Convergence Properties of Proximal (Sub)gradient Methods without Convexity or Smoothness of Any of the Functions