Unit testing is the practice of testing the smallest possible units of our code, functions. We run our tests and automatically verify that our functions do the thing we expect them to do. We assert that, given a set of inputs, our functions return the proper values and handle problems.
This boilerplate uses the Jest test
framework to run tests and make assertions. This library makes writing tests as easy as speaking - you
describe
a unit of your code and expect
it
to do the correct thing.
We use this glob pattern to find unit tests app/**/*.test.js
- this tells
Jest to run all files that end with .test.js
anywhere within the app
folder. Use this to your advantage, and put unit tests next to the files you
want to test so relevant files stay together!
For the sake of this guide, lets pretend we're testing this function. It's
situated in the add.js
file:
// add.js
export function add(x, y) {
return x + y;
}
Note: The
export
here is ES6 syntax, and you would need an ES6 transpiler (e.g. babel.js) to run this JavaScript.
The
export
makes our function available as a module, which we canimport
and use in other files. Continue below to see what that looks like.
Jest is our unit testing framework. Its API, which we write tests with, is speech like and easy to use.
Note: The official documentation for Jest can be found here.
We're going to add a second file called add.test.js
with our unit tests
inside.
First, we import
the function in our add.test.js
file:
// add.test.js
import { add } from './add.js';
Second, we describe
our function:
describe('add()', () => {
});
Note:
(arg1, arg2) => { }
is ES6 notation for anonymous functions, i.e. is the same thing asfunction(arg1, arg2) { }
Third, we tell Jest what it
(our function) should do:
describe('add()', () => {
it('adds two numbers', () => {
});
it("doesn't add the third number", () => {
});
});
Now, We're going to test that our little function correctly adds two numbers.
We are going to take some chosen inputs, and expect
the result toEqual
the
corresponding output:
// [...]
it('adds two numbers', () => {
expect(add(2, 3)).toEqual(5);
});
// [...]
Lets add the second test, which determines that our function doesn't add the third number if one is present:
// [...]
it("doesn't add the third number", () => {
expect(add(2, 3, 5)).toEqual(add(2, 3));
});
// [...]
Note: Notice that we call
add
intoEqual
. I won't tell you why, but just think about what would happen if we rewrote the expect asexpect(add(2, 3, 5)).toEqual(5)
and somebody broke something in the add function. What would this test actually... test?
Should our function work, Jest will show this output when running the tests:
add()
✓ adds two numbers
✓ doesn't add the third number
Lets say an unnamed colleague of ours breaks our function:
// add.js
export function add(x, y) {
return x * y;
}
Oh no, now our function doesn't add the numbers anymore, it multiplies them! Imagine the consequences to our code that uses the function!
Thankfully, we have unit tests in place. Because we run the unit tests before we deploy our application, we see this output:
● add() › adds two numbers
expect(received).toEqual(expected)
Expected value to equal:
5
Received:
6
add()
✕ adds two numbers
✓ doesn't add the third number
This tells us that something is broken in the add function before any users get the code! Congratulations, you just saved time and money!
Imagine a navigation bar, this is what its folder might look like:
NavBar # Wrapping folder
├── index.js # Actual component
├── actions.js # Actions
├── constants.js # Constants
├── reducer.js # Reducer
└── test # Folder of tests
├── actions.test.js # Actions tests
└── reducer.test.js # Reducer tests
This boilerplate uses Redux, partially because it turns our data flow into
testable (pure) functions. Using the NavBar
component above,
let's see what testing the actions and the reducer would look like.
This is what our NavBar
actions look like:
// actions.js
import { TOGGLE_NAV } from './constants.js';
export function toggleNav() {
return { type: TOGGLE_NAV };
}
with this reducer:
// reducer.js
import { TOGGLE_NAV } from './constants.js';
const initialState = {
open: false
};
function NavBarReducer(state = initialState, action) {
switch (action.type) {
case TOGGLE_NAV:
return Object.assign({}, state, {
open: !state.open
});
default:
return state;
}
}
export default NavBarReducer;
Lets test the reducer first!
First, we have to import the reducer and the constant.
// reducer.test.js
import NavBarReducer from '../reducer';
import { TOGGLE_NAV } from '../constants';
Then we describe
the reducer, and add two tests: we check that it returns the
initial state and that it handles the toggleNav
action.
describe('NavBarReducer', () => {
it('returns the initial state', () => {
});
it('handles the toggleNav action', () => {
});
});
Lets write the tests themselves! Since the reducer is just a function, we can
call it like any other function and expect
the output to equal something.
To test that it returns the initial state, we call it with a state of undefined
(the first argument), and an empty action (second argument). The reducer should
return the initial state of the NavBar
, which is
{
open: false
}
Lets put that into practice:
describe('NavBarReducer', () => {
it('returns the initial state', () => {
expect(NavBarReducer(undefined, {})).toEqual({
open: false
});
});
it('handles the toggleNav action', () => {
});
});
This works, but we have one problem: We also have to explicitly write the initial state itself. When somebody changes the initial state, they will also have to manually update this code to directly reflect it.
Instead, we can leverage Jest's new snapshots feature.
Jest has the ability to store serialized snapshots of most basic types of information (objects, arrays, etc). It then compares the stored version when later tests are run, to find any unexpected mismatches.
We can write the test like
describe('NavBarReducer', () => {
it('returns the initial state', () => {
expect(NavBarReducer(undefined, {})).toMatchSnapshot();
});
it('handles the toggleNav action', () => {
});
});
Jest is now the one responsible for tracking the definition of the initial state. When somebody changes it in the future, Jest will warn that the snapshot doesn't match and then allow them to update the snapshot with a single command. No more manual updates!
For more details on Jest snapshots, please view Kent Dodd's feature video.
Lets see how we can test actions next.
We have one action toggleNav
that changes the NavBar
open state.
A Redux action is a pure function, so testing it isn't more difficult than
testing our add
function from the first part of this guide!
The first step is to import the action to be tested, the constant it should
return and expect
:
// actions.test.js
import { toggleNav } from '../actions';
import { TOGGLE_NAV } from '../constants';
Then we describe
the actions:
describe('NavBar actions', () => {
describe('toggleNav', () => {
it('should return the correct constant', () => {
});
});
});
Note:
describe
s can be nested, which gives us nice output, as we'll see later.
And the last step is to add the assertion:
it('should return the correct constant', () => {
expect(toggleNav()).toEqual({
type: TOGGLE_NAV
});
});
If our toggleNav
action works correctly, this is the output Jest will show us:
NavBar actions
toggleNav
✓ should return the correct constant
And that's it, we now know when somebody breaks the toggleNav
action!
Continue to learn how to test your application with Component Testing!