Re: regression problems

[Jason Stover]

On Wed, Feb 08, 2006 at 10:31:12PM +0800, John Darrington wrote:
> I'm having a hard time getting the regression command to do anything
> sensible.  Below is  a simple example from a trusty spss for dummies book
> I have.
> 
> The first regression command quits with the error "memory exhausted"
> (note to self: should catch this and call custom handler).
> The second two both crash with different errors.  One with an
> assertion failure the other a plain segmentation fault.
> 
> J'
> 
> 
> DATA LIST LIST /sales * price * space *.
> BEGIN DATA.
> 15.00     2.10     1.00 
> 15.00     1.80     1.00 
> 21.00     2.20     1.00 
> 28.00     2.40     2.00 
> 30.00     2.50     2.00 
> 35.00     2.50     2.00 
> 40.00     2.60     2.00 
> 35.00     3.40     3.00 
> 30.00     2.50     3.00 
> 45.00     3.80     3.00 
> 50.00     4.40     4.00 
> 60.00     5.10     4.00 
> 45.00     3.90     5.00 
> 60.00     5.40     5.00 
> 50.00     5.50     6.00 
> END DATA.
> 
> REGRESSION 
>       /DEPENDENT sales .
> 
> REGRESSION /VARIABLES space
>        /DEPENDENT sales .
> 
> REGRESSION /VARIABLES price space
>        /DEPENDENT sales .
> 

At least one bug shows itself because of a lack of a VARIABLES
keyword, which causes n_variables to be 0 in run_regression(). When I fix
this stuff, should I just send you a patch to avoid any checkins?
(I may not be able to get to it for a couple of days.)

-Jason