Re: [Gretl-users] Nonlinear switching regression vs Loop constructs