使用AspectJ Compiler解决Spring @Transactional方法自调用失效

最新推荐文章于 2024-10-11 00:35:17 发布

转载最新推荐文章于 2024-10-11 00:35:17 发布 · 1.2k 阅读

1 ·

CC 4.0 BY-SA版权

原文链接：https://my.oschina.net/yqz/blog/1603556

文章标签：

#java

本文介绍如何通过配置Spring使用AspectJ来为私有方法提供事务管理功能，包括配置步骤及如何启用AspectJ编译器，并展示了AspectJ生成的代码示例。

为什么80%的码农都做不了架构师？>>>

If we want to use the @Transactional annotation on private methods, we cannot use Spring’s default proxy mode - it only works for public methods that are called from client code.

For self-invoked methods we need to use AspectJ.

An Example

This may be the case e.g. when we want to save incoming data and send a letter using that data within a transaction. If saving the data fails we do not want to send the letter. If sending the letter fails we do not want to loose the data that we received. Hence, we want to save the data in a separate transaction, by annotating our saveData(..) method with @Transactional(propagation = Propagation.REQUIRES_NEW).

Configuring Spring To Use AspectJ

On a @Configuration class we just need to set the AspectJ advice mode for the transaction management configuration:

@Configuration
@ComponentScan
@EnableTransactionManagement(mode = AdviceMode.ASPECTJ)
class ApplicationConfiguration {
    // ...
}

Configuring Maven To Run The AspectJ Compiler

Next we need to use the AspectJ compiler, so let’s add its needed dependencies:

<dependencies>
    <dependency>
        <groupId>org.aspectj</groupId>
        <artifactId>aspectjrt</artifactId>
        <version>1.7.4</version>
    </dependency>

    <!-- In this example we use JPA for persistence. -->
    <dependency>
        <groupId>org.hibernate.javax.persistence</groupId>
        <artifactId>hibernate-jpa-2.1-api</artifactId>
        <version>1.0.0.Final</version>
    </dependency>

    <dependency>
        <groupId>org.springframework</groupId>
        <artifactId>spring-aspects</artifactId>
        <version>4.0.6.RELEASE</version>
    </dependency>
</dependencies>

The AspectJ compiler itself can then be enabled with:

<build>
    <plugins>
        <plugin>
            <groupId>org.codehaus.mojo</groupId>
            <artifactId>aspectj-maven-plugin</artifactId>
            <version>1.6</version>
            <executions>
                <execution>
                    <goals>
                        <goal>compile</goal>
                        <!-- <goal>test-compile</goal> -->
                    </goals>
                </execution>
            </executions>
            <configuration>
                <complianceLevel>1.7</complianceLevel>
                <forceAjcCompile>true</forceAjcCompile>
                <weaveDirectories>
                    <weaveDirectory>${project.build.outputDirectory}</weaveDirectory>
                </weaveDirectories>
                <aspectLibraries>
                    <aspectLibrary>
                        <groupId>org.springframework</groupId>
                        <artifactId>spring-aspects</artifactId>
                    </aspectLibrary>
                </aspectLibraries>
            </configuration>
        </plugin>
    </plugin>
</build>

If we do not use <forceAjcCompile> and do not set the <weaveDirectory> AspectJ will not weave the transaction code into our generated Groovy bytecode. Instead, it will warn us with No sources found skipping aspectJ compile when Maven runs. That is as the aspectj-maven-plugin processes the src/main/java directory by default but not src/main/groovy.

How The AspectJ Generated Code Looks Like

Just for completeness, let’s look how AspectJ has wrapped the saveData(..) method with a transaction. Using the Procyon decompiler we can see:

@Transactional(propagation = Propagation.REQUIRES_NEW)
public void saveData(final SomeData data) {
    ((AbstractTransactionAspect)AnnotationTransactionAspect.aspectOf()).ajc$around$org_springframework_transaction_aspectj_AbstractTransactionAspect$1$2a73e96c((Object)this, (AroundClosure)new DataService$AjcClosure3(new Object[] { this, data }), DataService.ajc$tjp_1);
}

转载于:https://my.oschina.net/yqz/blog/1603556