In this post, we will see how does one make use of TestNG to kick off parallel UI tests using WebDriver.

So lets try doing this with a typical cooking recipe style :)

So here are the ingredients that are required.

  • A Factory class that will create WebDriver instances
  • A Manager class that can be accessed to retrieve a WebDriver instance
  • A TestNG listener that will be responsible for instantiating the WebDriver instance automatically

So without wasting any time lets see how this all blends in.

First lets look at our Factory class. This is a very simplified Factory class that will create instances of WebDriver based upon the browser flavour. I have purposefully kept it simple only for illustration purposes:
Here’s how the Factory class will look like:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
package organized.chaos;
 
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.chrome.ChromeDriver;
import org.openqa.selenium.firefox.FirefoxDriver;
import org.openqa.selenium.ie.InternetExplorerDriver;
 
class LocalDriverFactory {
    static WebDriver createInstance(String browserName) {
        WebDriver driver = null;
        if (browserName.toLowerCase().contains("firefox")) {
            driver = new FirefoxDriver();
            return driver;
        }
        if (browserName.toLowerCase().contains("internet")) {
            driver = new InternetExplorerDriver();
            return driver;
        }
        if (browserName.toLowerCase().contains("chrome")) {
            driver = new ChromeDriver();
            return driver;
        }
        return driver;
    }
}

As you can see its a very simple class with a static method that creates WebDriver instances. The one interesting part to be noted here is that the class has been purposefully given only package visibility [ notice how the keyword "public" is missing from the class declaration ]. One of the many aspects that are involved in designing APIs is “Hide what is not necessary to be visible to your user”. For you to be able to drive a car, you don’t need to know how the piston works or how the fuel injection happens do you :)

Now lets take a look at how our Manager class would look like. The Manager class essentially uses a concept in java called ThreadLocalvariables.

The code would look like below :

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
package organized.chaos;
 
import org.openqa.selenium.WebDriver;
 
public class LocalDriverManager {
    private static ThreadLocal<WebDriver> webDriver = new ThreadLocal<WebDriver>();
 
    public static WebDriver getDriver() {
        return webDriver.get();
    }
 
    static void setWebDriver(WebDriver driver) {
        webDriver.set(driver);
    }
}

Were you surprised that its such a small class ? :)
So as you can see we basically have a static ThreadLocal variable wherein we are setting webDriver instances and also querying webdriver instances as well.

Next comes the TestNG listener. The role of the TestNG listener is to perform “Automatic webdriver instantiation” behind the scenes without your test code even realising it. For this we will make use of IInvokedMethodListener so that the WebDriver gets instantiated right before a Test Method gets invoked and the webDriver gets automatically quit right after the Test method.
You can improvize this by incorporating custom annotations as well and parsing for your custom annotations [ The current implementation that you will see basically spawns a browser irrespective of whether you want to use it or not. That's not a nice idea all the time is it ]

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
package organized.chaos;
 
import org.openqa.selenium.WebDriver;
import org.testng.IInvokedMethod;
import org.testng.IInvokedMethodListener;
import org.testng.ITestResult;
 
public class WebDriverListener implements IInvokedMethodListener {
 
    @Override
    public void beforeInvocation(IInvokedMethod method, ITestResult testResult) {
        if (method.isTestMethod()) {
            String browserName = method.getTestMethod().getXmlTest().getLocalParameters().get("browserName");
            WebDriver driver = LocalDriverFactory.createInstance(browserName);
            LocalDriverManager.setWebDriver(driver);
        }
    }
 
    @Override
    public void afterInvocation(IInvokedMethod method, ITestResult testResult) {
        if (method.isTestMethod()) {
            WebDriver driver = LocalDriverManager.getDriver();
            if (driver != null) {
                driver.quit();
            }
        }
    }
}

Now that we have shown all of the ingredients, lets take a look at a sample test as well, which is going to use all of this.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
package organized.chaos;
 
import org.testng.annotations.Test;
 
public class ThreadLocalDemo {
    @Test
    public void testMethod1() {
        invokeBrowser("http://www.ndtv.com");
    }
 
    @Test
    public void testMethod2() {
        invokeBrowser("http://www.facebook.com");
 
    }
 
    private void invokeBrowser(String url) {
        System.out.println("Thread id = " + Thread.currentThread().getId());
        System.out.println("Hashcode of webDriver instance = " + LocalDriverManager.getDriver().hashCode());
        LocalDriverManager.getDriver().get(url);
 
    }
}

As you can see its a very simple test class with two test methods. Each of the test methods opens up a different website. I have also add print statements for printing the thread id [yes thats the only reliable way of figuring out if your test method is running in parallel or in sequential mode. If you see unique values for Thread.currentThread().getId() then you can rest assured that TestNG is invoking your test methods in parallel.
We are printing the hashCode() values for the browser to demonstrate the fact that there are unique and different webDriver instances being created for every test method. [ Remember hashCode() value for an object would always be unique ]

Now lets take a look at how our suite file looks like :

1
2
3
4
5
6
7
8
9
10
11
12
13
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE suite SYSTEM "http://testng.org/testng-1.0.dtd">
<suite name="Suite" parallel="methods">
<listeners>
<listener class-name="organized.chaos.WebDriverListener"></listener>
</listeners>
    <test name="Test">
        <parameter name="browserName" value="firefox"></parameter>
        <classes>
            <class name="organized.chaos.ThreadLocalDemo" />
        </classes>
    </test> <!-- Test -->
</suite> <!-- Suite -->

So when you run this test this is how your output would look like [ apart from you seeing two firefox windows popup on your desktop ]

1
2
3
4
5
6
7
8
9
10
11
12
[TestNG] Running:
  /githome/PlayGround/testbed/src/test/resources/threadLocalDem.xml
 
Thread id = 10
Hashcode of webDriver instance = 1921042184
Thread id = 9
Hashcode of webDriver instance = 2017986718
 
===============================================
Suite
Total tests run: 2, Failures: 0, Skips: 0
===============================================

And thus we have managed to leverage TestNG and run WebDriver tests in parallel without having to worry about race conditions or leaving browsers open etc.,

Hope that clears out some of the confusions and helps you get started with WebDriver automation powered by TestNG.

Parallel WebDriver executions using TestNG的更多相关文章

  1. Selenium2(webdriver)入门之TestNG的使用

    一.在Eclipse中安装TestNG 1.打开eclipse-->help-->Install New Software-->Add,输入Name和Location后,点击OK. ...

  2. Selenium2(webdriver)入门之TestNG的安装与简单使用

    上一篇已经搭建好了Eclipse+selenium2的环境,这一篇主要记录下TestNG的使用. 一.在Eclipse中安装TestNG 1.打开eclipse-->help-->Inst ...

  3. 基于WebDriver&TestNG 实现自己的Annotation @TakeScreenshotOnFailure

    相信用过Selenium WebDriver 的朋友都应该知道如何使用WebDriver API实现Take Screenshot的功能. 在这篇文章里,我主要来介绍对failed tests实现 t ...

  4. testng 失败自动截图

    testng执行case failed ,testng Listener会捕获执行失败,如果要实现失败自动截图,需要重写Listener的onTestFailure方法 那么首先新建一个Listene ...

  5. WebDriver - 添加失败截图

    WebDriver失败截图可以通过两种方式实现: 1. Use WebdriverEventListener 第一步:创建自己的WebDriverEventListener 创建自己的WebDrive ...

  6. TestNG实现日志输出

    这里介绍的是TestNG中的Report类来实现简单的log输出这个很简单直接看例子吧 package com.rrx.test; import java.io.IOException; import ...

  7. TestNG实现用例运行失败自动截图(转载)

    转载自:https://blog.csdn.net/galen2016/article/details/70193684 重写Listener的onTestFailure方法 package com. ...

  8. testng失败截图,注解方式调用。

    今天一整天都在研究testng失败截图的方法,参考网上的前辈们的资料,加上自己的理解,终于搞出来了. package com.dengnapianhuahai; /** * 自定义注释 * */ im ...

  9. maven+selenium+java+testng+jenkins自动化测试

    最近在公司搭建了一套基于maven+selenium+java+testng+jenkins的自动化测试框架,免得以后重写记录下 工程目录 pom.xml <project xmlns=&quo ...

随机推荐

  1. String对象中常用的方法

    String对象中常用的方法   1.charCodeAt方法返回一个整数,代表指定位置字符的Unicode编码.strObj.charCodeAt(index)说明:index将被处理字符的从零开始 ...

  2. php用户注册

    前言 网站用户注册与登录是很常用的一个功能,本节教材就以此来演示一下 PHP 中如何开发用户注册与登录模块. 本节需要用到的重点 PHP 基础知识: PHP 中预定义 $_POST 和 $_GET 全 ...

  3. DB天气app冲刺二阶段第七天

    又冲刺了一个礼拜了 今天收获应该算是不小了 虽然进度上来说还是一点也没前进 但是找到了好几个突破口 明天继续 今天先不多说了困了..

  4. JSTL标签总结

    一.JSTL简介: 1.JSP标准标签库JSTL(JSP Standard Tag Library)是一个JSP标签集合,它封装了JSP应用的通用核心功能. 2.JSTL支持通用的.结构化的任务.比如 ...

  5. js eval()执行传参函数的写法

    .cs public class Message<T> { // 数据总数 public int? Total { get; set; } // 关键数据 public List<T ...

  6. html利用锚点实现定位代码实例

    本章节介绍介绍一下如何利用锚点实现定位,使用锚点实现定位是html固有的功能,当然比较简单,也实现了基本的功能,但是功能相对简单一些,如果想要实现平滑的定位可以参阅jquery实现的点击页面动画方式平 ...

  7. 【BZOJ】【2253】【WC 2010 BeijingWC】纸箱堆叠

    树套树 Orz zyf 我的树套树不知道为啥一直WA……只好copy了zyf的写法TAT 这题还可以用CDQ分治来做……但是蒟蒻不会…… //y坐标的树状数组是按权值建的……所以需要离散化…… /** ...

  8. js 判断是否为chrome浏览器

    var isChrome =navigator.userAgent.indexOf("Chrome") !== -1 用 navigator.appVersion 不好使,因为al ...

  9. oracle 笔记

    1.Oracle认证,与其它数据库比较,安装 Oracle安装会自动的生成sys用户和system用户: (1)sys用户是超级用户,具有最高权限,具有sysdba角色,有create databas ...

  10. Winform 窗体的操作

    原文:http://www.cnblogs.com/Billy-rao/archive/2012/05/16/2503437.html 怎样能使winform窗体的大小固定住,不能调整其大小 窗体Fo ...