Time not tracked for adjoint with new differentiable physics

[tlroy]

DifferentiablePhysics does not track time history. So for e.g. in LiDO if your objective explicitly depends on time, the adjoint solve will be stuck at the final time.