I've been working with a React project using create-react-app and I have two options to start the project:
First way:
npm run start with the definition at the package.json like this:
"start": "react-scripts start",
Second way:
npm start
What is the difference between these two commands? And, what is the purpose of the react-scripts start?
I tried to find the definition, but I just found a package with this name. I still don't know what is the use of this command?
create-react-app and react-scripts
react-scripts is a set of scripts from the create-react-app starter pack. create-react-app helps you kick off projects without configuring, so you do not have to setup your project by yourself.
react-scripts start sets up the development environment and starts a server, as well as hot module reloading. You can read here to see what everything it does for you.
with create-react-app you have following features out of the box.
React, JSX, ES6, and Flow syntax support.
Language extras beyond ES6 like the object spread operator.
Autoprefixed CSS, so you don’t need -webkit- or other prefixes.
A fast interactive unit test runner with built-in support for coverage reporting.
A live development server that warns about common mistakes.
A build script to bundle JS, CSS, and images for production, with hashes and sourcemaps.
An offline-first service worker and a web app manifest, meeting all the Progressive Web App criteria.
Hassle-free updates for the above tools with a single dependency.
npm scripts
npm start is a shortcut for npm run start.
npm run is used to run scripts that you define in the scripts object of your package.json
if there is no start key in the scripts object, it will default to node server.js
Sometimes you want to do more than the react scripts gives you, in this case you can do react-scripts eject. This will transform your project from a "managed" state into a not managed state, where you have full control over dependencies, build scripts and other configurations.
As Sagiv b.g. pointed out, the npm start command is a shortcut for npm run start. I just wanted to add a real-life example to clarify it a bit more.
The setup below comes from the create-react-app github repo. The package.json defines a bunch of scripts which define the actual flow.
"scripts": {
"start": "npm-run-all -p watch-css start-js",
"build": "npm run build-css && react-scripts build",
"watch-css": "npm run build-css && node-sass-chokidar --include-path ./src --include-path ./node_modules src/ -o src/ --watch --recursive",
"build-css": "node-sass-chokidar --include-path ./src --include-path ./node_modules src/ -o src/",
"start-js": "react-scripts start"
},
For clarity, I added a diagram.
The blue boxes are references to scripts, all of which you could executed directly with an npm run <script-name> command. But as you can see, actually there are only 2 practical flows:
npm run start
npm run build
The grey boxes are commands which can be executed from the command line.
So, for instance, if you run npm start (or npm run start) that actually translate to the npm-run-all -p watch-css start-js command, which is executed from the commandline.
In my case, I have this special npm-run-all command, which is a popular plugin that searches for scripts that start with "build:", and executes all of those. I actually don't have any that match that pattern. But it can also be used to run multiple commands in parallel, which it does here, using the -p <command1> <command2> switch. So, here it executes 2 scripts, i.e. watch-css and start-js. (Those last mentioned scripts are watchers which monitor file changes, and will only finish when killed.)
The watch-css makes sure that the *.scss files are translated to *.cssfiles, and looks for future updates.
The start-js points to the react-scripts start which hosts the website in a development mode.
In conclusion, the npm start command is configurable. If you want to know what it does, then you have to check the package.json file. (and you may want to make a little diagram when things get complicated).
succinctly - it runs this
node node_modules/react-scripts/bin/react-scripts.js start
"start" is a name of a script, in npm you run scripts like this npm run scriptName, npm start is also a short for npm run start
As for "react-scripts" this is a script related specifically to create-react-app
npm start is the short form for npm run start
You can check about it here Difference between npm start and npm run start
react-scripts start
react-scripts is a set of scripts to support the creation, development and testing of react applications. It is used by create-react-app.
create-react-app is the officially supported way to create single-page React applications. create react app uses webpack to parse and bundle the application.
webpack parses the application and creates a dependency graph from its entry point specified in the webpack config file. while parsing, webpack uses babel to transpile the application to JavaScript, which has better support across browsers.
Webpack uses the generated dependency graph to create a single JavaScript file consisting of the application source code and modules used by the app, injects the file via script tag into public/index.html, and starts a development server on http://localhost:3000. Navigating to this URL in the browser will show a live, interactive instance of your application. Any changes saved to the source code will reflect in the running app instance automatically.
You can read more about this topic more on here
Related
I'm using create-react-app. When I run npm start (react-scripts start) it continuously builds the changes for me and does it magic. But what is the output folder for that? I know when I build it manually where the files go.
I want to use firebase emulator to serve the current version (the continuous build) of my react all but I don't understand where's the output folder or how to achieve it.
You could try this package https://github.com/Nargonath/cra-build-watch
Install it and add the script to your package.json
{
"scripts": {
"watch": "cra-build-watch"
}
}
and run it
npm run watch
more info here
https://ibraheem.ca/writings/cra-write-to-disk-in-dev/
and if you go to the react repo issue linked in the article you would find more workarounds
tl;dr
run npm run build, not npm run start
More Detail
react-scripts start runs webpack-dev-server internally. As a default setting, webpack-dev-server serves bundled files from memory and does not write files in directory.
If you want to write files with webpack-dev-sever, you could set writeToDisk option to true in your dev server configuration.
However, I dont think this is what you want to serve on firebase emulator. Webpack-dev-server does not build optimal app for production, and you also need to use react-app-rewired to customize dev server configuration in cra template.
What you want to do is npm run build to run react-scripts build, which builds optimized production app in /build directory.
I am wondering what would the difference be between npm run dev and npm run start.
To my surprise, I could not find much information online about this topic.
Specifically, I'd like to know in the context of React and Next JS.
I noticed that with React, you can start your app by running npm run start, without the need of running a build first. On the other hand, Next JS doesn't seem to behave in the same way (but I could have done something wrong with the setup).
I tried running a new Next app using npm run start, as it's a default script in package.json, but it didn't work. It shows this error: *Error: Could not find a production build*
Instead, running npm run dev created a .next folder, and started the server on port 3000 with no issues.
Can anyone help me understand how this works?
TL;DR: In Next.js, next dev is used to run the app in development mode. On the other hand, next start is used to run the app in production mode, but requires next build to be run first to generate an optimized production build.
Development
When running the Next.js app in development, you'll want to use next dev:
next dev starts the application in development mode with hot-code
reloading, error reporting, and more.
Production
When building the Next.js app for production, you'll want to use next build:
next build creates an optimized production build of your
application. The output displays information about each route.
Size – The number of assets downloaded when navigating to the page
client-side. The size for each route only includes its dependencies.
First Load JS – The number of assets downloaded when visiting the page
from the server. The amount of JS shared by all is shown as a separate
metric.
Followed by either next start, when you want to start the production server:
next start starts the application in production mode. The
application should be compiled with next build first.
Or next export, when exporting the app as static HTML:
next export allows you to export your app to static HTML, which can be
run standalone without the need of a Node.js server.
For more information refer to Next.js CLI docs.
Normally this depend on what is written in your package.json file. For example, in my case, within this file I got:
"scripts": {
"watch": "webpack --watch --watch-poll --progress --color",
"prod": "webpack -p",
"watch2": "webpack --watch --watch-poll --progress --color",
"build": "webpack --config=webpack.prod.config.js --progress --watch-poll -p"
},
so, if I run
npm run watch
I'll be compiling for development and it will be executed:
webpack --watch --watch-poll --progress --color
However, if I run
npm run build
it will be executed:
webpack --config=webpack.prod.config.js --progress --watch-poll -p
and it will compile for production.
To build an ExtJS application using sencha cmd I used the command below
sencha app build
But how I can build using open tooling? The docs is not clean about build application with open tooling.
npm run build should do the job
Check your package.json, you normally have a script section
You could build the app using the commands
npm run build:desktop
npm run build:phone
The above commands used based on the package.json file script section.Below is the snippet of code of script section of package.json file.
"scripts": {
"start": "npm run dev:desktop",
"clean": "rimraf build",
"dev:desktop": "webpack-dev-server --env.profile=desktop --env.browser=yes --env.verbose=no",
"dev:phone": "webpack-dev-server --env.profile=phone --env.browser=yes --env.verbose=no",
"build:desktop": "npm run clean && cross-env webpack --env.profile=desktop --env.environment=production --env.treeshake=yes",
"build:phone": "npm run clean && cross-env webpack --env.profile=phone --env.environment=production --env.treeshake=yes"
}
We were suffering a lot and spent significant effort to find the tooling what suits our needs the best. Finally we ended up using webpack because there is a large ecosystem around that so it opened up endless possibilities.
Although sencha did some webpack plugins but they are mostly just watching changes. Therefore we have created a webpack loader what is resolving ext's dependencies. This webpack loader will allow to use webpack to build your ext project. There is a small sample to help to configure it.
My company uses the three standard environments: Development, Test, and Production. My create-react-app based application is hosted as a content item within our CMS, so to get it into any environment, I need to run the npm run build command.
I've created a file, config.js, which exports a different configuration object based on variables in process.env, but the default behavior here has the limitation that npm run build is always considered production. This makes sense, I just need different behavior.
What I'd like to do is run a script like npm run build:dev, etc, which sets a process.env variable that I can switch on. Essentially I need to create an npm script that sets a dotenv variable, then calls npm run build.
What is the best way to accomplish this?
You can use cross-env package (from npm) to define a environment variable.
Just install the package:
npm install --save-dev cross-env
And create your custom script hwere do you define your variable, for example:
{
"scripts": {
"build": "cross-env NODE_ENV=production webpack --config build/webpack.config.js"
}
}
It worked like a charm on my projects.
More information here cross-env
We are working on a previously ejected create-react-app and now want to add flowtype.
We have followed the guide at: https://flow.org/en/docs/tools/create-react-app/
Should that work for an ejected app?
This has unfortunately caused the webpack-dev server launched with yarn start to stop automatically reloading on file updates.
Additionally, after adding // #flow to some files there is no output or indication of flow enforcing type checking.
Will we need to manually update the webpack configs?
Heres the package.json scripts
"scripts": {
"start": "node scripts/start.js",
"build": "yarn build-client && yarn build-server",
"build-client": "node scripts/build.js",
"build-server": "./node_modules/.bin/webpack --config ./config/webpack.server.config.js",
"test": "node scripts/test.js --env=jsdom",
},
The output for running yarn start is:
Compiled successfully!
You can now view cra in the browser.
Local: http://localhost:3000/
On Your Network: http://192.168.1.65:3000/
Note that the development build is not optimized.
To create a production build, use yarn build.
The doc you linked tells you how to install the flow-bin and to make a configuration file but don`t tells how to launch it.
Flow is separated tool that should be launch by own command (depends on how you wanna run it):
if you want to check types check manually, you need to add npm command on the "scripts" section of your package.js: "example-comand-flow": "flow". Then call it by npm run example-comand-flow and you`ll get errors directly on a terminal you running the script.
if you wanna have continuing type checking, you should find a manual how to configure it in your IDE. For example, in WebStorm you should go Preferences -> Languages & Frameworks -> JavaScript and set JavaScript language version to Flow and specify flow executable.