Bill Allombert on Sun, 09 May 2004 18:44:33 +0200


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: nffactormod


On Sun, May 09, 2004 at 05:10:09PM +0200, Elie_CALI wrote:
>  I use the nffactormod function on a PC with the version 2.2.6 of Pari.
>  It works well, but when I send program to a friend who has a Unix version
> (I
>  don't know the exact version but a quite recent one), I get a bug report.
>  The problem is an "unexpected charcater" problem. Indeed, I try to compute
> :
> 
>  nffactormod(bnf,p,id)[1][1],
> 
>  which works well with my version, while the Unix version seems to expect
>  something like :
> 
>  nffactormod(bnf,p,id)[1,1].

The correct way is nffactormod(bnf,p,id)[1,1]. Unfortunately there was a
bug in the developement version of PARI that caused nffactormod to
return a vector of vector instead of a matrix. In particular that bug
affect 2.2.6 and 2.2.7. This bug was fixed recently in the CVS branch:

   40- return type for nffactormod was not a factorization (had become
   t_VEC)
   
>  Do I do something wrong, or can I get a Pari version for my PC which works
>  the same as the Unix one?

Though there is a more recent Pari executable for Windows, (pari 2.2.7)
here: http://pari.math.u-bordeaux.fr/download/Pari.exe, it will not 
fix this problem.

Cheers,
Bill.