2 posts tagged with "tests"

At work, I'm part of a team of PHP developers where we work together to create a very large proprietary application with a web interface.

This application is developed using the Laravel framework and we use PHPUnit and Pest to run our unit tests.

However, in addition to the unit tests, we are also putting in place a tool that will simulate actions on our interface, such as accessing the login page, entering a login and password, simulating the click on the 'login' button, wait our main screen is displayed then, on the new page, assert a lot of things / do a lot of tasks.

It's a bit like asking a human to play out scenarios over and over again, every day, to make sure we haven't introduced any regressions in our latest developments, like a cool new feature whose code changes have broken a previous feature.

For this, we're using Behat. This is a PHP composer dependency you can add to your project (or a new one). The idea is to be able to write assertions in pure English (or French) like "I go to the ABC website then I click on the blog menu item and I should receive a list of blog items" or "In the search box, I type Docker then I click on the search button and I should receive articles having the Docker tags".

In this article, we'll learn more about Behat and how to use it for a first test.

If you are developing your own API (whatever the language) or if you need to consume some, Postman can be really handy.

Calling an API and getting the response is one thing, quite simple in fact, but a nice feature of Postman is the ability to validate the response like making sure the returned type is, f.i. application/json, the HTTP status code is 200, the response body is a JSON object (or an XML string), that the body contains some required information and so on.

You can also validate the response against a given schema to make sure the structure is well the one expected.

In this article, we'll use Postman like a unit test tool i.e. run checks on our own API and make a lot of assertions. This is improve the quality of your code by hightlighting potentials errors and, for any refactoring you'll do in the future, by running the tests again, you'll make sure you've not broken something; that you don't have any regression. Make sure you've not broken an API when you upgrade some code is gold.