On Jan 15, 4:22 pm, mabshoff <mabsh...@googlemail.com> wrote:
> Hi,
<SNIP>
> Anyway, please give it a spin and report back issues. It should be
> noted that this build is not ready for prime time!
A couple more remarks that I forgot to mention and just realized once
I looked at my OSX 64 bit port page:
* There are about half a dozen components missing, i.e. R, rpy,
flintqs, cvxopt and so on. I have build fixes for all of them, but I
was testing libSingular and eMPIRe with that build
* I am using gfortran 4.2.3 to build the Fortran components and this
seems to introduce issues with numpy/scipy since a lot of the
segfaults seem related to it. Having numerous segfaults on Solaris go
away by switching from gfortran 4.2.3 to 4.3.2 I think I will also go
this route on OSX, i.e. build my toolchain completely from scratch.
But this does require a bit of work since for example Python on OSX
uses *hard coded* flags in configure that are Apple gcc specific. This
is fixed in Python 2.6, but I can easily patch configure so Python
builds with gcc 4.3.2
* As mentioned above this build uses eMPIRe instead of gmp, so xgcd &
friends return different results which does cause about 12 doctests
failures even though the output is identical mathematically. Since we
are switching to it in 3.3 anyway this problem will be gone soon.
* I posted this binary because rpw asked for a 64 bit build in IRC
and this was the fastest way to get it done. Since he is mainly
interested in linalg it should work well for him
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
-~----------~----~----~----~------~----~------~--~---
