On Wed, Mar 10, 2010 at 08:18:40PM +, David Loeffler wrote:
> So it's a Pari bug. I see. I guess I was misled by the fact that the
> "trace" command only reports calls to Python functions. Thanks for
> clearing that up.
Thanks indeed! I feel better, though still curious why this involved
compu
So it's a Pari bug. I see. I guess I was misled by the fact that the
"trace" command only reports calls to Python functions. Thanks for
clearing that up.
David
On 10 March 2010 18:53, Robert Bradshaw wrote:
> On Mar 10, 2010, at 10:36 AM, daveloeffler wrote:
>
>> PS: I've realised that one can t
On Wed, Mar 10, 2010 at 10:53 AM, Robert Bradshaw
wrote:
> The solution is to upgrade pari, but for the moment I hacked around this
> pari bug. Maybe I should clean that up and submit a patch if pari 2.3.5
> isn't going to happen right away.
I think that 2.3.5 should be pretty straightforward.
-
On Mar 10, 2010, at 10:36 AM, daveloeffler wrote:
PS: I've realised that one can trigger more or less the same bug by
doing
sage: x = QQbar(-7).sqrt()
sage: y = QQbar(-7).sqrt()
sage: x == y
so embeddings of number fields aren't the problem, there's something
amiss with QQbar.
QQ(sqrt(-7)) i
