This is a fast, tiny and smart state management system. Based on simplest principles: you have a state and you can watch for the state changes. Was born during working on innet.

watch-state inspired by async-await pattern, you can image it like this:

state count = 0

watch {
  console.log(count)
}

Browser supports

Desktop

| Firefox | Chrome | Safari | Opera | Edge | |:-------:|:------:|:------:|:-----:|:----:| | 45+ | 49+ | 9+ | 36+ | 13+ |

Mobile

| Firefox | Chrome | Safari | Opera | |:-------:|:------:|:------:|:-----:| | 87+ | 90+ | 9+ | 62+ |

You can transpile it supporting old browsers, but the performance decreases.

Install

npm

npm i watch-state

yarn

yarn add watch-state

Use watchState to get any class from the library.

const {
  Watch,
  State,
  Cache,
} = watchState

Usage

Simple example:

You can create an instance of State and watch its value.

import { Watch, State } from 'watch-state'

const count = new State(0)

new Watch(() => console.log(count.value))
// console.log(0)

count.value++
// console.log(1)

count.value++
// console.log(2)

Update argument:

You can check if the watching ran first by update argument.

const count = new State(0)

new Watch(update => {
  console.log(update, count.value)
})
// console.log(false, 0)

count.value++
// console.log(true, 1)

count.value++
// console.log(true, 2)

As example, you can watch a state once

const count = new State(0)

new Watch(update => {
  
  if (!update) {

    // Watch this value
    count.value

  } else {

    // React on changes
    console.log('The value was changed')

  }

})

count.value++
// console.log('The value was changed')

count.value++
// nothing happenes

Force update of State

You can run watchers of a state with update method.

const count = new State(0)

new Watch(() => {
  console.log(count.value)
})
// console.log(0)

count.update()
// console.log(0)

Force update of Watch

You can run a watcher even when it's states are not updated.

const count = new State(0)

const watcher = new Watch(() => {
  console.log(count.value)
})
// console.log(0)

watcher.update()
// console.log(0)

destroy

You can stop watching by destroy method of Watch.

const count = new State(0)

const watcher = new Watch(() => {
  console.log(count.value)
})
// console.log(0)

count.value++
// console.log(1)

watcher.destroy()

count.value++
// nothing happens

onDestroy()

You can subscribe on destroy or update of watcher

const count = new State(0)
const watcher = new Watch(() => {
  console.log('count', count.value)
  // the order does not matter
  onDestroy(() => console.log('destructor'))
})
// console.log('count', 0)

count.value++
// console.log('destructor')
// console.log('count', 1)

watcher.destroy()
// console.log('destructor')

watcher.destroy()
count.value++
// nothing happens

Deep watch:

You can use Watch inside a watcher. Each watcher reacts on that states which used only inside it.

const watching = new State(true)
const state = new State(0)
let test = 0

new Watch(() => {
  test++
  if (watching.value) {
    new Watch(() => {
      console.log(state.value)
    })
  }
})
// console.log(0), test = 1

state.value++
// console.log(1), test = 1

watching.value = false
// test = 2

state.value++
// nothing happens

Cache:

You can cache computed state.
The watcher will not be triggered while new result is the same.

const name = new State('Foo')
const surname = new State('Bar')

const fullName = new Cache(() => (
  `${name.value} ${surname.value[0]}`
))

new Watch(() => {
  console.log(fullName.value)
})
// console.log('Foo B')

surname.value = 'Baz'
// nothing happens

surname.value = 'Quux'
// console.log('Foo Q')

You can force update the cache by update method.

fullName.update()
// console.log('Foo Q')

Cache will be immediately updated only if a watcher looks after the cache.

You can use destroy and onDestroy like you do it on a watcher.

fullName.destroy()

The computing will be triggered only when a state inside the cache will be changed. So you can modify data only when it's needed.

const list = new State(['foo', 'bar', 'baz'])

const sortedList = new Cache(() => {
  console.log('computing')
  return [...list.value].sort()
})
// nothing happens

const value = sortedList.value
// console.log('computing')

console.log(sortedList.value)
// console.log(['bar', 'baz', 'foo'])

console.log(value === sortedList.value)
// console.log(true)

list.value = ['b', 'c', 'a']
// nothing happens

console.log(sortedList.value)
// console.log('computing')
// console.log(['a', 'b', 'c'])

createEvent

You can create event function with createEvent

import { State, createEvent } from 'watch-state'

const count = new State(0)
const increase = createEvent(() => {
  console.log(count.value++)
})

new Watch(() => console.log(count.value))
// console.log(0)

increase()
// console.log(1)

increase()
// console.log(2)

Typescript:

Generic of State

const key = new State<string | number>()

key.value = false
// error, you can use only string or number

Generic of Cache

new Cache<string>(() => false)
// error, target of cache should return string

Performance

You can check a performance test with MobX, Effector, Storeon, Mazzard and Redux. Clone the repo, install packages and run this command

npm run speed

Links

You can find more tools here

Issues

If you find a bug or have a suggestion, please file an issue on GitHub