Dave,

It looks more like a linker bug/feature --- the same behaviour on
skynet's eno (Linux), but
no errors on skynet's mark (Solaris, which I believe uses Sun's
linker).

Dima


On Aug 2, 3:08 pm, "Dr. David Kirkby" <david.kir...@onetel.net> wrote:
> On 08/ 2/10 12:48 PM, Dima Pasechnik wrote:
> > Running sage's python on a python file gives ImportError:
>
> > sage subshell$ sage -python lp.py
> > Traceback (most recent call last):
> >    File "lp.py", line 3, in<module>
> >      from cvxopt import matrix
> >    File "/home/dima/sage/sage-4.5.1/local/lib/python2.6/site-packages/
> > cvxopt/__init__.py", line 30, in<module>
> >      import base
> > ImportError: /home/dima/sage/sage-4.5.1/local/lib/liblapack.so:
> > undefined symbol: __powidf2
>
> > --------------------------
> > on the other hand, load'ing this file in sage happily works:
>
> > sage subshell$ sage
> > ----------------------------------------------------------------------
> > | Sage Version 4.5.1, Release Date: 2010-07-19                       |
> > | Type notebook() for the GUI, and license() for information.        |
> > ----------------------------------------------------------------------
> > sage: load('lp.py')
> >       pcost       dcost       gap    pres   dres   k/t
> >   0: -8.1000e+00 -1.8300e+01  4e+00  0e+00  8e-01  1e+00
> > [....]
>
> > This all happens on skynet's taurus:
>
> > sage subshell$ uname -a
> > Linux taurus 2.6.32.12-115.fc12.x86_64 #1 SMP Fri Apr 30 19:46:25 UTC
> > 2010 x86_64 x86_64 x86_64 GNU/Linux
>
> > with Sage compiled using gcc-4.5.0
>
> > I was unable to reproduce this on systems that run gcc-4.x with x<5.
>
> > Any ideas where to look?
>
> Have you tried gcc 4.5.1? It could be a new bug in 4.5.0, which if you are 
> lucky
> will be corrected in 4.5.1.
>
> Note that wheres 4.5.0 had extra features compared to the 4.4 series, 4.5.1 is
> just a bug-fix release on 4.5.0, so it should have fewer bugs.
>
> Dave

-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to