Epic Web Conf late-bird tickets are available now, hurry!

Get your tickets here

Join the community and network with other great web devs.

Time's up. The sale is over

Common Testing Mistakes

November 12th, 2018 — 6 min read

by Jeremy Bishop
by Jeremy Bishop

Mistake Number 0

One of the biggest mistakes you could make would be missing out on my full Testing JS course. (see what I did there?)

Mistake Number 1: Testing Implementation Details

I harp on this a lot (read more). It's because it's a huge problem in testing and leads to tests that don't give nearly as much confidence as they could. Here's a very simple example of a test that's testing implementation details:

// counter.js
import * as React from 'react'

export class Counter extends React.Component {
  state = {count: 0}
  increment = () => this.setState(({count}) => ({count: count + 1}))
  render() {
    const {count} = this.state
    return <button onClick={this.increment}>{count}</button>
  }
}

// __tests__/counter.js
import * as React from 'react'
// (it's hard to test implementation details with React Testing Library,
//  so we'll use enzyme in this example 😅)
import {mount} from 'enzyme'
import {Counter} from '../counter'

test('the increment method increments count', () => {
  const wrapper = mount(<Counter />)
  // don't ever do this:
  expect(wrapper.instance().state.count).toBe(0)
  wrapper.instance().increment()
  expect(wrapper.instance().state.count).toBe(1)
})

So why is this testing implementation details? Why is it so bad to test implementation details? Here are two truths about tests that focus on implementation details like the test above:

  1. I can break the code and not the test (eg: I could make a typo in my button's onClick assignment)
  2. I can refactor the code and break the test (eg: I could rename increment to updateCount)

These kinds of tests are the worst to maintain because you're constantly updating them (due to point #2), and they don't even give you solid confidence (due to point #1).

In my course I'll show you the right way to write tests and avoid this common mistake.

Mistake Number 2: 100% code coverage

Trying to go for 100% code coverage for an application is a total mistake and I see this all the time. Interestingly I've normally seen this as a mandate from management, but wherever it's coming from it's coming out of a misunderstanding of what a code coverage report can and cannot tell you about the confidence you can have in your codebase.

What code coverage is telling you:

  • This code was run when your tests were run.

What code coverage is NOT telling you:

  • This code will work according to the business requirements.
  • This code works with all the other code in the application.
  • The application cannot get into a bad state

Another problem with code coverage reports is that every line of covered code adds just as much to the overall coverage report as any other line. What this means is that you can increase your code coverage just as much by adding tests to your "About us" page as you can by adding tests to your "Checkout" page. One of those things is more important than the other, and code coverage can't give you any insight into that for your codebase...

There's no one-size-fits-all solution for a good code coverage number to shoot for. Every application's needs are different. I concern myself less with the code coverage number and more with how confident I am that the important parts of my application are covered. I use the code coverage report to help me after I've already identified which parts of my application code are critical. It helps me to know if I'm missing some edge cases the code is covering but my tests are not.

I should note that for open source modules, going for 100% code coverage is totally appropriate because they're generally a lot easier to keep at 100% (because they're smaller and more isolated) and they're really important code due to the fact that they're shared in multiple projects.

I talked a bit about this in my livestream the other day, check it out!

Mistake Number 3: Repeat Testing

One of the biggest complaints people have about end-to-end (E2E) tests is how slow and brittle they are when compared to integration or unit tests. There's no way you'll ever get a single E2E test as fast or reliable as a single unit test. It's just never going to happen. That said a single E2E test will get you WAY more confidence than a single unit test. In fact, there are some corners of confidence that are impossible to get out of unit tests that E2E tests are great at, so it's definitely worth having them around!

But this doesn't mean that we can't make our E2E tests faster and more reliable than you've likely experienced in the past. Repeat testing is a common mistake that people make when writing E2E tests that contribute to the poor performance and reliability.

Tests should always work in isolation. So that means every test should be executed as a different user. So every test will need to register and login as a brand new user right? Right. So you need to have a few page objects for the registration and login pages because you'll be running through those pages in every test right? WRONG! That's the mistake!

Let's take a step back. Why are you writing tests? So you can ship your application with confidence that things wont break! Let's say you have 100 tests that need an authenticated user. How many times do you need to run through the "happy path" registration flow to be confident that flow works? 100 times or 1 time? I think it's safe to say that if it worked once, it should work every time. So those 99 extra runs don't give you any extra confidence. That's wasted effort.

So what do you do instead? I mean, we already established that your tests should work in isolation so you shouldn't be sharing a user between them. Here's what you do: make the same HTTP calls in your tests that your application makes when you register and log in a new user! Those requests will be MUCH faster than clicking and typing around the page and there's less of a chance for false negative failures. And as long as you keep one test around that actually does test the registration/login flow you haven't lost any confidence that this flow works.

Conclusion

Always remember the reason that you're testing is about confidence. If something your test is doing isn't bringing you more confidence, then consider whether you can stop doing it!

Good luck!

Testing JavaScript

Ship Apps with Confidence

Illustration of a trophy
Kent C. Dodds
Written by Kent C. Dodds

Kent C. Dodds is a JavaScript software engineer and teacher. Kent's taught hundreds of thousands of people how to make the world a better place with quality software development tools and practices. He lives with his wife and four kids in Utah.

Learn more about Kent

If you found this article helpful.

You will love these ones as well.