Re: This looks like a bug

[Bill Allombert <Bill.Allombert@math.u-bordeaux1.fr>]

On Thu, Jul 02, 2009 at 10:35:21AM -0400, Bill Daly wrote:
> Version 2.4.2 for Windows gives me this:
>
>    facq1 = nffactor(nfq1, q1);
>    rootsq1 = -subst(lift(facq1[,1]),x,0);

There is a function nfroots that does that a bit faster.
(and also looks at nfgaloisconj).

> 'facq1' is a 24*2 matrix (the output of nffactor), but 'rootsq1', which  
> should be t_COL [24,1] is instead t_VEC [1,24]. This is in the middle of  
> a longish piece of code, so it wouldn't surprise me if it's sporadic.  
> (It's easy enough to work around, but it looks sort of dumb.)

This was a bug in nffactor (it returns an invalid object). It seems to
be fixed in the SVN repository.

Cheers,
Bill.