X-IL: Exploring the Design Space of Imitation Learning Policies