Reinforcement Learning in Diffblue Cover