当第三方库请求ExecutorService时，请使用Scala ExecutionContext

Question

I am using Play Framework (Scala version) with the Amazon AWS Java SDK to integrate Amazon S3 into an application.

The AWS SDK has a TransferManager class that provides an abstraction to manage a thread pool for handling download/uploads to S3.

I am trying to determine if it is possible to integrate the core support Play has for custom ExecutionContexts into this object provided by the SDK. In particular, when instantiating the TransferManager provided by the AWS SDK, you can supply a custom ExecutorService as an optional parameter.

Scala's ExecutionClass binds the ExecutorService class via the "with" keyword in its class declaration, so I am wondering if there is some mechanism to get an ExecutorService object from the ExecutionContext like a method that transforms ExecutionContext => ExecutorService.

If not, is there any other approach? At the moment I am just instantiating a custom ExecutorService directly in a class outside of Play's standard approach which is outlined here:

https://www.playframework.com/documentation/2.3.x/ThreadPools

This feels messy and against the conventions provided by the framework.

Thanks for your time.

Answer 1

If you create your context like this (do not copy-paste this blindly - it's configured for blocking operations):

val blockingContext: ExecutionContext = {
    val executor = new ThreadPoolExecutor(100, 100, 1, TimeUnit.MINUTES, new LinkedBlockingQueue(1000))
    executor.allowCoreThreadTimeOut(true)
    ExecutionContext.fromExecutorService(executor) // main part
}

Then you will be able to get ExecutorService instance from it:

val executor: ExecutorService = blockingContext.prepare().asInstanceOf[ExecutorService]

Answer 2

Some implementations of ExecutionContext implement the ExecutionContextExecutorService trait, which is a sub-interface of both ExecutionContext and ExecutorService. You might check to see if the ExecutionContext you get from Play/Akka is one of those.