> Can William or someone who gets this error (I don't!) test what
> happens if you replace
>
> G=E.change_ring(GF(p)).abelian_group()
>
> by
>
> N=E.change_ring(GF(p)).cardinality()
>
> ? The cardinality will be provided by pari, so if it is a pari
> problem that should still cause a problem. The abelian group code is
> mine, so I would expect that is more likely to be the issue (!) We
> used to have a function abelian_group_prime_field() which only worked
> over prime fields and used pari, which was deleted at some point, but
> in fact that function just called a gp script of mine so was not
> really inedpendent.
>
> Secondly, add "debug=True" to the abelian_group() call and send me the
> output.
>
> John
Hi John,
I can reproduce this bug 100%, and some time ago you and I (the thread
should be findable, I didn't search thoroughly) already went together
through this topic with the following result (it's all written down in
#4181):
The bug is triggered by:
sage: for p in prime_range(31300, 32600): EllipticCurve(GF(p),
[0,1,1,10,13])
The bug is *not* triggered by (although the prime range is more than
an order of magnitude larger!):
sage: for p in prime_range(32768, 100000): EllipticCurve(GF(p),
[0,1,1,10,13])
So it seems very unlikely that the root cause is anything in
"arithmetic" parts of the code.
It is something very strange and deeply buried in the system, somehow
memory allocation / heap / whatever related. My thought back then was,
that it must be something with 16-Bit signed integers on Mac OS X / 32
Bit.
I'll investigate further as soon as I have access to a system with a
Sage install again.
Cheers,
gsw
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
