Environments for Lifelong Reinforcement Learning