On May 19, 1:24 pm, David Harvey <dmhar...@cims.nyu.edu> wrote:
> On May 19, 11:17 pm, David Harvey <dmhar...@cims.nyu.edu> wrote:
<SNIP>
> Sorry, in case (1), I should have said, there is no easy fix from
> within zn_poly itself. If Sage is not supposed to be compiling with
> unsigned long != mp_limb_t, then if you can figure out how to make it
> use unsigned long == mp_limb_t, everything should work fine.
Well, in this case it is completely ironic that the zn_poly 0.9 code
in FLINT is compiled, but not used since it causes a doctest failure
in the Monsky code. When using only FLINT it passes the doctest. The
failure is a different one compared to when FLINT shipped with zn_poly
0.8, but there are two possibilities:
(a) the glue code from FLINT to zn_poly or some other change Bill
might have made to zn_poly inside FLINT causes the bug
(b) zn_poly 0.9 has an undetected bug.
At the moment it is unclear and it might even be possible that (a)
*and* (b) apply ;). Last time this happened malb helped turn on some
debugging code that compared results from NTL and FLINT+zn_poly, but I
don't have the expertise or the time to get into this right now.
Anyway, even once the bug is fixed in FLINT we would still have to
deal with zn_poly. Fixing it in MPIR and/or GMP on PPC 64 would be the
way to go. I also believe that this happen on OSX also, so it isn't
Linux specific.
> david
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---
