A thread-pool abstraction for ES6 async operations
npm install threadpoolbash
npm install --save threadpool
`
Usage
Sort notation:
`javascript
// Fetch the dev names of all GitHub projects,
// executing no more than 10 active requests at any given time
let devNames = new Set()
let range = [...Array(100000).keys()] // (a neat ES6 trick for getting num ranges)
await ThreadPool.run(10, range.map(n => fetch(https://api.github.com/repositories?since=${n})
.then(res => res.json())
.then(json => devNames.add(json.name))))
console.log(Found ${devNames.size} distinct developers)
`
Full notation (allowes queueing, executing and waiting completion in different places):
`javascript
// Fetch the metadata of all GitHub projects,
// executing no more than 10 active requests at any given time
let devNames = new Set()
// init the ThreadPool supplying it the max number of parallel executions.
// an options object can also be supplied with a 'errorHandler' function,
// an error thrown from this function will halt ThreadPool execution
let tp = new ThreadPool(10, { errorHandler: err => { throw err } })
for (let i = 0; i < 10000000; i += 100) {
// queues sync / async functions or Promises (they don't start running yet)
tp.queue(async () => {
await fetch(https://api.github.com/repositories?since=${i})
.then((res) => res.json())
.then((json) => devNames.add(json.name))
})
}
// starts execution (queueing new tasks still possible)
tp.run()
// closes the ThreadPool for the queueing of new tasks (so it's completion can be awaited)
tp.close()
// blocks untill all queued tasks completed
try {
await tp.waitComplete()
} catch(err) {
// an error from a task, rethrown by 'errorHandler' can be handled here
console.error(Uncaught error, err)
}
console.log(Found ${devNames.size} distinct developers)
``