As the codebase grows it makes our test runs take longer and longer. Naturally people will run the tests less frequently to avoid waiting. With Jest’s powerful and interactive watch mode we can easily run the tests that we want to focus on without having to run all the tests in the codebase.

Configure Jest for Testing JavaScript Applications

Use Jest Watch Mode to Speed Up Development

Instructor: [00:00] As cool as it is that we have this test script, which will run jest for us and it just runs all of our test, it would be really nice if we didn't have to run that test every time we make a change to our code.

[00:10] Instead, we could just have this running all the time. Anytime we make a change, it runs our test automatically. That's with the jest watch flag test. If we add a script called test:watch and that is going to run jest--watch, then we can create this run `npm run test:watch`.

```js
"scripts": {
    "test": "jest",
    "test:watch": "jest --watch",
    "dev": "webpack-dev-server --mode=development",
    "build": "webpack --mode=production",
    "postbuild": "cp ./public/index.html ./dist/index.html",
    "start": "serve --no-clipboard --single --listen 8080 dist",
    "format": "prettier --ignore-path .gitignore --write \"**/*.+(js|json|css|html|md)\"",
    "lint": "eslint --ignore-path .gitignore .",
    "validate": "npm run lint && npm run test && npm run build",
    "setup": "npm install && npm run validate"
  },
```

[00:30] That runs jest--watch, and we enter jest watch mode, which is awesome. Here at the very start, it says, "No test found related to files change since last commit." By commit, it's talking about the git commit. Jest will actually look up and get what files have changed since the last commit.

[00:48] Based on those files that have changed, it will run the test that are relevant for those changed files. We can see an example of this by going to the autoscaling text, and I break something here. Maybe by default, will return two instead of one. That's going to run all the files that are affected by that change.

[01:05] We actually have four tests. There are three of those four tests that are affected by that change. These two tests pass just fine, no problem, but this test fails. That's because we're taking a snapshot. That snapshot changes with this change.

[01:18] With that, we have our snapshot summary. We can update our code changes by pressing the U key, and then we can take a look at other options by pressing the W key to show more so I'll press W. Now we can run all the tests, run only the failed tests.

[01:33] I have a bunch of other commands. One command in particular I want to take a look at is this U to update failing snapshots as indicated here, or press I to update failing snapshots interactively.

[01:45] If you have a whole bunch of snapshots that this change has impacted, and a bunch of them are changing, it can be hard to scroll through all this output and review each and every one of them. You can press the I key to update the failing snapshots interactively.

[01:58] That's what I'm going to do. I'll press I, and it's going to run each test one by one and show me the output for the failing snapshot for that test in particular. Here, I can decide whether or not to update that particular snapshot by pressing the U key to update the failing snapshots for this test in particular.

[02:16] I can also press S to skip or Q to quit interactive snapshot mode. I'll press S to skip. That'll give me a summary of what happened during this interactive snapshot session where in particular, I reviewed one, I skipped that one.

[02:31] Now, I can press R to restart interactive snapshot or Q to quit. I will press Q. Now it's going to run all the tests as I had before.

[02:40] Next, I'm going to press W. Here we're going to press F to run only the failing tests. You'll notice, we're running three tests. We have the autoscaling texts, the calculator display, and the calculator.

[02:51] If I press F, it will only run the calculator display, because that's the one that had a failing test since the last test run. This makes it a lot easier to focus in on the tests that are failing and reduce the output as you're debugging and adding console log statements.

[03:07] Here, I'll press the W key. If I want to get out of this mode, then I press the F key to get out of the failing test mode. I'll press F. That's going to run all of my tests again. I press W here.

[03:20] As a reminder, this is only running the tests that are affected by the change that I've made. If I change this to one again, now there's no changes since the last git commit, so it doesn't run any tests. I can now press A to run all the tests. If I press W, it'll show me all the options, so now I press A. That's going to run all four of the tests in my test suite.

[03:38] I'm going to press W. Now I can press O to run the tests related to the changed files. I'll do that. Again, I'm not running any tests because there were no files changed. Let's change this file again so we get those three tests running.

[03:51] I'm going to press W, and I'm going to press P to filter by a file named regex pattern. If I press P and I want to run auto, that'll run just my auto scaling text. Now I can press W and you can see that I have an active filters here. If I want to clear the active filters, I can press C. That will get me back to the previous mode that I was in before.

[04:13] I'll press W again and press P to filter by a different regex pattern. I'll say .calc.star.js. That's going to run my two calculator tests that I have in here. Now I can press W and I'll press C again to clear those filters. It'll run all the tests that have been changed.

[04:33] I'll note that if I press W here and then A to run all the tests again, it will run all the tests. If I press P now and do `util`, it will focus in on that utils file, even though that file has not been affected by any changes that I've made.

[04:49] Now, if I press W, and we're going to filter by the test name regex, this is similar to a file name regex, except if we look at our utils.gs under our test directory here, we have this test that formats of value. If I press T and then type formats, it's only going to run the test that matches both a util for the test file name and formats for the test name itself.

[05:17] We can see this if I add a test that it is titled, "It is Broken," and then throw a new error, hi.

```js
test('formats the value', () => {
  expect(getFormattedValue('1234.0')).toBe('1,234.0')
})

test('it is broken', () =>{
  throw new Error('hi')
})
```
I'll save that and it's going to automatically skip that test for me because we're filtering by the test name.

[05:32] If I press W and then press C, I'm going to clear all the filters that means the filename as well as the test name filters. We're going to notice we have that snapshot that's still broken and we have this broken test in the utils test file.

[05:47] Let's go ahead and get rid of this broken test. We'll save that. Now we're only running test relevant to the change that we made here in get scale. We scroll down, press W and there are two more things that I want to show you.

[06:00] First, I'm going to get rid of 2 changed back to 1, save that, we'll press W here and then press H around all of the test. We press W to show all of this and now we have press enter to trigger a test run. This one appears in pretty much every mode.

[06:15] Anytime you need to rerun the test for any reason, you can press enter and it will trigger a rerun of all the test. Finally, and probably most importantly, to get out of this mode, you can press Q at any time to exit just watch mode.

[06:29] Another thing that I'll note is while your in Jest watch mode, if you're running some test and you decide, "Oh. This is taking too long." Maybe you've got thousands of test in here or something, then, you can press any other key at any time and the test run will be interrupted. You can go ahead and change your watch mode as you see fit.

[06:47] In review, all we did to make this magic happen is in our package JSON, we added a test:watch script and that simply runs Jest with the watch flag.

[06:57] One thing to note here is that while NPM does have a run test and it has a test and it has a T command, NMP does not have a short command for test:watch. To run this script, you really have to type out the full NPM run test:watch.

```bash
npm run test:watch
```


Jest is a fully featured testing framework with a developer experience that is second to none. It’s remarkably simple and flexible at the same time. For simple use cases, you often don’t need to configure anything, install and enjoy the built-in coverage and watch mode support.

In a real-world application though, you’ll often have needs specific to your application, especially when testing browser-based applications. You'll need to handle Webpack loaders, dynamic imports, and custom module resolution which Node.js does not support.

In this course we’ll go over ways you can optimize your Jest configuration to make testing real-world JavaScript applications a delight. We’ll cover what’s already been mentioned in addition to Babel support, code coverage, how to make watch mode even more helpful, and how to run test suites with entirely different configurations.

Intro to Configure Jest for Testing JavaScript Applications

We start in a small, but real-world application that’s built with webpack and React. We’ll go over installing the Jest testing framework locally to our project, adding an npm script to make it easy to run, and adding an example test file.

Install and Run Jest

Jest automatically loads and applies our babel configuration. However, because our project takes advantage of tree shaking with webpack, our babel configuration disables compiling modules. For Jest to work on our code, we need to make sure that our babel configuration compiles modules during our tests. We’ll use the NODE_ENV environment variable to generate the proper configuration.

Compile Modules with Babel in Jest Tests

In our application we’re testing code that should run in the browser, but Jest is intended to test JavaScript that runs in the browser or in node. Let’s create a custom jest config file so we can customize the test environment.

Configure Jest’s Test Environment for Testing Node or Browser Code

In most real-world applications using webpack, you’re likely using a few loaders. In this lesson we’ll see how to make Jest load our modules properly with a moduleNameMapper so we can test our code that makes use of these loaders.

Support Importing CSS files with Jest’s moduleNameMapper

If you’re using CSS modules with webpack, then we can improve our moduleNameMapper to improve our experience and capabilities testing our components by including the css module property name in our tests using identity-obj-proxy.

Support using Webpack CSS Modules with Jest

Snapshot testing is a way to simplify writing and maintaining assertions. As noted in [the Jest documentation](http://jestjs.io/docs/en/snapshot-testing.html#snapshot-testing-with-jest): “The snapshot artifact should be committed alongside code changes, and reviewed as part of your code review process. Jest uses pretty-format to make snapshots human-readable during code review. On subsequent test runs Jest will simply compare the rendered output with the previous snapshot. If they match, the test will pass. If they don't match, either the test runner found a bug in your code that should be fixed, or the implementation has changed and the snapshot needs to be updated.” Let’s see how these work and in what situations we can and should use them.

This lesson uses a simple demo before hopping into the application code testing DOM nodes. This example code isn't available.

Generate a Serializable Value with Jest Snapshots

Part of the power of snapshots is the ability to provide custom serializers. Let’s check out how to use jest-emotion to include our emotion CSS styles in our React component snapshots so we can be made aware of the impact of our CSS changes on our components.

UPDATE: `jest-emotion` has been renamed to `@emotion/jest` and you add the serializer with `'@emotion/jest/serializer'` instead of simply `'jest-emotion'`.

Test an Emotion Styled UI with Custom Jest Snapshot Serializers

Webpack’s resolve.modules configuration is a great way to make common application utilities easily accessible throughout your application. We can emulate this same behavior in Jest using the moduleDirectories configuration option.

Support Custom Module Resolution with Jest moduleDirectories

Often you’ll have some setup that you want to run before your tests start. Let’s see how we can use the setupFilesAfterEnv configuration in Jest to take care of test boilerplate and establish a good testing environment for our tests.

Configure Jest to Run Setup for All Tests with Jest setupFilesAfterEnv

Every large testbase has common utilities that make testing easier to accomplish. Whether that be a custom render function or a way to generate random data. Let’s see how we can make accessing these custom utilities throughout the tests easier using Jest’s moduleDirectories feature.

Support a Test Utilities File with Jest moduleDirectories

Sometimes it can be a real challenge to determine what’s going on when testing your code. It can be really helpful to step through your code in a debugger. In this lesson we’ll see how to use Jest’s --runInBand flag with node’s --inspect-brk to debug our tests in Chrome’s debugger.

Step through Code in Jest using the Node.js Debugger and Chrome DevTools

Jest comes with code coverage reporting built-into the framework, let’s see how quick and easy it is to add code coverage reporting to our project and take a look at the generated report. 

Configure Jest to Report Code Coverage on Project Files

You may have an idea of what Code Coverage is, but let’s take a quick look at how it’s recorded so we can understand how to analyze code coverage reports.

Analyze Jest Code Coverage Reports

Wherever you are at with code coverage you generally don’t want that level to go down. Let’s add coverage thresholds globally as well as in specific files to ensure we never drop below a certain level of coverage.

Set a Code Coverage Threshold in Jest to Maintain Code Coverage Levels

The coverage report generated by Jest is fantastic, but it’d be great to track that coverage over time and be able to review that coverage at a glance, maybe even put it up on a display in the office! Codecov.io is a fantastic service that can consume our code coverage report and it integrates great with GitHub. Let’s see how we can extend our existing Travis build configuration to send the coverage report to Codecov.io.

Report Jest Test Coverage to Codecov through TravisCI

In CI, we don’t want to start the tests in watch mode, but locally we normally want to run the tests in watch mode. We can have separate scripts, but it’d be great to not have to remember which script to run locally. Let’s use is-ci-cli to run the right script in the right environment when running the test script.

Run Jest Watch Mode by Default Locally with is-ci-cli

Sometimes you may have situations where configuration needs to be different for certain tests. In this lesson we’ll take a look at how we could create a custom configuration for tests that are intended to run in a node environment.

Run Tests with a Different Configuration using Jest’s --config Flag and testMatch Option

Sometimes you may find it useful to have more than one configuration in a project (for example, running some tests in a node environment and others in the jsdom environment). In this lesson we’ll learn about Jest’s projects feature to have jest run both of these configurations at once.

Support Running Multiple Configurations with Jest’s Projects Feature

Jest is more than a testing framework. It’s a highly optimized, blazing fast platform with incredible parallelization for running tasks across many files in our project. It has a capability to run more than just tests. We can bring these features to our linting as well. Let’s see how we can bring our favorite Jest features (like watch mode) to ESLint with jest-runner-eslint.

Run ESLint with Jest using jest-runner-eslint

It’s great that we can run multiple projects in our watch mode and that we can scope the tests down to specific tests, but sometimes it’s nice to be able to quickly switch between projects in watch mode. Let’s see how this works with jest-watch-select-projects.

Test Specific Projects in Jest Watch Mode with jest-watch-select-projects

Jest’s watch mode is [pluggable](http://jestjs.io/docs/en/watch-plugins.html) and jest-watch-typeahead is one plugin that you definitely don’t want to live without. It enhances the watch mode experience to help you know which tests will be run based on your filter. Making it even easier to run only the tests you’re concerned with running as you develop your codebase.

Filter which Tests are Run with Typeahead Support in Jest Watch Mode

Running the project tests on commit is a great way to avoid breaking the application accidentally and leverage the mechanism for confidence you have from your testbase. However, as the testbase grows this can take a very long time and slow productivity down. Let’s see how Jest is capable of running only the tests and linting only the files that are affected by the files we’re committing with husky and lint-staged to speed up our local test runs as well as help us avoid accidentally committing code that breaks our application.

Use Jest Watch Mode to Speed Up Development

Transcript

Use Jest Watch Mode to Speed Up Development