This project was bootstrapped with Create React App.
Below you will find some information on how to perform common tasks.
You can find the most recent version of this guide here.
- Updating to New Releases
- Sending Feedback
- Folder Structure
- Available Scripts
- Displaying Lint Output in the Editor
- Installing a Dependency
- Importing a Component
- Adding a Stylesheet
- Post-Processing CSS
- Adding Images and Fonts
- Adding Bootstrap
- Adding Flow
- Adding Custom Environment Variables
- Integrating with a Node Backend
- Proxying API Requests in Development
- Adding
<meta>
Tags - Running Tests
- Deployment
- Something Missing?
Create React App is divided into two packages:
create-react-app
is a global command-line utility that you use to create new projects.react-scripts
is a development dependency in the generated projects (including this one).
You almost never need to update create-react-app
itself: it’s delegates all the setup to react-scripts
.
When you run create-react-app
, it always creates the project with the latest version of react-scripts
so you’ll get all the new features and improvements in newly created apps automatically.
To update an existing project to a new version of react-scripts
, open the changelog, find the version you’re currently on (check package.json
in this folder if you’re not sure), and apply the migration instructions for the newer versions.
In most cases bumping the react-scripts
version in package.json
and running npm install
in this folder should be enough, but it’s good to consult the changelog for potential breaking changes.
We commit to keeping the breaking changes minimal so you can upgrade react-scripts
painlessly.
We are always open to your feedback.
After creation, your project should look like this:
my-app/
README.md
index.html
favicon.ico
node_modules/
package.json
src/
App.css
App.js
index.css
index.js
logo.svg
For the project to build, these files must exist with exact filenames:
index.html
is the page template;favicon.ico
is the icon you see in the browser tab;src/index.js
is the JavaScript entry point.
You can delete or rename the other files.
You may create subdirectories inside src
. For faster rebuilds, only files inside src
are processed by Webpack.
You need to put any JS and CSS files inside src
, or Webpack won’t see them.
You can, however, create more top-level directories.
They will not be included in the production build so you can use them for things like documentation.
In the project directory, you can run:
Runs the app in the development mode.
Open http://localhost:3000 to view it in the browser.
The page will reload if you make edits.
You will also see any lint errors in the console.
Launches the test runner in the interactive watch mode.
See the section about running tests for more information.
Builds the app for production to the build
folder.
It correctly bundles React in production mode and optimizes the build for the best performance.
The build is minified and the filenames include the hashes.
Your app is ready to be deployed!
Note: this is a one-way operation. Once you eject
, you can’t go back!
If you aren’t satisfied with the build tool and configuration choices, you can eject
at any time. This command will remove the single build dependency from your project.
Instead, it will copy all the configuration files and the transitive dependencies (Webpack, Babel, ESLint, etc) right into your project so you have full control over them. All of the commands except eject
will still work, but they will point to the copied scripts so you can tweak them. At this point you’re on your own.
You don’t have to ever use eject
. The curated feature set is suitable for small and middle deployments, and you shouldn’t feel obligated to use this feature. However we understand that this tool wouldn’t be useful if you couldn’t customize it when you are ready for it.
Note: this feature is available with
react-scripts@0.2.0
and higher.
Some editors, including Sublime Text, Atom, and Visual Studio Code, provide plugins for ESLint.
They are not required for linting. You should see the linter output right in your terminal as well as the browser console. However, if you prefer the lint results to appear right in your editor, there are some extra steps you can do.
You would need to install an ESLint plugin for your editor first.
A note for Atom
linter-eslint
users
If you are using the Atom
linter-eslint
plugin, make sure that Use global ESLint installation option is checked:
Then make sure package.json
of your project ends with this block:
{
// ...
"eslintConfig": {
"extends": "./node_modules/react-scripts/config/eslint.js"
}
}
Projects generated with react-scripts@0.2.0
and higher should already have it.
If you don’t need ESLint integration with your editor, you can safely delete those three lines from your package.json
.
Finally, you will need to install some packages globally:
npm install -g eslint babel-eslint eslint-plugin-react eslint-plugin-import eslint-plugin-jsx-a11y eslint-plugin-flowtype
We recognize that this is suboptimal, but it is currently required due to the way we hide the ESLint dependency. The ESLint team is already working on a solution to this so this may become unnecessary in a couple of months.
The generated project includes React and ReactDOM as dependencies. It also includes a set of scripts used by Create React App as a development dependency. You may install other dependencies (for example, React Router) with npm
:
npm install --save <library-name>
This project setup supports ES6 modules thanks to Babel.
While you can still use require()
and module.exports
, we encourage you to use import
and export
instead.
For example:
import React, { Component } from 'react';
class Button extends Component {
render() {
// ...
}
}
export default Button; // Don’t forget to use export default!
import React, { Component } from 'react';
import Button from './Button'; // Import a component from another file
class DangerButton extends Component {
render() {
return <Button color="red" />;
}
}
export default DangerButton;
Be aware of the difference between default and named exports. It is a common source of mistakes.
We suggest that you stick to using default imports and exports when a module only exports a single thing (for example, a component). That’s what you get when you use export default Button
and import Button from './Button'
.
Named exports are useful for utility modules that export several functions. A module may have at most one default export and as many named exports as you like.
Learn more about ES6 modules:
This project setup uses Webpack for handling all assets. Webpack offers a custom way of “extending” the concept of import
beyond JavaScript. To express that a JavaScript file depends on a CSS file, you need to import the CSS from the JavaScript file:
.Button {
padding: 20px;
}
import React, { Component } from 'react';
import './Button.css'; // Tell Webpack that Button.js uses these styles
class Button extends Component {
render() {
// You can use them as regular CSS styles
return <div className="Button" />;
}
}
This is not required for React but many people find this feature convenient. You can read about the benefits of this approach here. However you should be aware that this makes your code less portable to other build tools and environments than Webpack.
In development, expressing dependencies this way allows your styles to be reloaded on the fly as you edit them. In production, all CSS files will be concatenated into a single minified .css
file in the build output.
If you are concerned about using Webpack-specific semantics, you can put all your CSS right into src/index.css
. It would still be imported from src/index.js
, but you could always remove that import if you later migrate to a different build tool.
This project setup minifies your CSS and adds vendor prefixes to it automatically through Autoprefixer so you don’t need to worry about it.
For example, this:
.App {
display: flex;
flex-direction: row;
align-items: center;
}
becomes this:
.App {
display: -webkit-box;
display: -ms-flexbox;
display: flex;
-webkit-box-orient: horizontal;
-webkit-box-direction: normal;
-ms-flex-direction: row;
flex-direction: row;
-webkit-box-align: center;
-ms-flex-align: center;
align-items: center;
}
There is currently no support for preprocessors such as Less, or for sharing variables across CSS files.
With Webpack, using static assets like images and fonts works similarly to CSS.
You can import
an image right in a JavaScript module. This tells Webpack to include that image in the bundle. Unlike CSS imports, importing an image or a font gives you a string value. This value is the final image path you can reference in your code.
Here is an example:
import React from 'react';
import logo from './logo.png'; // Tell Webpack this JS file uses this image
console.log(logo); // /logo.84287d09.png
function Header() {
// Import result is the URL of your image
return <img src={logo} alt="Logo" />;
}
export default function Header;
This works in CSS too:
.Logo {
background-image: url(./logo.png);
}
Webpack finds all relative module references in CSS (they start with ./
) and replaces them with the final paths from the compiled bundle. If you make a typo or accidentally delete an important file, you will see a compilation error, just like when you import a non-existent JavaScript module. The final filenames in the compiled bundle are generated by Webpack from content hashes. If the file content changes in the future, Webpack will give it a different name in production so you don’t need to worry about long-term caching of assets.
Please be advised that this is also a custom feature of Webpack.
It is not required for React but many people enjoy it (and React Native uses a similar mechanism for images). However it may not be portable to some other environments, such as Node.js and Browserify. If you prefer to reference static assets in a more traditional way outside the module system, please let us know in this issue, and we will consider support for this.
You don’t have to use React Bootstrap together with React but it is a popular library for integrating Bootstrap with React apps. If you need it, you can integrate it with Create React App by following these steps:
Install React Bootstrap and Bootstrap from NPM. React Bootstrap does not include Bootstrap CSS so this needs to be installed as well:
npm install react-bootstrap --save
npm install bootstrap@3 --save
Import Bootstrap CSS and optionally Bootstrap theme CSS in the src/index.js
file:
import 'bootstrap/dist/css/bootstrap.css';
import 'bootstrap/dist/css/bootstrap-theme.css';
Import required React Bootstrap components within src/App.js
file or your custom component files:
import { Navbar, Jumbotron, Button } from 'react-bootstrap';
Now you are ready to use the imported React Bootstrap components within your component hierarchy defined in the render method. Here is an example App.js
redone using React Bootstrap.
Flow typing is currently not supported out of the box with the default .flowconfig
generated by Flow. If you run it, you might get errors like this:
node_modules/fbjs/lib/Deferred.js.flow:60
60: Promise.prototype.done.apply(this._promise, arguments);
^^^^ property `done`. Property not found in
495: declare class Promise<+R> {
^ Promise. See lib: /private/tmp/flow/flowlib_34952d31/core.js:495
node_modules/fbjs/lib/shallowEqual.js.flow:29
29: return x !== 0 || 1 / (x: $FlowIssue) === 1 / (y: $FlowIssue);
^^^^^^^^^^ identifier `$FlowIssue`. Could not resolve name
src/App.js:3
3: import logo from './logo.svg';
^^^^^^^^^^^^ ./logo.svg. Required module not found
src/App.js:4
4: import './App.css';
^^^^^^^^^^^ ./App.css. Required module not found
src/index.js:5
5: import './index.css';
^^^^^^^^^^^^^ ./index.css. Required module not found
To fix this, change your .flowconfig
to look like this:
[libs]
./node_modules/fbjs/flow/lib
[options]
esproposal.class_static_fields=enable
esproposal.class_instance_fields=enable
module.name_mapper='^\(.*\)\.css$' -> 'react-scripts/config/flow/css'
module.name_mapper='^\(.*\)\.\(jpg\|png\|gif\|eot\|otf\|webp\|svg\|ttf\|woff\|woff2\|mp4\|webm\)$' -> 'react-scripts/config/flow/file'
suppress_type=$FlowIssue
suppress_type=$FlowFixMe
Re-run flow, and you shouldn’t get any extra issues.
If you later eject
, you’ll need to replace react-scripts
references with the <PROJECT_ROOT>
placeholder, for example:
module.name_mapper='^\(.*\)\.css$' -> '<PROJECT_ROOT>/config/flow/css'
module.name_mapper='^\(.*\)\.\(jpg\|png\|gif\|eot\|otf\|webp\|svg\|ttf\|woff\|woff2\|mp4\|webm\)$' -> '<PROJECT_ROOT>/config/flow/file'
We will consider integrating more tightly with Flow in the future so that you don’t have to do this.
Note: this feature is available with
react-scripts@0.2.3
and higher.
Your project can consume variables declared in your environment as if they were declared locally in your JS files. By
default you will have NODE_ENV
defined for you, and any other environment variables starting with
REACT_APP_
. These environment variables will be defined for you on process.env
. For example, having an environment
variable named REACT_APP_SECRET_CODE
will be exposed in your JS as process.env.REACT_APP_SECRET_CODE
, in addition
to process.env.NODE_ENV
.
These environment variables can be useful for displaying information conditionally based on where the project is deployed or consuming sensitive data that lives outside of version control.
First, you need to have environment variables defined, which can vary between OSes. For example, let's say you wanted to
consume a secret defined in the environment inside a <form>
:
render() {
return (
<div>
<small>You are running this application in <b>{process.env.NODE_ENV}</b> mode.</small>
<form>
<input type="hidden" defaultValue={process.env.REACT_APP_SECRET_CODE} />
</form>
</div>
);
}
The above form is looking for a variable called REACT_APP_SECRET_CODE
from the environment. In order to consume this
value, we need to have it defined in the environment:
set REACT_APP_SECRET_CODE=abcdef&&npm start
(Note: the lack of whitespace is intentional.)
REACT_APP_SECRET_CODE=abcdef npm start
Note: Defining environment variables in this manner is temporary for the life of the shell session. Setting permanent environment variables is outside the scope of these docs.
With our environment variable defined, we start the app and consume the values. Remember that the NODE_ENV
variable will be set for you automatically. When you load the app in the browser and inspect the <input>
, you will see
its value set to abcdef
, and the bold text will show the environment provided when using npm start
:
<div>
<small>You are running this application in <b>development</b> mode.</small>
<form>
<input type="hidden" value="abcdef" />
</form>
</div>
Having access to the NODE_ENV
is also useful for performing actions conditionally:
if (process.env.NODE_ENV !== 'production') {
analytics.disable();
}
Check out this tutorial for instructions on integrating an app with a Node backend running on another port, and using fetch()
to access it. You can find the companion GitHub repository here.
Note: this feature is available with
react-scripts@0.2.3
and higher.
People often serve the front-end React app from the same host and port as their backend implementation.
For example, a production setup might look like this after the app is deployed:
/ - static server returns index.html with React app
/todos - static server returns index.html with React app
/api/todos - server handles any /api/* requests using the backend implementation
Such setup is not required. However, if you do have a setup like this, it is convenient to write requests like fetch('/api/todos')
without worrying about redirecting them to another host or port during development.
To tell the development server to proxy any unknown requests to your API server in development, add a proxy
field to your package.json
, for example:
"proxy": "http://localhost:4000",
This way, when you fetch('/api/todos')
in development, the development server will recognize that it’s not a static asset, and will proxy your request to http://localhost:4000/api/todos
as a fallback.
Conveniently, this avoids CORS issues and error messages like this in development:
Fetch API cannot load http://localhost:4000/api/todos. No 'Access-Control-Allow-Origin' header is present on the requested resource. Origin 'http://localhost:3000' is therefore not allowed access. If an opaque response serves your needs, set the request's mode to 'no-cors' to fetch the resource with CORS disabled.
Keep in mind that proxy
only has effect in development (with npm start
), and it is up to you to ensure that URLs like /api/todos
point to the right thing in production. You don’t have to use the /api
prefix. Any unrecognized request will be redirected to the specified proxy
.
Currently the proxy
option only handles HTTP requests, and it won’t proxy WebSocket connections.
If the proxy
option is not flexible enough for you, alternatively you can:
- Enable CORS on your server (here’s how to do it for Express).
- Use environment variables to inject the right server host and port into your app.
You can edit the generated index.html
and add any tags you’d like to it. However, since Create React App doesn’t support server rendering, you might be wondering how to make <meta>
tags dynamic and reflect the current URL.
To solve this, we recommend to add placeholders into the HTML, like this:
<!doctype html>
<html lang="en">
<head>
<meta property="og:title" content="$OG_TITLE">
<meta property="og:description" content="$OG_DESCRIPTION">
Then, on the server, regardless of the backend you use, you can read index.html
into memory and replace $OG_TITLE
, $OG_DESCRIPTION
, and any other placeholders with values depending on the current URL. Just make sure to sanitize and escape the interpolated values so that they are safe to embed into HTML!
If you use a Node server, you can even share the route matching logic between the client and the server. However duplicating it also works fine in simple cases.
Note: this feature is available with
react-scripts@0.3.0
and higher.
Read the migration guide to learn how to enable it in older projects!
Create React App uses Jest as its test runner. To prepare for this integration, we did a major revamp of Jest so if you heard bad things about it, give it another try.
Jest is a Node-based runner. This means that the tests always run in a Node environment and not in a real browser. This lets us enable fast iteration speed and prevent flakiness.
While Jest provides browser globals such as window
thanks to jsdom, they are only approximations of the real browser behavior. Jest is intended to be used for unit tests of your logic and your components rather than the DOM quirks.
We recommend that you use a separate tool for browser end-to-end tests if you need them. They are beyond the scope of Create React App.
Jest will look for test files with any of the following popular naming conventions:
- Files with
.js
suffix in__tests__
folders. - Files with
.test.js
suffix. - Files with
.spec.js
suffix.
The .test.js
/ .spec.js
files (or the __tests__
folders) can be located at any depth under the src
top level folder.
We recommend to put the test files (or __tests__
folders) next to the code they are testing so that relative imports appear shorter. For example, if App.test.js
and App.js
are in the same folder, the test just needs to import App from './App'
instead of a long relative path.
When you run npm test
, Jest will launch in the watch mode. Every time you save a file, it will re-run the tests, just like npm start
recompiles the code.
The watcher includes an interactive command-line interface with the ability to run all tests, or focus on a search pattern. It is designed this way so that you could keep it open and enjoy fast re-runs. You can learn the commands from the “Watch Usage” note that the watcher prints after every run:
By default, when you run npm test
, Jest will only run the tests related to files changed since last commit. This is an optimization designed to make your tests runs fast regardless of how many tests you have. However it assumes that you don’t often commit the code that doesn’t pass the tests.
Jest will always explicitly mention that it only ran tests related to the files changed since the last commit. You can also press a
in the watch mode to force Jest to run all tests.
Jest will always run all tests on a continuous integration server or if the project is not inside a Git or Mercurial repository.
To create tests, add it()
blocks with the name of the test and its code. You may optionally wrap them in describe()
blocks for logical grouping but this is neither required nor recommended.
Jest provides a built-in expect()
global function for making assertions. A basic test could look like this:
import sum from './sum';
it('sums numbers', () => {
expect(sum(1, 2)).toEqual(3);
expect(sum(2, 2)).toEqual(4);
});
All expect()
matchers supported by Jest are extensively documented here.
You can also use jest.fn()
and expect(fn).toBeCalled()
to create “spies” or mock functions.
There is a broad spectrum of component testing techniques. They range from a “smoke test” verifying that a component renders without throwing, to shallow rendering and testing some of the output, to full rendering and testing component lifecycle and state changes.
Different projects choose different testing tradeoffs based on how often components change, and how much logic they contain. If you haven’t decided on a testing strategy yet, we recommend that you start with creating simple smoke tests for your components:
import React from 'react';
import ReactDOM from 'react-dom';
import App from './App';
it('renders without crashing', () => {
const div = document.createElement('div');
ReactDOM.render(<App />, div);
});
This test mounts a component and makes sure that it didn’t throw during rendering. Tests like this provide a lot value with very little effort so they are great as a starting point, and this is the test you will find in src/App.test.js
.
When you encounter bugs caused by changing components, you will gain a deeper insight into which parts of them are worth testing in your application. This might be a good time to introduce more specific tests asserting specific expected output or behavior.
If you’d like to test components in isolation from the child components they render, we recommend using shallow()
rendering API from Enzyme. You can write a smoke test with it too:
npm install --save-dev enzyme react-addons-test-utils
import React from 'react';
import { shallow } from 'enzyme';
import App from './App';
it('renders without crashing', () => {
shallow(<App />);
});
Unlike the previous smoke test using ReactDOM.render()
, this test only renders <App>
and doesn’t go deeper. For example, even if <App>
itself renders a <Button>
that throws, this test will pass. Shallow rendering is great for isolated unit tests, but you may still want to create some full rendering tests to ensure the components integrate correctly. Enzyme supports full rendering with mount()
, and you can also use it for testing state changes and component lifecyle.
You can read the Enzyme documentation for more testing techniques. Enzyme documentation uses Chai and Sinon for assertions but you don’t have to use them because Jest provides built-in expect()
and jest.fn()
for spies.
Here is an example from Enzyme documentation that asserts specific output, rewritten to use Jest matchers:
import React from 'react';
import { shallow } from 'enzyme';
import App from './App';
it('renders welcome message', () => {
const wrapper = shallow(<App />);
const welcome = <h2>Welcome to React</h2>;
// expect(wrapper.contains(welcome)).to.equal(true);
expect(wrapper.contains(welcome)).toEqual(true);
});
All Jest matchers are extensively documented here.
Nevertheless you can use a third-party assertion library like Chai if you want to, as described below.
We recommend that you use expect()
for assertions and jest.fn()
for spies. If you are having issues with them please file those against Jest, and we’ll fix them. We intend to keep making them better for React, supporting, for example, pretty-printing React elements as JSX.
However, if you are used to other libraries, such as Chai and Sinon, or if you have existing code using them that you’d like to port over, you can import them normally like this:
import sinon from 'sinon';
import { expect } from 'chai';
and then use them in your tests like you normally do.
Jest has an integrated coverage reporter that works well with ES6 and requires no configuration.
Run npm test -- --coverage
(note extra --
in the middle) to include a coverage report like this:
Note that tests run much slower with coverage so it is recommended to run it separately from your normal workflow.
By default npm test
runs the watcher with interactive CLI. However, you can force it to run tests once and finish the process by setting an environment variable called CI
. Popular CI servers already set it by default but you can do this yourself too:
set CI=true&&npm test
(Note: the lack of whitespace is intentional.)
CI=true npm test
This way Jest will run tests once instead of launching the watcher.
If you find yourself doing this often in development, please file an issue to tell us about your use case because we want to make watcher the best experience and are open to changing how it works to accommodate more workflows.
By default, the package.json
of the generated project looks like this:
// ...
"scripts": {
// ...
"test": "react-scripts test --env=jsdom"
}
If you know that none of your tests depend on jsdom, you can safely remove --env=jsdom
, and your tests will run faster.
To help you make up your mind, here is a list of APIs that need jsdom:
- Any browser globals like
window
anddocument
ReactDOM.render()
TestUtils.renderIntoDocument()
(a shortcut for the above)mount()
in Enzyme
In contrast, jsdom is not needed for the following APIs:
TestUtils.createRenderer()
(shallow rendering)shallow()
in Enzyme
Finally, jsdom is also not needed for snapshot testing. Longer term, this is the direction we are interested in exploring, but snapshot testing is not fully baked yet so we don’t officially encourage its usage yet.
Snapshot testing is a new feature of Jest that automatically generates text snapshots of your components and saves them on the disk so if the UI output changes, you get notified without manually writing any assertions on the component output.
This feature is experimental and still has major usage issues so we only encourage you to use it if you like experimental technology. We intend to gradually improve it over time and eventually offer it as the default solution for testing React components, but this will take time. Read more about snapshot testing.
By default, Create React App produces a build assuming your app is hosted at the server root.
To override this, specify the homepage
in your package.json
, for example:
"homepage": "http://mywebsite.com/relativepath",
This will let Create React App correctly infer the root path to use in the generated HTML file.
See this example for a zero-configuration single-command deployment with now.
Use the Heroku Buildpack for Create React App.
You can find instructions in Deploying React with Zero Configuration.
Install the Surge CLI if you haven't already by running npm install -g surge
. Run the surge
command and log in you or create a new account. You just need to specify the build folder and your custom domain, and you are done.
email: email@domain.com
password: ********
project path: /path/to/project/build
size: 7 files, 1.8 MB
domain: create-react-app.surge.sh
upload: [====================] 100%, eta: 0.0s
propagate on CDN: [====================] 100%
plan: Free
users: email@domain.com
IP Address: X.X.X.X
Success! Project is published and running at create-react-app.surge.sh
Note that in order to support routers that use html5 pushState
API, you may want to rename the index.html
in your build folder to 200.html
before deploying to Surge. This ensures that every URL falls back to that file.
Note: this feature is available with
react-scripts@0.2.0
and higher.
Open your package.json
and add a homepage
field:
"homepage": "http://myusername.github.io/my-app",
The above step is important!
Create React App uses the homepage
field to determine the root URL in the built HTML file.
Now, whenever you run npm run build
, you will see a cheat sheet with a sequence of commands to deploy to GitHub pages:
git commit -am "Save local changes"
git checkout -B gh-pages
git add -f build
git commit -am "Rebuild website"
git filter-branch -f --prune-empty --subdirectory-filter build
git push -f origin gh-pages
git checkout -
You may copy and paste them, or put them into a custom shell script. You may also customize them for another hosting provider.
Note that GitHub Pages doesn't support routers that use the HTML5 pushState
history API under the hood (for example, React Router using browserHistory
). This is because when there is a fresh page load for a url like http://user.github.io/todomvc/todos/42
, where /todos/42
is a frontend route, the GitHub Pages server returns 404 because it knows nothing of /todos/42
. If you want to add a router to a project hosted on GitHub Pages, here are a couple of solutions:
- You could switch from using HTML5 history API to routing with hashes. If you use React Router, you can switch to
hashHistory
for this effect, but the URL will be longer and more verbose (for example,http://user.github.io/todomvc/#/todos/42?_k=yknaj
). Read more about different history implementations in React Router. - Alternatively, you can use a trick to teach GitHub Pages to handle 404 by redirecting to your
index.html
page with a special redirect parameter. You would need to add a404.html
file with the redirection code to thebuild
folder before deploying your project, and you’ll need to add code handling the redirect parameter toindex.html
. You can find a detailed explanation of this technique in this guide.
If you have ideas for more “How To” recipes that should be on this page, let us know or contribute some!