Re: rnfkummer() regression

[Karim BELABAS <Karim.Belabas@math.u-psud.fr>]

On Thu, 30 Jan 2003, Igor Schein wrote:
> ? rnfkummer(bnrinit(bnf=bnfinit(quadpoly(689,y),1),2,1),Mat(2));
>   ***   incorrect type in hnf_invimage.
>
> The following change apears to have caused this regression:
>
> revision 1.593
> date: 2002/11/14 20:35:26;  author: karim;  state: Exp;  lines: +1 -0
> 28- new algorithm for conductor() [much faster, does not need bnr.gen]

Yes, I had not treated correctly the [trivial] case of trivial conductors
when class group is non-trivial.

Trivial fix. :-)

    Karim.

P.S: The _real_ underlying problem is that pari can represent 0 x n matrices
[ cgetg(n+1, t_MAT) + cgetg(1, t_COL) columns ], but not n x 0 matrices
[ uniformly treated by cgetg(1, t_MAT) ]. So a truly generic program may break
on degenerate cases, as here.
-- 
Karim Belabas                    Tel: (+33) (0)1 69 15 57 48
Dép. de Mathématiques, Bât. 425  Fax: (+33) (0)1 69 15 60 19
Université Paris-Sud             Email: Karim.Belabas@math.u-psud.fr
F-91405 Orsay (France)           http://www.math.u-psud.fr/~belabas/
--
PARI/GP Home Page: http://www.parigp-home.de/