一文搞懂JMeter engine中HashTree的配置问题_Java教程

一、前言

之前介绍了JMeter engine启动原理，但是里面涉及到HashTree这个类结构没有给大家详细介绍，这边文章就详细介绍JMeter engine里面的HashTree结构具体用来做什么
大家看到下面是JMeter控制台配置截图，是一个标准的菜单形式；菜单形式其实就类似于“树型”的数据结构，而HashTree其实就是一个树型数据结构

一文搞懂JMeter engine中HashTree的配置问题

我们在JMeter控制台导出的jmx文件，是一个xml结构的数据，他其实就是由HashTree生成的，后面我们会讲到

一文搞懂JMeter engine中HashTree的配置问题

二、HashTree的用法

首先通过HashTree类介绍，它一个集合类；具备Map结构的功能，而且是一种树型结构

/**
* This class is used to create a tree structure of objects. Each element in the
* tree is also a key to the next node down in the tree. It provides many ways
* to add objects and branches, as well as many ways to retrieve.
* <p>
* HashTree implements the Map interface for convenience reasons. The main
* difference between a Map and a HashTree is that the HashTree organizes the
* data into a recursive tree structure, and provides the means to manipulate
* that structure.
* <p>
* Of special interest is the {@link #traverse(HashTreeTraverser)} method, which
* provides an expedient way to traverse any HashTree by implementing the
* {@link HashTreeTraverser} interface in order to perform some operation on the
* tree, or to extract information from the tree.
*
* @see HashTreeTraverser
* @see SearchByClass
*/
public class HashTree implements Serializable, Map<Object, HashTree>, Cloneable {
}

JMeter常用的HashTree方法（以下图配置为例）

一文搞懂JMeter engine中HashTree的配置问题

//ListedHashTree是HashTree的继承类，可以保证HashTree的顺序性
HashTree tree = new ListedHashTree();
//TestPlan对象，测试计划
TestPlan plan = new TestPlan();
//ThreadGroup对象，线程组
ThreadGroup group = new ThreadGroup();
//创建线程组数结构的对象groupTree
HashTree groupTree = new ListedHashTree();
//表示取样器中的HTTP请求
HTTPSamplerProxy sampler = new HTTPSamplerProxy();
//创建HTTP请求的数结构对象samplerTree
//调用put方法相当于在plan（测试计划）菜单对象下添加group（线程组）子菜单，这样就形成了一种树型结构
HashTree samplerTree = new ListedHashTree();
samplerTree.put(sampler,new ListedHashTree())
//groupTree树结构添加子树samplerTree
groupTree.put(group,samplerTree)
//tree树结构为测试计划对象，添加子树groupTree，这样就形成了上图的层级形式
tree.put(plan, groupTree)
//调用add方法相当于在tree菜单对象下添加同级菜单
tree.add(Object key)

三、JMeter源码导出jmx脚本文件介绍

首先在JMeter控制台所有点击事件，都会被ActionRouter中performaAction方法进行监听执行，点击导出按钮，会进入到如图方法通过反射由Save类执行

一文搞懂JMeter engine中HashTree的配置问题

在Save类中执行doAction主要是获取到配置的HashTree

一文搞懂JMeter engine中HashTree的配置问题

当你点击保存的时候，它会创建一个空文件，此时文件没有任何内容

一文搞懂JMeter engine中HashTree的配置问题 Save

类的doAction方法最后会调用backupAndSave(e, subTree, fullSave, updateFile)这个是来将创建的空文件写入xml内容的

一文搞懂JMeter engine中HashTree的配置问题

在SaveService中saveTree方法，其中JMXSAVER是XStream对象，对应的maven坐标如下

<!-- https://mvnrepository.com/artifact/com.thoughtworks.xstream/xstream -->
<dependency>
  <groupId>com.thoughtworks.xstream</groupId>
  <artifactId>xstream</artifactId>
  <version>1.4.15</version>
</dependency>

一文搞懂JMeter engine中HashTree的配置问题

四、自定义HashTree生成JMeter脚本

首先maven引入以下几个坐标<jmeter.version>5.3</jmeter.version>

      <dependency>
          <groupId>org.apache.jmeter</groupId>
          <artifactId>ApacheJMeter_http</artifactId>
          <version>${jmeter.version}</version>
          <exclusions>
              <exclusion>
                  <groupId>org.apache.logging.log4j</groupId>
                  <artifactId>log4j-slf4j-impl</artifactId>
              </exclusion>
          </exclusions>
      </dependency>
      <dependency>
          <groupId>org.apache.jmeter</groupId>
          <artifactId>ApacheJMeter_functions</artifactId>
          <version>${jmeter.version}</version>
      </dependency>
      <dependency>
          <groupId>org.apache.jmeter</groupId>
          <artifactId>ApacheJMeter_jdbc</artifactId>
          <version>${jmeter.version}</version>
      </dependency>

      <dependency>
          <groupId>org.apache.jmeter</groupId>
          <artifactId>ApacheJMeter_tcp</artifactId>
          <version>${jmeter.version}</version>
      </dependency>

先创建一个取样器，然后写成HashTree的数据结构

public static ThreadGroup threadGroup;
//创建一个标准的线程组
private static void initThreadGroup(){
  LoopController loopController = new LoopController();
  loopController.setName("LoopController");
  loopController.setProperty(TestElement.TEST_CLASS, LoopController.class.getName());
  loopController.setProperty(TestElement.GUI_CLASS, JMeterUtil.readSaveProperties("LoopControlPanel"));
  loopController.setEnabled(true);
  loopController.setLoops(1);
  ThreadGroup group = new ThreadGroup();
  group.setEnabled(true);
  group.setName("ThreadGroup");
  group.setProperty(TestElement.TEST_CLASS, JMeterUtil.readSaveProperties("ThreadGroup"));
  group.setProperty(TestElement.GUI_CLASS, JMeterUtil.readSaveProperties("ThreadGroupGui"));
  group.setProperty(ThreadGroup.ON_SAMPLE_ERROR,"continue");
  group.setProperty(ThreadGroup.IS_SAME_USER_ON_NEXT_ITERATION,true);
  group.setProperty(TestElement.COMMENTS,"");
  group.setNumThreads(1);
  group.setRampUp(1);
  group.setDelay(0);
  group.setDuration(0);
  group.setProperty(ThreadGroup.ON_SAMPLE_ERROR, ThreadGroup.ON_SAMPLE_ERROR_CONTINUE);
  group.setScheduler(false);
  group.setSamplerController(loopController);
  threadGroup = group;
}

创建一个标准的线程组

public static ThreadGroup threadGroup;
//创建一个标准的线程组
private static void initThreadGroup(){
  LoopController loopController = new LoopController();
  loopController.setName("LoopController");
  loopController.setProperty(TestElement.TEST_CLASS, LoopController.class.getName());
  loopController.setProperty(TestElement.GUI_CLASS, JMeterUtil.readSaveProperties("LoopControlPanel"));
  loopController.setEnabled(true);
  loopController.setLoops(1);
  ThreadGroup group = new ThreadGroup();
  group.setEnabled(true);
  group.setName("ThreadGroup");
  group.setProperty(TestElement.TEST_CLASS, JMeterUtil.readSaveProperties("ThreadGroup"));
  group.setProperty(TestElement.GUI_CLASS, JMeterUtil.readSaveProperties("ThreadGroupGui"));
  group.setProperty(ThreadGroup.ON_SAMPLE_ERROR,"continue");
  group.setProperty(ThreadGroup.IS_SAME_USER_ON_NEXT_ITERATION,true);
  group.setProperty(TestElement.COMMENTS,"");
  group.setNumThreads(1);
  group.setRampUp(1);
  group.setDelay(0);
  group.setDuration(0);
  group.setProperty(ThreadGroup.ON_SAMPLE_ERROR, ThreadGroup.ON_SAMPLE_ERROR_CONTINUE);
  group.setScheduler(false);
  group.setSamplerController(loopController);
  threadGroup = group;
}

创建一个标准的测试计划

public static TestPlan testPlan;
//创建一个标准的测试计划
private static void initTestPlan() {
  TestPlan plan = new TestPlan();
  //设置测试计划属性及内容，最后都会转为xml标签的属性及内容
  plan.setProperty(TestElement.NAME, "测试计划");
  plan.setProperty(TestElement.TEST_CLASS, JMeterUtil.readSaveProperties("TestPlan"));
  plan.setProperty(TestElement.GUI_CLASS, JMeterUtil.readSaveProperties("TestPlanGui"));
  plan.setEnabled(true);
  plan.setComment("");
  plan.setFunctionalMode(false);
  plan.setTearDownOnShutdown(true);
  plan.setSerialized(false);
  plan.setProperty("TestPlan.user_define_classpath","");
  plan.setProperty("TestPlan.user_defined_variables","");
  plan.setUserDefinedVariables(new Arguments());
  testPlan = plan;
}

开始封装成一个HashTree的配置

//先创建一个测试计划hashtree对象
HashTree hashTree = new ListedHashTree();
//在创建一个线程组threaddGroupTree对象
HashTree threadGroupTree = new ListedHashTree();
//HttpRequestConfig为HTTP对应的请求头、请求体等信息数据，传入httpToHashTree静态方法获取到取样器的HashTree数据结构，源码上图已分享
HashTree httpConfigTree = XXClass.httpToHashTree(HttpRequestConfig httpRequestData)
//threadGroupTree添加子菜单httpConfigTree对象
threadGroupTree.put(group, httpConfigTree);
//测试计划hashTree添加子菜单threadGroupTree对象
hashTree.put(JMeterTestPlanConfigService.testPlan, threadGroupTree);

HashTree写好后，调用JMeter原生方法SaveService.saveTree(hashTree,outStream);生成对应的xml

如果直接调用的话生成的xml格式会形成如下图所示，而非JMeter原生导出jmx形式，这种文件结构JMeter控制台读取会报错，识别不了

一文搞懂JMeter engine中HashTree的配置问题

后面阅读SaveService源码才明白，生成xml文件之前会先初始化静态代码块内容，初始化属性

一文搞懂JMeter engine中HashTree的配置问题

过程中会调用JMeterUtils中的findFile方法来寻找saveservice.properties文件

一文搞懂JMeter engine中HashTree的配置问题

由于SaveService 中都是静态方法无法重写，所以根据最后调用JMeterUtils中的findFile方法来寻找saveservice.properties有两种解决方案

方案一：不推荐，在项目根目录下存放saveservice.properties，这样findFile方法就能拿到，但是这样不好，因为maven打包的时候该文件会打不进去，至少我springboot项目是遇到这样的问题

方案二：推荐，创建一个临时文件命名为saveservice.properties，然后提前将saveservice.properties配置读取到临时文件中，这样在调用JMeterUtils中的findFile方法同样能够找到配置，成功解决SaveService初始化属性导致的问题，具体代码如下

private void hashTreeToXML(HashTree hashTree,PressureConfigInfo configInfo){
  FileOutputStream outStream = null;
  File file = new File("temp.jmx");
  File tempFile = null;
  try {
  	//创建一个临时的saveservice.properties文件
      tempFile = new File("saveservice.properties");
      InputStream is = JMeterUtil.class.getResource("/jmeter/saveservice.properties").openStream();
      //将配置文件写入临时文件中
      FileUtil.writeFromStream(is,tempFile);
      outStream = new FileOutputStream(file);
      //调用saveTree成功转为xml
      SaveService.saveTree(hashTree,outStream);
      String xmlContent = FileUtil.readUtf8String(file);
      configInfo.setFile(xmlContent.getBytes());
  } catch (IOException e) {
      e.printStackTrace();
  }finally {
      try {
          FileUtils.forceDelete(file);
          FileUtils.forceDelete(tempFile);
      } catch (IOException e) {
          e.printStackTrace();
      }
  }
}

最后生成的xml文件结构如下图，通过JMeter控制台也能成功打开识别

一文搞懂JMeter engine中HashTree的配置问题