@ksryy/parallel-async

parallel-async

Moved to @ksryy/parallel-async due to OTP misconfiguration (may move back in a few days)

A zero-dependency async parallel task runner. Powered by Promise.

NOTE: This package is designed for both Webpack and Node.js usage, but if you are using a enviroment that does not support Promise (such as IE) you need to polyfill it by yourself.

Install

npm install @ksryy/parallel-async

Then import it use either CommonJS:

const parallel = require('parallel-async');

or ESModule:

import parallel from 'parallel-async'

Usage (more of a mixture of guideline and design idea)

First you need to pass two parameters to the parallel function. One object and one callback function.

The object contains two things: the task list (property tasks) and a optional parameter that will be passed to the outer layer function.

The tasks property should be an Array. It contains the task that you want to run.

You can see that I emphasized the "outer layer function". This library is powered by Promise, so the task should orginally recieve only two parameter, resolve and reject as in Using promises at MDN, but what if you want to give your task something from outside that only works in the task function's scope? The solution is to wrap your function that will return a function that actually become the Promise. The outer layer will recieve that thing as a parameter, and then the outer layer will retern the inner layer function that actually executes in the promise. Because of closure, the inner function can still access the variables in the outer function's scope. Here's a example:

functon eg(param) {
  return function (resolve,reject) {}
}

The returned function is what actually run in the Promise.

The only way to pass in the task list is to use an Array now, because I no longer use for...of loops for better polyfillability

There is also something tricky about the callback function that you need to provide. The callback function won't be called more than once now, but if an error occured, all results of previously runned tasks will be discarded, no matter they successed or not.