AUTOMATING WITH SELENIUM 2 and the PageObject Design Model SELENIUM YOU CAN’T BE GOOD AT EVERYTHING How Selenium Remote Control works You launch a server on your test machine. Your tests connect to that server via IP. The server launches a browser, with selenium CORE embedded as javascript into the page. Selenium CORE simulates user actions with javascript. THE JAVASCRIPT SANDBOX THE GOOD Doesn’t steal your mouse/keyboard. Works with any browser that uses javascript Works for any OS that supports java. Very fast page interactions. Large API Supports a variety of programming languages. Can be run on remote machines THE BAD Can’t see anything outside the page object. Not all browsers support all functionality. Some pages aren’t automatable. Can’t see inside third party apps XSS Limitations WEBDRIVER INTRODUCTION A different way of automating the browser. Create a browser-specific driver to control the browser directly. Have to do this for each browser! Object oriented API Doesn’t need a real browser No javascript limitations No need for a server. Isn’t as delicate as selenium. SELENIUM 2 Went into Beta Dec 24th. WebDriver + Selenium The two projects have merged (literally) into Selenium 2.0 Large browser support and no javascript limitations. No server The old API’s are still available. New API’s are becoming available. AUTOMATING WITH SELENIUM 2 You have 2 options: IWebDriver This is just the WebDriver api Doesn’t support a lot of browsers. Will need to change all your tests. WebDriverBackedSelenium Uses the old Selenium 1 API But uses WebDriver to run things if possible Can use selenium 1 if the browser isn’t supported. THE WEBDRIVER API Object Oriented Doesn’t break nearly as often Handles pageloads automatically Fewer problems automating Somewhat more complicated API selenium.type(“password”,”thisIsMyPassword”); driver.findElement(By.id("password")).sendKeys(“thisIsMyPassword"); By.Id, By.Xpath, By.Name, By.ClassName, By.PartialLinkText All the supported browsers work really well Can extend the API to add custom functionality. Works well with a Page Object design model. PAGE OBJECT DESIGN PATTERN Adds a layer of abstraction into your code. Helps to organize your code once it grows large. All automation is automatically reusable and shareable. A way to separate tests from re-usable functions. A way to store information about how the system works. A way to specify what page functions start on, and what page they end on. A way to programmatically break your tests when functionality changes. Makes code maintenance easier. There is even a PageFactory class available to automatically create them. HOW DOES IT WORK? Each page is defined as it’s own class. Actions (including navigation) are represented as functions for a class. Each function returns a new Page object, signifying what page the actions stops on. Your tests “know” what page you are on, and will only give you access to functions available to that class. Tests only talk to the page objects. Page objects only talk to the driver. Elements on the page are stored as variables for the page object. Automatic page validations can be stored in the constructor for each page object. Tests become a string of well defined functions, not meaningless gibberish. Tests can be grouped by namespace. Class Inheritance can be used to define functionality to a set of pages. We can make functional logic transparent to the tests by returning different inherited classes. BAD TESTS globalVars.logDescription = "log in"; globalVars.selenium.Open(globalVars.mobiUrl); functions.type("userId", globalVars.studName + "1"); functions.type("password", globalVars.studPass + "1"); functions.clickAndWait("signInBtn"); selenium.click("//a[@id='discussions']/span"); selenium.click("//a[@id='thingsToKnow']/span"); globalVars.logDescription = "Checking elements on happenings:by date page"; selenium.waitForElementNotVisible("//div[@id='THSContainer']//span[@class='ajaxLoadingHeader']"); selenium.waitForElementVisible("//div[@id='THSContainer']/ul[1]/li[1]"); selenium.click("//div[@id='THSContainer']//span[@class='replytext']"); selenium.click("backButton"); selenium.waitForElementVisible("//div[@id='TTHContainer']/ul[1]/li[1]"); selenium.click("//div[@id='TTHContainer']//span[@class='replytext']"); selenium.click("backButton"); globalVars.selenium.Select("byDateFilter", "label=Things Happening Soon"); selenium.waitForElementVisible("//div[@id='THSContainer']/ul[1]/li[1]"); selenium.click("//div[@id='THSContainer']//span[@class='replytext']"); selenium.click("backButton"); globalVars.selenium.Select("byDateFilter", "label=Things That Happened"); BETTER TESTS [Test] public void testByDateTab() { funtions.loginMobi(); selenium.click("//a[@id='discussions']/span"); selenium.click("//a[@id='thingsToKnow']/span"); functions.verifyThingsToKnow(); functions.verifyThingsHappeningSoon(); selenium.Select("byDateFilter", "label=Things That Happened"); functions.verifyThingsThatHappened(); } GOOD TESTS [Test] public void testByDateTab() { selenium.Open(Moby_Common.MobyLoginUrl); LoginPage loginPage = new LoginPage(selenium); HappeningsPage happeningsPage = loginPage.loginAs(Common.stud1Name, Common.stud1Password); happeningsPage.waitToLoad(); Assert.That(!happeningsPage.isByTypePageLoaded()); Assert.That(happeningsPage.isByDatePageLoaded()); Assert.That(!happeningsPage.isByCoursePageLoaded()); happeningsPage.filterResults("byDateFilter","Things That Happened"); Assert.That(happeningsPage.isVisible("TTHContainer")); happeningsPage.filterResults("byDateFilter", "Things Happening Soon"); Assert.That(happeningsPage.isVisible("THSContainer")); happeningsPage.filterResults("byDateFilter", "All Types"); Assert.That(happeningsPage.isVisible("TTHContainer")); Assert.That(happeningsPage.isVisible("THSContainer")); } THE PAGE OBJECT: public class HappeningsPage : WebPageBaseClass { private string _loadingImage = "//span[@class='ajaxLoadingHeader']"; private string _moreLink = "more"; public HappeningsPage(ISelenium selenium) { this.selenium = selenium; this.title = "Happenings"; this.url = "index.html"; assertPageLoadedCorrectly(); } public HappeningsPage waitToLoad() { waitForElementNotVisible(_loadingImage ); return new HappeningsPage(selenium); } public ContentItemPage goToItem(string type, string name) { click("//div[@id='" + type + "']//span[text()=\"" + name + "\"]"); return new ContentItemPage(selenium); } public HappeningsPage clickMoreLink() { click(_moreLink); return new HappeningsPage(selenium); } public HappeningsPage filterResults(string id, string name) { selectDropdown(id, name); return new HappeningsPage(selenium); } }