@server/next v0.20.6

Server @ Next

⚠️ WIP This is an experimental library right now!

Documentation Here

A fully-fledged web server for Bun and Node.js, with all the basics covered for you:

import server from "@server/next";

// Create a running instance of the server
export default server(options)
  .get("/books", () => Book.list())
  .post("/books", { body: BookSchema }, (ctx) => {
    return Book.create(ctx.body).save();
  });

It includes all the things you would expect from a modern Server framework, like routing, static file serving, body+file parsing, gzip+brotli, streaming, testing, error handling, websockets, etc. We also have integrations with these:

• KV Stores: in-memory, Redis, Consul, DynamoDB.
• Buckets: AWS S3, Cloudflare R2, Backblaze B2.
• Validation libraries: Zod, Joi, Yup.

// Easy testing as well - index.test.js
import app from "./";

it("can retrieve the homepage", async () => {
  const res = await app.fetch(new Request("http://localhost:3000/books/"));
  const books = await res.json();
  expect(books[0]).toEqual({ id: 0, name: 'The Catcher In The Rye', ... });
});

Upgrading server

Why? We live in the era of multi-cloud (Heroku, Workers, Lambda, etc) and multi-runtimes (Node.js, Bun, WinterGC, etc). Desired improvements (WIP!):

• Tiny footprint with few dependencies. Installing and using the full library takes under 10kb (target limit).
• Faster! Reimplemented from scratch for speed. With raw ES6+ and a tiny code footprint, your server will fly.
• Modern ES6+ESM syntax for both the library and examples.
• Not using express underneath anymore. Considering keeping the compatibility layer anyway (since Express itself is a thin layer).
• Changed the reply logic greatly, including the removal of render(). This is the main reason express was removed. Many servers don't need render() at all.
• security Removed mandatory CSRF token, since this is only useful for server-rendered pages and not for SPA. Now we provide an auth module instead.

Progress

• Router has all verbs, as well as URL pattern matches
• Full URL parsing, including query and params in ctx.url.
• Body and Files parsing is working (need testing)
• The middleware can return:
  • A number and it'll be set as the status code
  • A string and it'll be sent as plain text or html (if it starts with "<")
  • A readStream and it'll be piped to the response
  • An object with status, body and headers and it'll be set raw.
• Response compression works

Examples

Streams

Creating a 100x100px thumbnail on the fly with Sharp:

// createThumbnail.js
import sharp from "sharp";

export default function createThumbnail(ctx) {
  // Return a pipe, which will be streamed to the output
  return sharp(ctx.url.params.name).resize(100, 100, { fit: "cover" }).png();
}

Breaking Changes

import, export and routing are the main changes from your point of view:

import server, { status, type, ...reply } from 'server';

export default server({ ...options })
  .use(mid1)
  .get('/', cb1)
  .get('/b', mid2, cb2)
  .routes({ get: [['/c', mid3, cb3]] });

status() now it's always partial:

// OLD
return 404;
return status(404).send("Not here..."); // treated as partial
return status(404); // treated as final

// NEW
return 404;
return status(404).send("Not here..."); // GOOD
return status(404).send(); // GOOD

// DON'T DO:
return status(404); // INVALID

import server from "server";

export default server()
  .get("/", () => "Hello world")
  .post("/", (ctx) => {
    console.log(ctx.body);
    return 201;
  });