On May 19, 1:17 pm, David Harvey <dmhar...@cims.nyu.edu> wrote:
> On May 19, 1:58 pm, Ralf-Philipp Weinmann <crypto....@gmail.com>
> wrote:
>
> > zn_poly/pack.c:86:2: error: #error Not nails-safe yet
> > zn_poly/pack.c:168:2: error: #error Not nails-safe yet
> > zn_poly/pack.c:252:2: error: #error Not nails-safe yet
> > zn_poly/pack.c:351:2: error: #error Not nails-safe yet
> > zn_poly/pack.c:433:2: error: #error Not nails-safe yet
>
> > Is there any easy fix to this? Apparently it's compiling in 32-bit
> > mode. If you want full logs, I can put them up.
>
> This message indicates that zn_poly thinks either that GMP is compiled
> with nails support (which I doubt), or that unsigned long is a
> different width from mp_limb_t. I can think of several possible
> causes....
Ok.
> (1) Sage is actually compiling with unsigned long !=
> mp_limb_t, or (2) GMP/MPIR is not defining things like GMP_NUMB_BITS
> correctly, etc. In case (1), there is no easy fix (the next version of
> zn_poly will be able to handle mp_limb_t != unsigned long, but won't
> be ready for release for a few months yet). In case (2), hopefully the
> MPIR guys can help debug.
Note that GMP as well as MPIR on ppc64 use ABI=mode64 (or something
along those lines) and it is not the exact same thing as ABI=64. This
caused various problems with pari for example as discussed on the GMP
mailing list. I believe there is a MPIR trac ticket to deal with this,
so we will see how it plays out. As is Sage only works on PPC 32 and
not PPC64 since for example also pari will not work correctly due to
the above bug.
> 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
-~----------~----~----~----~------~----~------~--~---
