[Gretl-users] pre-release testing for gretl 1.8.0

[Allin Cottrell]

On Wed, 14 Jan 2009, Yuniarto Hadiwibowo wrote:

> If the data contains variable named i, I get the error messages.
> It seems to me that the loop variable i is mixed up with
> variable i in the data.
>
> I get this problem only with cvs I downloaded in Jan 13. Using
> 1.7.9 or cvs in December, the script runs OK.

Thanks for the clarification, now I see the point.  The new
behavior is not a a bug, it's a bug-fix!  If you have a series
named "i" in the dataset, you can't also define a scalar named
"i", even if only implicitly, as the index variable for a loop.
The old behavior was sloppy and liable to lead to errors.

(I got no error because I ran your script using the longley
dataset, which does not contain a series "i" although it does
contain series with the names used in your ADF tests.)

Allin.