The Book of Ruby

Fibers

Ruby 1.9 introduces a new class called a Fiber, which is a bit like a thread and a bit like a block. Fibers are intended for the implementation of “lightweight concurrency.” This broadly means they operate like blocks (see Chapter 10) whose execution can be paused and restarted just as you can with threads. Unlike threads, however, the execution of fibers is not scheduled by the Ruby virtual machine; it has to be controlled explicitly by the programmer. Another difference between threads and fibers is that threads run automatically when they are created; fibers do not. To start a fiber, you must call its resume method. To yield control to code outside the fiber, you must call the yield method.

Let’s look at some simple examples:

fiber_test.rb ...