(..))")// the pointcut expression
private void anyOldTransfer() {}// the pointcut signature

The pointcut expression that forms the value of the @Pointcut annotation is a regular AspectJ 5 pointcut expression. For a full discussion of AspectJ’s pointcut language, see the AspectJ Programming Guide (and for extensions, the AspectJ 5 Developers Notebook) or one of the books on AspectJ such as "Eclipse AspectJ" by Colyer et. al. or "AspectJ in Action" by Ramnivas Laddad.

Supported Pointcut Designators

Spring AOP supports the following AspectJ pointcut designators (PCD) for use in pointcut expressions:

Other pointcut types

The full AspectJ pointcut language supports additional pointcut designators that are not supported in Spring. These are: call, get, set, preinitialization, staticinitialization, initialization, handler, adviceexecution, withincode, cflow, cflowbelow, if, @this, and @withincode. Use of these pointcut designators in pointcut expressions interpreted by Spring AOP will result in an IllegalArgumentException being thrown.

The set of pointcut designators supported by Spring AOP may be extended in future releases to support more of the AspectJ pointcut designators.

• execution - for matching method execution join points, this is the primary pointcut designator you will use when working with Spring AOP
• within - limits matching to join points within certain types (simply the execution of a method declared within a matching type when using Spring AOP)
• this - limits matching to join points (the execution of methods when using Spring AOP) where the bean reference (Spring AOP proxy) is an instance of the given type
• target - limits matching to join points (the execution of methods when using Spring AOP) where the target object (application object being proxied) is an instance of the given type
• args - limits matching to join points (the execution of methods when using Spring AOP) where the arguments are instances of the given types
• @target - limits matching to join points (the execution of methods when using Spring AOP) where the class of the executing object has an annotation of the given type
• @args - limits matching to join points (the execution of methods when using Spring AOP) where the runtime type of the actual arguments passed have annotations of the given type(s)
• @within - limits matching to join points within types that have the given annotation (the execution of methods declared in types with the given annotation when using Spring AOP)
• @annotation - limits matching to join points where the subject of the join point (method being executed in Spring AOP) has the given annotation

Because Spring AOP limits matching to only method execution join points, the discussion of the pointcut designators above gives a narrower definition than you will find in the AspectJ programming guide. In addition, AspectJ itself has type-based semantics and at an execution join point both this and target refer to the same object - the object executing the method. Spring AOP is a proxy-based system and differentiates between the proxy object itself (bound to this) and the target object behind the proxy (bound to target).

Due to the proxy-based nature of Spring’s AOP framework, calls within the target object are by definition not intercepted. For JDK proxies, only public interface method calls on the proxy can be intercepted. With CGLIB, public and protected method calls on the proxy will be intercepted, and even package-visible methods if necessary. However, common interactions through proxies should always be designed through public signatures. Note that pointcut definitions are generally matched against any intercepted method. If a pointcut is strictly meant to be public-only, even in a CGLIB proxy scenario with potential non-public interactions through proxies, it needs to be defined accordingly. If your interception needs include method calls or even constructors within the target class, consider the use of Spring-driven native AspectJ weaving instead of Spring’s proxy-based AOP framework. This constitutes a different mode of AOP usage with different characteristics, so be sure to make yourself familiar with weaving first before making a decision.

Spring AOP also supports an additional PCD named bean. This PCD allows you to limit the matching of join points to a particular named Spring bean, or to a set of named Spring beans (when using wildcards). The bean PCD has the following form:

bean(idOrNameOfBean)

The idOrNameOfBean token can be the name of any Spring bean: limited wildcard support using the * character is provided, so if you establish some naming conventions for your Spring beans you can quite easily write a bean PCD expression to pick them out. As is the case with other pointcut designators, the bean PCD can be &&'ed, ||'ed, and ! (negated) too.

Please note that the bean PCD is only supported in Spring AOP - and not in native AspectJ weaving. It is a Spring-specific extension to the standard PCDs that AspectJ defines and therefore not available for aspects declared in the @Aspect model. The bean PCD operates at the instance level (building on the Spring bean name concept) rather than at the type level only (which is what weaving-based AOP is limited to). Instance-based pointcut designators are a special capability of Spring’s proxy-based AOP framework and its close integration with the Spring bean factory, where it is natural and straightforward to identify specific beans by name.

Combining pointcut expressions

Pointcut expressions can be combined using '&&', '||' and '!'. It is also possible to refer to pointcut expressions by name. The following example shows three pointcut expressions: anyPublicOperation (which matches if a method execution join point represents the execution of any public method); inTrading (which matches if a method execution is in the trading module), and tradingOperation (which matches if a method execution represents any public method in the trading module).

@Pointcut("execution(public * *(..))")
private void anyPublicOperation() {}

@Pointcut("within(com.xyz.someapp.trading..*)")
private void inTrading() {}

@Pointcut("anyPublicOperation() && inTrading()")
private void tradingOperation() {}

It is a best practice to build more complex pointcut expressions out of smaller named components as shown above. When referring to pointcuts by name, normal Java visibility rules apply (you can see private pointcuts in the same type, protected pointcuts in the hierarchy, public pointcuts anywhere and so on). Visibility does not affect pointcutmatching.

Sharing common pointcut definitions

When working with enterprise applications, you often want to refer to modules of the application and particular sets of operations from within several aspects. We recommend defining a "SystemArchitecture" aspect that captures common pointcut expressions for this purpose. A typical such aspect would look as follows:

package com.xyz.someapp;

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Pointcut;

@Aspect
public class SystemArchitecture {

    /**
     * A join point is in the web layer if the method is defined
     * in a type in the com.xyz.someapp.web package or any sub-package
     * under that.
     */
    @Pointcut("within(com.xyz.someapp.web..*)")
    public void inWebLayer() {}

    /**
     * A join point is in the service layer if the method is defined
     * in a type in the com.xyz.someapp.service package or any sub-package
     * under that.
     */
    @Pointcut("within(com.xyz.someapp.service..*)")
    public void inServiceLayer() {}

    /**
     * A join point is in the data access layer if the method is defined
     * in a type in the com.xyz.someapp.dao package or any sub-package
     * under that.
     */
    @Pointcut("within(com.xyz.someapp.dao..*)")
    public void inDataAccessLayer() {}

    /**
     * A business service is the execution of any method defined on a service
     * interface. This definition assumes that interfaces are placed in the
     * "service" package, and that implementation types are in sub-packages.
     *
     * If you group service interfaces by functional area (for example,
     * in packages com.xyz.someapp.abc.service and com.xyz.someapp.def.service) then
     * the pointcut expression "execution(* com.xyz.someapp..service.*.*(..))"
     * could be used instead.
     *
     * Alternatively, you can write the expression using the 'bean'
     * PCD, like so "bean(*Service)". (This assumes that you have
     * named your Spring service beans in a consistent fashion.)
     */
    @Pointcut("execution(* com.xyz.someapp..service.*.*(..))")
    public void businessService() {}

    /**
     * A data access operation is the execution of any method defined on a
     * dao interface. This definition assumes that interfaces are placed in the
     * "dao" package, and that implementation types are in sub-packages.
     */
    @Pointcut("execution(* com.xyz.someapp.dao.*.*(..))")
    public void dataAccessOperation() {}

}

The pointcuts defined in such an aspect can be referred to anywhere that you need a pointcut expression. For example, to make the service layer transactional, you could write:

    
        pointcut="com.xyz.someapp.SystemArchitecture.businessService()"
        advice-ref="tx-advice"/>


 id="tx-advice">
    
         name="*" propagation="REQUIRED"/>
    

The  and  elements are discussed in Section 11.3, “Schema-based AOP support”. The transaction elements are discussed in Chapter 17, Transaction Management.

Examples

Spring AOP users are likely to use the execution pointcut designator the most often. The format of an execution expression is:

execution(modifiers-pattern? ret-type-pattern declaring-type-pattern?name-pattern(param-pattern)
            throws-pattern?)

All parts except the returning type pattern (ret-type-pattern in the snippet above), name pattern, and parameters pattern are optional. The returning type pattern determines what the return type of the method must be in order for a join point to be matched. Most frequently you will use * as the returning type pattern, which matches any return type. A fully-qualified type name will match only when the method returns the given type. The name pattern matches the method name. You can use the * wildcard as all or part of a name pattern. If specifying a declaring type pattern then include a trailing . to join it to the name pattern component. The parameters pattern is slightly more complex: () matches a method that takes no parameters, whereas (..) matches any number of parameters (zero or more). The pattern (*) matches a method taking one parameter of any type,(*,String) matches a method taking two parameters, the first can be of any type, the second must be a String. Consult the Language Semantics section of the AspectJ Programming Guide for more information.

Some examples of common pointcut expressions are given below.

• the execution of any public method:

execution(public * *(..))

• the execution of any method with a name beginning with "set":

execution(* set*(..))

• the execution of any method defined by the AccountService interface:

execution(* com.xyz.service.AccountService.*(..))

• the execution of any method defined in the service package:

execution(* com.xyz.service.*.*(..))

• the execution of any method defined in the service package or a sub-package:

execution(* com.xyz.service..*.*(..))

• any join point (method execution only in Spring AOP) within the service package:

within(com.xyz.service.*)

• any join point (method execution only in Spring AOP) within the service package or a sub-package:

within(com.xyz.service..*)

• any join point (method execution only in Spring AOP) where the proxy implements the AccountService interface:

this(com.xyz.service.AccountService)

	'this' is more commonly used in a binding form :- see the following section on advice for how to make the proxy object available in the advice body.

• any join point (method execution only in Spring AOP) where the target object implements the AccountService interface:

target(com.xyz.service.AccountService)

	'target' is more commonly used in a binding form :- see the following section on advice for how to make the target object available in the advice body.

• any join point (method execution only in Spring AOP) which takes a single parameter, and where the argument passed at runtime is Serializable:

args(java.io.Serializable)

	'args' is more commonly used in a binding form :- see the following section on advice for how to make the method arguments available in the advice body.

Note that the pointcut given in this example is different to execution(* *(java.io.Serializable)): the args version matches if the argument passed at runtime is Serializable, the execution version matches if the method signature declares a single parameter of type Serializable.

• any join point (method execution only in Spring AOP) where the target object has an @Transactional annotation:

@target(org.springframework.transaction.annotation.Transactional)

	'@target' can also be used in a binding form :- see the following section on advice for how to make the annotation object available in the advice body.

• any join point (method execution only in Spring AOP) where the declared type of the target object has an @Transactional annotation:

@within(org.springframework.transaction.annotation.Transactional)

	'@within' can also be used in a binding form :- see the following section on advice for how to make the annotation object available in the advice body.

• any join point (method execution only in Spring AOP) where the executing method has an @Transactional annotation:

@annotation(org.springframework.transaction.annotation.Transactional)

	'@annotation' can also be used in a binding form :- see the following section on advice for how to make the annotation object available in the advice body.

• any join point (method execution only in Spring AOP) which takes a single parameter, and where the runtime type of the argument passed has the @Classifiedannotation:

@args(com.xyz.security.Classified)

	'@args' can also be used in a binding form :- see the following section on advice for how to make the annotation object(s) available in the advice body.

• any join point (method execution only in Spring AOP) on a Spring bean named tradeService:

bean(tradeService)

• any join point (method execution only in Spring AOP) on Spring beans having names that match the wildcard expression *Service:

bean(*Service)

Writing good pointcuts

During compilation, AspectJ processes pointcuts in order to try and optimize matching performance. Examining code and determining if each join point matches (statically or dynamically) a given pointcut is a costly process. (A dynamic match means the match cannot be fully determined from static analysis and a test will be placed in the code to determine if there is an actual match when the code is running). On first encountering a pointcut declaration, AspectJ will rewrite it into an optimal form for the matching process. What does this mean? Basically pointcuts are rewritten in DNF (Disjunctive Normal Form) and the components of the pointcut are sorted such that those components that are cheaper to evaluate are checked first. This means you do not have to worry about understanding the performance of various pointcut designators and may supply them in any order in a pointcut declaration.

However, AspectJ can only work with what it is told, and for optimal performance of matching you should think about what they are trying to achieve and narrow the search space for matches as much as possible in the definition. The existing designators naturally fall into one of three groups: kinded, scoping and context:

• Kinded designators are those which select a particular kind of join point. For example: execution, get, set, call, handler
• Scoping designators are those which select a group of join points of interest (of probably many kinds). For example: within, withincode
• Contextual designators are those that match (and optionally bind) based on context. For example: this, target, @annotation

A well written pointcut should try and include at least the first two types (kinded and scoping), whilst the contextual designators may be included if wishing to match based on join point context, or bind that context for use in the advice. Supplying either just a kinded designator or just a contextual designator will work but could affect weaving performance (time and memory used) due to all the extra processing and analysis. Scoping designators are very fast to match and their usage means AspectJ can very quickly dismiss groups of join points that should not be further processed - that is why a good pointcut should always include one if possible.

11.2.4 Declaring advice

Advice is associated with a pointcut expression, and runs before, after, or around method executions matched by the pointcut. The pointcut expression may be either a simple reference to a named pointcut, or a pointcut expression declared in place.

Before advice

Before advice is declared in an aspect using the @Before annotation:

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Before;

@Aspect
public class BeforeExample {

    @Before("com.xyz.myapp.SystemArchitecture.dataAccessOperation()")
    public void doAccessCheck() {
        // ...
    }

}

If using an in-place pointcut expression we could rewrite the above example as:

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Before;

@Aspect
public class BeforeExample {

    @Before("execution(* com.xyz.myapp.dao.*.*(..))")
    public void doAccessCheck() {
        // ...
    }

}

After returning advice

After returning advice runs when a matched method execution returns normally. It is declared using the @AfterReturning annotation:

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.AfterReturning;

@Aspect
public class AfterReturningExample {

    @AfterReturning("com.xyz.myapp.SystemArchitecture.dataAccessOperation()")
    public void doAccessCheck() {
        // ...
    }

}

	Note: it is of course possible to have multiple advice declarations, and other members as well, all inside the same aspect. We’re just showing a single advice declaration in these examples to focus on the issue under discussion at the time.

Sometimes you need access in the advice body to the actual value that was returned. You can use the form of @AfterReturning that binds the return value for this:

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.AfterReturning;

@Aspect
public class AfterReturningExample {

    @AfterReturning(
        pointcut="com.xyz.myapp.SystemArchitecture.dataAccessOperation()",
        returning="retVal")
    public void doAccessCheck(Object retVal) {
        // ...
    }

}

The name used in the returning attribute must correspond to the name of a parameter in the advice method. When a method execution returns, the return value will be passed to the advice method as the corresponding argument value. A returning clause also restricts matching to only those method executions that return a value of the specified type ( Object in this case, which will match any return value).

Please note that it is not possible to return a totally different reference when using after-returning advice.

After throwing advice

After throwing advice runs when a matched method execution exits by throwing an exception. It is declared using the @AfterThrowing annotation:

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.AfterThrowing;

@Aspect
public class AfterThrowingExample {

    @AfterThrowing("com.xyz.myapp.SystemArchitecture.dataAccessOperation()")
    public void doRecoveryActions() {
        // ...
    }

}

Often you want the advice to run only when exceptions of a given type are thrown, and you also often need access to the thrown exception in the advice body. Use thethrowing attribute to both restrict matching (if desired, use Throwable as the exception type otherwise) and bind the thrown exception to an advice parameter.

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.AfterThrowing;

@Aspect
public class AfterThrowingExample {

    @AfterThrowing(
        pointcut="com.xyz.myapp.SystemArchitecture.dataAccessOperation()",
        throwing="ex")
    public void doRecoveryActions(DataAccessException ex) {
        // ...
    }

}

The name used in the throwing attribute must correspond to the name of a parameter in the advice method. When a method execution exits by throwing an exception, the exception will be passed to the advice method as the corresponding argument value. A throwing clause also restricts matching to only those method executions that throw an exception of the specified type ( DataAccessException in this case).

After (finally) advice

After (finally) advice runs however a matched method execution exits. It is declared using the @After annotation. After advice must be prepared to handle both normal and exception return conditions. It is typically used for releasing resources, etc.

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.After;

@Aspect
public class AfterFinallyExample {

    @After("com.xyz.myapp.SystemArchitecture.dataAccessOperation()")
    public void doReleaseLock() {
        // ...
    }

}

Around advice

The final kind of advice is around advice. Around advice runs "around" a matched method execution. It has the opportunity to do work both before and after the method executes, and to determine when, how, and even if, the method actually gets to execute at all. Around advice is often used if you need to share state before and after a method execution in a thread-safe manner (starting and stopping a timer for example). Always use the least powerful form of advice that meets your requirements (i.e. don’t use around advice if simple before advice would do).

Around advice is declared using the @Around annotation. The first parameter of the advice method must be of type ProceedingJoinPoint. Within the body of the advice, calling proceed() on the ProceedingJoinPoint causes the underlying method to execute. The proceed method may also be called passing in an Object[] - the values in the array will be used as the arguments to the method execution when it proceeds.

The behavior of proceed when called with an Object[] is a little different than the behavior of proceed for around advice compiled by the AspectJ compiler. For around advice written using the traditional AspectJ language, the number of arguments passed to proceed must match the number of arguments passed to the around advice (not the number of arguments taken by the underlying join point), and the value passed to proceed in a given argument position supplants the original value at the join point for the entity the value was bound to (Don’t worry if this doesn’t make sense right now!). The approach taken by Spring is simpler and a better match to its proxy-based, execution only semantics. You only need to be aware of this difference if you are compiling @AspectJ aspects written for Spring and using proceed with arguments with the AspectJ compiler and weaver. There is a way to write such aspects that is 100% compatible across both Spring AOP and AspectJ, and this is discussed in the following section on advice parameters.

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Around;
import org.aspectj.lang.ProceedingJoinPoint;

@Aspect
public class AroundExample {

    @Around("com.xyz.myapp.SystemArchitecture.businessService()")
    public Object doBasicProfiling(ProceedingJoinPoint pjp) throws Throwable {
        // start stopwatch
        Object retVal = pjp.proceed();
        // stop stopwatch
        return retVal;
    }

}

The value returned by the around advice will be the return value seen by the caller of the method. A simple caching aspect for example could return a value from a cache if it has one, and invoke proceed() if it does not. Note that proceed may be invoked once, many times, or not at all within the body of the around advice, all of these are quite legal.

Advice parameters

Spring offers fully typed advice - meaning that you declare the parameters you need in the advice signature (as we saw for the returning and throwing examples above) rather than work with Object[] arrays all the time. We’ll see how to make argument and other contextual values available to the advice body in a moment. First let’s take a look at how to write generic advice that can find out about the method the advice is currently advising.

Access to the current JoinPoint

Any advice method may declare as its first parameter, a parameter of type org.aspectj.lang.JoinPoint (please note that around advice is required to declare a first parameter of type ProceedingJoinPoint, which is a subclass of JoinPoint. The JoinPoint interface provides a number of useful methods such as getArgs()(returns the method arguments), getThis() (returns the proxy object), getTarget() (returns the target object), getSignature() (returns a description of the method that is being advised) and toString() (prints a useful description of the method being advised). Please do consult the javadocs for full details.

Passing parameters to advice

We’ve already seen how to bind the returned value or exception value (using after returning and after throwing advice). To make argument values available to the advice body, you can use the binding form of args. If a parameter name is used in place of a type name in an args expression, then the value of the corresponding argument will be passed as the parameter value when the advice is invoked. An example should make this clearer. Suppose you want to advise the execution of dao operations that take an Account object as the first parameter, and you need access to the account in the advice body. You could write the following:

@Before("com.xyz.myapp.SystemArchitecture.dataAccessOperation() && args(account,..)")
public void validateAccount(Account account) {
    // ...
}

The args(account,..) part of the pointcut expression serves two purposes: firstly, it restricts matching to only those method executions where the method takes at least one parameter, and the argument passed to that parameter is an instance of Account; secondly, it makes the actual Account object available to the advice via the account parameter.

Another way of writing this is to declare a pointcut that "provides" the Account object value when it matches a join point, and then just refer to the named pointcut from the advice. This would look as follows:

@Pointcut("com.xyz.myapp.SystemArchitecture.dataAccessOperation() && args(account,..)")
private void accountDataAccessOperation(Account account) {}

@Before("accountDataAccessOperation(account)")
public void validateAccount(Account account) {
    // ...
}

The interested reader is once more referred to the AspectJ programming guide for more details.

The proxy object ( this), target object ( target), and annotations ( @within, @target, @annotation, @args) can all be bound in a similar fashion. The following example shows how you could match the execution of methods annotated with an @Auditable annotation, and extract the audit code.

First the definition of the @Auditable annotation:

@Retention(RetentionPolicy.RUNTIME)
@Target(ElementType.METHOD)
public @interface Auditable {
    AuditCode value();
}

And then the advice that matches the execution of @Auditable methods:

@Before("com.xyz.lib.Pointcuts.anyPublicMethod() && @annotation(auditable)")
public void audit(Auditable auditable) {
    AuditCode code = auditable.value();
    // ...
}

Advice parameters and generics

Spring AOP can handle generics used in class declarations and method parameters. Suppose you have a generic type like this:

public interface Sample {
    void sampleGenericMethod(T param);
    void sampleGenericCollectionMethod(Collection param);
}

You can restrict interception of method types to certain parameter types by simply typing the advice parameter to the parameter type you want to intercept the method for:

@Before("execution(* ..Sample+.sampleGenericMethod(*)) && args(param)")
public void beforeSampleMethod(MyType param) {
    // Advice implementation
}

That this works is pretty obvious as we already discussed above. However, it’s worth pointing out that this won’t work for generic collections. So you cannot define a pointcut like this:

@Before("execution(* ..Sample+.sampleGenericCollectionMethod(*)) && args(param)")
public void beforeSampleMethod(Collection param) {
    // Advice implementation
}

To make this work we would have to inspect every element of the collection, which is not reasonable as we also cannot decide how to treat null values in general. To achieve something similar to this you have to type the parameter to Collection and manually check the type of the elements.

Determining argument names

The parameter binding in advice invocations relies on matching names used in pointcut expressions to declared parameter names in (advice and pointcut) method signatures. Parameter names are not available through Java reflection, so Spring AOP uses the following strategies to determine parameter names:

• If the parameter names have been specified by the user explicitly, then the specified parameter names are used: both the advice and the pointcut annotations have an optional "argNames" attribute which can be used to specify the argument names of the annotated method - these argument names are available at runtime. For example:

@Before(value="com.xyz.lib.Pointcuts.anyPublicMethod() && target(bean) && @annotation(auditable)",
        argNames="bean,auditable")
public void audit(Object bean, Auditable auditable) {
    AuditCode code = auditable.value();
    // ... use code and bean
}

If the first parameter is of the JoinPoint, ProceedingJoinPoint, or JoinPoint.StaticPart type, you may leave out the name of the parameter from the value of the "argNames" attribute. For example, if you modify the preceding advice to receive the join point object, the "argNames" attribute need not include it:

@Before(value="com.xyz.lib.Pointcuts.anyPublicMethod() && target(bean) && @annotation(auditable)",
        argNames="bean,auditable")
public void audit(JoinPoint jp, Object bean, Auditable auditable) {
    AuditCode code = auditable.value();
    // ... use code, bean, and jp
}

The special treatment given to the first parameter of the JoinPoint, ProceedingJoinPoint, and JoinPoint.StaticPart types is particularly convenient for advice that do not collect any other join point context. In such situations, you may simply omit the "argNames" attribute. For example, the following advice need not declare the "argNames" attribute:

@Before("com.xyz.lib.Pointcuts.anyPublicMethod()")
public void audit(JoinPoint jp) {
    // ... use jp
}

• Using the 'argNames' attribute is a little clumsy, so if the 'argNames' attribute has not been specified, then Spring AOP will look at the debug information for the class and try to determine the parameter names from the local variable table. This information will be present as long as the classes have been compiled with debug information ( '-g:vars' at a minimum). The consequences of compiling with this flag on are: (1) your code will be slightly easier to understand (reverse engineer), (2) the class file sizes will be very slightly bigger (typically inconsequential), (3) the optimization to remove unused local variables will not be applied by your compiler. In other words, you should encounter no difficulties building with this flag on.

	If an @AspectJ aspect has been compiled by the AspectJ compiler (ajc) even without the debug information then there is no need to add the argNames attribute as the compiler will retain the needed information.

• If the code has been compiled without the necessary debug information, then Spring AOP will attempt to deduce the pairing of binding variables to parameters (for example, if only one variable is bound in the pointcut expression, and the advice method only takes one parameter, the pairing is obvious!). If the binding of variables is ambiguous given the available information, then an AmbiguousBindingException will be thrown.
• If all of the above strategies fail then an IllegalArgumentException will be thrown.

Proceeding with arguments

We remarked earlier that we would describe how to write a proceed call with arguments that works consistently across Spring AOP and AspectJ. The solution is simply to ensure that the advice signature binds each of the method parameters in order. For example:

@Around("execution(List find*(..)) && " +
        "com.xyz.myapp.SystemArchitecture.inDataAccessLayer() && " +
        "args(accountHolderNamePattern)")
public Object preProcessQueryPattern(ProceedingJoinPoint pjp,
        String accountHolderNamePattern) throws Throwable {
    String newPattern = preProcess(accountHolderNamePattern);
    return pjp.proceed(new Object[] {newPattern});
}

In many cases you will be doing this binding anyway (as in the example above).

Advice ordering

What happens when multiple pieces of advice all want to run at the same join point? Spring AOP follows the same precedence rules as AspectJ to determine the order of advice execution. The highest precedence advice runs first "on the way in" (so given two pieces of before advice, the one with highest precedence runs first). "On the way out" from a join point, the highest precedence advice runs last (so given two pieces of after advice, the one with the highest precedence will run second).

When two pieces of advice defined in different aspects both need to run at the same join point, unless you specify otherwise the order of execution is undefined. You can control the order of execution by specifying precedence. This is done in the normal Spring way by either implementing the org.springframework.core.Ordered interface in the aspect class or annotating it with the Order annotation. Given two aspects, the aspect returning the lower value from Ordered.getValue() (or the annotation value) has the higher precedence.

When two pieces of advice defined in the same aspect both need to run at the same join point, the ordering is undefined (since there is no way to retrieve the declaration order via reflection for javac-compiled classes). Consider collapsing such advice methods into one advice method per join point in each aspect class, or refactor the pieces of advice into separate aspect classes - which can be ordered at the aspect level.

11.2.5 Introductions

Introductions (known as inter-type declarations in AspectJ) enable an aspect to declare that advised objects implement a given interface, and to provide an implementation of that interface on behalf of those objects.

An introduction is made using the @DeclareParents annotation. This annotation is used to declare that matching types have a new parent (hence the name). For example, given an interface UsageTracked, and an implementation of that interface DefaultUsageTracked, the following aspect declares that all implementors of service interfaces also implement the UsageTracked interface. (In order to expose statistics via JMX for example.)

@Aspect
public class UsageTracking {

    @DeclareParents(value="com.xzy.myapp.service.*+", defaultImpl=DefaultUsageTracked.class)
    public static UsageTracked mixin;

    @Before("com.xyz.myapp.SystemArchitecture.businessService() && this(usageTracked)")
    public void recordUsage(UsageTracked usageTracked) {
        usageTracked.incrementUseCount();
    }

}

The interface to be implemented is determined by the type of the annotated field. The value attribute of the @DeclareParents annotation is an AspectJ type pattern :- any bean of a matching type will implement the UsageTracked interface. Note that in the before advice of the above example, service beans can be directly used as implementations of the UsageTracked interface. If accessing a bean programmatically you would write the following:

UsageTracked usageTracked = (UsageTracked) context.getBean("myService");

11.2.6 Aspect instantiation models

	(This is an advanced topic, so if you are just starting out with AOP you can safely skip it until later.)

By default there will be a single instance of each aspect within the application context. AspectJ calls this the singleton instantiation model. It is possible to define aspects with alternate lifecycles :- Spring supports AspectJ’s perthis and pertarget instantiation models ( percflow, percflowbelow, and pertypewithin are not currently supported).

A "perthis" aspect is declared by specifying a perthis clause in the @Aspect annotation. Let’s look at an example, and then we’ll explain how it works.

@Aspect("perthis(com.xyz.myapp.SystemArchitecture.businessService())")
public class MyAspect {

    private int someState;

    @Before(com.xyz.myapp.SystemArchitecture.businessService())
    public void recordServiceUsage() {
        // ...
    }

}

The effect of the 'perthis' clause is that one aspect instance will be created for each unique service object executing a business service (each unique object bound to 'this' at join points matched by the pointcut expression). The aspect instance is created the first time that a method is invoked on the service object. The aspect goes out of scope when the service object goes out of scope. Before the aspect instance is created, none of the advice within it executes. As soon as the aspect instance has been created, the advice declared within it will execute at matched join points, but only when the service object is the one this aspect is associated with. See the AspectJ programming guide for more information on per-clauses.

The 'pertarget' instantiation model works in exactly the same way as perthis, but creates one aspect instance for each unique target object at matched join points.

11.2.7 Example

Now that you have seen how all the constituent parts work, let’s put them together to do something useful!

The execution of business services can sometimes fail due to concurrency issues (for example, deadlock loser). If the operation is retried, it is quite likely to succeed next time round. For business services where it is appropriate to retry in such conditions (idempotent operations that don’t need to go back to the user for conflict resolution), we’d like to transparently retry the operation to avoid the client seeing a PessimisticLockingFailureException. This is a requirement that clearly cuts across multiple services in the service layer, and hence is ideal for implementing via an aspect.

Because we want to retry the operation, we will need to use around advice so that we can call proceed multiple times. Here’s how the basic aspect implementation looks:

@Aspect
public class ConcurrentOperationExecutor implements Ordered {

    private static final int DEFAULT_MAX_RETRIES = 2;

    private int maxRetries = DEFAULT_MAX_RETRIES;
    private int order = 1;

    public void setMaxRetries(int maxRetries) {
        this.maxRetries = maxRetries;
    }

    public int getOrder() {
        return this.order;
    }

    public void setOrder(int order) {
        this.order = order;
    }

    @Around("com.xyz.myapp.SystemArchitecture.businessService()")
    public Object doConcurrentOperation(ProceedingJoinPoint pjp) throws Throwable {
        int numAttempts = 0;
        PessimisticLockingFailureException lockFailureException;
        do {
            numAttempts++;
            try {
                return pjp.proceed();
            }
            catch(PessimisticLockingFailureException ex) {
                lockFailureException = ex;
            }
        } while(numAttempts <= this.maxRetries);
        throw lockFailureException;
    }

}

Note that the aspect implements the Ordered interface so we can set the precedence of the aspect higher than the transaction advice (we want a fresh transaction each time we retry). The maxRetries and order properties will both be configured by Spring. The main action happens in the doConcurrentOperation around advice. Notice that for the moment we’re applying the retry logic to all businessService()s. We try to proceed, and if we fail with an PessimisticLockingFailureException we simply try again unless we have exhausted all of our retry attempts.

The corresponding Spring configuration is:

 id="concurrentOperationExecutor" class="com.xyz.myapp.service.impl.ConcurrentOperationExecutor">
     name="maxRetries" value="3"/>
     name="order" value="100"/>

To refine the aspect so that it only retries idempotent operations, we might define an Idempotent annotation:

@Retention(RetentionPolicy.RUNTIME)
public @interface Idempotent {
    // marker annotation
}

and use the annotation to annotate the implementation of service operations. The change to the aspect to only retry idempotent operations simply involves refining the pointcut expression so that only @Idempotent operations match:

@Around("com.xyz.myapp.SystemArchitecture.businessService() && " +
        "@annotation(com.xyz.myapp.service.Idempotent)")
public Object doConcurrentOperation(ProceedingJoinPoint pjp) throws Throwable {
    ...
}

11.3 Schema-based AOP support

If you prefer an XML-based format, then Spring also offers support for defining aspects using the new "aop" namespace tags. The exact same pointcut expressions and advice kinds are supported as when using the @AspectJ style, hence in this section we will focus on the new syntax and refer the reader to the discussion in the previous section (Section 11.2, “@AspectJ support”) for an understanding of writing pointcut expressions and the binding of advice parameters.

To use the aop namespace tags described in this section, you need to import the spring-aop schema as described in Chapter 41, XML Schema-based configuration. See Section 41.2.7, “the aop schema” for how to import the tags in the aop namespace.

Within your Spring configurations, all aspect and advisor elements must be placed within an  element (you can have more than one  element in an application context configuration). An  element can contain pointcut, advisor, and aspect elements (note these must be declared in that order).

	The  style of configuration makes heavy use of Spring’s auto-proxying mechanism. This can cause issues (such as advice not being woven) if you are already using explicit auto-proxying via the use of BeanNameAutoProxyCreator or suchlike. The recommended usage pattern is to use either just the  style, or just the AutoProxyCreator style.

11.3.1 Declaring an aspect

Using the schema support, an aspect is simply a regular Java object defined as a bean in your Spring application context. The state and behavior is captured in the fields and methods of the object, and the pointcut and advice information is captured in the XML.

An aspect is declared using the element, and the backing bean is referenced using the ref attribute:

     id="myAspect" ref="aBean">
        ...
    


 id="aBean" class="...">
    ...

The bean backing the aspect (" `aBean`" in this case) can of course be configured and dependency injected just like any other Spring bean.

11.3.2 Declaring a pointcut

A named pointcut can be declared inside an element, enabling the pointcut definition to be shared across several aspects and advisors.

A pointcut representing the execution of any business service in the service layer could be defined as follows:

     id="businessService"
        expression="execution(* com.xyz.myapp.service.*.*(..))"/>

Note that the pointcut expression itself is using the same AspectJ pointcut expression language as described in Section 11.2, “@AspectJ support”. If you are using the schema based declaration style, you can refer to named pointcuts defined in types (@Aspects) within the pointcut expression. Another way of defining the above pointcut would be:

     id="businessService"
        expression="com.xyz.myapp.SystemArchitecture.businessService()"/>

Assuming you have a SystemArchitecture aspect as described in the section called “Sharing common pointcut definitions”.

Declaring a pointcut inside an aspect is very similar to declaring a top-level pointcut:

     id="myAspect" ref="aBean">

         id="businessService"
            expression="execution(* com.xyz.myapp.service.*.*(..))"/>

        ...

    

Much the same way in an @AspectJ aspect, pointcuts declared using the schema based definition style may collect join point context. For example, the following pointcut collects the 'this' object as the join point context and passes it to advice:

     id="myAspect" ref="aBean">

         id="businessService"
            expression="execution(* com.xyz.myapp.service.*.*(..)) && this(service)"/>

         pointcut-ref="businessService" method="monitor"/>

        ...

    

The advice must be declared to receive the collected join point context by including parameters of the matching names:

public void monitor(Object service) {
    ...
}

When combining pointcut sub-expressions, '&&' is awkward within an XML document, and so the keywords 'and', 'or' and 'not' can be used in place of '&&', '||' and '!' respectively. For example, the previous pointcut may be better written as:

     id="myAspect" ref="aBean">

         id="businessService"
            expression="execution(* com.xyz.myapp.service.*.*(..)) **and** this(service)"/>

         pointcut-ref="businessService" method="monitor"/>

        ...
    

Note that pointcuts defined in this way are referred to by their XML id and cannot be used as named pointcuts to form composite pointcuts. The named pointcut support in the schema based definition style is thus more limited than that offered by the @AspectJ style.

11.3.3 Declaring advice

The same five advice kinds are supported as for the @AspectJ style, and they have exactly the same semantics.

Before advice

Before advice runs before a matched method execution. It is declared inside an  using the element.

 id="beforeExample" ref="aBean">

    
        pointcut-ref="dataAccessOperation"
        method="doAccessCheck"/>

    ...

Here dataAccessOperation is the id of a pointcut defined at the top ( ) level. To define the pointcut inline instead, replace the pointcut-ref attribute with a pointcut attribute:

 id="beforeExample" ref="aBean">

    
        pointcut="execution(* com.xyz.myapp.dao.*.*(..))"
        method="doAccessCheck"/>

    ...

As we noted in the discussion of the @AspectJ style, using named pointcuts can significantly improve the readability of your code.

The method attribute identifies a method ( doAccessCheck) that provides the body of the advice. This method must be defined for the bean referenced by the aspect element containing the advice. Before a data access operation is executed (a method execution join point matched by the pointcut expression), the "doAccessCheck" method on the aspect bean will be invoked.

After returning advice

After returning advice runs when a matched method execution completes normally. It is declared inside an  in the same way as before advice. For example:

 id="afterReturningExample" ref="aBean">

    
        pointcut-ref="dataAccessOperation"
        method="doAccessCheck"/>

    ...

Just as in the @AspectJ style, it is possible to get hold of the return value within the advice body. Use the returning attribute to specify the name of the parameter to which the return value should be passed:

 id="afterReturningExample" ref="aBean">

    
        pointcut-ref="dataAccessOperation"
        returning="retVal"
        method="doAccessCheck"/>

    ...

The doAccessCheck method must declare a parameter named retVal. The type of this parameter constrains matching in the same way as described for @AfterReturning. For example, the method signature may be declared as:

public void doAccessCheck(Object retVal) {...

After throwing advice

After throwing advice executes when a matched method execution exits by throwing an exception. It is declared inside an  using the after-throwing element:

 id="afterThrowingExample" ref="aBean">

    
        pointcut-ref="dataAccessOperation"
        method="doRecoveryActions"/>

    ...

Just as in the @AspectJ style, it is possible to get hold of the thrown exception within the advice body. Use the throwing attribute to specify the name of the parameter to which the exception should be passed:

 id="afterThrowingExample" ref="aBean">

    
        pointcut-ref="dataAccessOperation"
        throwing="dataAccessEx"
        method="doRecoveryActions"/>

    ...

The doRecoveryActions method must declare a parameter named dataAccessEx. The type of this parameter constrains matching in the same way as described for @AfterThrowing. For example, the method signature may be declared as:

public void doRecoveryActions(DataAccessException dataAccessEx) {...

After (finally) advice

After (finally) advice runs however a matched method execution exits. It is declared using the after element:

 id="afterFinallyExample" ref="aBean">

    
        pointcut-ref="dataAccessOperation"
        method="doReleaseLock"/>

    ...

Around advice

The final kind of advice is around advice. Around advice runs "around" a matched method execution. It has the opportunity to do work both before and after the method executes, and to determine when, how, and even if, the method actually gets to execute at all. Around advice is often used if you need to share state before and after a method execution in a thread-safe manner (starting and stopping a timer for example). Always use the least powerful form of advice that meets your requirements; don’t use around advice if simple before advice would do.

Around advice is declared using the aop:around element. The first parameter of the advice method must be of type ProceedingJoinPoint. Within the body of the advice, calling proceed() on the ProceedingJoinPoint causes the underlying method to execute. The proceed method may also be calling passing in an Object[] - the values in the array will be used as the arguments to the method execution when it proceeds. See the section called “Around advice” for notes on calling proceed with an Object[].

 id="aroundExample" ref="aBean">

    
        pointcut-ref="businessService"
        method="doBasicProfiling"/>

    ...

The implementation of the doBasicProfiling advice would be exactly the same as in the @AspectJ example (minus the annotation of course):

public Object doBasicProfiling(ProceedingJoinPoint pjp) throws Throwable {
    // start stopwatch
    Object retVal = pjp.proceed();
    // stop stopwatch
    return retVal;
}

Advice parameters

The schema based declaration style supports fully typed advice in the same way as described for the @AspectJ support - by matching pointcut parameters by name against advice method parameters. See the section called “Advice parameters” for details. If you wish to explicitly specify argument names for the advice methods (not relying on the detection strategies previously described) then this is done using the arg-names attribute of the advice element, which is treated in the same manner to the "argNames" attribute in an advice annotation as described in the section called “Determining argument names”. For example:

    pointcut="com.xyz.lib.Pointcuts.anyPublicMethod() and @annotation(auditable)"
    method="audit"
    arg-names="auditable"/>

The arg-names attribute accepts a comma-delimited list of parameter names.

Find below a slightly more involved example of the XSD-based approach that illustrates some around advice used in conjunction with a number of strongly typed parameters.

package x.y.service;

public interface FooService {

    Foo getFoo(String fooName, int age);
}

public class DefaultFooService implements FooService {

    public Foo getFoo(String name, int age) {
        return new Foo(name, age);
    }
}

Next up is the aspect. Notice the fact that the profile(..) method accepts a number of strongly-typed parameters, the first of which happens to be the join point used to proceed with the method call: the presence of this parameter is an indication that the profile(..) is to be used as around advice:

package x.y;

import org.aspectj.lang.ProceedingJoinPoint;
import org.springframework.util.StopWatch;

public class SimpleProfiler {

    public Object profile(ProceedingJoinPoint call, String name, int age) throws Throwable {
        StopWatch clock = new StopWatch("Profiling for '" + name + "' and '" + age + "'");
        try {
            clock.start(call.toShortString());
            return call.proceed();
        } finally {
            clock.stop();
            System.out.println(clock.prettyPrint());
        }
    }
}

Finally, here is the XML configuration that is required to effect the execution of the above advice for a particular join point:

 xmlns="http://www.springframework.org/schema/beans"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:aop="http://www.springframework.org/schema/aop"
    xsi:schemaLocation="
        http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans.xsd
        http://www.springframework.org/schema/aop http://www.springframework.org/schema/aop/spring-aop.xsd">

    
     id="fooService" class="x.y.service.DefaultFooService"/>

    
     id="profiler" class="x.y.SimpleProfiler"/>

    
         ref="profiler">

             id="theExecutionOfSomeFooServiceMethod"
                expression="execution(* x.y.service.FooService.getFoo(String,int))
                and args(name, age)"/>

             pointcut-ref="theExecutionOfSomeFooServiceMethod"
                method="profile"/>

        
    

If we had the following driver script, we would get output something like this on standard output:

import org.springframework.beans.factory.BeanFactory;
import org.springframework.context.support.ClassPathXmlApplicationContext;
import x.y.service.FooService;

public final class Boot {

    public static void main(final String[] args) throws Exception {
        BeanFactory ctx = new ClassPathXmlApplicationContext("x/y/plain.xml");
        FooService foo = (FooService) ctx.getBean("fooService");
        foo.getFoo("Pengo", 12);
    }
}

StopWatch 'Profiling for 'Pengo' and '12'': running time (millis) = 0
-----------------------------------------
ms     %     Task name
-----------------------------------------
00000  ?  execution(getFoo)

Advice ordering

When multiple advice needs to execute at the same join point (executing method) the ordering rules are as described in the section called “Advice ordering”. The precedence between aspects is determined by either adding the Order annotation to the bean backing the aspect or by having the bean implement the Ordered interface.

11.3.4 Introductions

Introductions (known as inter-type declarations in AspectJ) enable an aspect to declare that advised objects implement a given interface, and to provide an implementation of that interface on behalf of those objects.

An introduction is made using the aop:declare-parents element inside an aop:aspect This element is used to declare that matching types have a new parent (hence the name). For example, given an interface UsageTracked, and an implementation of that interface DefaultUsageTracked, the following aspect declares that all implementors of service interfaces also implement the UsageTracked interface. (In order to expose statistics via JMX for example.)

 id="usageTrackerAspect" ref="usageTracking">

    
        types-matching="com.xzy.myapp.service.*+"
        implement-interface="com.xyz.myapp.service.tracking.UsageTracked"
        default-impl="com.xyz.myapp.service.tracking.DefaultUsageTracked"/>

    
        pointcut="com.xyz.myapp.SystemArchitecture.businessService()
            and this(usageTracked)"
            method="recordUsage"/>

The class backing the usageTracking bean would contain the method:

public void recordUsage(UsageTracked usageTracked) {
    usageTracked.incrementUseCount();
}

The interface to be implemented is determined by implement-interface attribute. The value of the types-matching attribute is an AspectJ type pattern :- any bean of a matching type will implement the UsageTracked interface. Note that in the before advice of the above example, service beans can be directly used as implementations of the UsageTracked interface. If accessing a bean programmatically you would write the following:

UsageTracked usageTracked = (UsageTracked) context.getBean("myService");

11.3.5 Aspect instantiation models

The only supported instantiation model for schema-defined aspects is the singleton model. Other instantiation models may be supported in future releases.

11.3.6 Advisors

The concept of "advisors" is brought forward from the AOP support defined in Spring 1.2 and does not have a direct equivalent in AspectJ. An advisor is like a small self-contained aspect that has a single piece of advice. The advice itself is represented by a bean, and must implement one of the advice interfaces described in Section 12.3.2, “Advice types in Spring”. Advisors can take advantage of AspectJ pointcut expressions though.

Spring supports the advisor concept with the  element. You will most commonly see it used in conjunction with transactional advice, which also has its own namespace support in Spring. Here’s how it looks:

     id="businessService"
        expression="execution(* com.xyz.myapp.service.*.*(..))"/>

    
        pointcut-ref="businessService"
        advice-ref="tx-advice"/>



 id="tx-advice">
    
         name="*" propagation="REQUIRED"/>
    

As well as the pointcut-ref attribute used in the above example, you can also use the pointcut attribute to define a pointcut expression inline.

To define the precedence of an advisor so that the advice can participate in ordering, use the order attribute to define the Ordered value of the advisor.

11.3.7 Example

Let’s see how the concurrent locking failure retry example from Section 11.2.7, “Example” looks when rewritten using the schema support.

The execution of business services can sometimes fail due to concurrency issues (for example, deadlock loser). If the operation is retried, it is quite likely it will succeed next time round. For business services where it is appropriate to retry in such conditions (idempotent operations that don’t need to go back to the user for conflict resolution), we’d like to transparently retry the operation to avoid the client seeing a PessimisticLockingFailureException. This is a requirement that clearly cuts across multiple services in the service layer, and hence is ideal for implementing via an aspect.

Because we want to retry the operation, we’ll need to use around advice so that we can call proceed multiple times. Here’s how the basic aspect implementation looks (it’s just a regular Java class using the schema support):

public class ConcurrentOperationExecutor implements Ordered {

    private static final int DEFAULT_MAX_RETRIES = 2;

    private int maxRetries = DEFAULT_MAX_RETRIES;
    private int order = 1;

    public void setMaxRetries(int maxRetries) {
        this.maxRetries = maxRetries;
    }

    public int getOrder() {
        return this.order;
    }

    public void setOrder(int order) {
        this.order = order;
    }

    public Object doConcurrentOperation(ProceedingJoinPoint pjp) throws Throwable {
        int numAttempts = 0;
        PessimisticLockingFailureException lockFailureException;
        do {
            numAttempts++;
            try {
                return pjp.proceed();
            }
            catch(PessimisticLockingFailureException ex) {
                lockFailureException = ex;
            }
        } while(numAttempts <= this.maxRetries);
        throw lockFailureException;
    }

}

Note that the aspect implements the Ordered interface so we can set the precedence of the aspect higher than the transaction advice (we want a fresh transaction each time we retry). The maxRetries and order properties will both be configured by Spring. The main action happens in the doConcurrentOperation around advice method. We try to proceed, and if we fail with a PessimisticLockingFailureException we simply try again unless we have exhausted all of our retry attempts.

	This class is identical to the one used in the @AspectJ example, but with the annotations removed.

The corresponding Spring configuration is:

     id="concurrentOperationRetry" ref="concurrentOperationExecutor">

         id="idempotentOperation"
            expression="execution(* com.xyz.myapp.service.*.*(..))"/>

        
            pointcut-ref="idempotentOperation"
            method="doConcurrentOperation"/>

    



 id="concurrentOperationExecutor"
    class="com.xyz.myapp.service.impl.ConcurrentOperationExecutor">
         name="maxRetries" value="3"/>
         name="order" value="100"/>

Notice that for the time being we assume that all business services are idempotent. If this is not the case we can refine the aspect so that it only retries genuinely idempotent operations, by introducing an Idempotent annotation:

@Retention(RetentionPolicy.RUNTIME)
public @interface Idempotent {
    // marker annotation
}

and using the annotation to annotate the implementation of service operations. The change to the aspect to retry only idempotent operations simply involves refining the pointcut expression so that only @Idempotent operations match:

 id="idempotentOperation"
        expression="execution(* com.xyz.myapp.service.*.*(..)) and
        @annotation(com.xyz.myapp.service.Idempotent)"/>

11.4 Choosing which AOP declaration style to use

Once you have decided that an aspect is the best approach for implementing a given requirement, how do you decide between using Spring AOP or AspectJ, and between the Aspect language (code) style, @AspectJ annotation style, or the Spring XML style? These decisions are influenced by a number of factors including application requirements, development tools, and team familiarity with AOP.

11.4.1 Spring AOP or full AspectJ?

Use the simplest thing that can work. Spring AOP is simpler than using full AspectJ as there is no requirement to introduce the AspectJ compiler / weaver into your development and build processes. If you only need to advise the execution of operations on Spring beans, then Spring AOP is the right choice. If you need to advise objects not managed by the Spring container (such as domain objects typically), then you will need to use AspectJ. You will also need to use AspectJ if you wish to advise join points other than simple method executions (for example, field get or set join points, and so on).

When using AspectJ, you have the choice of the AspectJ language syntax (also known as the "code style") or the @AspectJ annotation style. Clearly, if you are not using Java 5+ then the choice has been made for you…​ use the code style. If aspects play a large role in your design, and you are able to use the AspectJ Development Tools (AJDT)plugin for Eclipse, then the AspectJ language syntax is the preferred option: it is cleaner and simpler because the language was purposefully  designed for writing aspects. If you are not using Eclipse, or have only a few aspects that do not play a major role in your application, then you may want to consider using the @AspectJ style and sticking with a regular Java compilation in your IDE, and adding an aspect weaving phase to your build script.

11.4.2 @AspectJ or XML for Spring AOP?

If you have chosen to use Spring AOP, then you have a choice of @AspectJ or XML style. There are various tradeoffs to consider.

The XML style will be most familiar to existing Spring users and it is backed by genuine POJOs. When using AOP as a tool to configure enterprise services then XML can be a good choice (a good test is whether you consider the pointcut expression to be a part of your configuration you might want to change independently). With the XML style arguably it is clearer from your configuration what aspects are present in the system.

The XML style has two disadvantages. Firstly it does not fully encapsulate the implementation of the requirement it addresses in a single place. The DRY principle says that there should be a single, unambiguous, authoritative representation of any piece of knowledge within a system. When using the XML style, the knowledge of how a requirement is implemented is split across the declaration of the backing bean class, and the XML in the configuration file. When using the @AspectJ style there is a single module - the aspect - in which this information is encapsulated. Secondly, the XML style is slightly more limited in what it can express than the @AspectJ style: only the "singleton" aspect instantiation model is supported, and it is not possible to combine named pointcuts declared in XML. For example, in the @AspectJ style you can write something like:

@Pointcut(execution(* get*()))
public void propertyAccess() {}

@Pointcut(execution(org.xyz.Account+ *(..))
public void operationReturningAnAccount() {}

@Pointcut(propertyAccess() && operationReturningAnAccount())
public void accountPropertyAccess() {}

In the XML style I can declare the first two pointcuts:

 id="propertyAccess"
        expression="execution(* get*())"/>

 id="operationReturningAnAccount"
        expression="execution(org.xyz.Account+ *(..))"/>

The downside of the XML approach is that you cannot define the accountPropertyAccess pointcut by combining these definitions.

The @AspectJ style supports additional instantiation models, and richer pointcut composition. It has the advantage of keeping the aspect as a modular unit. It also has the advantage the @AspectJ aspects can be understood (and thus consumed) both by Spring AOP and by AspectJ - so if you later decide you need the capabilities of AspectJ to implement additional requirements then it is very easy to migrate to an AspectJ-based approach. On balance the Spring team prefer the @AspectJ style whenever you have aspects that do more than simple "configuration" of enterprise services.

11.5 Mixing aspect types

It is perfectly possible to mix @AspectJ style aspects using the autoproxying support, schema-defined  aspects,  declared advisors and even proxies and interceptors defined using the Spring 1.2 style in the same configuration. All of these are implemented using the same underlying support mechanism and will co-exist without any difficulty.

11.6 Proxying mechanisms

Spring AOP uses either JDK dynamic proxies or CGLIB to create the proxy for a given target object. (JDK dynamic proxies are preferred whenever you have a choice).

If the target object to be proxied implements at least one interface then a JDK dynamic proxy will be used. All of the interfaces implemented by the target type will be proxied. If the target object does not implement any interfaces then a CGLIB proxy will be created.

If you want to force the use of CGLIB proxying (for example, to proxy every method defined for the target object, not just those implemented by its interfaces) you can do so. However, there are some issues to consider:

• final methods cannot be advised, as they cannot be overridden.
• As of Spring 3.2, it is no longer necessary to add CGLIB to your project classpath, as CGLIB classes are repackaged under org.springframework and included directly in the spring-core JAR. This means that CGLIB-based proxy support 'just works' in the same way that JDK dynamic proxies always have.
• As of Spring 4.0, the constructor of your proxied object will NOT be called twice anymore since the CGLIB proxy instance will be created via Objenesis. Only if your JVM does not allow for constructor bypassing, you might see double invocations and corresponding debug log entries from Spring’s AOP support.

To force the use of CGLIB proxies set the value of the proxy-target-class attribute of the  element to true:

 proxy-target-class="true">
    

To force CGLIB proxying when using the @AspectJ autoproxy support, set the 'proxy-target-class' attribute of the  element to true:

 proxy-target-class="true"/>

	Multiple  sections are collapsed into a single unified auto-proxy creator at runtime, which applies the strongest proxy settings that any of the sections (typically from different XML bean definition files) specified. This also applies to the  and  elements. To be clear: using proxy-target-class="true" on ,  or  elements will force the use of CGLIB proxies for all three of them.

11.6.1 Understanding AOP proxies

Spring AOP is proxy-based. It is vitally important that you grasp the semantics of what that last statement actually means before you write your own aspects or use any of the Spring AOP-based aspects supplied with the Spring Framework.

Consider first the scenario where you have a plain-vanilla, un-proxied, nothing-special-about-it, straight object reference, as illustrated by the following code snippet.

public class SimplePojo implements Pojo {

    public void foo() {
        // this next method invocation is a direct call on the 'this' reference
        this.bar();
    }

    public void bar() {
        // some logic...
    }
}

If you invoke a method on an object reference, the method is invoked directly on that object reference, as can be seen below.

public class Main {

    public static void main(String[] args) {

        Pojo pojo = new SimplePojo();

        // this is a direct method call on the 'pojo' reference
        pojo.foo();
    }
}

Things change slightly when the reference that client code has is a proxy. Consider the following diagram and code snippet.

public class Main {

    public static void main(String[] args) {

        ProxyFactory factory = new ProxyFactory(new SimplePojo());
        factory.addInterface(Pojo.class);
        factory.addAdvice(new RetryAdvice());

        Pojo pojo = (Pojo) factory.getProxy();

        // this is a method call on the proxy!
        pojo.foo();
    }
}

The key thing to understand here is that the client code inside the main(..) of the Main class has a reference to the proxy. This means that method calls on that object reference will be calls on the proxy, and as such the proxy will be able to delegate to all of the interceptors (advice) that are relevant to that particular method call. However, once the call has finally reached the target object, the SimplePojo reference in this case, any method calls that it may make on itself, such as this.bar() or this.foo(), are going to be invoked against the this reference, and not the proxy. This has important implications. It means that self-invocation is not going to result in the advice associated with a method invocation getting a chance to execute.

Okay, so what is to be done about this? The best approach (the term best is used loosely here) is to refactor your code such that the self-invocation does not happen. For sure, this does entail some work on your part, but it is the best, least-invasive approach. The next approach is absolutely horrendous, and I am almost reticent to point it out precisely because it is so horrendous. You can (choke!) totally tie the logic within your class to Spring AOP by doing this:

public class SimplePojo implements Pojo {

    public void foo() {
        // this works, but... gah!
        ((Pojo) AopContext.currentProxy()).bar();
    }

    public void bar() {
        // some logic...
    }
}

This totally couples your code to Spring AOP, and it makes the class itself aware of the fact that it is being used in an AOP context, which flies in the face of AOP. It also requires some additional configuration when the proxy is being created:

public class Main {

    public static void main(String[] args) {

        ProxyFactory factory = new ProxyFactory(new SimplePojo());
        factory.adddInterface(Pojo.class);
        factory.addAdvice(new RetryAdvice());
        factory.setExposeProxy(true);

        Pojo pojo = (Pojo) factory.getProxy();

        // this is a method call on the proxy!
        pojo.foo();
    }
}

Finally, it must be noted that AspectJ does not have this self-invocation issue because it is not a proxy-based AOP framework.

11.7 Programmatic creation of @AspectJ Proxies

In addition to declaring aspects in your configuration using either  or , it is also possible programmatically to create proxies that advise target objects. For the full details of Spring’s AOP API, see the next chapter. Here we want to focus on the ability to automatically create proxies using @AspectJ aspects.

The class org.springframework.aop.aspectj.annotation.AspectJProxyFactory can be used to create a proxy for a target object that is advised by one or more @AspectJ aspects. Basic usage for this class is very simple, as illustrated below. See the javadocs for full information.

// create a factory that can generate a proxy for the given target object
AspectJProxyFactory factory = new AspectJProxyFactory(targetObject);

// add an aspect, the class must be an @AspectJ aspect
// you can call this as many times as you need with different aspects
factory.addAspect(SecurityManager.class);

// you can also add existing aspect instances, the type of the object supplied must be an @AspectJ aspect
factory.addAspect(usageTracker);

// now get the proxy object...
MyInterfaceType proxy = factory.getProxy();

11.8 Using AspectJ with Spring applications

Everything we’ve covered so far in this chapter is pure Spring AOP. In this section, we’re going to look at how you can use the AspectJ compiler/weaver instead of, or in addition to, Spring AOP if your needs go beyond the facilities offered by Spring AOP alone.

Spring ships with a small AspectJ aspect library, which is available standalone in your distribution as spring-aspects.jar; you’ll need to add this to your classpath in order to use the aspects in it. Section 11.8.1, “Using AspectJ to dependency inject domain objects with Spring” and Section 11.8.2, “Other Spring aspects for AspectJ” discuss the content of this library and how you can use it. Section 11.8.3, “Configuring AspectJ aspects using Spring IoC” discusses how to dependency inject AspectJ aspects that are woven using the AspectJ compiler. Finally, Section 11.8.4, “Load-time weaving with AspectJ in the Spring Framework” provides an introduction to load-time weaving for Spring applications using AspectJ.

11.8.1 Using AspectJ to dependency inject domain objects with Spring

The Spring container instantiates and configures beans defined in your application context. It is also possible to ask a bean factory to configure a pre-existing object given the name of a bean definition containing the configuration to be applied. The spring-aspects.jar contains an annotation-driven aspect that exploits this capability to allow dependency injection of any object. The support is intended to be used for objects created outside of the control of any container. Domain objects often fall into this category because they are often created programmatically using the new operator, or by an ORM tool as a result of a database query.

The @Configurable annotation marks a class as eligible for Spring-driven configuration. In the simplest case it can be used just as a marker annotation:

package com.xyz.myapp.domain;

import org.springframework.beans.factory.annotation.Configurable;

@Configurable
public class Account {
    // ...
}

When used as a marker interface in this way, Spring will configure new instances of the annotated type ( Account in this case) using a bean definition (typically prototype-scoped) with the same name as the fully-qualified type name ( com.xyz.myapp.domain.Account). Since the default name for a bean is the fully-qualified name of its type, a convenient way to declare the prototype definition is simply to omit the id attribute:

 class="com.xyz.myapp.domain.Account" scope="prototype">
     name="fundsTransferService" ref="fundsTransferService"/>

If you want to explicitly specify the name of the prototype bean definition to use, you can do so directly in the annotation:

package com.xyz.myapp.domain;

import org.springframework.beans.factory.annotation.Configurable;

@Configurable("account")
public class Account {
    // ...
}

Spring will now look for a bean definition named "account" and use that as the definition to configure new Account instances.

You can also use autowiring to avoid having to specify a dedicated bean definition at all. To have Spring apply autowiring use the autowire property of the @Configurableannotation: specify either @Configurable(autowire=Autowire.BY_TYPE) or @Configurable(autowire=Autowire.BY_NAME for autowiring by type or by name respectively. As an alternative, as of Spring 2.5 it is preferable to specify explicit, annotation-driven dependency injection for your @Configurable beans by using@Autowired or @Inject at the field or method level (see Section 7.9, “Annotation-based container configuration” for further details).

Finally you can enable Spring dependency checking for the object references in the newly created and configured object by using the dependencyCheck attribute (for example: @Configurable(autowire=Autowire.BY_NAME,dependencyCheck=true)). If this attribute is set to true, then Spring will validate after configuration that all properties (which are not primitives or collections) have been set.

Using the annotation on its own does nothing of course. It is the AnnotationBeanConfigurerAspect in spring-aspects.jar that acts on the presence of the annotation. In essence the aspect says "after returning from the initialization of a new object of a type annotated with @Configurable, configure the newly created object using Spring in accordance with the properties of the annotation". In this context, initialization refers to newly instantiated objects (e.g., objects instantiated with the newoperator) as well as to Serializable objects that are undergoing deserialization (e.g., via readResolve()).

One of the key phrases in the above paragraph is 'in essence'. For most cases, the exact semantics of 'after returning from the initialization of a new object' will be fine…​ in this context, 'after initialization' means that the dependencies will be injected after the object has been constructed - this means that the dependencies will not be available for use in the constructor bodies of the class. If you want the dependencies to be injected before the constructor bodies execute, and thus be available for use in the body of the constructors, then you need to define this on the @Configurable declaration like so: @Configurable(preConstruction=true) You can find out more information about the language semantics of the various pointcut types in AspectJ in this appendix of the AspectJ Programming Guide.

For this to work the annotated types must be woven with the AspectJ weaver - you can either use a build-time Ant or Maven task to do this (see for example the AspectJ Development Environment Guide) or load-time weaving (see Section 11.8.4, “Load-time weaving with AspectJ in the Spring Framework”). TheAnnotationBeanConfigurerAspect itself needs configuring by Spring (in order to obtain a reference to the bean factory that is to be used to configure new objects). If you are using Java based configuration simply add @EnableSpringConfigured to any @Configuration class.

@Configuration
@EnableSpringConfigured
public class AppConfig {

}

If you prefer XML based configuration, the Spring context namespace defines a convenient context:spring-configured element:

Instances of @Configurable objects created before the aspect has been configured will result in a message being issued to the debug log and no configuration of the object taking place. An example might be a bean in the Spring configuration that creates domain objects when it is initialized by Spring. In this case you can use the "depends-on" bean attribute to manually specify that the bean depends on the configuration aspect.

 id="myService"
        class="com.xzy.myapp.service.MyService"
        depends-on="org.springframework.beans.factory.aspectj.AnnotationBeanConfigurerAspect">

    

	Do not activate @Configurable processing through the bean configurer aspect unless you really mean to rely on its semantics at runtime. In particular, make sure that you do not use @Configurable on bean classes which are registered as regular Spring beans with the container: You would get double initialization otherwise, once through the container and once through the aspect.

Unit testing @Configurable objects

One of the goals of the @Configurable support is to enable independent unit testing of domain objects without the difficulties associated with hard-coded lookups. If@Configurable types have not been woven by AspectJ then the annotation has no affect during unit testing, and you can simply set mock or stub property references in the object under test and proceed as normal. If @Configurable types have been woven by AspectJ then you can still unit test outside of the container as normal, but you will see a warning message each time that you construct an @Configurable object indicating that it has not been configured by Spring.

Working with multiple application contexts

The AnnotationBeanConfigurerAspect used to implement the @Configurable support is an AspectJ singleton aspect. The scope of a singleton aspect is the same as the scope of static members, that is to say there is one aspect instance per classloader that defines the type. This means that if you define multiple application contexts within the same classloader hierarchy you need to consider where to define the @EnableSpringConfigured bean and where to place spring-aspects.jar on the classpath.

Consider a typical Spring web-app configuration with a shared parent application context defining common business services and everything needed to support them, and one child application context per servlet containing definitions particular to that servlet. All of these contexts will co-exist within the same classloader hierarchy, and so theAnnotationBeanConfigurerAspect can only hold a reference to one of them. In this case we recommend defining the @EnableSpringConfigured bean in the shared (parent) application context: this defines the services that you are likely to want to inject into domain objects. A consequence is that you cannot configure domain objects with references to beans defined in the child (servlet-specific) contexts using the @Configurable mechanism (probably not something you want to do anyway!).

When deploying multiple web-apps within the same container, ensure that each web-application loads the types in spring-aspects.jar using its own classloader (for example, by placing spring-aspects.jar in 'WEB-INF/lib'). If spring-aspects.jar is only added to the container wide classpath (and hence loaded by the shared parent classloader), all web applications will share the same aspect instance which is probably not what you want.

11.8.2 Other Spring aspects for AspectJ

In addition to the @Configurable aspect, spring-aspects.jar contains an AspectJ aspect that can be used to drive Spring’s transaction management for types and methods annotated with the @Transactional annotation. This is primarily intended for users who want to use the Spring Framework’s transaction support outside of the Spring container.

The aspect that interprets @Transactional annotations is the AnnotationTransactionAspect. When using this aspect, you must annotate the implementation class (and/or methods within that class), not the interface (if any) that the class implements. AspectJ follows Java’s rule that annotations on interfaces are not inherited.

A @Transactional annotation on a class specifies the default transaction semantics for the execution of any public operation in the class.

A @Transactional annotation on a method within the class overrides the default transaction semantics given by the class annotation (if present). Methods of any visibility may be annotated, including private methods. Annotating non-public methods directly is the only way to get transaction demarcation for the execution of such methods.

	Since Spring Framework 4.2, spring-aspects provides a similar aspect that offers the exact same features for the standard javax.transaction.Transactional annotation. Check JtaAnnotationTransactionAspect for more details.

For AspectJ programmers that want to use the Spring configuration and transaction management support but don’t want to (or cannot) use annotations, spring-aspects.jar also contains abstract aspects you can extend to provide your own pointcut definitions. See the sources for the AbstractBeanConfigurerAspect and AbstractTransactionAspect aspects for more information. As an example, the following excerpt shows how you could write an aspect to configure all instances of objects defined in the domain model using prototype bean definitions that match the fully-qualified class names:

public aspect DomainObjectConfiguration extends AbstractBeanConfigurerAspect {

    public DomainObjectConfiguration() {
        setBeanWiringInfoResolver(new ClassNameBeanWiringInfoResolver());
    }

    // the creation of a new bean (any object in the domain model)
    protected pointcut beanCreation(Object beanInstance) :
        initialization(new(..)) &&
        SystemArchitecture.inDomainModel() &&
        this(beanInstance);

}

11.8.3 Configuring AspectJ aspects using Spring IoC

When using AspectJ aspects with Spring applications, it is natural to both want and expect to be able to configure such aspects using Spring. The AspectJ runtime itself is responsible for aspect creation, and the means of configuring the AspectJ created aspects via Spring depends on the AspectJ instantiation model (the per-xxx clause) used by the aspect.

The majority of AspectJ aspects are singleton aspects. Configuration of these aspects is very easy: simply create a bean definition referencing the aspect type as normal, and include the bean attribute 'factory-method="aspectOf"'. This ensures that Spring obtains the aspect instance by asking AspectJ for it rather than trying to create an instance itself. For example:

 id="profiler" class="com.xyz.profiler.Profiler"
        factory-method="aspectOf">

     name="profilingStrategy" ref="jamonProfilingStrategy"/>

Non-singleton aspects are harder to configure: however it is possible to do so by creating prototype bean definitions and using the @Configurable support fromspring-aspects.jar to configure the aspect instances once they have bean created by the AspectJ runtime.

If you have some @AspectJ aspects that you want to weave with AspectJ (for example, using load-time weaving for domain model types) and other @AspectJ aspects that you want to use with Spring AOP, and these aspects are all configured using Spring, then you will need to tell the Spring AOP @AspectJ autoproxying support which exact subset of the @AspectJ aspects defined in the configuration should be used for autoproxying. You can do this by using one or more  elements inside the  declaration. Each  element specifies a name pattern, and only beans with names matched by at least one of the patterns will be used for Spring AOP autoproxy configuration:

     name="thisBean"/>
     name="thatBean"/>

	Do not be misled by the name of the  element: using it will result in the creation of Spring AOP proxies. The @AspectJ style of aspect declaration is just being used here, but the AspectJ runtime is not involved.

11.8.4 Load-time weaving with AspectJ in the Spring Framework

Load-time weaving (LTW) refers to the process of weaving AspectJ aspects into an application’s class files as they are being loaded into the Java virtual machine (JVM). The focus of this section is on configuring and using LTW in the specific context of the Spring Framework: this section is not an introduction to LTW though. For full details on the specifics of LTW and configuring LTW with just AspectJ (with Spring not being involved at all), see the LTW section of the AspectJ Development Environment Guide.

The value-add that the Spring Framework brings to AspectJ LTW is in enabling much finer-grained control over the weaving process. 'Vanilla' AspectJ LTW is effected using a Java (5+) agent, which is switched on by specifying a VM argument when starting up a JVM. It is thus a JVM-wide setting, which may be fine in some situations, but often is a little too coarse. Spring-enabled LTW enables you to switch on LTW on a per-ClassLoader basis, which obviously is more fine-grained and which can make more sense in a 'single-JVM-multiple-application' environment (such as is found in a typical application server environment).

Further, in certain environments, this support enables load-time weaving without making any modifications to the application server’s launch script that will be needed to add -javaagent:path/to/aspectjweaver.jar or (as we describe later in this section) -javaagent:path/to/org.springframework.instrument-{version}.jar(previously named spring-agent.jar). Developers simply modify one or more files that form the application context to enable load-time weaving instead of relying on administrators who typically are in charge of the deployment configuration such as the launch script.

Now that the sales pitch is over, let us first walk through a quick example of AspectJ LTW using Spring, followed by detailed specifics about elements introduced in the following example. For a complete example, please see the Petclinic sample application.

A first example

Let us assume that you are an application developer who has been tasked with diagnosing the cause of some performance problems in a system. Rather than break out a profiling tool, what we are going to do is switch on a simple profiling aspect that will enable us to very quickly get some performance metrics, so that we can then apply a finer-grained profiling tool to that specific area immediately afterwards.

	The example presented here uses XML style configuration, it is also possible to configure and use @AspectJ with Java Configuration. Specifically the@EnableLoadTimeWeaving annotation can be used as an alternative to  (see below for details).

Here is the profiling aspect. Nothing too fancy, just a quick-and-dirty time-based profiler, using the @AspectJ-style of aspect declaration.

package foo;

import org.aspectj.lang.ProceedingJoinPoint;
import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Around;
import org.aspectj.lang.annotation.Pointcut;
import org.springframework.util.StopWatch;
import org.springframework.core.annotation.Order;

@Aspect
public class ProfilingAspect {

    @Around("methodsToBeProfiled()")
    public Object profile(ProceedingJoinPoint pjp) throws Throwable {
        StopWatch sw = new StopWatch(getClass().getSimpleName());
        try {
            sw.start(pjp.getSignature().getName());
            return pjp.proceed();
        } finally {
            sw.stop();
            System.out.println(sw.prettyPrint());
        }
    }

    @Pointcut("execution(public * foo..*.*(..))")
    public void methodsToBeProfiled(){}
}

We will also need to create an META-INF/aop.xml file, to inform the AspectJ weaver that we want to weave our ProfilingAspect into our classes. This file convention, namely the presence of a file (or files) on the Java classpath called META-INF/aop.xml is standard AspectJ.

    
        
         within="foo.*"/>
    

    
        
         name="foo.ProfilingAspect"/>
    

Now to the Spring-specific portion of the configuration. We need to configure a LoadTimeWeaver (all explained later, just take it on trust for now). This load-time weaver is the essential component responsible for weaving the aspect configuration in one or more META-INF/aop.xml files into the classes in your application. The good thing is that it does not require a lot of configuration, as can be seen below (there are some more options that you can specify, but these are detailed later).

 xmlns="http://www.springframework.org/schema/beans"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:context="http://www.springframework.org/schema/context"
    xsi:schemaLocation="
        http://www.springframework.org/schema/beans
        http://www.springframework.org/schema/beans/spring-beans.xsd
        http://www.springframework.org/schema/context
        http://www.springframework.org/schema/context/spring-context.xsd">

    
     id="entitlementCalculationService"
            class="foo.StubEntitlementCalculationService"/>

    
    

Now that all the required artifacts are in place - the aspect, the META-INF/aop.xml file, and the Spring configuration -, let us create a simple driver class with a main(..)method to demonstrate the LTW in action.

package foo;

import org.springframework.context.support.ClassPathXmlApplicationContext;

public final class Main {

    public static void main(String[] args) {

        ApplicationContext ctx = new ClassPathXmlApplicationContext("beans.xml", Main.class);

        EntitlementCalculationService entitlementCalculationService
            = (EntitlementCalculationService) ctx.getBean("entitlementCalculationService");

        // the profiling aspect is 'woven' around this method execution
        entitlementCalculationService.calculateEntitlement();
    }
}

There is one last thing to do. The introduction to this section did say that one could switch on LTW selectively on a per- ClassLoader basis with Spring, and this is true. However, just for this example, we are going to use a Java agent (supplied with Spring) to switch on the LTW. This is the command line we will use to run the above Mainclass:

java -javaagent:C:/projects/foo/lib/global/spring-instrument.jar foo.Main

The -javaagent is a flag for specifying and enabling agents to instrument programs running on the JVM. The Spring Framework ships with such an agent, the InstrumentationSavingAgent, which is packaged in the spring-instrument.jar that was supplied as the value of the -javaagent argument in the above example.

The output from the execution of the Main program will look something like that below. (I have introduced a Thread.sleep(..) statement into the calculateEntitlement() implementation so that the profiler actually captures something other than 0 milliseconds - the 01234 milliseconds is not an overhead introduced by the AOP :) )

Calculating entitlement

StopWatch 'ProfilingAspect': running time (millis) = 1234
------ ----- ----------------------------
ms     %     Task name
------ ----- ----------------------------
01234  100%  calculateEntitlement

Since this LTW is effected using full-blown AspectJ, we are not just limited to advising Spring beans; the following slight variation on the Main program will yield the same result.

package foo;

import org.springframework.context.support.ClassPathXmlApplicationContext;

public final class Main {

    public static void main(String[] args) {

        new ClassPathXmlApplicationContext("beans.xml", Main.class);

        EntitlementCalculationService entitlementCalculationService =
            new StubEntitlementCalculationService();

        // the profiling aspect will be 'woven' around this method execution
        entitlementCalculationService.calculateEntitlement();
    }
}

Notice how in the above program we are simply bootstrapping the Spring container, and then creating a new instance of the StubEntitlementCalculationService totally outside the context of Spring…​ the profiling advice still gets woven in.

The example admittedly is simplistic…​ however the basics of the LTW support in Spring have all been introduced in the above example, and the rest of this section will explain the 'why' behind each bit of configuration and usage in detail.

	The ProfilingAspect used in this example may be basic, but it is quite useful. It is a nice example of a development-time aspect that developers can use during development (of course), and then quite easily exclude from builds of the application being deployed into UAT or production.

Aspects

The aspects that you use in LTW have to be AspectJ aspects. They can be written in either the AspectJ language itself or you can write your aspects in the @AspectJ-style. It means that your aspects are then both valid AspectJ and Spring AOP aspects. Furthermore, the compiled aspect classes need to be available on the classpath.

'META-INF/aop.xml'

The AspectJ LTW infrastructure is configured using one or more META-INF/aop.xml files, that are on the Java classpath (either directly, or more typically in jar files).

The structure and contents of this file is detailed in the main AspectJ reference documentation, and the interested reader is referred to that resource. (I appreciate that this section is brief, but the aop.xml file is 100% AspectJ - there is no Spring-specific information or semantics that apply to it, and so there is no extra value that I can contribute either as a result), so rather than rehash the quite satisfactory section that the AspectJ developers wrote, I am just directing you there.)

Required libraries (JARS)

At a minimum you will need the following libraries to use the Spring Framework’s support for AspectJ LTW:

• spring-aop.jar (version 2.5 or later, plus all mandatory dependencies)
• aspectjweaver.jar (version 1.6.8 or later)

If you are using the Spring-provided agent to enable instrumentation, you will also need:

• spring-instrument.jar

Spring configuration

The key component in Spring’s LTW support is the LoadTimeWeaver interface (in the org.springframework.instrument.classloading package), and the numerous implementations of it that ship with the Spring distribution. A LoadTimeWeaver is responsible for adding one or more java.lang.instrument.ClassFileTransformers to a ClassLoader at runtime, which opens the door to all manner of interesting applications, one of which happens to be the LTW of aspects.

If you are unfamiliar with the idea of runtime class file transformation, you are encouraged to read the javadoc API documentation for the java.lang.instrument package before continuing. This is not a huge chore because there is - rather annoyingly - precious little documentation there…​ the key interfaces and classes will at least be laid out in front of you for reference as you read through this section.

Configuring a LoadTimeWeaver for a particular ApplicationContext can be as easy as adding one line. (Please note that you almost certainly will need to be using anApplicationContext as your Spring container - typically a BeanFactory will not be enough because the LTW support makes use of BeanFactoryPostProcessors.)

To enable the Spring Framework’s LTW support, you need to configure a LoadTimeWeaver, which typically is done using the @EnableLoadTimeWeaving annotation.

@Configuration
@EnableLoadTimeWeaving
public class AppConfig {

}

Alternatively, if you prefer XML based configuration, use the  element. Note that the element is defined in the context namespace.

 xmlns="http://www.springframework.org/schema/beans"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:context="http://www.springframework.org/schema/context"
    xsi:schemaLocation="
        http://www.springframework.org/schema/beans
        http://www.springframework.org/schema/beans/spring-beans.xsd
        http://www.springframework.org/schema/context
        http://www.springframework.org/schema/context/spring-context.xsd">

    

The above configuration will define and register a number of LTW-specific infrastructure beans for you automatically, such as a LoadTimeWeaver and an AspectJWeavingEnabler. The default LoadTimeWeaver is the DefaultContextLoadTimeWeaver class, which attempts to decorate an automatically detected LoadTimeWeaver: the exact type of LoadTimeWeaver that will be 'automatically detected' is dependent upon your runtime environment (summarized in the following table).

Table 11.1. DefaultContextLoadTimeWeaver LoadTimeWeavers

Runtime Environment	LoadTimeWeaver implementation
Running in Oracle’s WebLogic	WebLogicLoadTimeWeaver
Running in Oracle’s GlassFish	GlassFishLoadTimeWeaver
Running in Apache Tomcat	TomcatLoadTimeWeaver
Running in Red Hat’s JBoss AS or WildFly	JBossLoadTimeWeaver
Running in IBM’s WebSphere	WebSphereLoadTimeWeaver
JVM started with Spring InstrumentationSavingAgent (java -javaagent:path/to/spring-instrument.jar)	InstrumentationLoadTimeWeaver
Fallback, expecting the underlying ClassLoader to follow common conventions (e.g. applicable to TomcatInstrumentableClassLoader and Resin)	ReflectiveLoadTimeWeaver

Note that these are just the LoadTimeWeavers that are autodetected when using the DefaultContextLoadTimeWeaver: it is of course possible to specify exactly whichLoadTimeWeaver implementation that you wish to use.

To specify a specific LoadTimeWeaver with Java configuration implement the LoadTimeWeavingConfigurer interface and override the getLoadTimeWeaver()method:

@Configuration
@EnableLoadTimeWeaving
public class AppConfig implements LoadTimeWeavingConfigurer {

    @Override
    public LoadTimeWeaver getLoadTimeWeaver() {
        return new ReflectiveLoadTimeWeaver();
    }
}

If you are using XML based configuration you can specify the fully-qualified classname as the value of the weaver-class attribute on the element:

 xmlns="http://www.springframework.org/schema/beans"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:context="http://www.springframework.org/schema/context"
    xsi:schemaLocation="
        http://www.springframework.org/schema/beans
        http://www.springframework.org/schema/beans/spring-beans.xsd
        http://www.springframework.org/schema/context
        http://www.springframework.org/schema/context/spring-context.xsd">

    
            weaver-class="org.springframework.instrument.classloading.ReflectiveLoadTimeWeaver"/>

The LoadTimeWeaver that is defined and registered by the configuration can be later retrieved from the Spring container using the well-known name loadTimeWeaver. Remember that the LoadTimeWeaver exists just as a mechanism for Spring’s LTW infrastructure to add one or more ClassFileTransformers. The actualClassFileTransformer that does the LTW is the ClassPreProcessorAgentAdapter (from the org.aspectj.weaver.loadtime package) class. See the class-level javadocs of the ClassPreProcessorAgentAdapter class for further details, because the specifics of how the weaving is actually effected is beyond the scope of this section.

There is one final attribute of the configuration left to discuss: the aspectjWeaving attribute (or aspectj-weaving if you are using XML). This is a simple attribute that controls whether LTW is enabled or not; it is as simple as that. It accepts one of three possible values, summarized below, with the default value being autodetect if the attribute is not present.

Table 11.2. AspectJ weaving attribute values

Annotation Value	XML Value	Explanation
ENABLED	on	AspectJ weaving is on, and aspects will be woven at load-time as appropriate.
DISABLED	off	LTW is off…​ no aspect will be woven at load-time.
AUTODETECT	autodetect	If the Spring LTW infrastructure can find at least one META-INF/aop.xml file, then AspectJ weaving is on, else it is off. This is the default value.

Environment-specific configuration

This last section contains any additional settings and configuration that you will need when using Spring’s LTW support in environments such as application servers and web containers.

Tomcat

Historically, Apache Tomcat's default class loader did not support class transformation which is why Spring provides an enhanced implementation that addresses this need. Named TomcatInstrumentableClassLoader, the loader works on Tomcat 6.0 and above.

	Do not define TomcatInstrumentableClassLoader anymore on Tomcat 8.0 and higher. Instead, let Spring automatically use Tomcat’s new native InstrumentableClassLoader facility through the TomcatLoadTimeWeaver strategy.

If you still need to use TomcatInstrumentableClassLoader, it can be registered individually for each web application as follows:

• Copy org.springframework.instrument.tomcat.jar into $CATALINA_HOME/lib, where $CATALINA_HOME represents the root of the Tomcat installation)
• Instruct Tomcat to use the custom class loader (instead of the default) by editing the web application context file:

 path="/myWebApp" docBase="/my/webApp/location">
    
        loaderClass="org.springframework.instrument.classloading.tomcat.TomcatInstrumentableClassLoader"/>

Apache Tomcat (6.0+) supports several context locations:

• server configuration file - $CATALINA_HOME/conf/server.xml
• default context configuration - $CATALINA_HOME/conf/context.xml - that affects all deployed web applications
• per-web application configuration which can be deployed either on the server-side at $CATALINA_HOME/conf/[enginename]/[hostname]/[webapp]-context.xml or embedded inside the web-app archive at META-INF/context.xml

For efficiency, the embedded per-web-app configuration style is recommended because it will impact only applications that use the custom class loader and does not require any changes to the server configuration. See the Tomcat 6.0.x documentation for more details about available context locations.

Alternatively, consider the use of the Spring-provided generic VM agent, to be specified in Tomcat’s launch script (see above). This will make instrumentation available to all deployed web applications, no matter what ClassLoader they happen to run on.

WebLogic, WebSphere, Resin, GlassFish, JBoss

Recent versions of WebLogic Server (version 10 and above), IBM WebSphere Application Server (version 7 and above), Resin (3.1 and above) and JBoss (6.x or above) provide a ClassLoader that is capable of local instrumentation. Spring’s native LTW leverages such ClassLoaders to enable AspectJ weaving. You can enable LTW by simply activating load-time weaving as described earlier. Specifically, you do not need to modify the launch script to add -javaagent:path/to/spring-instrument.jar.

Note that GlassFish instrumentation-capable ClassLoader is available only in its EAR environment. For GlassFish web applications, follow the Tomcat setup instructions as outlined above.

Note that on JBoss 6.x, the app server scanning needs to be disabled to prevent it from loading the classes before the application actually starts. A quick workaround is to add to your artifact a file named WEB-INF/jboss-scanning.xml with the following content:

 xmlns="urn:jboss:scanning:1.0"/>

Generic Java applications

When class instrumentation is required in environments that do not support or are not supported by the existing LoadTimeWeaver implementations, a JDK agent can be the only solution. For such cases, Spring provides InstrumentationLoadTimeWeaver, which requires a Spring-specific (but very general) VM agent,org.springframework.instrument-{version}.jar (previously named spring-agent.jar).

To use it, you must start the virtual machine with the Spring agent, by supplying the following JVM options:

-javaagent:/path/to/org.springframework.instrument-{version}.jar

Note that this requires modification of the VM launch script which may prevent you from using this in application server environments (depending on your operation policies). Additionally, the JDK agent will instrument the entire VM which can prove expensive.

For performance reasons, it is recommended to use this configuration only if your target environment (such as Jetty) does not have (or does not support) a dedicated LTW.

11.9 Further Resources

More information on AspectJ can be found on the AspectJ website.

The book Eclipse AspectJ by Adrian Colyer et. al. (Addison-Wesley, 2005) provides a comprehensive introduction and reference for the AspectJ language.

The book AspectJ in Action, Second Edition by Ramnivas Laddad (Manning, 2009) comes highly recommended; the focus of the book is on AspectJ, but a lot of general AOP themes are explored (in some depth).

12. Spring AOP APIs

12.1 Introduction

The previous chapter described the Spring’s support for AOP using @AspectJ and schema-based aspect definitions. In this chapter we discuss the lower-level Spring AOP APIs and the AOP support used in Spring 1.2 applications. For new applications, we recommend the use of the Spring 2.0 and later AOP support described in the previous chapter, but when working with existing applications, or when reading books and articles, you may come across Spring 1.2 style examples. Spring 4.0 is backwards compatible with Spring 1.2 and everything described in this chapter is fully supported in Spring 4.0.

12.2 Pointcut API in Spring

Let’s look at how Spring handles the crucial pointcut concept.

12.2.1 Concepts

Spring’s pointcut model enables pointcut reuse independent of advice types. It’s possible to target different advice using the same pointcut.

The org.springframework.aop.Pointcut interface is the central interface, used to target advices to particular classes and methods. The complete interface is shown below:

public interface Pointcut {

    ClassFilter getClassFilter();

    MethodMatcher getMethodMatcher();

}

Splitting the Pointcut interface into two parts allows reuse of class and method matching parts, and fine-grained composition operations (such as performing a "union" with another method matcher).

The ClassFilter interface is used to restrict the pointcut to a given set of target classes. If the matches() method always returns true, all target classes will be matched:

public interface ClassFilter {

    boolean matches(Class clazz);
}

The MethodMatcher interface is normally more important. The complete interface is shown below:

public interface MethodMatcher {

    boolean matches(Method m, Class targetClass);

    boolean isRuntime();

    boolean matches(Method m, Class targetClass, Object[] args);
}

The matches(Method, Class) method is used to test whether this pointcut will ever match a given method on a target class. This evaluation can be performed when an AOP proxy is created, to avoid the need for a test on every method invocation. If the 2-argument matches method returns true for a given method, and the isRuntime()method for the MethodMatcher returns true, the 3-argument matches method will be invoked on every method invocation. This enables a pointcut to look at the arguments passed to the method invocation immediately before the target advice is to execute.

Most MethodMatchers are static, meaning that their isRuntime() method returns false. In this case, the 3-argument matches method will never be invoked.

	If possible, try to make pointcuts static, allowing the AOP framework to cache the results of pointcut evaluation when an AOP proxy is created.

12.2.2 Operations on pointcuts

Spring supports operations on pointcuts: notably, union and intersection.

• Union means the methods that either pointcut matches.
• Intersection means the methods that both pointcuts match.
• Union is usually more useful.
• Pointcuts can be composed using the static methods in the org.springframework.aop.support.Pointcuts class, or using the ComposablePointcut class in the same package. However, using AspectJ pointcut expressions is usually a simpler approach.

12.2.3 AspectJ expression pointcuts

Since 2.0, the most important type of pointcut used by Spring is org.springframework.aop.aspectj.AspectJExpressionPointcut. This is a pointcut that uses an AspectJ supplied library to parse an AspectJ pointcut expression string.

See the previous chapter for a discussion of supported AspectJ pointcut primitives.

12.2.4 Convenience pointcut implementations

Spring provides several convenient pointcut implementations. Some can be used out of the box; others are intended to be subclassed in application-specific pointcuts.

Static pointcuts

Static pointcuts are based on method and target class, and cannot take into account the method’s arguments. Static pointcuts are sufficient - and best - for most usages. It’s possible for Spring to evaluate a static pointcut only once, when a method is first invoked: after that, there is no need to evaluate the pointcut again with each method invocation.

Let’s consider some static pointcut implementations included with Spring.

Regular expression pointcuts

One obvious way to specify static pointcuts is regular expressions. Several AOP frameworks besides Spring make this possible.org.springframework.aop.support.JdkRegexpMethodPointcut is a generic regular expression pointcut, using the regular expression support in JDK 1.4+.

Using the JdkRegexpMethodPointcut class, you can provide a list of pattern Strings. If any of these is a match, the pointcut will evaluate to true. (So the result is effectively the union of these pointcuts.)

The usage is shown below:

 id="settersAndAbsquatulatePointcut"
        class="org.springframework.aop.support.JdkRegexpMethodPointcut">
     name="patterns">
        
            .*set.*
            .*absquatulate
        
    

Spring provides a convenience class, RegexpMethodPointcutAdvisor, that allows us to also reference an Advice (remember that an Advice can be an interceptor, before advice, throws advice etc.). Behind the scenes, Spring will use a JdkRegexpMethodPointcut. Using RegexpMethodPointcutAdvisor simplifies wiring, as the one bean encapsulates both pointcut and advice, as shown below:

 id="settersAndAbsquatulateAdvisor"
        class="org.springframework.aop.support.RegexpMethodPointcutAdvisor">
     name="advice">
         bean="beanNameOfAopAllianceInterceptor"/>
    
     name="patterns">
        
            .*set.*
            .*absquatulate
        
    

RegexpMethodPointcutAdvisor can be used with any Advice type.

Attribute-driven pointcuts

An important type of static pointcut is a metadata-driven pointcut. This uses the values of metadata attributes: typically, source-level metadata.

Dynamic pointcuts

Dynamic pointcuts are costlier to evaluate than static pointcuts. They take into account method arguments, as well as static information. This means that they must be evaluated with every method invocation; the result cannot be cached, as arguments will vary.

The main example is the control flow pointcut.

Control flow pointcuts

Spring control flow pointcuts are conceptually similar to AspectJ cflow pointcuts, although less powerful. (There is currently no way to specify that a pointcut executes below a join point matched by another pointcut.) A control flow pointcut matches the current call stack. For example, it might fire if the join point was invoked by a method in the com.mycompany.web package, or by the SomeCaller class. Control flow pointcuts are specified using the org.springframework.aop.support.ControlFlowPointcut class.

	Control flow pointcuts are significantly more expensive to evaluate at runtime than even other dynamic pointcuts. In Java 1.4, the cost is about 5 times that of other dynamic pointcuts.

12.2.5 Pointcut superclasses

Spring provides useful pointcut superclasses to help you to implement your own pointcuts.

Because static pointcuts are most useful, you’ll probably subclass StaticMethodMatcherPointcut, as shown below. This requires implementing just one abstract method (although it’s possible to override other methods to customize behavior):

class TestStaticPointcut extends StaticMethodMatcherPointcut {

    public boolean matches(Method m, Class targetClass) {
        // return true if custom criteria match
    }
}

There are also superclasses for dynamic pointcuts.

You can use custom pointcuts with any advice type in Spring 1.0 RC2 and above.

12.2.6 Custom pointcuts

Because pointcuts in Spring AOP are Java classes, rather than language features (as in AspectJ) it’s possible to declare custom pointcuts, whether static or dynamic. Custom pointcuts in Spring can be arbitrarily complex. However, using the AspectJ pointcut expression language is recommended if possible.

	Later versions of Spring may offer support for "semantic pointcuts" as offered by JAC: for example, "all methods that change instance variables in the target object."

12.3 Advice API in Spring

Let’s now look at how Spring AOP handles advice.

12.3.1 Advice lifecycles

Each advice is a Spring bean. An advice instance can be shared across all advised objects, or unique to each advised object. This corresponds to per-class or per-instance advice.

Per-class advice is used most often. It is appropriate for generic advice such as transaction advisors. These do not depend on the state of the proxied object or add new state; they merely act on the method and arguments.

Per-instance advice is appropriate for introductions, to support mixins. In this case, the advice adds state to the proxied object.

It’s possible to use a mix of shared and per-instance advice in the same AOP proxy.

12.3.2 Advice types in Spring

Spring provides several advice types out of the box, and is extensible to support arbitrary advice types. Let us look at the basic concepts and standard advice types.

Interception around advice

The most fundamental advice type in Spring is interception around advice.

Spring is compliant with the AOP Alliance interface for around advice using method interception. MethodInterceptors implementing around advice should implement the following interface:

public interface MethodInterceptor extends Interceptor {

    Object invoke(MethodInvocation invocation) throws Throwable;
}

The MethodInvocation argument to the invoke() method exposes the method being invoked; the target join point; the AOP proxy; and the arguments to the method. Theinvoke() method should return the invocation’s result: the return value of the join point.

A simple MethodInterceptor implementation looks as follows:

public class DebugInterceptor implements MethodInterceptor {

    public Object invoke(MethodInvocation invocation) throws Throwable {
        System.out.println("Before: invocation=[" + invocation + "]");
        Object rval = invocation.proceed();
        System.out.println("Invocation returned");
        return rval;
    }
}

Note the call to the MethodInvocation’s proceed() method. This proceeds down the interceptor chain towards the join point. Most interceptors will invoke this method, and return its return value. However, a MethodInterceptor, like any around advice, can return a different value or throw an exception rather than invoke the proceed method. However, you don’t want to do this without good reason!

MethodInterceptors offer interoperability with other AOP Alliance-compliant AOP implementations. The other advice types discussed in the remainder of this section implement common AOP concepts, but in a Spring-specific way. While there is an advantage in using the most specific advice type, stick with MethodInterceptor around advice if you are likely to want to run the aspect in another AOP framework. Note that pointcuts are not currently interoperable between frameworks, and the AOP Alliance does not currently define pointcut interfaces.

Before advice

A simpler advice type is a before advice. This does not need a MethodInvocation object, since it will only be called before entering the method.

The main advantage of a before advice is that there is no need to invoke the proceed() method, and therefore no possibility of inadvertently failing to proceed down the interceptor chain.

The MethodBeforeAdvice interface is shown below. (Spring’s API design would allow for field before advice, although the usual objects apply to field interception and it’s unlikely that Spring will ever implement it).

public interface MethodBeforeAdvice extends BeforeAdvice {

    void before(Method m, Object[] args, Object target) throws Throwable;
}

Note the return type is void. Before advice can insert custom behavior before the join point executes, but cannot change the return value. If a before advice throws an exception, this will abort further execution of the interceptor chain. The exception will propagate back up the interceptor chain. If it is unchecked, or on the signature of the invoked method, it will be passed directly to the client; otherwise it will be wrapped in an unchecked exception by the AOP proxy.

An example of a before advice in Spring, which counts all method invocations:

public class CountingBeforeAdvice implements MethodBeforeAdvice {

    private int count;

    public void before(Method m, Object[] args, Object target) throws Throwable {
        ++count;
    }

    public int getCount() {
        return count;
    }
}

	Before advice can be used with any pointcut.

Throws advice

Throws advice is invoked after the return of the join point if the join point threw an exception. Spring offers typed throws advice. Note that this means that theorg.springframework.aop.ThrowsAdvice interface does not contain any methods: It is a tag interface identifying that the given object implements one or more typed throws advice methods. These should be in the form of:

afterThrowing([Method, args, target], subclassOfThrowable)

Only the last argument is required. The method signatures may have either one or four arguments, depending on whether the advice method is interested in the method and arguments. The following classes are examples of throws advice.

The advice below is invoked if a RemoteException is thrown (including subclasses):

public class RemoteThrowsAdvice implements ThrowsAdvice {

    public void afterThrowing(RemoteException ex) throws Throwable {
        // Do something with remote exception
    }
}

The following advice is invoked if a ServletException is thrown. Unlike the above advice, it declares 4 arguments, so that it has access to the invoked method, method arguments and target object:

public class ServletThrowsAdviceWithArguments implements ThrowsAdvice {

    public void afterThrowing(Method m, Object[] args, Object target, ServletException ex) {
        // Do something with all arguments
    }
}

The final example illustrates how these two methods could be used in a single class, which handles both RemoteException and ServletException. Any number of throws advice methods can be combined in a single class.

public static class CombinedThrowsAdvice implements ThrowsAdvice {

    public void afterThrowing(RemoteException ex) throws Throwable {
        // Do something with remote exception
    }

    public void afterThrowing(Method m, Object[] args, Object target, ServletException ex) {
        // Do something with all arguments
    }
}

If a throws-advice method throws an exception itself, it will override the original exception (i.e. change the exception thrown to the user). The overriding exception will typically be a RuntimeException; this is compatible with any method signature. However, if a throws-advice method throws a checked exception, it will have to match the declared exceptions of the target method and is hence to some degree coupled to specific target method signatures. Do not throw an undeclared checked exception that is incompatible with the target method’s signature!

	Throws advice can be used with any pointcut.

After Returning advice

An after returning advice in Spring must implement the org.springframework.aop.AfterReturningAdvice interface, shown below:

public interface AfterReturningAdvice extends Advice {

    void afterReturning(Object returnValue, Method m, Object[] args, Object target)
            throws Throwable;
}

An after returning advice has access to the return value (which it cannot modify), invoked method, methods arguments and target.

The following after returning advice counts all successful method invocations that have not thrown exceptions:

public class CountingAfterReturningAdvice implements AfterReturningAdvice {

    private int count;

    public void afterReturning(Object returnValue, Method m, Object[] args, Object target)
            throws Throwable {
        ++count;
    }

    public int getCount() {
        return count;
    }
}

This advice doesn’t change the execution path. If it throws an exception, this will be thrown up the interceptor chain instead of the return value.

	After returning advice can be used with any pointcut.

Introduction advice

Spring treats introduction advice as a special kind of interception advice.

Introduction requires an IntroductionAdvisor, and an IntroductionInterceptor, implementing the following interface:

public interface IntroductionInterceptor extends MethodInterceptor {

    boolean implementsInterface(Class intf);
}

The invoke() method inherited from the AOP Alliance MethodInterceptor interface must implement the introduction: that is, if the invoked method is on an introduced interface, the introduction interceptor is responsible for handling the method call - it cannot invoke proceed().

Introduction advice cannot be used with any pointcut, as it applies only at class, rather than method, level. You can only use introduction advice with theIntroductionAdvisor, which has the following methods:

public interface IntroductionAdvisor extends Advisor, IntroductionInfo {

    ClassFilter getClassFilter();

    void validateInterfaces() throws IllegalArgumentException;
}

public interface IntroductionInfo {

    Class[] getInterfaces();
}

There is no MethodMatcher, and hence no Pointcut, associated with introduction advice. Only class filtering is logical.

The getInterfaces() method returns the interfaces introduced by this advisor.

The validateInterfaces() method is used internally to see whether or not the introduced interfaces can be implemented by the configured IntroductionInterceptor.

Let’s look at a simple example from the Spring test suite. Let’s suppose we want to introduce the following interface to one or more objects:

public interface Lockable {
    void lock();
    void unlock();
    boolean locked();
}

This illustrates a mixin. We want to be able to cast advised objects to Lockable, whatever their type, and call lock and unlock methods. If we call the lock() method, we want all setter methods to throw a LockedException. Thus we can add an aspect that provides the ability to make objects immutable, without them having any knowledge of it: a good example of AOP.

Firstly, we’ll need an IntroductionInterceptor that does the heavy lifting. In this case, we extend the org.springframework.aop.support.DelegatingIntroductionInterceptor convenience class. We could implement IntroductionInterceptor directly, but usingDelegatingIntroductionInterceptor is best for most cases.

The DelegatingIntroductionInterceptor is designed to delegate an introduction to an actual implementation of the introduced interface(s), concealing the use of interception to do so. The delegate can be set to any object using a constructor argument; the default delegate (when the no-arg constructor is used) is this. Thus in the example below, the delegate is the LockMixin subclass of DelegatingIntroductionInterceptor. Given a delegate (by default itself), a DelegatingIntroductionInterceptor instance looks for all interfaces implemented by the delegate (other than IntroductionInterceptor), and will support introductions against any of them. It’s possible for subclasses such as LockMixin to call the suppressInterface(Class intf) method to suppress interfaces that should not be exposed. However, no matter how many interfaces an IntroductionInterceptor is prepared to support, the IntroductionAdvisor used will control which interfaces are actually exposed. An introduced interface will conceal any implementation of the same interface by the target.

Thus LockMixin extends DelegatingIntroductionInterceptor and implements Lockable itself. The superclass automatically picks up that Lockable can be supported for introduction, so we don’t need to specify that. We could introduce any number of interfaces in this way.

Note the use of the locked instance variable. This effectively adds additional state to that held in the target object.

public class LockMixin extends DelegatingIntroductionInterceptor implements Lockable {

    private boolean locked;

    public void lock() {
        this.locked = true;
    }

    public void unlock() {
        this.locked = false;
    }

    public boolean locked() {
        return this.locked;
    }

    public Object invoke(MethodInvocation invocation) throws Throwable {
        if (locked() && invocation.getMethod().getName().indexOf("set") == 0) {
            throw new LockedException();
        }
        return super.invoke(invocation);
    }

}

Often it isn’t necessary to override the invoke() method: the DelegatingIntroductionInterceptor implementation - which calls the delegate method if the method is introduced, otherwise proceeds towards the join point - is usually sufficient. In the present case, we need to add a check: no setter method can be invoked if in locked mode.

The introduction advisor required is simple. All it needs to do is hold a distinct LockMixin instance, and specify the introduced interfaces - in this case, just Lockable. A more complex example might take a reference to the introduction interceptor (which would be defined as a prototype): in this case, there’s no configuration relevant for a LockMixin, so we simply create it using new.

public class LockMixinAdvisor extends DefaultIntroductionAdvisor {

    public LockMixinAdvisor() {
        super(new LockMixin(), Lockable.class);
    }
}

We can apply this advisor very simply: it requires no configuration. (However, it is necessary: It’s impossible to use an IntroductionInterceptor without anIntroductionAdvisor.) As usual with introductions, the advisor must be per-instance, as it is stateful. We need a different instance of LockMixinAdvisor, and henceLockMixin, for each advised object. The advisor comprises part of the advised object’s state.

We can apply this advisor programmatically, using the Advised.addAdvisor() method, or (the recommended way) in XML configuration, like any other advisor. All proxy creation choices discussed below, including "auto proxy creators," correctly handle introductions and stateful mixins.

12.4 Advisor API in Spring

In Spring, an Advisor is an aspect that contains just a single advice object associated with a pointcut expression.

Apart from the special case of introductions, any advisor can be used with any advice. org.springframework.aop.support.DefaultPointcutAdvisor is the most commonly used advisor class. For example, it can be used with a MethodInterceptor, BeforeAdvice or ThrowsAdvice.

It is possible to mix advisor and advice types in Spring in the same AOP proxy. For example, you could use a interception around advice, throws advice and before advice in one proxy configuration: Spring will automatically create the necessary interceptor chain.

12.5 Using the ProxyFactoryBean to create AOP proxies

If you’re using the Spring IoC container (an ApplicationContext or BeanFactory) for your business objects - and you should be! - you will want to use one of Spring’s AOP FactoryBeans. (Remember that a factory bean introduces a layer of indirection, enabling it to create objects of a different type.)

	The Spring AOP support also uses factory beans under the covers.

The basic way to create an AOP proxy in Spring is to use the org.springframework.aop.framework.ProxyFactoryBean. This gives complete control over the pointcuts and advice that will apply, and their ordering. However, there are simpler options that are preferable if you don’t need such control.

12.5.1 Basics

The ProxyFactoryBean, like other Spring FactoryBean implementations, introduces a level of indirection. If you define a ProxyFactoryBean with name foo, what objects referencing foo see is not the ProxyFactoryBean instance itself, but an object created by the ProxyFactoryBean's implementation of the getObject()method. This method will create an AOP proxy wrapping a target object.

One of the most important benefits of using a ProxyFactoryBean or another IoC-aware class to create AOP proxies, is that it means that advices and pointcuts can also be managed by IoC. This is a powerful feature, enabling certain approaches that are hard to achieve with other AOP frameworks. For example, an advice may itself reference application objects (besides the target, which should be available in any AOP framework), benefiting from all the pluggability provided by Dependency Injection.

12.5.2 JavaBean properties

In common with most FactoryBean implementations provided with Spring, the ProxyFactoryBean class is itself a JavaBean. Its properties are used to:

• Specify the target you want to proxy.
• Specify whether to use CGLIB (see below and also Section 12.5.3, “JDK- and CGLIB-based proxies”).

Some key properties are inherited from org.springframework.aop.framework.ProxyConfig (the superclass for all AOP proxy factories in Spring). These key properties include:

• proxyTargetClass: true if the target class is to be proxied, rather than the target class' interfaces. If this property value is set to true, then CGLIB proxies will be created (but see also Section 12.5.3, “JDK- and CGLIB-based proxies”).
• optimize: controls whether or not aggressive optimizations are applied to proxies created via CGLIB. One should not blithely use this setting unless one fully understands how the relevant AOP proxy handles optimization. This is currently used only for CGLIB proxies; it has no effect with JDK dynamic proxies.
• frozen: if a proxy configuration is frozen, then changes to the configuration are no longer allowed. This is useful both as a slight optimization and for those cases when you don’t want callers to be able to manipulate the proxy (via the Advised interface) after the proxy has been created. The default value of this property is false, so changes such as adding additional advice are allowed.
• exposeProxy: determines whether or not the current proxy should be exposed in a ThreadLocal so that it can be accessed by the target. If a target needs to obtain the proxy and the exposeProxy property is set to true, the target can use the AopContext.currentProxy() method.

Other properties specific to ProxyFactoryBean include:

• proxyInterfaces: array of String interface names. If this isn’t supplied, a CGLIB proxy for the target class will be used (but see also Section 12.5.3, “JDK- and CGLIB-based proxies”).
• interceptorNames: String array of Advisor, interceptor or other advice names to apply. Ordering is significant, on a first come-first served basis. That is to say that the first interceptor in the list will be the first to be able to intercept the invocation.

The names are bean names in the current factory, including bean names from ancestor factories. You can’t mention bean references here since doing so would result in theProxyFactoryBean ignoring the singleton setting of the advice.

You can append an interceptor name with an asterisk ( *). This will result in the application of all advisor beans with names starting with the part before the asterisk to be applied. An example of using this feature can be found in Section 12.5.6, “Using 'global' advisors”.

• singleton: whether or not the factory should return a single object, no matter how often the getObject() method is called. Several FactoryBean implementations offer such a method. The default value is true. If you want to use stateful advice - for example, for stateful mixins - use prototype advices along with a singleton value offalse.

12.5.3 JDK- and CGLIB-based proxies

This section serves as the definitive documentation on how the ProxyFactoryBean chooses to create one of either a JDK- and CGLIB-based proxy for a particular target object (that is to be proxied).

	The behavior of the ProxyFactoryBean with regard to creating JDK- or CGLIB-based proxies changed between versions 1.2.x and 2.0 of Spring. The ProxyFactoryBean now exhibits similar semantics with regard to auto-detecting interfaces as those of the TransactionProxyFactoryBean class.

If the class of a target object that is to be proxied (hereafter simply referred to as the target class) doesn’t implement any interfaces, then a CGLIB-based proxy will be created. This is the easiest scenario, because JDK proxies are interface based, and no interfaces means JDK proxying isn’t even possible. One simply plugs in the target bean, and specifies the list of interceptors via the interceptorNames property. Note that a CGLIB-based proxy will be created even if the proxyTargetClass property of theProxyFactoryBean has been set to false. (Obviously this makes no sense, and is best removed from the bean definition because it is at best redundant, and at worst confusing.)

If the target class implements one (or more) interfaces, then the type of proxy that is created depends on the configuration of the ProxyFactoryBean.

If the proxyTargetClass property of the ProxyFactoryBean has been set to true, then a CGLIB-based proxy will be created. This makes sense, and is in keeping with the principle of least surprise. Even if the proxyInterfaces property of the ProxyFactoryBean has been set to one or more fully qualified interface names, the fact that the proxyTargetClass property is set to true will cause CGLIB-based proxying to be in effect.

If the proxyInterfaces property of the ProxyFactoryBean has been set to one or more fully qualified interface names, then a JDK-based proxy will be created. The created proxy will implement all of the interfaces that were specified in the proxyInterfaces property; if the target class happens to implement a whole lot more interfaces than those specified in the proxyInterfaces property, that is all well and good but those additional interfaces will not be implemented by the returned proxy.

If the proxyInterfaces property of the ProxyFactoryBean has not been set, but the target class does implement one (or more) interfaces, then theProxyFactoryBean will auto-detect the fact that the target class does actually implement at least one interface, and a JDK-based proxy will be created. The interfaces that are actually proxied will be all of the interfaces that the target class implements; in effect, this is the same as simply supplying a list of each and every interface that the target class implements to the proxyInterfaces property. However, it is significantly less work, and less prone to typos.

12.5.4 Proxying interfaces

Let’s look at a simple example of ProxyFactoryBean in action. This example involves:

• A target bean that will be proxied. This is the "personTarget" bean definition in the example below.
• An Advisor and an Interceptor used to provide advice.
• An AOP proxy bean definition specifying the target object (the personTarget bean) and the interfaces to proxy, along with the advices to apply.

 id="personTarget" class="com.mycompany.PersonImpl">
     name="name" value="Tony"/>
     name="age" value="51"/>


 id="myAdvisor" class="com.mycompany.MyAdvisor">
     name="someProperty" value="Custom string property value"/>


 id="debugInterceptor" class="org.springframework.aop.interceptor.DebugInterceptor">


 id="person"
    class="org.springframework.aop.framework.ProxyFactoryBean">
     name="proxyInterfaces" value="com.mycompany.Person"/>

     name="target" ref="personTarget"/>
     name="interceptorNames">
        
            myAdvisor
            debugInterceptor
        
    

Note that the interceptorNames property takes a list of String: the bean names of the interceptor or advisors in the current factory. Advisors, interceptors, before, after returning and throws advice objects can be used. The ordering of advisors is significant.

You might be wondering why the list doesn’t hold bean references. The reason for this is that if the ProxyFactoryBean’s singleton property is set to false, it must be able to return independent proxy instances. If any of the advisors is itself a prototype, an independent instance would need to be returned, so it’s necessary to be able to obtain an instance of the prototype from the factory; holding a reference isn’t sufficient.

The "person" bean definition above can be used in place of a Person implementation, as follows:

Person person = (Person) factory.getBean("person");

Other beans in the same IoC context can express a strongly typed dependency on it, as with an ordinary Java object:

 id="personUser" class="com.mycompany.PersonUser">
     name="person"> bean="person"/>

The PersonUser class in this example would expose a property of type Person. As far as it’s concerned, the AOP proxy can be used transparently in place of a "real" person implementation. However, its class would be a dynamic proxy class. It would be possible to cast it to the Advised interface (discussed below).

It’s possible to conceal the distinction between target and proxy using an anonymous inner bean, as follows. Only the ProxyFactoryBean definition is different; the advice is included only for completeness:

 id="myAdvisor" class="com.mycompany.MyAdvisor">
     name="someProperty" value="Custom string property value"/>


 id="debugInterceptor" class="org.springframework.aop.interceptor.DebugInterceptor"/>

 id="person" class="org.springframework.aop.framework.ProxyFactoryBean">
     name="proxyInterfaces" value="com.mycompany.Person"/>
    
     name="target">
         class="com.mycompany.PersonImpl">
             name="name" value="Tony"/>
             name="age" value="51"/>
        
    
     name="interceptorNames">
        
            myAdvisor
            debugInterceptor
        
    

This has the advantage that there’s only one object of type Person: useful if we want to prevent users of the application context from obtaining a reference to the un-advised object, or need to avoid any ambiguity with Spring IoC autowiring. There’s also arguably an advantage in that the ProxyFactoryBean definition is self-contained. However, there are times when being able to obtain the un-advised target from the factory might actually be an advantage: for example, in certain test scenarios.

12.5.5 Proxying classes

What if you need to proxy a class, rather than one or more interfaces?

Imagine that in our example above, there was no Person interface: we needed to advise a class called Person that didn’t implement any business interface. In this case, you can configure Spring to use CGLIB proxying, rather than dynamic proxies. Simply set the proxyTargetClass property on the ProxyFactoryBean above to true. While it’s best to program to interfaces, rather than classes, the ability to advise classes that don’t implement interfaces can be useful when working with legacy code. (In general, Spring isn’t prescriptive. While it makes it easy to apply good practices, it avoids forcing a particular approach.)

If you want to, you can force the use of CGLIB in any case, even if you do have interfaces.

CGLIB proxying works by generating a subclass of the target class at runtime. Spring configures this generated subclass to delegate method calls to the original target: the subclass is used to implement the Decorator pattern, weaving in the advice.

CGLIB proxying should generally be transparent to users. However, there are some issues to consider:

• Final methods can’t be advised, as they can’t be overridden.
• There is no need to add CGLIB to your classpath. As of Spring 3.2, CGLIB is repackaged and included in the spring-core JAR. In other words, CGLIB-based AOP will work "out of the box" just as do JDK dynamic proxies.

There’s little performance difference between CGLIB proxying and dynamic proxies. As of Spring 1.0, dynamic proxies are slightly faster. However, this may change in the future. Performance should not be a decisive consideration in this case.

12.5.6 Using 'global' advisors

By appending an asterisk to an interceptor name, all advisors with bean names matching the part before the asterisk, will be added to the advisor chain. This can come in handy if you need to add a standard set of 'global' advisors:

 id="proxy" class="org.springframework.aop.framework.ProxyFactoryBean">
     name="target" ref="service"/>
     name="interceptorNames">
        
            global*
        
    


 id="global_debug" class="org.springframework.aop.interceptor.DebugInterceptor"/>
 id="global_performance" class="org.springframework.aop.interceptor.PerformanceMonitorInterceptor"/>

12.6 Concise proxy definitions

Especially when defining transactional proxies, you may end up with many similar proxy definitions. The use of parent and child bean definitions, along with inner bean definitions, can result in much cleaner and more concise proxy definitions.

First a parent, template, bean definition is created for the proxy:

 id="txProxyTemplate" abstract="true"
        class="org.springframework.transaction.interceptor.TransactionProxyFactoryBean">
     name="transactionManager" ref="transactionManager"/>
     name="transactionAttributes">
        
             key="*">PROPAGATION_REQUIRED
        
    

This will never be instantiated itself, so may actually be incomplete. Then each proxy which needs to be created is just a child bean definition, which wraps the target of the proxy as an inner bean definition, since the target will never be used on its own anyway.

 id="myService" parent="txProxyTemplate">
     name="target">
         class="org.springframework.samples.MyServiceImpl">
        
    

It is of course possible to override properties from the parent template, such as in this case, the transaction propagation settings:

 id="mySpecialService" parent="txProxyTemplate">
     name="target">
         class="org.springframework.samples.MySpecialServiceImpl">
        
    
     name="transactionAttributes">
        
             key="get*">PROPAGATION_REQUIRED,readOnly
             key="find*">PROPAGATION_REQUIRED,readOnly
             key="load*">PROPAGATION_REQUIRED,readOnly
             key="store*">PROPAGATION_REQUIRED
        
    

Note that in the example above, we have explicitly marked the parent bean definition as abstract by using the abstract attribute, as described previously, so that it may not actually ever be instantiated. Application contexts (but not simple bean factories) will by default pre-instantiate all singletons. It is therefore important (at least for singleton beans) that if you have a (parent) bean definition which you intend to use only as a template, and this definition specifies a class, you must make sure to set the abstractattribute to true, otherwise the application context will actually try to pre-instantiate it.

12.7 Creating AOP proxies programmatically with the ProxyFactory

It’s easy to create AOP proxies programmatically using Spring. This enables you to use Spring AOP without dependency on Spring IoC.

The following listing shows creation of a proxy for a target object, with one interceptor and one advisor. The interfaces implemented by the target object will automatically be proxied:

ProxyFactory factory = new ProxyFactory(myBusinessInterfaceImpl);
factory.addAdvice(myMethodInterceptor);
factory.addAdvisor(myAdvisor);
MyBusinessInterface tb = (MyBusinessInterface) factory.getProxy();

The first step is to construct an object of type org.springframework.aop.framework.ProxyFactory. You can create this with a target object, as in the above example, or specify the interfaces to be proxied in an alternate constructor.

You can add advices (with interceptors as a specialized kind of advice) and/or advisors, and manipulate them for the life of the ProxyFactory. If you add an IntroductionInterceptionAroundAdvisor, you can cause the proxy to implement additional interfaces.

There are also convenience methods on ProxyFactory (inherited from AdvisedSupport) which allow you to add other advice types such as before and throws advice. AdvisedSupport is the superclass of both ProxyFactory and ProxyFactoryBean.

	Integrating AOP proxy creation with the IoC framework is best practice in most applications. We recommend that you externalize configuration from Java code with AOP, as in general.

12.8 Manipulating advised objects

However you create AOP proxies, you can manipulate them using the org.springframework.aop.framework.Advised interface. Any AOP proxy can be cast to this interface, whichever other interfaces it implements. This interface includes the following methods:

Advisor[] getAdvisors();

void addAdvice(Advice advice) throws AopConfigException;

void addAdvice(int pos, Advice advice) throws AopConfigException;

void addAdvisor(Advisor advisor) throws AopConfigException;

void addAdvisor(int pos, Advisor advisor) throws AopConfigException;

int indexOf(Advisor advisor);

boolean removeAdvisor(Advisor advisor) throws AopConfigException;

void removeAdvisor(int index) throws AopConfigException;

boolean replaceAdvisor(Advisor a, Advisor b) throws AopConfigException;

boolean isFrozen();

The getAdvisors() method will return an Advisor for every advisor, interceptor or other advice type that has been added to the factory. If you added an Advisor, the returned advisor at this index will be the object that you added. If you added an interceptor or other advice type, Spring will have wrapped this in an advisor with a pointcut that always returns true. Thus if you added a MethodInterceptor, the advisor returned for this index will be an DefaultPointcutAdvisor returning your MethodInterceptor and a pointcut that matches all classes and methods.

The addAdvisor() methods can be used to add any Advisor. Usually the advisor holding pointcut and advice will be the generic DefaultPointcutAdvisor, which can be used with any advice or pointcut (but not for introductions).

By default, it’s possible to add or remove advisors or interceptors even once a proxy has been created. The only restriction is that it’s impossible to add or remove an introduction advisor, as existing proxies from the factory will not show the interface change. (You can obtain a new proxy from the factory to avoid this problem.)

A simple example of casting an AOP proxy to the Advised interface and examining and manipulating its advice:

Advised advised = (Advised) myObject;
Advisor[] advisors = advised.getAdvisors();
int oldAdvisorCount = advisors.length;
System.out.println(oldAdvisorCount + " advisors");

// Add an advice like an interceptor without a pointcut
// Will match all proxied methods
// Can use for interceptors, before, after returning or throws advice
advised.addAdvice(new DebugInterceptor());

// Add selective advice using a pointcut
advised.addAdvisor(new DefaultPointcutAdvisor(mySpecialPointcut, myAdvice));

assertEquals("Added two advisors", oldAdvisorCount + 2, advised.getAdvisors().length);

It’s questionable whether it’s advisable (no pun intended) to modify advice on a business object in production, although there are no doubt legitimate usage cases. However, it can be very useful in development: for example, in tests. I have sometimes found it very useful to be able to add test code in the form of an interceptor or other advice, getting inside a method invocation I want to test. (For example, the advice can get inside a transaction created for that method: for example, to run SQL to check that a database was correctly updated, before marking the transaction for roll back.)

Depending on how you created the proxy, you can usually set a frozen flag, in which case the Advised isFrozen() method will return true, and any attempts to modify advice through addition or removal will result in an AopConfigException. The ability to freeze the state of an advised object is useful in some cases, for example, to prevent calling code removing a security interceptor. It may also be used in Spring 1.1 to allow aggressive optimization if runtime advice modification is known not to be required.

12.9 Using the "auto-proxy" facility

So far we’ve considered explicit creation of AOP proxies using a ProxyFactoryBean or similar factory bean.

Spring also allows us to use "auto-proxy" bean definitions, which can automatically proxy selected bean definitions. This is built on Spring "bean post processor" infrastructure, which enables modification of any bean definition as the container loads.

In this model, you set up some special bean definitions in your XML bean definition file to configure the auto proxy infrastructure. This allows you just to declare the targets eligible for auto-proxying: you don’t need to use ProxyFactoryBean.

There are two ways to do this:

• Using an auto-proxy creator that refers to specific beans in the current context.
• A special case of auto-proxy creation that deserves to be considered separately; auto-proxy creation driven by source-level metadata attributes.

12.9.1 Autoproxy bean definitions

The org.springframework.aop.framework.autoproxy package provides the following standard auto-proxy creators.

BeanNameAutoProxyCreator

The BeanNameAutoProxyCreator class is a BeanPostProcessor that automatically creates AOP proxies for beans with names matching literal values or wildcards.

 class="org.springframework.aop.framework.autoproxy.BeanNameAutoProxyCreator">
     name="beanNames" value="jdk*,onlyJdk"/>
     name="interceptorNames">
        
            myInterceptor
        
    

As with ProxyFactoryBean, there is an interceptorNames property rather than a list of interceptors, to allow correct behavior for prototype advisors. Named "interceptors" can be advisors or any advice type.

As with auto proxying in general, the main point of using BeanNameAutoProxyCreator is to apply the same configuration consistently to multiple objects, with minimal volume of configuration. It is a popular choice for applying declarative transactions to multiple objects.

Bean definitions whose names match, such as "jdkMyBean" and "onlyJdk" in the above example, are plain old bean definitions with the target class. An AOP proxy will be created automatically by the BeanNameAutoProxyCreator. The same advice will be applied to all matching beans. Note that if advisors are used (rather than the interceptor in the above example), the pointcuts may apply differently to different beans.

DefaultAdvisorAutoProxyCreator

A more general and extremely powerful auto proxy creator is DefaultAdvisorAutoProxyCreator. This will automagically apply eligible advisors in the current context, without the need to include specific bean names in the auto-proxy advisor’s bean definition. It offers the same merit of consistent configuration and avoidance of duplication as BeanNameAutoProxyCreator.

Using this mechanism involves:

• Specifying a DefaultAdvisorAutoProxyCreator bean definition.
• Specifying any number of Advisors in the same or related contexts. Note that these must be Advisors, not just interceptors or other advices. This is necessary because there must be a pointcut to evaluate, to check the eligibility of each advice to candidate bean definitions.

The DefaultAdvisorAutoProxyCreator will automatically evaluate the pointcut contained in each advisor, to see what (if any) advice it should apply to each business object (such as "businessObject1" and "businessObject2" in the example).

This means that any number of advisors can be applied automatically to each business object. If no pointcut in any of the advisors matches any method in a business object, the object will not be proxied. As bean definitions are added for new business objects, they will automatically be proxied if necessary.

Autoproxying in general has the advantage of making it impossible for callers or dependencies to obtain an un-advised object. Calling getBean("businessObject1") on this ApplicationContext will return an AOP proxy, not the target business object. (The "inner bean" idiom shown earlier also offers this benefit.)

 class="org.springframework.aop.framework.autoproxy.DefaultAdvisorAutoProxyCreator"/>

 class="org.springframework.transaction.interceptor.TransactionAttributeSourceAdvisor">
     name="transactionInterceptor" ref="transactionInterceptor"/>


 id="customAdvisor" class="com.mycompany.MyAdvisor"/>

 id="businessObject1" class="com.mycompany.BusinessObject1">
    


 id="businessObject2" class="com.mycompany.BusinessObject2"/>

The DefaultAdvisorAutoProxyCreator is very useful if you want to apply the same advice consistently to many business objects. Once the infrastructure definitions are in place, you can simply add new business objects without including specific proxy configuration. You can also drop in additional aspects very easily - for example, tracing or performance monitoring aspects - with minimal change to configuration.

The DefaultAdvisorAutoProxyCreator offers support for filtering (using a naming convention so that only certain advisors are evaluated, allowing use of multiple, differently configured, AdvisorAutoProxyCreators in the same factory) and ordering. Advisors can implement the org.springframework.core.Ordered interface to ensure correct ordering if this is an issue. The TransactionAttributeSourceAdvisor used in the above example has a configurable order value; the default setting is unordered.

AbstractAdvisorAutoProxyCreator

This is the superclass of DefaultAdvisorAutoProxyCreator. You can create your own auto-proxy creators by subclassing this class, in the unlikely event that advisor definitions offer insufficient customization to the behavior of the framework DefaultAdvisorAutoProxyCreator.

12.9.2 Using metadata-driven auto-proxying

A particularly important type of auto-proxying is driven by metadata. This produces a similar programming model to .NET ServicedComponents. Instead of defining metadata in XML descriptors, configuration for transaction management and other enterprise services is held in source-level attributes.

In this case, you use the DefaultAdvisorAutoProxyCreator, in combination with Advisors that understand metadata attributes. The metadata specifics are held in the pointcut part of the candidate advisors, rather than in the auto-proxy creation class itself.

This is really a special case of the DefaultAdvisorAutoProxyCreator, but deserves consideration on its own. (The metadata-aware code is in the pointcuts contained in the advisors, not the AOP framework itself.)

The /attributes directory of the JPetStore sample application shows the use of attribute-driven auto-proxying. In this case, there’s no need to use theTransactionProxyFactoryBean. Simply defining transactional attributes on business objects is sufficient, because of the use of metadata-aware pointcuts. The bean definitions include the following code, in /WEB-INF/declarativeServices.xml. Note that this is generic, and can be used outside the JPetStore:

 class="org.springframework.aop.framework.autoproxy.DefaultAdvisorAutoProxyCreator"/>

 class="org.springframework.transaction.interceptor.TransactionAttributeSourceAdvisor">
     name="transactionInterceptor" ref="transactionInterceptor"/>


 id="transactionInterceptor"
        class="org.springframework.transaction.interceptor.TransactionInterceptor">
     name="transactionManager" ref="transactionManager"/>
     name="transactionAttributeSource">
         class="org.springframework.transaction.interceptor.AttributesTransactionAttributeSource">
             name="attributes" ref="attributes"/>
        
    


 id="attributes" class="org.springframework.metadata.commons.CommonsAttributes"/>

The DefaultAdvisorAutoProxyCreator bean definition (the name is not significant, hence it can even be omitted) will pick up all eligible pointcuts in the current application context. In this case, the "transactionAdvisor" bean definition, of type TransactionAttributeSourceAdvisor, will apply to classes or methods carrying a transaction attribute. The TransactionAttributeSourceAdvisor depends on a TransactionInterceptor, via constructor dependency. The example resolves this via autowiring. The AttributesTransactionAttributeSource depends on an implementation of the org.springframework.metadata.Attributes interface. In this fragment, the "attributes" bean satisfies this, using the Jakarta Commons Attributes API to obtain attribute information. (The application code must have been compiled using the Commons Attributes compilation task.)

The /annotation directory of the JPetStore sample application contains an analogous example for auto-proxying driven by JDK 1.5+ annotations. The following configuration enables automatic detection of Spring’s Transactional annotation, leading to implicit proxies for beans containing that annotation:

 class="org.springframework.aop.framework.autoproxy.DefaultAdvisorAutoProxyCreator"/>

 class="org.springframework.transaction.interceptor.TransactionAttributeSourceAdvisor">
     name="transactionInterceptor" ref="transactionInterceptor"/>


 id="transactionInterceptor"
        class="org.springframework.transaction.interceptor.TransactionInterceptor">
     name="transactionManager" ref="transactionManager"/>
     name="transactionAttributeSource">
         class="org.springframework.transaction.annotation.AnnotationTransactionAttributeSource"/>
    

The TransactionInterceptor defined here depends on a PlatformTransactionManager definition, which is not included in this generic file (although it could be) because it will be specific to the application’s transaction requirements (typically JTA, as in this example, or Hibernate, JDO or JDBC):

 id="transactionManager"
        class="org.springframework.transaction.jta.JtaTransactionManager"/>

	If you require only declarative transaction management, using these generic XML definitions will result in Spring automatically proxying all classes or methods with transaction attributes. You won’t need to work directly with AOP, and the programming model is similar to that of .NET ServicedComponents.

This mechanism is extensible. It’s possible to do auto-proxying based on custom attributes. You need to:

• Define your custom attribute.
• Specify an Advisor with the necessary advice, including a pointcut that is triggered by the presence of the custom attribute on a class or method. You may be able to use an existing advice, merely implementing a static pointcut that picks up the custom attribute.

It’s possible for such advisors to be unique to each advised class (for example, mixins): they simply need to be defined as prototype, rather than singleton, bean definitions. For example, the LockMixin introduction interceptor from the Spring test suite, shown above, could be used in conjunction with a generic DefaultIntroductionAdvisor:

 id="lockMixin" class="test.mixin.LockMixin" scope="prototype"/>

 id="lockableAdvisor" class="org.springframework.aop.support.DefaultIntroductionAdvisor"
        scope="prototype">
     ref="lockMixin"/>

Note that both lockMixin and lockableAdvisor are defined as prototypes.

12.10 Using TargetSources

Spring offers the concept of a TargetSource, expressed in the org.springframework.aop.TargetSource interface. This interface is responsible for returning the "target object" implementing the join point. The TargetSource implementation is asked for a target instance each time the AOP proxy handles a method invocation.

Developers using Spring AOP don’t normally need to work directly with TargetSources, but this provides a powerful means of supporting pooling, hot swappable and other sophisticated targets. For example, a pooling TargetSource can return a different target instance for each invocation, using a pool to manage instances.

If you do not specify a TargetSource, a default implementation is used that wraps a local object. The same target is returned for each invocation (as you would expect).

Let’s look at the standard target sources provided with Spring, and how you can use them.

	When using a custom target source, your target will usually need to be a prototype rather than a singleton bean definition. This allows Spring to create a new target instance when required.

12.10.1 Hot swappable target sources

The org.springframework.aop.target.HotSwappableTargetSource exists to allow the target of an AOP proxy to be switched while allowing callers to keep their references to it.

Changing the target source’s target takes effect immediately. The HotSwappableTargetSource is threadsafe.

You can change the target via the swap() method on HotSwappableTargetSource as follows:

HotSwappableTargetSource swapper = (HotSwappableTargetSource) beanFactory.getBean("swapper");
Object oldTarget = swapper.swap(newTarget);

The XML definitions required look as follows:

 id="initialTarget" class="mycompany.OldTarget"/>

 id="swapper" class="org.springframework.aop.target.HotSwappableTargetSource">
     ref="initialTarget"/>


 id="swappable" class="org.springframework.aop.framework.ProxyFactoryBean">
     name="targetSource" ref="swapper"/>

The above swap() call changes the target of the swappable bean. Clients who hold a reference to that bean will be unaware of the change, but will immediately start hitting the new target.

Although this example doesn’t add any advice - and it’s not necessary to add advice to use a TargetSource - of course any TargetSource can be used in conjunction with arbitrary advice.

12.10.2 Pooling target sources

Using a pooling target source provides a similar programming model to stateless session EJBs, in which a pool of identical instances is maintained, with method invocations going to free objects in the pool.

A crucial difference between Spring pooling and SLSB pooling is that Spring pooling can be applied to any POJO. As with Spring in general, this service can be applied in a non-invasive way.

Spring provides out-of-the-box support for Commons Pool 2.2, which provides a fairly efficient pooling implementation. You’ll need the commons-pool Jar on your application’s classpath to use this feature. It’s also possible to subclass org.springframework.aop.target.AbstractPoolingTargetSource to support any other pooling API.

	Commons Pool 1.5+ is also supported but deprecated as of Spring Framework 4.2.

Sample configuration is shown below:

 id="businessObjectTarget" class="com.mycompany.MyBusinessObject"
        scope="prototype">
    ... properties omitted


 id="poolTargetSource" class="org.springframework.aop.target.CommonsPool2TargetSource">
     name="targetBeanName" value="businessObjectTarget"/>
     name="maxSize" value="25"/>


 id="businessObject" class="org.springframework.aop.framework.ProxyFactoryBean">
     name="targetSource" ref="poolTargetSource"/>
     name="interceptorNames" value="myInterceptor"/>

Note that the target object - "businessObjectTarget" in the example - must be a prototype. This allows the PoolingTargetSource implementation to create new instances of the target to grow the pool as necessary. See the javadocs of AbstractPoolingTargetSource and the concrete subclass you wish to use for information about its properties: "maxSize" is the most basic, and always guaranteed to be present.

In this case, "myInterceptor" is the name of an interceptor that would need to be defined in the same IoC context. However, it isn’t necessary to specify interceptors to use pooling. If you want only pooling, and no other advice, don’t set the interceptorNames property at all.

It’s possible to configure Spring so as to be able to cast any pooled object to the org.springframework.aop.target.PoolingConfig interface, which exposes information about the configuration and current size of the pool through an introduction. You’ll need to define an advisor like this:

 id="poolConfigAdvisor" class="org.springframework.beans.factory.config.MethodInvokingFactoryBean">
     name="targetObject" ref="poolTargetSource"/>
     name="targetMethod" value="getPoolingConfigMixin"/>

This advisor is obtained by calling a convenience method on the AbstractPoolingTargetSource class, hence the use of MethodInvokingFactoryBean. This advisor’s name ("poolConfigAdvisor" here) must be in the list of interceptors names in the ProxyFactoryBean exposing the pooled object.

The cast will look as follows:

PoolingConfig conf = (PoolingConfig) beanFactory.getBean("businessObject");
System.out.println("Max pool size is " + conf.getMaxSize());

	Pooling stateless service objects is not usually necessary. We don’t believe it should be the default choice, as most stateless objects are naturally thread safe, and instance pooling is problematic if resources are cached.

Simpler pooling is available using auto-proxying. It’s possible to set the TargetSources used by any auto-proxy creator.

12.10.3 Prototype target sources

Setting up a "prototype" target source is similar to a pooling TargetSource. In this case, a new instance of the target will be created on every method invocation. Although the cost of creating a new object isn’t high in a modern JVM, the cost of wiring up the new object (satisfying its IoC dependencies) may be more expensive. Thus you shouldn’t use this approach without very good reason.

To do this, you could modify the poolTargetSource definition shown above as follows. (I’ve also changed the name, for clarity.)

 id="prototypeTargetSource" class="org.springframework.aop.target.PrototypeTargetSource">
     name="targetBeanName" ref="businessObjectTarget"/>

There’s only one property: the name of the target bean. Inheritance is used in the TargetSource implementations to ensure consistent naming. As with the pooling target source, the target bean must be a prototype bean definition.

12.10.4 ThreadLocal target sources

ThreadLocal target sources are useful if you need an object to be created for each incoming request (per thread that is). The concept of a ThreadLocal provide a JDK-wide facility to transparently store resource alongside a thread. Setting up a ThreadLocalTargetSource is pretty much the same as was explained for the other types of target source:

 id="threadlocalTargetSource" class="org.springframework.aop.target.ThreadLocalTargetSource">
     name="targetBeanName" value="businessObjectTarget"/>

ThreadLocals come with serious issues (potentially resulting in memory leaks) when incorrectly using them in a multi-threaded and multi-classloader environments. One should always consider wrapping a threadlocal in some other class and never directly use the ThreadLocal itself (except of course in the wrapper class). Also, one should always remember to correctly set and unset (where the latter simply involved a call to ThreadLocal.set(null)) the resource local to the thread. Unsetting should be done in any case since not unsetting it might result in problematic behavior. Spring’s ThreadLocal support does this for you and should always be considered in favor of using ThreadLocals without other proper handling code.

12.11 Defining new Advice types

Spring AOP is designed to be extensible. While the interception implementation strategy is presently used internally, it is possible to support arbitrary advice types in addition to the out-of-the-box interception around advice, before, throws advice and after returning advice.

The org.springframework.aop.framework.adapter package is an SPI package allowing support for new custom advice types to be added without changing the core framework. The only constraint on a custom Advice type is that it must implement the org.aopalliance.aop.Advice marker interface.

Please refer to the org.springframework.aop.framework.adapter javadocs for further information.

12.12 Further resources

Please refer to the Spring sample applications for further examples of Spring AOP:

• The JPetStore’s default configuration illustrates the use of the TransactionProxyFactoryBean for declarative transaction management.
• The /attributes directory of the JPetStore illustrates the use of attribute-driven declarative transaction management.

Part IV. Testing