Bill Allombert on Wed, 6 Nov 2002 10:57:45 +0100


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

Re: bug in bnrstark


On Tue, Nov 05, 2002 at 07:16:47PM +0100, Karim BELABAS wrote:
> On Tue, 5 Nov 2002, Igor Schein wrote:
> > On Tue, Nov 05, 2002 at 06:32:57PM +0100, Karim BELABAS wrote:
> >> On Tue, 5 Nov 2002, Bill Allombert wrote:
> >>> There is a bug in bnrstark:
> >>> ? allocatemem(2^25);
> >>> ? B=bnfinit(quadpoly(13,a));
> >>> ? C=bnrinit(B,409,1);
> >>> ? bnrstark(B,Mat([5]))
> >>>   ***   forbidden multiplication t_VEC * t_VEC.
> >>
> >> Assuming you meant
> >>
> >>   ? bnrstark(C,Mat([5]))
> >>
> >> in that last line [ I get "incorrect bigray field otherwise" ], I can't
> >> reproduce your problem in 2.2.5 (development CHANGES-1.580) [Linux + ix86].
> >> I get a stack overflow, with the specified stack size, and a (presumably
> >> correct) result otherwise in about 15 minutes (100MB stack).
> >
> > It's most likely seed-dependent.
> 
> Possibly.
> 
> By the way, rnfkummer(C, Mat([5])) succeeds in 0.5s, default stack.

Well I cannot reproduce this bug, but I found out the hard disk was full, and
so the build has failed and gp was not CVS current, so it may be an old bug
already fixed.

Cheers,
Bill.