Remember what your tests are trying to verify

Lately, I’ve been working a lot on Selenium-based test automation solutions. And even though I’m still not overly enthusiastic about creating lots of user interface-driven automated tests, now that I’m getting more skilled at creating robust and stable Selenium tests, I am starting to appreciate the tool and what you can do with it more and more. As everybody working with Selenium can tell you, there are situations where things can get, well, interesting. And by interesting, I mean tricky. Dealing with dynamic front end frameworks and unpredictable modals and overlays can ask a lot of you in terms of debugging and exception handling skills.

Not yet being fluent in Selenese, I find myself on Google and (subsequently) StackOverflow a lot, trying to find suitable solutions for problems I encounter. While doing so, there’s one thing that I see a lot of people do, yet strikes me as really odd, given what I think is the purpose of these user interface-driven tests:

Forcing your tests to do something your users can’t.

From what I’ve understood, studying and working in the test automation field for a while now, user interface-driven tests, such as these Selenium tests, should be used to verify that the user of your user interface is able to complete a sequence of predefined actions. If you’re working in a hip and happening environment, these are often called ‘customer journeys’ or ‘user journeys’. But that’s not my point. What my point is, is that quite often I see workarounds suggested that go beyond what a regular user could do with his or her keyboard and / or mouse.

For example, take an element that is invisible until you hover over it with your mouse. If you just try to do a click(), your test will probably throw an exception stating that the element was not visible. Now, there are (at least) two ways to deal with this situation:

  1. Use the Actions class to simulate a mouseover, then click the element.
  2. Use a JavaScriptExecutor to perform the click.

While both approaches might result in a passing test, I am of the opinion that one is useful and the other is a horrible idea. Based on what I’ve written so far, can you guess which of the two options I’d suggest?

Indeed, option #1 is the way to go, for two reasons:

  • User interface-driven tests should mimic actual user interaction as closely as possible. I’ve never seen a user execute some JavaScript on a page to make an element visible. Either that, or I’m hanging around the wrong group of users..
  • What happens if a front-end developer makes a mistake (I know, they never do, but let’s assume so anyway) which causes the element not to become visible, even on a mouseover? With option #1, your test will fail, for the right reason. With #2, hello false negative!

There are some exceptions to the rule, though. The prime example I can think of is handling file uploads by directly sending the absolute path of the file to be uploaded to the input element responsible for the file upload using sendKeys(), instead of clicking on it and handling the file dialog. I’ve tried the latter before, and it’s a pain, first because you can’t do it with the standard Selenium API (because the file dialog is native to the operating system), second because different browsers use different file dialog layouts, resulting in a lot of pesky code that easily breaks down. In this case, I prefer to bypass the file dialog altogether (it’s probably not the subject of my test anyway).

In (almost) all other cases though, I’d advise you to stick to simulating your end user behavior as closely as possible. The job of your user interface-driven tests, and therefore of you as its creator, is not to force a pass, but to simulate end user interaction and see if that leads to a successfully executed scenario. Don’t fool yourself and your stakeholders by underwater tricks that obscure potential user interface issues.

Remember what your tests are trying to verify.

Let your test automation talk to you

Most of my current projects involve me paving the automation way by discussing requirements and needs with stakeholders, deciding on an approach and the tools used, setting up a solution and some initial tests and then instructing others and handing over the project. I love doing this type of projects, as it allows me (who’s bored quite easily and quickly) to move from project to project and from client to client regularly (sometimes once every couple of weeks). One of the most important factors in facilitating an easy handover of ‘my’ (it isn’t really mine, of course) automation solution to those who are going to continue and expand on it is by making the automation setup as self-explanatory as possible.

This is especially so when those that are going to take over don’t have as much experience with the tools and architecture that has been decided upon (or, in some cases, that I’ve decided to use). Let’s take a look at two ways to make it easy to hand over automation (results) to your successor (stakeholder), or, as I’d like to call it, to have your automation talk to you about what it’s doing and what the results of executing the automated tests are. Not literally talk to you (though that would be nice!), but you know what I mean..

Through the code
One of the most effective ways of making your automation solution easily explainable to others is through the code. Not only should your code be well structured and maintainable, good code requires little to no comments (who’s got time or motivation to write comments anyway?) because it is self-explanatory. Some prime examples of self describing code that I’ve used or seen are:

Having your Page Object methods in Selenium return Page Objects.
When you write your Page Object Methods like this:

public LoginPage SetEmailAddressTo(string emailAddress)
{
    PElements.SendKeys(_driver, textfieldEmailAddress, emailAddress);
    return this;
}

public LoginPage SetPasswordTo(string password)
{
    PElements.SendKeys(_driver, textfieldPassword, password);
    return this;
}

public void ClickLoginButton()
{
    PElements.Click(_driver, buttonLogin);
}

you can write your tests (or your step definition implementations, if you’re using Cucumber or SpecFlow) like this:

[When(@"he logs in using his credentials")]
public void WhenHeLogsInUsingHisCredentials()
{
    new LoginPage().
        SetEmailAddressTo("user@example.com").
        SetPassword("mypassword").
        ClickLoginButton();
}

It’s instantly clear what your tests is doing here, all by means of allowing method chaining through well-chosen method return types and expressive method names.

Choosing to work with libraries that offer fluent APIs
Two prime examples I often use are REST Assured and WireMock. For example, REST Assured allows you to create powerful, yet highly readable tests for RESTful APIs, while abstracting away boilerplate code like this:

@Test
public void verifyCountryForZipCode() {
						
	given().
	when().
		get("http://localhost:9876/us/90210").
	then().
		assertThat().
		body("country", equalTo("United States"));
}

WireMock does something similar, but for creating over-the-wire mocks:

public void setupExampleStub() {

	stubFor(post(urlEqualTo("/pingpong"))
		.withRequestBody(matching("<input>PING</input>"))
		.willReturn(aResponse()
			.withStatus(200)
			.withHeader("Content-Type", "application/xml")
			.withBody("<output>PONG</output>")));
}

One of the main reasons I like to work with both is that it’s so easy to read the code and explain to others what it does, without loss of power and features.

Create fluent assertions
Arguably the most important part of your test code is where the assertions are being made. Of course you need readable arrange (given) and act (when) sections too, but when you’re explaining or demonstrating your code to others, having readable assert (then) sections will be of great help, simply because that’s where the magic happens (so to say). You can make your assertions pretty much self-explanatory by using libraries such as Hamcrest when you’re using Java, or FluentAssertions when you’re working with C#. The latter even allows you to create better readable error messages when an assertion fails:

IEnumerable collection = new[] { 1, 2, 3 };
collection.Should().HaveCount(4, "because we thought we put four items in the collection"));

results in the following error message:

Expected <4> items because we thought we put four items in the collection, but found <3>.

Through the reporting
Now that you’ve got your code all cleaned up and readable, it’s time to shift our attention towards the output of the test execution: your reporting. We’ve seen the first step towards clearly readable and therefore useful reporting above: readable error messages. But good reporting is more than that. First, you need to think of the audience for your report. Who’s going to be informed by your report?

  • Is it a developer? In that case you might want to include details about the test data used, steps to reproduce the erroneous situation and detailed information about the failure (such as stack traces and screenshots) in the report. Pretty much everything that makes it easy for your devs to analyze the failure, so to say. In my experience, more is better here.
  • Is it a manager or a product owner? In that case he or she is probably only interested in the overall outcome (did the tests pass?). Maybe also in test coverage (what is it that we covered with these automated tests?). They’ll probably be less interested in stack traces, though.
  • Is it a build tool, such as Jenkins or Bamboo? In that case the highest priority is that the results are presented in a way that can be interpreted automatically by that tool. A prime example is the xUnit test output format that frameworks such as JUnit and TestNG produce. These can be picked up, parsed and presented by Jenkins, Bamboo and any other CI tool that’s worth its salt without any additional configuration or programming.

Again, think about the audience for your test automation reports, and include the information that’s valuable and useful to them. Nothing less, but nothing more either. This might require you to talk with these stakeholders. This might even require you to create more than a single report for every test run. Doesn’t matter. Have your automation talk to you and to your stakeholders by means of proper reporting.

Stop sweeping your failing tests under the RUG

Hello and welcome to this week’s rant on bad practices in test automation! Today’s serving of automation bitterness is inspired by a question I saw (and could not NOT reply to) on LinkedIn. It went something like this:

My tests are sometimes failing for no apparent reason. How can I implement a mechanism that automatically retries running the failing tests to get them to pass?

It’s questions like this that make the hairs in my neck stand on end. Instead of sweeping your test results under the RUG (Retry Until Green), how about diving into your tests and fixing the root cause of the failure?

First of all, there is ALWAYS a reason your test fails. It might be the application under test (your test did its work, congratulations!), but it might just as well be your test itself that’s causing the failure. The fact that the reason for the failure is not apparent does not mean you can simply ignore it and try running your test a second time to see if it passes then. No, it means there’s work for you to do. It might not be fun work: dealing with and catching with all kinds of exceptions that can be thrown by a Selenium test can be very tricky. The task also might not be suitable for you: maybe you’re inexperienced and therefore think ‘forget debugging, I’ll just retry the test, that’s way easier’. That’s OK, we’ve all been inexperienced at some point in our career. In a lot of ways, most of us still are. And I myself have not exactly been innocent of this behavior in the past either.

But at some point in time, it’s time to get over complaining about flaky tests and doing something about it. That means diving deep into your tests, how they interact with your application under test, getting to the root cause of the error or exception being thrown and fixing it, for once and for all. Here’s a real world example from a project I’m currently working on.

In my tests, I need to fill out a form to create a new savings account. Because the application needs to be sure that all information entered is valid, there’s a lot of front-end input validation going on (zip code needs to exist, email address should be formatted correctly, etc.). Whenever the application is busy validating or processing input, a modal appears that indicates to the end user that the site is busy processing input, and that therefore you should wait a little before proceeding. Sounds like a good idea, right? However, when you want your tests to fill in these forms automatically, you’ll sometimes run into the issue that you’re trying to click a button or complete a text field while it is being blocked by the modal. Cue WebDriverException (“other object would receive the click”) and failing test.

Now, there are two ways to deal with this:

  1. Sweep the test result under the RUG and retry until that pesky modal does not block your script from completing, or
  2. Catch the WebDriverException, wait until the modal is no longer there and do your click or sendKeys again. Writing wrappers around the Selenium API calls is a great way of achieving this, by the way.

Option 1. is the easy way. Option 2. is the right way. You choose. Just know that every failing test is trying to tell you something. Most of the time, it’s telling you to write a better test.

One more argument in favour of NOT sweeping your flaky tests under the RUG, but preventing them from happening in the future: some day, your organization might start, you know, actually relying on these test results. For example as part of a go / no go decision for deployment into production. If I were to call the shots, I’d make sure that all my tests that I rely on for making that decision were:

Really, it’s time to quit tolerating flaky tests. Repair them or throw them away, because what’s the added value of an unreliable test?. Just don’t sweep your failing tests under the RUG.