fun-component
A functional approach to authoring performant HTML components using plugins. Syntactic suggar on top of nanocomponent. Pass in a function and get another function back that handles rerendering as needed when called upon.
Usage
The most straightforward use is to pass in a function and have the default shallow diff figure out whether to rerender the component on consecutive calls.
// button.js
const html = require('bel')
const component = require('fun-component')
module.exports = component(function button (ctx, clicks, onclick) {
return html`
<button onclick=${onclick}>
Clicked ${clicks} times
</button>
`
})// app.js
const choo = require('choo')
const html = require('choo/html')
const button = require('./button')
const app = choo()
app.route('/', view)
app.mount('body')
function view (state, emit) {
return html`
<body>
${button(state.clicks, () => emit('click'))}
</body>
`
}
app.use(function (state, emitter) {
state.clicks = 0
emitter.on('click', () => {
state.clicks += 1
emitter.emit(state.events.RENDER)
})
})Standalone
Though fun-component was authored with choo in mind it works just as well standalone!
const button = require('./button')
let clicks = 0
function onclick () {
clicks += 1
button(clicks, onclick)
}
document.body.appendChild(button(clicks, onclick))API
component(render<function>)
component(name<string>, render<function>)
Create a new component context. Either takes a function as an only argument or a name and a function. Returns a function that renders the element. If no name is supplied the name is derrived from the functions name property.
Warning: implicit function names are most probably mangled during minification. If name consistency is important to your implementation, use the explicit name syntax.
component('hello', () => html`<h1>Hi there!</h1>`)Context
The component context (ctx) is prefixed to the arguments of all lifecycle hooks and the render function itself. The context object can be used to access the underlying nanocomponent instance.
// Exposing nanocomponent inner workings
component(function echo(ctx) {
return html`<h1>I'm ${ctx._name} on the ${ctx._hasWindow ? 'client' : 'server'}</h1>`
})Lifecycle
All the lifecycle hooks of nanocomponent are supported, i.e. beforerender, load, unload, afterupdate, and afterreorder. Lifecycle hooks are declared on the element itself (with an "on" prefix) and are forwarded to the underlying nanocomponent instance.
All lifecycle hooks are called with the context object and the latest arguments used to call the component.
Update
fun-component comes with a baked in default update function that performs a shallow diff of arguments to determine whether to update the component. Setting onupdate on the element overrides this default behavior.
onupdate(ctx<object>, args<array>, prev<array>)
Opposed to how nanocomponent calls the update function to detemrine whether to rerender the component, fun-component not only supplies the next arguments but also the previous arguments. These two can then be compared to determine whether to update.
Tip: Using ES2015 deconstruction makes this a breeze.
component(function greeting(ctx, title) {
return html`<h1 onupdate=${update}>Hello ${title}!</h1>`
})
// Deconstruct arguments and compare `title`
function update(ctx, [next], [prev]) {
return next !== prev
}Example
Using every lifecycle hook available. The rendered element can be accessed on the context object as ctx.element.
component(function hooks(ctx, title) {
return html`
<div onupdate=${update} onbeforerender=${beforerender} onload=${load} onunload=${unload} onafterupdate=${afterupdate} onafterreorder=${afterreorder}>
Hello ${title}!
</div>
`
})
function update(ctx, [title], [prev]) {
console.log(`diffing ${title} and ${prev}`)
return title !== prev
}
function beforerender(ctx, title) {
console.log(`will render with ${title}`)
}
function load(ctx, title) {
console.log(ctx.element, `mounted in DOM with ${title}`)
}
function unload(ctx, title) {
console.log(ctx.element, `removed from DOM with ${title}`)
}
function afterupdate(ctx, title) {
console.log(`updated with ${title}`)
}
function afterreorder(ctx, title) {
console.log(`reordered with ${title}`)
}Plugins
Plugins are middleware functions that are called just before the component is rendered or updated. A plugin can inspect the arguments, modify the context object or even return another context object that is to be used for rendering the component.
const html = require('bel')
const component = require('fun-component')
const greeter = component(function greeting(ctx, title) {
return html`<h1>Hello ${title}!</h1>`
})
greeter.use(function log(ctx, title) {
console.log(`Rendering ${ctx._ncID} with ${title}`)
return ctx
})
document.body.appendChild(greeter('world'))fun-component is bundled with with a handfull of plugins that cover the most common scenarios. Have you written a plugin you want featured in this list? Fork, add, and make a pull request!
- spawn – Spawn component contexts on demand and discard on unload.
- restate – Add state object and state management to the context object.
- logger – Add a logger (using nanologger) to the context object.
- cache – Cache element and reuse on consecutive mounts.
Examples
For example implementations, see /examples. Either spin them up locally or visit the link.
- Mapbox (using cache)
npm run example:mapbox- https://fun-component-mapbox.now.sh
- List (using spawn)
npm run example:list- https://fun-component-list.now.sh
- Expandable (using spawn and restate)
npm run example:expandable- https://fun-component-expandable.now.sh
Why tho?
Authoring a component should be as easy as writing a function. Using arguments and scope to handle state is both implicit and transparent. Worrying about calling context, and stashing things on this makes for cognitive overhead.
Not for you? If object oriented programming is your thing, use nanocomponent. If you're more into events, maybe microcomponent is a good fit.