The project's README from github.com/jamiebuilds/react-loadable. As one markdown file: README.md.

A higher order component for loading components with dynamic imports.
yarn add react-loadable
import Loadable from 'react-loadable';
import Loading from './my-loading-component';
const LoadableComponent = Loadable({
loader: () => import('./my-component'),
loading: Loading,
});
export default class App extends React.Component {
render() {
return <LoadableComponent/>;
}
}
If your company or project is using React Loadable, please open a PR and add yourself to this list (in alphabetical order please)
react-loadable-visibility - Building on top of and keeping the same API as react-loadable, this library enables you to load content that is visible on the screen.
react-loadable-ssr-addon - Server Side Render add-on for react-loadable. Discover & load automatically dynamically all files dependencies, e.g. splitted chunks, css, etc.
So you've got your React app, you're bundling it with Webpack, and things are going smooth. But then one day you notice your app's bundle is getting so big that it's slowing things down.
It's time to start code-splitting your app!

Code-splitting is the process of taking one large bundle containing your entire app, and splitting them up into multiple smaller bundles which contain separate parts of your app.
This might seem difficult to do, but tools like Webpack have this built in, and React Loadable is designed to make it super simple.
A common piece of advice you will see is to break your app into separate routes and load each one asynchronously. This seems to work well enough for many apps– as a user, clicking a link and waiting for a page to load is a familiar experience on the web.
But we can do better than that.
Using most routing tools for React, a route is simply a component. There's nothing particularly special about them (Sorry Ryan and Michael– you're what's special). So what if we optimized for splitting around components instead of routes? What would that get us?

As it turns out: Quite a lot. There are many more places than just routes where you can pretty easily split apart your app. Modals, tabs, and many more UI components hide content until the user has done something to reveal it.
Example: Maybe your app has a map buried inside of a tab component. Why would you load a massive mapping library for the parent route every time when the user may never go to that tab?
Not to mention all the places where you can defer loading content until higher priority content is finished loading. That component at the bottom of your page which loads a bunch of libraries: Why should that be loaded at the same time as the content at the top?
And because routes are just components, we can still easily code-split at the route level.
Introducing new code-splitting points in your app should be so easy that you don't think twice about it. It should be a matter of changing a few lines of code and everything else should be automated.
React Loadable is a small library that makes component-centric code splitting incredibly easy in React.
Loadable is a higher-order component (a function that creates a component)
which lets you dynamically load any module before rendering it into your app.
Let's imagine two components, one that imports and renders another.
import Bar from './components/Bar';
class Foo extends React.Component {
render() {
return <Bar/>;
}
}
Right now we're depending on Bar being imported synchronously via import,
but we don't need it until we go to render it. So why don't we just defer that?
Using a dynamic import (a tc39 proposal currently at Stage 3)
we can modify our component to load Bar asynchronously.
class MyComponent extends React.Component {
state = {
Bar: null
};
componentWillMount() {
import('./components/Bar').then(Bar => {
this.setState({ Bar: Bar.default });
});
}
render() {
let {Bar} = this.state;
if (!Bar) {
return <div>Loading...</div>;
} else {
return <Bar/>;
};
}
}
But that's a whole bunch of work, and it doesn't even handle a bunch of cases.
What about when import() fails? What about server-side rendering?
Instead you can use Loadable to abstract away the problem.
import Loadable from 'react-loadable';
const LoadableBar = Loadable({
loader: () => import('./components/Bar'),
loading() {
return <div>Loading...</div>
}
});
class MyComponent extends React.Component {
render() {
return <LoadableBar/>;
}
}
import()When you use import() with Webpack 2+, it will
automatically code-split for
you with no additional configuration.
This means that you can easily experiment with new code splitting points just
by switching to import() and using React Loadable. Figure out what performs
best for your app.
Rendering a static "Loading..." doesn't communicate enough to the user. You also need to think about error states, timeouts, and making it a nice experience.
function Loading() {
return <div>Loading...</div>;
}
Loadable({
loader: () => import('./WillFailToLoad'), // oh no!
loading: Loading,
});
To make this all nice, your loading component receives a couple different props.
When your loader fails, your loading component
will receive an error prop which will be an Error object (otherwise it
will be null).
function Loading(props) {
if (props.error) {
return <div>Error! <button onClick={ props.retry }>Retry</button></div>;
} else {
return <div>Loading...</div>;
}
}
Sometimes components load really quickly (<200ms) and the loading screen only quickly flashes on the screen.
A number of user studies have proven that this causes users to perceive things taking longer than they really have. If you don't show anything, users perceive it as being faster.
So your loading component will also get a pastDelay prop
which will only be true once the component has taken longer to load than a set
delay.
function Loading(props) {
if (props.error) {
return <div>Error! <button onClick={ props.retry }>Retry</button></div>;
} else if (props.pastDelay) {
return <div>Loading...</div>;
} else {
return null;
}
}
This delay defaults to 200ms but you can also customize the
delay in Loadable.
Loadable({
loader: () => import('./components/Bar'),
loading: Loading,
delay: 300, // 0.3 seconds
});
loader is taking too longSometimes network connections suck and never resolve or fail, they just hang there forever. This sucks for the user because they won't know if it should always take this long, or if they should try refreshing.
The loading component will receive a
timedOut prop which will be set to true when the
loader has timed out.
function Loading(props) {
if (props.error) {
return <div>Error! <button onClick={ props.retry }>Retry</button></div>;
} else if (props.timedOut) {
return <div>Taking a long time... <button onClick={ props.retry }>Retry</button></div>;
} else if (props.pastDelay) {
return <div>Loading...</div>;
} else {
return null;
}
}
However, this feature is disabled by default. To turn it on, you can pass a
timeout option to Loadable.
Loadable({
loader: () => import('./components/Bar'),
loading: Loading,
timeout: 10000, // 10 seconds
});
By default Loadable will render the default export of the returned module.
If you want to customize this behavior you can use the
render option.
Loadable({
loader: () => import('./my-component'),
render(loaded, props) {
let Component = loaded.namedExport;
return <Component {...props}/>;
}
});
Technically you can do whatever you want within loader() as long as it
returns a promise and you're able to render something.
But writing it out can be a bit annoying.
To make it easier to load multiple resources in parallel, you can use
Loadable.Map.
Loadable.Map({
loader: {
Bar: () => import('./Bar'),
i18n: () => fetch('./i18n/bar.json').then(res => res.json()),
},
render(loaded, props) {
let Bar = loaded.Bar.default;
let i18n = loaded.i18n;
return <Bar {...props} i18n={i18n}/>;
},
});
When using Loadable.Map the render() method is required. It
will be passed a loaded param which will be an object matching the shape of
your loader.
As an optimization, you can also decide to preload a component before it gets rendered.
For example, if you need to load a new component when a button gets pressed, you could start preloading the component when the user hovers over the button.
The component created by Loadable exposes a
static preload method which does exactly this.
const LoadableBar = Loadable({
loader: () => import('./Bar'),
loading: Loading,
});
class MyComponent extends React.Component {
state = { showBar: false };
onClick = () => {
this.setState({ showBar: true });
};
onMouseOver = () => {
LoadableBar.preload();
};
render() {
return (
<div>
<button
onClick={this.onClick}
onMouseOver={this.onMouseOver}>
Show Bar
</button>
{this.state.showBar && <LoadableBar/>}
</div>
)
}
}
When you go to render all these dynamically loaded components, what you'll get is a whole bunch of loading screens.
This really sucks, but the good news is that React Loadable is designed to make server-side rendering work as if nothing is being loaded dynamically.
Here's our starting server using Express.
import express from 'express';
import React from 'react';
import ReactDOMServer from 'react-dom/server';
import App from './components/App';
const app = express();
app.get('/', (req, res) => {
res.send(`
<!doctype html>
<html lang="en">
<head>...</head>
<body>
<div id="app">${ReactDOMServer.renderToString(<App/>)}</div>
<script src="/dist/main.js"></script>
</body>
</html>
`);
});
app.listen(3000, () => {
console.log('Running on http://localhost:3000/');
});
The first step to rendering the correct content from the server is to make sure that all of your loadable components are already loaded when you go to render them.
To do this, you can use the Loadable.preloadAll
method. It returns a promise that will resolve when all your loadable
components are ready.
Loadable.preloadAll().then(() => {
app.listen(3000, () => {
console.log('Running on http://localhost:3000/');
});
});
This is where things get a little bit tricky. So let's prepare ourselves little bit.
In order for us to pick up what was rendered from the server we need to have all the same code that was used to render on the server.
To do this, we first need our loadable components telling us which modules they are rendering.
There are two options in Loadable and
Loadable.Map which are used to tell us which modules our
component is trying to load: opts.modules and
opts.webpack.
Loadable({
loader: () => import('./Bar'),
modules: ['./Bar'],
webpack: () => [require.resolveWeak('./Bar')],
});
But don't worry too much about these options. React Loadable includes a Babel plugin to add them for you.
Just add the react-loadable/babel plugin to your Babel config:
{
"plugins": [
"react-loadable/babel"
]
}
Now these options will automatically be provided.
For typescript you can use react-loadable-ts-transformer which is a ts analog of react-loadable/babel plugin.
Next we need to find out which modules were actually rendered when a request comes in.
For this, there is Loadable.Capture component which can
be used to collect all the modules that were rendered.
import Loadable from 'react-loadable';
app.get('/', (req, res) => {
let modules = [];
let html = ReactDOMServer.renderToString(
<Loadable.Capture report={moduleName => modules.push(moduleName)}>
<App/>
</Loadable.Capture>
);
console.log(modules);
res.send(`...${html}...`);
});
In order to make sure that the client loads all the modules that were rendered server-side, we'll need to map them to the bundles that Webpack created.
This comes in two parts.
First we need Webpack to tell us which bundles each module lives inside. For this there is the React Loadable Webpack plugin.
Import the ReactLoadablePlugin from react-loadable/webpack and include it
in your webpack config. Pass it a filename for where to store the JSON data
about our bundles.
// webpack.config.js
import { ReactLoadablePlugin } from 'react-loadable/webpack';
export default {
plugins: [
new ReactLoadablePlugin({
filename: './dist/react-loadable.json',
}),
],
};
Then we'll go back to our server and use this data to convert our modules to bundles.
To convert from modules to bundles, import the getBundles
method from react-loadable/webpack and the data from Webpack.
import Loadable from 'react-loadable';
import { getBundles } from 'react-loadable/webpack'
import stats from './dist/react-loadable.json';
app.get('/', (req, res) => {
let modules = [];
let html = ReactDOMServer.renderToString(
<Loadable.Capture report={moduleName => modules.push(moduleName)}>
<App/>
</Loadable.Capture>
);
let bundles = getBundles(stats, modules);
// ...
});
We can then render these bundles into <script> tags in our HTML.
It is important that the bundles are included before the main bundle, so that they can be loaded by the browser prior to the app rendering.
However, as the Webpack manifest (including the logic for parsing bundles) lives in the main bundle, it will need to be extracted into its own chunk.
This is easy to do with the CommonsChunkPlugin
// webpack.config.js
export default {
plugins: [
new webpack.optimize.CommonsChunkPlugin({
name: 'manifest',
minChunks: Infinity
})
]
}
Notice: As of Webpack 4 the CommonsChunkPlugin has been removed and the manifest doesn't need to be extracted anymore.
let bundles = getBundles(stats, modules);
res.send(`
<!doctype html>
<html lang="en">
<head>...</head>
<body>
<div id="app">${html}</div>
<script src="/dist/manifest.js"></script>
<script src="/dist/main.js"></script>
${bundles.map(bundle => {
return `<script src="/dist/${bundle.file}"></script>`
// alternatively if you are using publicPath option in webpack config
// you can use the publicPath value from bundle, e.g:
// return `<script src="${bundle.publicPath}"></script>`
}).join('\n')}
<script>window.main();</script>
</body>
</html>
`);
We can use the Loadable.preloadReady() method on the
client to preload the loadable components that were include