Kuinka asentaa TypeScript-projekti

Perusteellinen opas aloittelijoille, jotka tekevät verkkosovelluksia Reactin avulla

Kaikkina kehittäjävuosieni aikana en ole koskaan kohdannut yhtä ylivoimaista erämaata kuin JavaScript-maailma. Se on hämmästyttävän monimutkainen maailma, jossa hyvin yksinkertaisen projektin tekeminen näyttää edellyttävän lukuisten työkalujen asentamista, useiden tekstitiedostojen muokkaamista, jotka liittävät kaikki nämä työkalut yhteen, ja suorittamaan joukko päätekomentoja.

On joitain työkaluja, jotka yrittävät piilottaa monimutkaisuuden sinulta vaihtelevalla menestyksellä. Mutta niin kauan kuin noilla työkaluilla ei ole yleismaailmallista adoptiota, ne vain näyttävät minusta pitävän vielä enemmän asioista, jotka sinun on opittava kaiken muun lisäksi.

Minulle suurin ärsytyslähde on se, että useimmissa opetusohjelmissa oletetaan, että olet jo perehtynyt ekosysteemiin, joten he eivät vaivaudu selittämään perusasioita. Tämän pahentamiseksi monet opetusohjelmat yrittävät työntää sinulle joukon ylimääräisiä työkaluja - kuten Webpack, Bower, NVM ja Redux - vain vähän selityksiä.

Se on ironista, koska JavaScript itse on jo asennettu käytännöllisesti katsoen kaikkiin tietokoneisiin maailmassa, mukaan lukien puhelimet. Miksi sovelluksen kirjoittamisen ammattimaisella tavalla on oltava niin monimutkaista kuin HTML-tiedoston kirjoittaminen, jossa on JavaScript-koodia?

Jos sinulla, kuten minulla, on synnynnäinen tarve ymmärtää, mitä tapahtuu, ja:

  • jos et voi sietää sokeasti komentojen kopioimista päätteisiin ja tekstitiedostoihin
  • jos haluat varmistaa, että tarvitset työkalua ennen sen asentamista
  • jos mietit, miksi npm-pohjainen projekti on 50 Mt, ennen kuin olet kirjoittanut ensimmäisen koodirivisi

tervetuloa! Olet tullut oikeaan paikkaan.

Toisaalta, jos haluat aloittaa ohjelmoinnin viiden minuutin kuluttua, tiedän siihen temppun: ohita esittely täällä ja aloita lukemisesta lähestymistavasta A osassa 2. Tai jos luulet antavan sinulle liikaa tietoa, Ohita vain osat, joista et halua oppia, kun jatkat.

Oletan, että tässä opetusohjelmassa sinulla on jonkin verran ohjelmointikokemusta HTML: stä, CSS: stä ja JavaScriptistä, mutta ei kokemusta TypeScriptistä, Reactista tai Node.js: stä.

Annan sinulle yleiskuvan JavaScript-ekosysteemistä ymmärtäessäni. Selitän, miksi mielestäni TypeScript ja React (tai Preact) ovat paras panoksesi verkkosovellusten tekemiseen. Ja autan sinua aloittamaan projektin ilman tarpeettomia ekstroja.

Kohdassa 2 keskustellaan siitä, miten ja miksi lisätä ekstroja projektiisi, jos päätät, että haluat ne.

Sisällysluettelo

Osa 1: Katsaus JavaScript-ekosysteemiin

Osa 2: Projektin perustaminen

  • Yhteiset vaiheet
  • Lähestymistapa A: helppo tapa
  • Muut lähestymistavat
  • Lähestymistapa B: harvimpien työkalujen tapa
  • Lähestymistapa C: Webpack Way
  • Yhteenveto

Osa 1: Katsaus JavaScript-ekosysteemiin

Monille ohjelmointikielille on olemassa tietty tapa tehdä asioita, joista kaikki tietävät.

Esimerkiksi, jos haluat tehdä C # -sovelluksen, asennat Visual Studion, luot Windows Forms -projektin muutamalla hiiren napsautuksella, napsauta vihreää "toista" -painiketta uuden ohjelman ajamiseksi ja aloita sitten kirjoittaa koodia sille. Paketinhallinta (NuGet) on sisäänrakennettu ja virheenkorjaaja Just Works. Toki, IDE: n asentaminen voi viedä muutaman tunnin, ja WPF on suunnilleen yhtä hauskaa kuin lyödä päätäsi tiiliseinää vasten, mutta ainakin aloittaminen on helppoa. (Paitsi jos et käytä Windowsia, se on täysin erilainen, mutta poikkeaen.)

Sen sijaan JavaScriptissä on niin paljon kilpailevia kirjastoja ja työkaluja melkein kaikkiin kehitysprosessin osa-alueisiin. Tämä työkalujen räjähdys voi tulla ylivoimainen ennen kuin kirjoitat ensimmäisen koodirivisi! Kun siirryt Googlelle "verkkosovelluksen kirjoittamiseen", jokainen vierailemasi verkkosivusto näyttää antavan erilaisia ​​neuvoja.

Yksi asia, josta useimmat ihmiset näyttävät olevan yhtä mieltä, on Node Package Managerin (NPM) käyttäminen JavaScript-kirjastojen (sekä palvelin- että selain) lataamiseen. Mutta jopa täällä jotkut ihmiset käyttävät lankaa, joka on yhteensopiva npm, tai mahdollisesti Boweria.

NPM sisältyy Node.js-verkkopalvelimeen, jota hallitset kokonaan JavaScript-koodilla. NPM on tiiviisti integroitu solmuun. Esimerkiksi npm startkomento suoritetaan node server.jsoletuksena.

Vaikka aiot käyttää toista verkkopalvelinta (tai käyttää mitään verkkopalvelinta ja kaksoisnapsauttaa vain HTML-tiedostoa), kaikki näyttävät oletavan, että Node.js on asennettu. Joten voit myös mennä eteenpäin ja asentaa node.js, joka antaa sinulle npmsivuvaikutuksen.

Node.js ei ole vain verkkopalvelin - se voi myös suorittaa komentorivisovelluksia, jotka on kirjoitettu JavaScript-muodossa. Tässä mielessä TypeScript-kääntäjä on Node.js-sovellus!

NPM: n lisäksi sinulla on useita vaihtoehtoja:

Minkä JavaScript-makun haluat?

JavaScriptin virallinen nimi on itse asiassa ECMAScript, ja laajimmin käytetty versio on lyhyesti ECMAScript 6 tai ES6. Vanhat selaimet, erityisesti Internet Explorer, tukevat vain ES5: tä.

ES6 lisää paljon hyödyllisiä ja tärkeitä uusia ominaisuuksia, kuten moduulit, let, const, nuolitoiminnot (tai lambda-toiminnot), luokat ja tehtävän uudelleenjärjestely.

ES7 lisää muutaman lisäominaisuuden, etenkin jotain, jota kutsutaan nimellä async / await.

Jos sinun ei tarvitse tukea vanhoja selaimia ja koodisi ei ole kovin suuri, koodin suorittaminen suoraan selaimessa on houkutteleva vaihtoehto, koska sinun ei tarvitse "kääntää" JavaScriptiäsi ennen kuin avaat sen selaimessa.

Mutta kääntövaiheen käyttämiseen on monia syitä:

  • Jos haluat tukea vanhoja selaimia, tarvitset "transpilerin", jotta voit käyttää JavaScriptin uusia ominaisuuksia vanhoissa selaimissa. Transpiler on kääntäjä, jonka lähtökoodi on korkean tason kieli, tässä tapauksessa JavaScript. Luulisin, että suosituin kääntäjä on Babel, jonka toisella sijalla on TypeScript.
  • Jos haluat käyttää suosittua React-kehystä (mutta ilman TypeScriptiä), kirjoitat todennäköisesti "JSX" -koodia - fragmentteja XML: stä JavaScript-koodin sisään. Selaimet eivät tue JSX: ää, joten se vaatii esiprosessorin (tyypillisesti Babel).
  • Jos haluat "pienentää" koodiasi niin, että se käyttää vähemmän kaistanleveyttä (tai on hämmentynyt), tarvitset "minifier" -prosessorin. Suosittuja pienentäjiä ovat UglifyJS, JSMin ja Closure Compiler.
  • Jos haluat tyyppitarkistuksen tai laadukkaan koodin täydennyksen (tunnetaan myös nimellä IntelliSense), sinun on käytettävä TypeScriptiä, joka on JavaScriptiä (eli jokainen JavaScript-tiedosto on myös TypeScript-tiedosto ... näennäisesti). TypeScript tukee sekä ES7-ominaisuuksia että JSX: tä, ja sen lähtö on ES5- tai ES6-koodi. Kun TypeScript- ja JSX-koodeja käytetään yhdessä, tiedostotunnisteen on oltava .tsx. Jotkut ihmiset käyttävät eri kieltä, joka on samankaltainen kuin TypeScript, nimeltään Flow.
  • Jos et pidä JavaScriptistä, voit kokeilla täysin erilaista kieltä, joka siirretään JavaScriptiin, kuten Elm, ClojureScript tai Dart.

Onneksi on mahdollista automatisoida kääntäminen siten, että koodisi käännetään uudelleen aina, kun tallennat tiedoston.

Tässä opetusohjelmassa käytetään TypeScriptiä, JavaScriptiä, joka sisältää kattavan tyyppijärjestelmän. TypeScriptin edut ovat seuraavat:

  1. Saat kääntäjän virheilmoitukset, kun teet tyyppiin liittyviä virheitä (sen sijaan, että löytäisit virheitä epäsuorasti, kun ohjelmasi käyttäytyy virheellisesti). IDE: ssä, kuten Visual Studio Code, virheesi on alleviivattu punaisella.
  2. Voit saada refactoring-ominaisuuksia. Esimerkiksi Visual Studio -koodissa F2voit nimetä toiminnon tai muuttujan useille tiedostoille uudelleen vaikuttamatta muihin samannimisiin asioihin.
  3. Tyypit sallivat IDE: iden tarjota koodin täydennysikkunoita, jotka tunnetaan myös nimellä IntelliSense, mikä tekee ohjelmoinnista paljon helpompaa, koska sinun ei tarvitse muistaa kaikkia kutsumiesi toimintojen nimiä ja odotettuja argumentteja:

Vinkki : Jos haluat pelata TypeScriptillä asentamatta mitään, käy sen leikkikentällä.

Asiakas vs. palvelin

Your can run code in a client (front-end browser), a server (Node.js back-end), or both. The client is not under your control. The user might use Firefox, Chrome, Safari, Opera, Edge, or in the worst case, Internet Explorer.

For security reasons, keep in mind that the user can modify a browser’s behavior using browser extensions or the F12 developer tools. You can’t even be sure that your code is running in a real browser.

Developers used to rely on the jQuery library to get consistent behavior in different browsers, but these days you can rely on different browsers to behave the same way in most cases (except perhaps Internet Explorer).

In this tutorial, we’ll run all the important code in the browser, but we’ll also set up a simple Node.js server to serve the app to the browser. Many other servers are available, such as Apache, IIS, or a static server like Jekyll.

But Node.js has become a sort of standard, likely because Node.js and NPM are bundled together.

User Interface frameworks

HTML and CSS alone are great for plain-old articles with images, or simple forms. If that’s all you’re doing, there’s probably no need for JavaScript at all. CSS can even do some things that once required JavaScript, such as pull-down menus, pages that completely reformat themselves for small/mobile browsers or printing, and animations.

If you need something more complex than that, or if your pages are generated dynamically from raw data, you’ll probably want to use JavaScript with an optional user-interface library or framework. I’ll show you later how to use React, which has earned a position as the most popular UI framework, and its little cousin Preact.

The “large” popular alternatives include Angular 2 and Vue.js, while the “small” ones include D3, Mithril and an old classic called jQuery.

If your web server runs JavaScript (Node.js), you can run React on the server to pre-generate the initial appearance of the page.

Build tools

Several tools for “building” and “packaging” your code are available — Webpack, Grunt, Browserify, Gulp, Parcel — but all these things are optional. I’ll show you how make do with just npm and, if you want, Parcel or Webpack.

CSS Flavors

In this article we’ll use plain CSS. If you’re going to have a compile step anyway, you might want to try SCSS, an “improved” derivative of CSS with extra features. Or you could use SASS, which is conceptually identical to SCSS but has a more concise syntax.

Either way you’ll need the Sass preprocessor. And as always in the JavaScript World, there are a bunch of alternatives, notably LESS.

Unit testing

The popular unit testing libraries are Mocha, Jasmine and Jest. See here for more. NPM has a special command for testing, npm test (which is short for npm run test).

Other libraries

Besides Redux, other popular JavaScript libraries include Lodash, Ramda, Underscore, and GraphQL.

The most popular linting utility is ESLint.

Bootstrap is a popular CSS library but it requires a JavaScript part (and it’s really SASS, not CSS).

When you see $ in JavaScript code, it typically refers to jQuery. When you see _ it typically refers to either Lodash or Underscore.

And perhaps it’s worth mentioning popular templating libraries: Jade, Pug, Mustache and Handlebars.

Non-web apps

I won’t say anything more about this, but TypeScript and JavaScript can be used outside the web.

With Electron you can write cross-platform desktop apps. With React Native you can write JavaScript apps for Android/iOS devices that have a “native” user interface. You can also write command-line apps with Node.js.

Module types

For the longest time, all JavaScript code ran in a single global namespace. This caused conflicts between unrelated code libraries, so various kinds of “module definitions” were invented to simulate what other languages call packages or modules.

Node.js uses CommonJS modules, which involves a magic function called require('module-name') to import modules and a magic variable called module.exports to create modules. To write modules that work in both browsers and Node.js, one can use Universal Module Definition (UMD modules). Modules that can be asynchronously loaded use AMD.

ES6 introduced a module system involving import and export keywords, but Node.js and some browsers still don’t support it. Here’s a primer on the various module types.

Polyfills & Prototypes

As an experienced developer, I can think of only two words (other than the names of libraries and tools) that are used only in JavaScript Land: polyfill and prototype.

Polyfills are backward-compatibility helpers. They are pieces of code written in JavaScript that allow you to use new features in old browsers. For example, the expression "food".startsWith('F') tests whether the String 'food' starts with F (for the record, that’s false - it starts with f, not F.) But startsWith is a new feature of JavaScript that is not available in older browsers.

You can “polyfill it” with this code:

String.prototype.startsWith = String.prototype.startsWith || function(search, pos) ;

This code has the form X = X || function(...) {...}, which means “if X is defined, set X to itself (don’t change it), otherwise set X to be this function.” The function shown here behaves the way startsWith is supposed to.

This code refers to one of the other unique things about JavaScript, the idea of prototypes. Prototypes correspond roughly to classes in other languages, so what this code is doing is actually changing the definition of the built-in String data type. Afterward when you write 'string'.startsWith() it will call this polyfill (if String.prototype.startsWith was not already defined). There are various articles out there to teach you about prototypes and prototypical inheritance, like this one.

Even some advanced browser features have polyfills. Have you heard of WebAssembly, which lets you run C and C++ code in a browser? There’s a JavaScript polyfill for it!

Credit

I’d like to thank the State of Javascript survey and State of JavaScript frameworks for much of the information above! For a few items I used npm-stat to measure popularity. See also this other new survey.

Section 2: Actually Setting Up the Project

Hey there! Still awake? Now we will go on a tour of the JavaScript tools ecosystem. This part is not about React (we’ll get to that later) but it includes a simple React component.

This is somewhat of a grand tour, so we’ll talk about writing your app in three different ways (with a summary afterward):

  • A. The Easiest Way (with Parcel)
  • B. The Way of Fewest Tools (or the do-it-yourself way)
  • C. The Webpack way

The first six steps are the same in all three approaches, so let’s get started!

Step 1: Install Node.js/npm

If you haven’t yet, go install Node.js which will also install the command-line package manager, npm.

If you want to deploy your app on some other web server, I recommend worrying about how to do that later.

Step 2: Install Visual Studio Code or other editor

One of the main reasons to use TypeScript instead of JavaScript is that it supports code completion features.

To enjoy this benefit, you’ll need to edit your TypeScript .ts files in a compatible editor such as Visual Studio Code — which is free and multi-platform. It’s also the most popular text editor for JavaScript apps. Alternatives include Atom and Sublime Text.

Visual Studio Code (VS Code) is folder-oriented: you open a folder in VS Code and that folder will be treated like the current “project”. During installation (on Windows, anyway) it will offer a checkbox to add an “Open with Code” action for folders (directories). I recommend using that option as an easy way to start VS Code from any folder:

Create an empty folder for your app, then open that folder in VS Code. Notice that VS Code has a built-in terminal so you won’t need a separate terminal window.

Step 3: Set up package.json

The package.json file will represent your project configuration. This includes its name, build commands, the list of npm modules used by your project, and more.

If you haven’t done so yet, create an empty folder for your app and open a terminal window in that folder.

In the terminal, run npm init.

npm init will ask you some questions in order to produce package.json. Leave a field blank to accept the default suggestion.

I wanted to make a small educational app to draw some graphs demonstrating how climate science explains the 20th century temperature record.

So I called my app climate-app:

C:\Dev\climate-app>npm initThis utility will walk you through creating a package.json file.It only covers the most common items, and tries to guess sensible defaults.[....]
package name: (climate-app)version: (1.0.0)description: Demo to visualize climate dataentry point: (index.js) test command:git repository:keywords:author: David Piepgrasslicense: (ISC) MIT
About to write to C:\Dev\climate-app\package.json:{ "name": "climate-app", "version": "1.0.0", "description": "Demo to visualize climate data", "main": "index.js", "scripts": { "test": "echo \"Error: no test specified\" && exit 1" }, "author": "David Piepgrass", "license": "MIT"}
Is this ok? (yes)

Notice the reference to index.js. Oddly, this file does not need to exist and we won’t be using it. It is used only if you share your code via npm.

Step 4: Install Typescript

VS Code reportedly has TypeScript “language support” rather than a TypeScript compiler, so now we need to install the compiler.

There are two ways to install TypeScript with npm. Either use

npm install --global typescript

or

npm install --save-dev typescript

If you use the --global option, then the TypeScript compiler tsc will be available in all projects on the same machine. It will also be available as a terminal command, but it will not be added to your package.json file. Therefore, if you share your code with others, TypeScript will not be installed when another person gets your code and runs npm install.

If you use --save-dev, TypeScript will be added to package.json and installed in your project’s node_modules folder (current size: 34.2 MB), but it will not be available directly as a terminal command.

You can still run it from the terminal as ./node_modules/.bin/tsc, and you can still use tsc inside the npm"scripts" section of package.json.

Fun fact: the TypeScript compiler is multiplatform because it is written in TypeScript — and compiled to JavaScript.

Step 5: Install React or Preact

To add React to your project:

npm install react react-domnpm install --save-dev @types/react @types/react-dom

Note: --save-dev marks things as “used for development” while --save (which is the default, and therefore optional) means “used by the program when it is deployed”.

@types packages provide type information to TypeScript, but they are not used when your code is running/deployed.

If you forget --save-dev or if you use it on the wrong package, your project will still work. The distinction is only important if you share your project as an npm package.

Alternately you can use Preact, which is almost the same as React but more than 10 times smaller. Preact has TypeScript type definitions built-in, so you only need a single command to install it:

npm install preact

Tip:npm i is a shortcut for npm install, and -D is a short for --save-dev.

Note: do not to install preact and @types/react in the same project, or tsc will go insane and give you about 150 errors (see preact issue #639). If this happens, uninstall the React types with npm uninstall @types/react @types/react-dom

Step 6: Write some React code

Make a file called app.tsx with this small React program:

Note: in order for the embedded JSX (HTML/XML) to work, the file extension must be tsx, not ts.

If you have any trouble making your code work, try this code instead — it’s the simplest possible React program:

import * as ReactDOM from 'react-dom';import * as React from 'react';
ReactDOM.render(React.createElement("h2", null, "Hello, world!"), document.body);

We’ll discuss how the code works later. For now, let’s focus on making it run.

If you’re using Preact, change the first two lines like so:

import * as React from 'preact';import * as ReactDOM from 'preact';

Some notes about Preact:

  • There is a preact-compat library which allows you to use preact with zero changes to your React code. Usage instructions exist for users of Webpack/Browserify/Babel/Brunch, and this page shows how to use preact-compat with Parcel.
  • There are rumors that in Preact you should write /** @jsx h */ at the top of the file, which tells TypeScript to call h() instead of the default React.createElement. In this case you must not do that or you’ll get a error in your browser that h is not defined (React.h, however, is defined). In fact Preact defines createElement as an alias for h, and since our import statement assigns 'preact' to React, React.createElement exists and works just fine.

Optional: running TypeScript scripts

This tutorial is focused on making web pages that run TypeScript code. If you would like to run a TypeScript file directly from the command prompt, the easiest way is to use ts-node:

npm install --global ts-node

After installing ts-node, run ts-node X.ts where X.ts is the name of a script you want to run. In the script you can call console.log("Hello") to write text to the terminal (reading text from a user is more complicated). On Linux systems you can put a “shebang” at the top of the script if you would like to be able to run ./X.ts directly (without mentioning ts-node):

#!/usr/bin/env ts-node

Note: if you don’t need to run .ts files from a terminal then you don’t need to install ts-node.

Running your project, Approach A: The Easy Way

I discovered Parcel when I was mostly done writing this article. Honestly, if I knew about Parcel from the beginning I might not have bothered writing about the other approaches. Don’t get me started on how easy Parcel is! It deserves a medal!

It’s very large, though (81.9 MB), so you should install it as a global:

npm install --global parcel-bundler

The truth is I’ve been lying to you. Parcel is so easy, you don’t even need all six steps above! You only really need Steps 1, 2 and 6 (install Node, install an editor, and write some code) because Parcel will do steps 3, 4, and 5 for you automatically.

So all we have to do now is to create an index.html file that refers to our app.tsx file, like this:

Then, simply open a terminal in the same folder and run the command parcel index.html.

This can’t run directly in a browser, of course, so Parcel:

  1. Automatically compiles app.tsx
  2. Installs React or Preact if it wasn’t already installed, because it notices that you’re using it
  3. Bundles your app with its dependencies into a single file called app.dd451710.js (or some other funny name)
  4. Creates a modified index.html that refers to the compiled and bundled app
  5. Puts these new files in a folder called dist.

And then it does everything else for you:

  1. It runs your app on a mini web server at //127.0.0.1:1234 — viewable on a web browser on the same machine
  2. It watches for changes to your code (app.tsx and index.html) and recompiles when you change them
  3. As if that wasn’t enough, when your files change, it will send a command to your web browser to automatically refresh!
  4. Even better, it updates the page without fully reloading it using its Hot Module Replacement feature

It can be challenging to set up a conventional build that does all of these things. This tutorial only covers how to do #1 and #2 in a conventional build, with only code recompilation (not HTML).

To learn about more features of Parcel, have a look at the Parcel documentation.

One limitation of Parcel is that it doesn’t perform type checking (your code is translated to JavaScript, but type errors are not detected).

For small projects, this is not a big problem because Visual Studio Code performs its own type checking. It gives you red squiggly underlines to indicate errors and all errors are listed in the “Problems” pane (press Ctrl+Shift+M to show it). But if you want, you can npm install parcel-plugin-typescript for enhanced TypeScript support including type checking (not currently working for me).

Other approaches

The other approaches are more well-known and are standard practice in the JavaScript community. We will be creating a folder with the following files inside:

  • app/index.html
  • app/app.tsx
  • package.json
  • tsconfig.json
  • server.js
  • webpack.config.js (optional)

As a matter of communicating to other people who look at your code later, it is useful to separate your program’s front-end code from its build configuration and app server.

The root folder of a project tends to become cluttered with extra files over time (such as .gitignore if you use git, README and LICENSE files, appveyor/travis files if you use continuous integration.) Therefore, we should separate the code of our front-end into a different folder.

In addition to the files we create, TypeScript will compile app.tsx into app.js and app.js.map, while npm creates a folder called node_modules and a file called package-lock.json . I can’t imagine why it’s called “lock”, but this page explains why it exists.

So please begin by creating an app folder and putting your app.tsx there.

Running your project, Approach B: The Way of Fewest Tools

It seems like everybody’s JavaScript project uses a dozen tools plus the kitchen sink. Is it possible to make a small program without any extra tools? It certainly is! Here’s how.

Step B1: Create tsconfig.json

Create a text file called tsconfig.json in your root folder, and put this code in it:

This file marks the folder as a TypeScript project and enables build commands in VSCode with Ctrl+Shift+B (the tsc: watch command is useful — it will automatically recompile your code whenever you save it.)

Silly fact: tsc allows comments in .json files but npm does not.

This file is very important because if the settings aren’t right, something may go wrong and mysterious errors may punch you in the face. Here is the documentation of tsconfig.json, but compiler options are documented separately.).

Step B2: Add a build script

To allow npm to build your TypeScript code, you must also add entries in the scripts part of package.json. Modify that section so it looks like this:

"scripts": { "test": "echo \"Error: no tests installed\" && exit 1", "build": "tsc", "start": "node server.js"},

The build script simply runs tsc which compiles your code according to the options in tsconfig.json. To invoke this script you write npm run build on the command line.

“But wait!” you may be thinking. “It’s really much easier to type tsc than npm run build!” That’s true, but there are two reasons to define a build script:

  1. If you installed TypeScript with --save-dev but not --global, you can’t run tsc directly from the command line because it’s not in the PATH.
  2. There’s a good chance your build process will become more complicated later. By creating a build script you can easily add other commands to the build process later.

Note:npm runs the prestart script automatically whenever someone runs the start script, so you could add this additional an additional script:

"prestart": "npm run build",

This would build your project whenever you start your server with npm start or npm run start.

But this has two disadvantages:

  1. tsc is a bit slow
  2. if tsc finds type errors then your server won’t start

When TypeScript detects type errors, that doesn’t stop it from writing JavaScript output files, and you may find it is occasionally useful to run your code even with type errors.

The default behavior of npm start is to run node server.js, so it seems redundant to include "start": "node server.js". However, if your server is written in TypeScript you’ll need this line because server.js doesn’t exist until server.ts is compiled. And if server.js doesn’t exist, npm start will give the error missing script: start unless you include this line.

Step B3: Make a simple server

To make sure Node.js is working, create a text file called server.js and put this code in it:

const http = require('http');
http.createServer(function (request, response) { // Send HTTP headers and body with status 200 (meaning success) response.writeHead(200, {'Content-Type': 'text/html'}); response.end(`  

Hello, world!

You asked for: ${request.url} `);}).listen(1234);

Run npm start to start it, visit //127.0.0.1:1234/index.html to make sure it works, then press Ctrl+C to stop the server.

To get IntelliSense for Node.js, you need to install type information for it with this command:

npm install @types/node --save-dev

Then in VS Code, type http. to make sure it works:

Behind the scenes, VS Code uses the TypeScript engine for this. However, if you rename your file to server.ts, IntelliSense doesn’t work! Is TypeScript broken in Node.js? Not really. TypeScript can still compile it, it just doesn’t grok require in a .ts context. So in TypeScript files, you should use import instead of require:

import * as http from 'http';

Note: this is confusingly different from Node’s .mjs files, which require import http from 'http'; (Details)

TypeScript then converts import to require in its output (because of the "module": "umd" option in tsconfig.json).

Now let’s change our server so it can serve any file from our /app folder:

You’ll notice that this code has some funny… nesting. That’s because Node.js functions are normally asynchronous. When you call functions in fs, instead of returning a result, they pause your program until they are done and then they call a function provided by you, sending that function either an error (err) or some information (fileInfo).

For example, instead of returning information about a file, fs.statsends information to a callback.

A fishy thing about this web server is that it ignores request.method, treating every request as if it were a GET. But it works well enough to get started.

Step B4 (optional): Use Express

If you want your server side to do any “routing” that is more complicated than serving a few files, you should probably learn about the most popular Node.js server framework: Express.

If we use Express, our server code will be much shorter.

Just install it with npm install express and put the following code in server.js:

const express = require('express');const app = express();
app.use('/node_modules', express.static('node_modules'));app.use('/', express.static('app'));app.listen(1234, () => console.log( 'Express server running at //127.0.0.1:1234'));

Step B5: Make a web page to hold your app

Finally, in your app folder, create an index.html file in there to load your app:

This page includes both React (react.development.js and react-dom.development.js) and Preact (preact.dev.js) so I don’t need to give you separate instructions for each one. You can remove whichever one you aren’t using, but the page can still load with unresolved script elements.

At this point you should be able to build your code (npm run build), start your server (npm start) and visit //127.0.0.1:1234 to view your app!

Remember, you can recompile your code automatically in VS Code: press Ctrl+Shift+B and choose tsc: watch.

Note: It’s important to load app.js at the end of the body, or React will say Error: Target container is not a DOM element because app.js would be calling document.getElementById('app') before the app element exists.

At this point it’s worth noting that this code is a little hacky. Especially this part:

 module = {exports:{}}; exports = {}; window.require = function(name) { return window[name]; }; window['react'] = window['React']; window['react-dom'] = window['ReactDOM'];

What’s this for? The short answer is that if your code contains import, TypeScript cannot produce code that “just works” in a browser, and this is one of many possible workarounds for that problem.

The long answer? First of all, remember that the JavaScript ecosystem has multiple module systems. Right now, your tsconfig.json file uses the "module": "umd" option, because "module": "umd" and "module": "commonjs" are the only modes that can be used in both Node.js and a web browser.

I asked you to make a server.js (not server.ts) file, but by using "module": "umd" you could write your server code in TypeScript if you want to.

UMD is the natural choice since it’s supposed to make a “universal” module definition, but TypeScript doesn’t really try to be universal — it simply won’t attempt to work in a web browser unaided.

Instead, it expects to find predefined symbols either for an AMD module system or a CommonJS (or Node.js) module system. If neither of these is defined, the module exits without even logging an error message.

Even if we could use the "module": "es6" option, which keeps import commands unchanged in the output file, it wouldn’t work because Chrome somehow still doesn’t support import in 2018. Also, the URLs of our modules have little in common with the string in our import statements, and I have learned that TypeScript path mapping aliases can’t solve the problem because they don’t change the compiler’s output.

TypeScript’s CommonJS implementation requires the require to be defined, of course — it’s used to import modules. But it also looks for exports and module.exports, even though our module doesn’t export anything. So our little hack must define all three.

The UMD versions of React, ReactDOM, and Preact set global variables called React, ReactDOM and preact respectively. But “global” variables in a browser are actually members of a special object called window. And in JavaScript, window.something means exactly the same thing as window['something'] except that you can use special characters, such as dashes, in the latter form. Therefore, window['preact'] and/or window['React'] already exist. So by defining a require function that simply returns window[name], it allows React or Preact to be imported.

However, we also need to create lowercase aliases 'react' and 'react-dom' because those are the names we must use in our TypeScript code (those names are recognized by the TypeScript compiler because those are the names of the folders in node_modules).

There’s another thing in our index.html that is a bit… unfortunate:

What makes this code less than ideal?

  1. We already have import statements in our app.tsx file, so it’s unfortunate that we need a separate command to load the modules in our index.html.
  2. We’re specifically referring to the development versions of the code, which include comments and are a lot more readable than minified versions. But if we roll out our web site to a large audience we’ll want to switch to the minified versions so that pages load faster. It would be nice if we could do that without losing the debugging benefits of the development versions.
  3. It assumes we can access files in node_modules, which is an unusual way to set up a server.

All the disadvantages described here lead us to want some kind of additional tool to help us deploy code to our web browser. We discussed Parcel already, but the most popular one is Webpack.

Running your project, Approach C: The Webpack Way

The most popular thing to do with front-end apps is to “pack” all the modules (React + your code + anything else you need) into a single file. This is comparable to what they call “linking” in some other languages, such as C++. That’s basically what Parcel and Webpack are built to do (Gulp is not — it requires extra tools installed separately.)

Steps C1 & C2: Create tsconfig.json and server.js

If you skipped approach B, please do steps B1 and B4 now.

Step C3: Install webpack

You could install it like this:

npm install --save-dev webpack webpack-cli

Unfortunately, Webpack is over-sized: these two packages have 735 dependencies weighing in at 50.9 MB (13,198 files in 1868 folders). And for some reason, webpack-cli requires the Webpack package but doesn’t mark it as a dependency, hence you must install both of them explicitly.

And although webpack-cli is ostensibly “just” the command-line interface for Webpack’s APIs, it is disproportionately large for some reason (Webpack alone is only 13.6 MB).

Due to its size, it probably makes more sense to install it as a global:

npm install --global webpack webpack-cli 

When using --global, keep in mind that if you share your code with someone else, the other person won’t get Webpack automatically when they type npm install, so you’ll want to explain how to install in your README file.

If you change your mind and want to switch from --save-dev to --global, just run the --global installation command and then use npm uninstall webpack webpack-cli to delete the local copy.

Step C4: Add build scripts

To allow npm to build and serve your project, add entries in the "scripts" section of package.json.

You could modify that section so it looks like this:

"scripts": { "test": "echo \"Error: no tests installed\" && exit 1", "build": "tsc && webpack app/app.js -o app/app.bundle.js --mode=production", "build:dev": "tsc && webpack app/app.js -o app/app.bundle.js --mode=development", "start": "node server.js"},

With these scripts, you would use either npm run build to build a minified production version, or npm run build:dev to build a development version with full symbols and comments. However, this is inconvenient, because when you change your code, you have to manually repeat the npm run build:dev command.

In Approach B we could use tsc: watch in VS Code, but that won’t work this time because we also need to run Webpack — and tsc doesn’t know that.

Can we set it up to rebuild automatically when our code changes? Yes, but we will need a Webpack plugin to accomplish this. One of the plugins that can do the job is called awesome-typescript-loader. Install it like this:

npm install awesome-typescript-loader --save-dev

Then in package.json, change your "scripts" section to look lke this:

This makes webpack fully responsible for building our TypeScript code, and therefore we can use its --watch option to watch for code changes. The command to build and watch for code changes is npm run watch.

Step C5: Start server and Webpack

You’ll need two separate terminals, one for your build system (npm run watch) and one for your server (npm start). If your server is written in TypeScript, then you need to run npm run watch first, otherwise it doesn’t matter which one you start first.

It’s worth noting that VS Code can keep track of multiple terminals. You can create two terminals and run one command in each, like this:

Step C6: Create index.html and load it

In Approach C, your index.html file is much simpler than in Approach B:

 App  

Mini React app ❤

Visit //127.0.0.1:1234 and the page should load. You’re done!

Step C7: Create a webpack.config.js file (optional)

Our build command is getting rather long, and is very similar for our three modes. Also, we’ve only set up the tsx file extension so webpack doesn’t know how to compile ts files yet.

Suosituin tapa käyttää Webpackia on erityinen kokoonpanotiedosto erillään package.json. "build"Käsikirjoitus yläpuolella tulee seuraava webpack.config.jstiedosto:

module.exports = { entry: __dirname+'/app/app.tsx', output: { path: __dirname+'/app', filename: 'app.bundle.js' }, module: { rules: [ tsx)$/, loader: 'awesome-typescript-loader'  ] }};

Kun olet luonut tiedoston, muuttaa scriptsvuonna package.jsonseuraavasti:

Kuten aiemmin, voit rakentaa muutoksia ja seurata muutoksia npm run watchtai käyttää npm run buildpienennettyä tuotantorakennetta.

Olet valmis!

Se siitä! Napsauta tästä saadaksesi yhteenvedon kaikista yllä olevista vaiheista ja täältä jatkaaksesi TypeScript-oppimista. Kysymyksiä?