[Gretl-users] Something wrong with sample rectriction in panel data