Use "gfloat" and "gdouble", rather than "float" and "double", as the