• The need to constantly update test scripts

In web testing, every time developers change selectors (element class names in the DOM), the tests have to be fixed, even if the changes are not visible in the interface. Manual scenarios don't have this problem.

• Interpreting and analyzing test results

As you've already figured out from the previous practical assignment, analyzing test results is a separate job in itself, and it usually takes a team up to a whole day when running a regression cycle.

• Automation works only in a well-formalized environment.

Your project needs quality documentation to support the automated tests.

Key properties an automated testing system should have

• A dedicated test repository that supports multi-user work and version control of the changes made, since tests change frequently (a version control system such as Git or SVN)
• A central repository of test resources.
• A system for managing test resources.
• A system for standalone functional testing
• Tools for building reports and quantitatively assessing the quality of the current product version.
• A system for distributing the testing process.

Test Automation Lab

An automation lab is a configured system for running tests and analyzing their results. Building it takes the joint effort of a system administrator, an automation systems architect, test designers, and automation engineers (developers of automated scenarios). And today you'll play the role of automated scenario designers. Let's look at how a continuous integration and test execution system works.

Let's walk through the process of launching automated tests and getting their results. This diagram shows the classic approach to the automation process. Let's make it clear right away that only test automation architects can set up and manage a system like this; we will just get familiar with the principles of how it works. The components are described in the same order in which the lab is built.

1. WebDriver is a framework for web automation. It provides methods for interacting with a web page in popular programming languages such as Java/Python/C#... which let you perform actions the way a user would: selecting an element, clicking it, reading its properties, and so on.

2. Test scripts. These are small programs that implement an automated test case in a high-level programming language. They use the methods of the WebDriver framework and implement the test steps as commands. The structure of the site under test is described separately in the Page Object format and plugged into the test.

Example of implementing a test case step:

This example is written in Python. driver is an instance of a WebDriver implemented in a specific driver, for example ChromeDriver, PlaylistPage is a Page Object, and the assert command performs the check.

Selenium Grid used to be a standalone product. Now there is physically just one product, Selenium Server, but it has several launch modes: it can run as a standalone server, as the hub of a cluster, or as a node of a cluster, depending on the startup parameters.

4. As for specific implementations of the WebDriver framework, they exist for all the major browsers. A driver is a piece of code that knows how to control its browser using the browser's native JavaScript commands. Some of them are developed by the Selenium team, and some by the browser makers themselves, such as Google and Opera (it's still alive! Even though its time has passed).

5. Repository and version control system. This is the universal code storage now used for all software development projects. Besides providing multi-user access, a version control system lets you keep several versions of the same document, roll back to earlier versions when needed, see who made a given change and when, and much more. The most popular VCS today is GIT

6. The last and key element of the automation lab is the continuous integration system, which ties all the previous components together and provides remote, continuous test execution. It automates the part of the software development process that doesn't require human involvement, providing continuous integration functions. It runs inside a servlet container, for example Apache Tomcat. It supports version control systems tools, including CVS, Git, and others. It can build projects using Apache Ant and Apache Maven, as well as run arbitrary shell scripts and Windows batch files. Builds and test runs can be triggered in different ways, for example after a version is pushed, on a schedule, on a request to a specific URL, or after another build in the queue finishes.

Performance Testing Lab

It's important to understand that load testing (we'll study it only in the next level) belongs exclusively to automated testing, as a separate sub-type of its own. After all, it's hard to even imagine putting load on a server by manually sending hundreds of requests a minute. For that kind of fun you would have to hire a thousand or two extras to use the target resource. Instead, one tester emulates the behavior of a single user (records a test script) and then uses that automated test to simulate load. The script is then used by load generator machines to test how the server behaves in a given mode. A load testing lab looks like this:

It's like a pyramid scheme, only cooler. This is Google's test pyramid, which they often recommend as a first approximation for distributing tests in a 70/20/10 ratio: 70% unit tests, 20% integration tests, and only 10% of all tests being E2E.

While reading all this super useful information about automated tests, I came up with some questions. Please help me sort them out. As always, the Master will give a hint in a minute.

1) To extend the functionality of the IDE, you can plug in an additional custom module or create your own.

2) You can use JavaScript in tests;

This makes it possible to:

• use conditional statements (if) in tests.
• use unconditional jumps (go to).
• use loops.
• create and work with variables inside a test.
• harness the power of the JavaScript language.

To use logic in Selenium IDE, you need to know JavaScript

Disadvantages

1. The library for the Safari browser is still in development.

2. Profile support exists only in the Firefox browser.

3. WebDriver still doesn't support native windows (Basic HTTP authentication).

4. WebDriver can't check design (layout); screenshot-based testing doesn't count :)

5. You need to know one of the languages: Java, C#, Ruby, Python, Perl, PHP

1) Locators written for Firefox may not work in Chrome and Internet Explorer (both the element ids and the overall xpath can change)

2) By default, focus on an input field works differently in different browsers.

3) Firefox inserts text at the end, other browsers at the beginning.

4) If an element has no id or name set, the automated test crumbles at the slightest change, sometimes even one made in a completely different feature.

5) If an element has no id or name set, different browsers may produce different xpath and cssid values; as a result, cross-browser compatibility is lost and you have to write separate automated tests for every browser

When working with Selenium WebDriver, if a site's page is very complex, has a nested frame structure, or the site opens new windows, you need to be very careful. In such cases the most common error is Element not found, and it happens not because your locator is wrong, but because you're looking in the wrong frame or window.

// Reset frames, switch focus back to the page

driver.SwitchTo().DefaultContent();

// Switch to the frame you need

driver.SwitchTo().Frame("main");

Selenium IDE Tutorial Video

Selenium locators, or how to find the elements that commands are applied to

Locators are used to find the elements that commands act on. The list of Selenium locators is quite long:

• id – uses the element's id attribute (identifier)
• name – uses the element's name attribute
• identifier – uses the element's id attribute; if no element is found by id, the search falls back to the name attribute
• dom – used to find an element by a DOM expression, which must start with document.
• xpath – used to find an element by an XPath expression, which must start with //
• link – used to find links with the specified text.