This starter kit is designed to get you up and running with a bunch of awesome new front-end technologies, all on top of a configurable, feature-rich webpack build system that's already setup to provide hot reloading, CSS modules with Sass support, unit testing, code coverage reports, bundle splitting, and a whole lot more.
The primary goal of this project is to remain as unopinionated as possible. Its purpose is not to dictate your project structure or to demonstrate a complete sample application, but to provide a set of tools intended to make front-end development robust, easy, and, most importantly, fun. Check out the full feature list below!
Finally, This project wouldn't be possible without the help of our many contributors, so thank you for all of your help.
- Requirements
- Features
- Getting Started
- Usage
- CLI Generators
- Structure
- Webpack
- Server
- Styles
- Testing
- Deployment
- Troubleshooting
- Thank You
- node
^4.2.0
- npm
^3.0.0
- React (
^15.0.0
) - Redux (
^3.0.0
)- react-redux (
^4.0.0
) - redux-thunk middleware
- react-redux (
- react-router (
^2.0.0
)- Asynchronous routes configured with dependencies and reducers
- react-router-redux (
^4.0.0
) - Webpack
- Vanilla HMR using
module.hot
andwebpack-dev-middleware
- Code-splitting using
react-router
route configuration - Bundle splitting and CSS extraction
- Sass w/ CSS modules, autoprefixer, and minification
- Vanilla HMR using
- Koa (
^2.0.0-alpha
) - Karma
- Mocha w/ chai, sinon-chai, and chai-as-promised, and chai-enzyme
- PhantomJS
- Code coverage reports/instrumentation with isparta
- Flow (
^0.22.0
) - Babel (
^6.3.0
)- babel-plugin-transform-runtime so transforms aren't inlined
- babel-plugin-transform-react-constant-elements save some memory allocation
- babel-plugin-transform-react-remove-prop-types remove
PropTypes
- ESLint
- Uses Standard Style by default, but you're welcome to change this.
Just clone the repo and install the necessary node modules:
$ git clone https://github.com/davezuko/react-redux-starter-kit.git
$ cd react-redux-starter-kit
$ npm install # Install Node modules listed in ./package.json (may take a while the first time)
$ npm start # Compile and launch
We recommend using the Redux DevTools Chrome Extension.
Using the chrome extension allows your monitors to run on a separate thread and affords better performance and functionality. It comes with several of the most popular monitors, is easy to configure, filters actions, and doesn’t require installing any packages.
However, adding the DevTools components to your project is simple, first grab the packages from npm:
npm i --D redux-devtools redux-devtools-log-monitor redux-devtools-dock-monitor
Then follow the manual integration walkthrough.
First, I highly suggest checking out a new project by SpencerCDixon: redux-cli. This tool integrates extremely well with this project and offers added benefits such as generators (components, redux modules, etc.) and config/template management. It's still a work in progress, but give it a shot and file bugs to help make the project more robust.
Alternatively, if you just want to stick with this project and want to start a fresh project without having to clean up the example code in master
, you can do the following after cloning the repo:
git fetch origin new-project # Make sure you've fetched the latest copy of this branch from remote
git checkout new-project # Checkout the new-project branch
$ rm -rf .git && git init # Start a new git repository
Great, you now have a fresh project! There are a few titles you'll probably want to update, and then you're good to go:
~/package.json
- package name~/src/index.html
- template title tag
Before delving into the descriptions of each available npm script, here's a brief summary of the three which will most likely be your bread and butter:
- Doing live development? Use
npm start
to spin up the dev server. - Compiling the application to disk? Use
npm run compile
. - Deploying to an environment?
npm run deploy
can help with that.
NOTE: This package makes use of debug to improve your debugging experience. For convenience, all of messages are prefixed with app:*
. If you'd like to to change what debug statements are displayed, you can override the DEBUG
environment variable via the CLI (e.g. DEBUG=app:* npm start
) or tweak the npm scripts (betterScripts
in package.json
).
Great, now that introductions have been made here's everything in full detail:
npm run... |
Description |
---|---|
start |
Spins up Koa server to serve your app at localhost:3000 . HMR will be enabled in development. |
compile |
Compiles the application to disk (~/dist by default). |
dev |
Same as npm start , but enables nodemon to automatically restart the server when server-related code is changed. |
dev:nw |
Same as npm run dev , but opens the redux devtools in a new window. |
dev:no-debug |
Same as npm run dev but disables redux devtools. |
test |
Runs unit tests with Karma and generates a coverage report. |
test:dev |
Runs Karma and watches for changes to re-run tests; does not generate coverage reports. |
deploy |
Runs linter, tests, and then, on success, compiles your application to disk. |
deploy:dev |
Same as deploy but overrides NODE_ENV to "development". |
deploy:prod |
Same as deploy but overrides NODE_ENV to "production". |
flow:check |
Analyzes the project for type errors. |
lint |
Lint all .js files. |
lint:fix |
Lint and fix all .js files. Read more on this. |
NOTE: Deploying to a specific environment? Make sure to specify your target NODE_ENV
so webpack will use the correct configuration. For example: NODE_ENV=production npm run compile
will compile your application with ~/config/_production.js
.
Basic project configuration can be found in ~/config/_base.js
. Here you'll be able to redefine your src
and dist
directories, adjust compilation settings, tweak your vendor dependencies, and more. For the most part, you should be able to make changes in here without ever having to touch the webpack build configuration.
If you need environment-specific overrides (useful for dynamically setting API endpoints, for example), create a file with the name of target NODE_ENV
prefixed by an _
in ~/config
(e.g. ~/config/_production.js
). This can be entirely arbitrary, such as NODE_ENV=staging
where the config file is ~/config/_staging.js
.
Common configuration options:
Key | Description |
---|---|
dir_src |
application source code base path |
dir_dist |
path to build compiled application to |
server_host |
hostname for the Koa server |
server_port |
port for the Koa server |
compiler_css_modules |
whether or not to enable CSS modules |
compiler_devtool |
what type of source-maps to generate (set to false /null to disable) |
compiler_vendor |
packages to separate into to the vendor bundle |
This project integrates with Redux CLI out of the box. If you used it to generate this project you have immediate access to the generators listed below (if you cloned/forked the project you have these features as well, but make sure to install the CLI first!).
Script | Description | Options |
---|---|---|
redux g dumb <comp name> |
generates a dumb component and test file | |
redux g smart <smart name> |
generates a smart connected component and test file | |
redux g layout <comp name> |
generates functional layout component | |
redux g view <comp name> |
generates a view component | |
redux g form <form name> |
generates a form component (assumes redux-form) | |
redux g duck <duck name> |
generates a redux duck and test file | |
redux g blueprint <new blueprint> |
generates an empty blueprint for you to make | |
NOTE: redux-form is not a dependency by default. If you wish to use it make sure to npm i --save redux-form , or if you wish to modify the skeleton you can update the blueprint in ~/blueprints/form/files/... . |
All of these blueprints are available (and can be overriden) in the ~/blueprints
folder so you can customize the default generators for your project's specific needs. If you have an existing app you can run redux init
to set up the CLI, then make sure to copy over the blueprints
folder in this project for starter-kit specific generators.
See the Redux CLI github repo for more information on how to create and use blueprints.
The folder structure provided is only meant to serve as a guide, it is by no means prescriptive.
.
├── bin # Build/Start scripts
├── blueprints # Blueprint files for redux-cli
├── build # All build-related configuration
│ └── webpack # Environment-specific configuration files for webpack
├── config # Project configuration settings
├── interfaces # Type declarations for Flow
├── server # Koa application (uses webpack middleware)
│ └── main.js # Server application entry point
├── src # Application source code
│ ├── main.js # Application bootstrap and rendering
│ ├── components # Reusable Presentational Components
│ ├── containers # Reusable Container Components
│ ├── layouts # Components that dictate major page structure
│ ├── static # Static assets (not imported anywhere in source code)
│ ├── styles # Application-wide styles (generally settings)
│ ├── store # Redux-specific pieces
│ │ ├── createStore.js # Create and instrument redux store
│ │ └── reducers.js # Reducer registry and injection
│ └── routes # Main route definitions and async split points
│ ├── index.js # Bootstrap main application routes with store
│ ├── Root.js # Wrapper component for context-aware providers
│ ├── Home # Fractal route
│ │ ├── index.js # Route definitions and async split points
│ │ ├── assets # Assets required to render components
│ │ ├── components # Presentational React Components
│ │ ├── container # Connect components to actions and store
│ │ ├── modules # Collections of reducers/constants/actions
│ │ └── routes ** # Fractal sub-routes (** optional)
│ └── NotFound # Capture unknown routes in component
└── tests # Unit tests
Also known as: Self-Contained Apps, Recursive Route Hierarchy, Providers, etc
Small applications can be built using a flat directory structure, with folders for components
, containers
, etc. However, this structure does not scale and can seriously affect development velocity as your project grows. Starting with a fractal structure allows your application to organically drive its own architecture from day one.
We use react-router
route definitions (<route>/index.js
) to define units of logic within our application. Additional child routes can be nested in a fractal hierarchy.
This provides many benefits that may not be immediately obvious:
- Routes can be be bundled into "chunks" using webpack's code splitting and merging algorithm. This means that the entire dependency tree for each route can be omitted from the initial bundle and then loaded on demand.
- Since logic is self-contained, routes can easily be broken into separate repositories and referenced with webpack's DLL plugin for flexible, high-performance development and scalability.
Large, mature apps tend to naturally organize themselves in this way—analogous to large, mature trees (as in actual trees 🌲). The trunk is the router, branches are route bundles, and leaves are views composed of common/shared components/containers. Global application and UI state should be placed on or close to the trunk (or perhaps at the base of a huge branch, eg. /app
route).
- Stateless components that dictate major page structure
- Useful for composing
react-router
named components into views
- Prefer stateless function components
- eg:
const HelloMessage = ({ name }) => <div>Hello {name}</div>
- eg:
- Top-level
components
andcontainers
directories contain reusable components
- Containers only
connect
presentational components to actions/state- Rule of thumb: no JSX in containers!
- One or many container components can be composed in a stateless function component
- Tip: props injected by
react-router
can be accessed usingconnect
:// CounterWithMusicContainer.js import { connect } from 'react-redux' import Counter from 'components/Counter' export const mapStateToProps = (state, ownProps) => ({ counter: state.counter, music: ownProps.location.query.music // why not }) export default connect(mapStateToProps)(Counter) // Location -> 'localhost:3000/counter?music=reggae' // Counter.props = { counter: 0, music: 'reggae' }
- A route directory...
- Must contain an
index.js
that returns route definition - Optional: assets, components, containers, redux modules, nested child routes
- Additional child routes can be nested within
routes
directory in a fractal hierarchy
- Must contain an
Note: This structure is designed to provide a flexible foundation for module bundling and dynamic loading. Using a fractal structure is optional, smaller apps might benefit from a flat routes directory, which is totally cool! Webpack creates split points based on static analysis of require
during compilation; the recursive hierarchy folder structure is simply for organizational purposes.
You can redefine which packages to bundle separately by modifying compiler_vendor
in ~/config/_base.js
. These default to:
[
'history',
'react',
'react-redux',
'react-router',
'react-router-redux',
'redux'
]
Webpack is configured to make use of resolve.root, which lets you import local packages as if you were traversing from the root of your ~/src
directory. Here's an example:
// current file: ~/src/views/some/nested/View.js
// What used to be this:
import SomeComponent from '../../../components/SomeComponent'
// Can now be this:
import SomeComponent from 'components/SomeComponent' // Hooray!
These are global variables available to you anywhere in your source code. If you wish to modify them, they can be found as the globals
key in ~/config/_base.js
. When adding new globals, also add them to ~/.eslintrc
.
Variable | Description |
---|---|
process.env.NODE_ENV |
the active NODE_ENV when the build started |
__DEV__ |
True when process.env.NODE_ENV is development |
__PROD__ |
True when process.env.NODE_ENV is production |
__TEST__ |
True when process.env.NODE_ENV is test |
__DEBUG__ |
True when process.env.NODE_ENV is development and cli arg --no_debug is not set (npm run dev:no-debug ) |
__BASENAME__ |
npm history basename option |
This starter kit comes packaged with an Koa server. It's important to note that the sole purpose of this server is to provide webpack-dev-middleware
and webpack-hot-middleware
for hot module replacement. Using a custom Koa app in place of webpack-dev-server will hopefully make it easier for users to extend the starter kit to include functionality such as back-end API's, isomorphic/universal rendering, and more -- all without bloating the base boilerplate. Because of this, it should be noted that the provided server is not production-ready. If you're deploying to production, take a look at the deployment section.
Both .scss
and .css
file extensions are supported out of the box and are configured to use CSS Modules. After being imported, styles will be processed with PostCSS for minification and autoprefixing, and will be extracted to a .css
file during production builds.
NOTE: If you're importing styles from a base styles directory (useful for generic, app-wide styles), you can make use of the styles
alias, e.g.:
// current file: ~/src/components/some/nested/component/index.jsx
import 'styles/core.scss' // this imports ~/src/styles/core.scss
Furthermore, this styles
directory is aliased for sass imports, which further eliminates manual directory traversing; this is especially useful for importing variables/mixins.
Here's an example:
// current file: ~/src/styles/some/nested/style.scss
// what used to be this (where base is ~/src/styles/_base.scss):
@import '../../base';
// can now be this:
@import 'base';
To add a unit test, simply create a .spec.js
file anywhere in ~/tests
. Karma will pick up on these files automatically, and Mocha and Chai will be available within your test without the need to import them. If you are using redux-cli
, test files should automatically be generated when you create a component or redux module (duck).
Coverage reports will be compiled to ~/coverage
by default. If you wish to change what reporters are used and where reports are compiled, you can do so by modifying coverage_reporters
in ~/config/_base.js
.
Out of the box, this starter kit is deployable by serving the ~/dist
folder generated by npm run compile
(make sure to specify your target NODE_ENV
as well). This project does not concern itself with the details of server-side rendering or API structure, since that demands an opinionated structure that makes it difficult to extend the starter kit. However, if you do need help with more advanced deployment strategies, here are a few tips:
If you are serving the application via a web server such as nginx, make sure to direct incoming routes to the root ~/dist/index.html
file and let react-router take care of the rest. The Koa server that comes with the starter kit is able to be extended to serve as an API or whatever else you need, but that's entirely up to you.
Have more questions? Feel free to submit an issue or join the Gitter chat!
After installing npm dependencies, open .eslintrc
, change the semi
rule from never
to always
, and then run npm run lint:fix
-- Easy as that! Alternatively, use the same npm script after installing and extending your preferred ESLint configuration; it's easy to customize the project's code style to suit your team's needs. See, we can coexist peacefully.
This is most likely because the new window has been blocked by your popup blocker, so make sure it's disabled before trying again.
Reference: issue 110
Running into issues with Babel? Babel 6 can be tricky, please either report an issue or try out the stable v0.18.1 release with Babel 5. If you do report an issue, please try to include relevant debugging information such as your node, npm, and babel versions.
By default this repo does not bundle the babel polyfill in order to reduce bundle size. If you want to include it, you can use this commit from jokeyrhyme as a reference.
In keeping with the goals of this project, no internationalization support is provided out of the box. However, juanda99 has been kind enough to maintain a fork of this repo with internationalization support, check it out!
Make sure that your environment is installing both dependencies and devDependencies, since the latter are required to build the application. You can also reference this issue for more details.
While this is common to any sizable application, it's worth noting for those who may not know: if you happen to notice higher CPU usage in your editor after compiling the application, you may need to tell your editor not to process the dist folder. For example, in Sublime you can add:
"folder_exclude_patterns": [".svn", ".git", ".hg", "CVS", "node_modules", "dist"]
This project wouldn't be possible without help from the community, so I'd like to highlight some of its biggest contributors. Thank you all for your hard work, you've made my life a lot easier and taught me a lot in the process.
- Justin Greenberg - For all of your PR's, getting us to Babel 6, and constant work improving our patterns.
- Roman Pearah - For your bug reports, help in triaging issues, and PR contributions.
- Spencer Dixin - For your creation of redux-cli.
- Jonas Matser - For your help in triaging issues and unending support in our Gitter channel.
And to everyone else who has contributed, even if you are not listed here your work is appreciated.