In this article I’m going to show you how to write tests for your NodeJS application using Mocha, Chai and Sinon.
One of the cool things is you can choose your own assertion style when writing Mocha tests. In this article I will use Chai to do my assertions.
Chai supports 3 assertion styles. Should, expect and assert. This makes Mocha and Chai the ultimate combination to make your testing suite completely fit your own project needs and desires.
As a mocking framework I choose Sinon since it integrates neatly with Mocha and Chai and dozens of other test frameworks.
In order to start with writing your tests we first need to install Mocha, Chai and Sinon. Since I use Mocha for multiple projects I choose to install Mocha globally.
To be sure the consumers of my node package also have mocha installed I also add it to the dev dependencies. Since I installed Mocha globally it won’t be installed in my package folder again. Now we can actually start writing our tests.
Lets assume we have the following folder structure for our package.
First off all I want to configure Chai so we won’t have to define the configuration in each test. To do so I create a file chai.js in the specs/helpers folder.
In the above configuration we configure Chai to include the stacktrace on failing assertions. We also made the Chai expect and assert methods available globally. We will require this helper file when executing our tests using the mocha test runner.
Next we can write our tests / specifications to cover our code from lib. Lets assume we have a calculator in our lib we want to unit test.
The calculator supports 3 ways to show us the result of the calculation. Synchronously via a return value, and asynchronously via a callback or an event. First I will show you how to write a synchronous test for testing the return value.
This test is a pretty straight forward test. We describe the test case, then we create a new calculator before we specify our assertion. In this example we used the expect style of chai. In the next example I will show you how to test asynchronous code. Therefore I will use the callback approach in my test since it is delaying the result by 1 second.
In the above test I also used a different assertion style, “assert”, to make my assertion. Also notice the “done” callback parameter. This parameter is used to tell Mocha the test has completed.
As a last example I want to show you how you can test if a certain function is called using a Sinon stub and spy. To do so I’m going to test the calculator using the event approach. In the test I want to make sure the result event is emitted before the callback is invoked.
Since the callback is invoked after one second I want to make sure my assertions are executed when all the code of my calculation has executed. Therefore we use the Sinon fakeTimers. By doing so we are able to test if the callback is called and we are able to test if the emit method is called before the callback. If we would have skipped the clock then the assert on the spy would fail because the code hasn’t executed yet. Using the stub we can monitor calls to the emit method which our object inherits from the EventEmitter.
Now it is time to learn how we can execute those tests using Mocha. Using the following command we execute all our tests from our specs folder.
or use following command to have more visual feedback, since above command only shows you dots.
As a bonus I would like to give you my gruntfile configuration to execute the tests automatically using the grunt watch job. Also note I will only include the files ending with .spec.js. This makes sure the helper scripts won’t be used as source files to our tests. As a convention I choose to have all my tests/specifications files end on .spec.js.
Add the mochaTest task to your watch job tasks so your tests will be executed automatically. Don’t forget to install the grunt-mocha-test package to your package.json
Last, but not least I would like to give you some helpful links to the webpages of all three frameworks.