This is yet another DCPU-16 bytecode interpreter written in Ruby.

I wrote this for fun over half a weekend, without taking hints from any other existing implementation, as part of my deliberate practice routine.

I just wanted to see what kind of design would emerge when tackling such a low-level problem while keeping a focus on testability and methods working at a single level of abstraction.

Although the interpreter is functional, this is just a toy project with no aim to compete against any of the fully featured packages out there.

The code is surely not as clean as I want it to be, parts of the VM are still under construction

• The external API will change (especially code loading and the ugly parameters to the run method)
• Instrumentation must to be better documented, perhaps provider an observer class skeletton
• Verify the correctness of the cycles calculations
• ...

The interpreted supports the following proprietary features:

This new non-basic opcode will invoke a user-defined hook in the interpreter (there can up to 0x40 of them).

In Ruby:

cpu = DCPU16::CPU.new
cpu.interrupts[0x12] = Proc.new { |cpu| cpu.regset(:X, 42) }

in DASM16:

  ; ...
  INT 0x12
  IFE X, 42   ; true, will perform the next instruction
  ; ...

It's still very crude at the moment, programs need to be input as arrays of words.

A simple test script running a couple of programs for a few hundred cycles:

ruby examples/test.rb

You know the deal:

1. bundle install
2. bundle exec rspec

Copyright © 2012, Xavier Defrang

This code is released under the MIT license:

• www.opensource.org/licenses/MIT