Expressive router for nodejs and the browser. Rill brings cascading middleware to the browser and enables a familiar routing solution for web applications.
Rill provides the minimum for abstractions over nodejs and the browser enabling things like routing (with redirecting, refreshes and more), cookies, and middleware with the same api.
It supports many view engines including Marko, React, Svelte and even html only template engines such as Pug.
Installation
npm install rill
Browser support
All modern browsers are supported including IE10 and above. Older browsers will need to polyfill the Promise API, checkout es6-promise for a good polyfill, babel-polyfill also covers this.
Community
Articles
- Universal web application framework - Interview with Dylan Piercey
- Isomorphic Javascript, let’s make it easier.
- How to make universal JavaScript applications — Part 1
- Browsers, Servers, and APIs
- Why Everyone is Talking About Isomorphic
- Isomorphic JavaScript: The Future of Web Apps
Why Rill?
Rill is the answer to a simple question; Can I run my Express style router in the browser? Turns out you can and it works awesome.
It brings a common interface to many typical app like features in both the browser and nodejs. Many isomorphic frameworks and routers have crazy abstractions and learning curves but with Rill, if you understand Express or Koa, you already know how the routing works! In Rill you get to program much of your application logic using the same api (client or server) including routing, rendering, data fetching and more are easily shared.
Rill also works perfectly as a stand alone router for nodejs or in the browser. This allows for easy progressive enhancement. If all is well the browser can handle much of your application logic and if JavaScript fails for any reason your server knows exactly what to do.
How does this thing work?
If you look at the source for Rill here you will quickly notice there is ZERO browser specific code. This is all thanks to @rill/http which is node's HTTP.createServer ported to the browser.
In the browser it works by listening for internal link clicks, form submissions and browser history changes. It will then create a Rill Context for each of these events and emit it through the router, similar to how receiving a request works in nodejs.
It supports everything you'd expect from a client side nodejs server. This includes redirects, refreshes, cookies, scrolling and url updates using the History API.
Example
Create an app
/** * The following code can run 100% in the browser or in nodejs. * Examples use es2015/2016 with Babel and JSX but this is optional. */ const app = // You can call Rill without new, but autocomplete will not work.
Setup middleware
// Universal form data parsing middleware.app // Universal react rendering middleware.app // Example Loggerapp
Setup a page
// Respond to a GET request.app
Handle a form submission
// Respond to a POST request.app
Start app
// Start a regular http server.// In the browser any form submissions or link clicks will intercepted by @rill/http.app
See Also
- isbrowser - A browserify transform to remove server-side code.
- isomorphic-fetch - Universal http requests using WHATWG fetch.
- isomorphic-form-data - Send multipart form data universally (able to send files and works with fetch).
- scroll-behavior - @rill/http will automatically try to use the "smooth" scroll-behavior when scrolling to targets on link clicks. This will polyfill that across modern browsers.
- submit-form - Manually trigger Rill navigation in the browser.
Prior Art
- koa-client - Koa clone that runs in the browser, inspired this package.
- monorouter - Another isomorphic router that partially inspired this package.
Contributions
- Use
npm test
to build and run tests.