Re: [Gretl-devel] small bug with const series

On Thu, 20 Mar 2014, Ignacio Diaz-Emparanza wrote: > A colleague here tried to do this (please don't ask why): > > series c1 = const - 0.9*const(-1) We don't need to ask -- whoever dreamt this up either has a severe comprehension problem or just wanted to see what would happen when you torture gretl ;-) > obtaining this strange result: > > c1 > > 1959 > 1960 0.1000 > 1961 0.9100000 > 1962 0.1810000 > 1963 0.8371000 > 1964 0.2466100 > 1965 0.7780510 > 1966 0.2997541 > 1967 0.7302213 > 1968 0.3428008 > 1969 0.6914793 Not only that, but after this exercise "const" is messed up. In this instance gretl did not pass the torture test. We'll fix that shortly.