README

@chriscdn/promise-semaphore

Limit or throttle the simultaneous execution of asynchronous code in separate iterations of the event loop.

Installing

Using npm:

$ npm install @chriscdn/promise-semaphore

Using yarn:

$ yarn add @chriscdn/promise-semaphore

API

Create an instance

const Semaphore = require('@chriscdn/promise-semaphore')
const semaphore = new Semaphore([maxConcurrent])

The maxConcurrent parameter is optional, and defaults to 1 (making it an exclusive lock or binary semaphore). Use an integer value greater than one to limit how many times the code block can be simultaneously executing from separate iterations of the event loop.

Acquire a lock

semaphore.acquire([key])

This returns a Promise, which resolves once a lock has been acquired. The key parameter is optional and permits the same Semaphore instance to be used in different contexts. See the second example.

Release a lock

semaphore.release([key])

The release call should be executed from a finally block (whether using promises or a try/catch block) to guarantee it gets called.

Example 1

const Semaphore = require('@chriscdn/promise-semaphore')
const semaphore = new Semaphore()

// using promises
semaphore.acquire()
    .then(() => {
        // This block executes once a lock has been acquired.  If already locked
        // then this block will wait and execute once all locks preceeding it have been
        // released.
    })
    .finally(() => {
        // release the lock permitting the next queued process to continue
        semaphore.release()
    })

// or, using async/await
await semaphore.acquire()

try {
    // do your stuff here
} finally {
    semaphore.release()
}

Example 2

Say you have an asynchronous function to download a file and cache it to disk:

async function downloadAndCache(url) {

    // cacheFilePath could be based on a hash of the url
    const cacheFilePath = getCacheFilePath(url)

    if (!await pathExists(cacheFilePath)) {
        await downloadToFile(url, cacheFilePath)
    }

    return cacheFilePath
}

This works until a process calls downloadAndCache() in short succession with the same url parameter. This can cause multiple simultaneous downloads that attempt to write to the same cached file.

This can be resolved with a Semaphore instance using the key parameter:

const Semaphore = require('@chriscdn/promise-semaphore')
const semaphore = new Semaphore()

async function downloadAndCache(url) {

    await semaphore.acquire(url)

    // This block continues once a lock on url is acquired.  This permits
    // multiple simulataneous downloads for unique url values.

    try {
        const cacheFileName = getCacheFilePath(url)

        if (!await pathExists(cacheFilePath)) {
            await downloadToFile(url, cacheFilePath)	
        }

        return cacheFilePath

    } finally {
        semaphore.release(url)
    }
}

License

MIT

Usage no npm install needed!