Description

A work queue is designed to coordinate work between a producer and a pool of worker threads. When some task needs to be performed, the producer adds an object containing the task routine to the work queue. Eventually, one of the worker threads removes the object from the work queue and executes the routine. If the work queue is empty, the worker thread will exit, if more jobs are added to the queue additional worker threads will be created up to the configured maximum.

Work queues are useful for several reasons:

• To easily perform tasks asynchronously and concurrently in your application;

• To let you focus on the work you actually want to perform without having to worry about the thread creation and management;

• To minimize overhead, by reusing previously constructed threads rather than creating new ones;

• To bound resource use, by setting a limit on the maximum number of simultaneously executing threads;

Requirements

JRuby

Usage

Install the gem:

gem install simple_work_queue

Run the code:

require 'rubygems'
require 'simple_work_queue'
wq = SimpleWorkQueue.new
wq.enqueue_b { puts "Hello from the SimpleWorkQueue" }
wq.join

Note that you generally want to bound the number of worker threads:

# Limit the maximum number of simultaneous worker threads
SimpleWorkQueue.new(10)

History

SimpleWorkQueue is a fork of WorkQueue by Miguel Fonseca. SimpleWorkQueue removes the time out and bounded task queue. The timeout created issues under the concurrency implementation in JRuby. The time out mechanism was replaced using a non-blocking queue from the java.util.concurrent library. If a worker thread attempts to remove a task from the task queue and it is empty it will exit the worker thread.