On the dependence of the solutions and optimal solutions of control problems on the control constraint set