[Numpy-discussion] Error while upgrading to numpy 1.4.1

Benjamin Root ben.root@ou....
Mon Jun 14 09:39:50 CDT 2010


On Mon, Jun 14, 2010 at 1:09 AM, David Cournapeau <cournape@gmail.com>wrote:

> On Mon, Jun 14, 2010 at 3:00 PM, Vishal Rana <ranavishal@gmail.com> wrote:
> > Hi,
> > I am getting this error while upgrading to numpy 1.4.1 from 1.3 on ubuntu
> > 10.04 using easy_install -U numpy command, I even tried downloading /
> > installing numpy 1.4.1 but no luck, help appreciated.
>
> First, do not use easy_install to upgrade the numpy included with
> ubuntu. You should never do that, because easy_install will break the
> state of your installed packages w.r.t. apt database.
>
> It looks like you do not have development tools installed (the crt.o
> is a C runtime file which is provided by gcc) - install gcc, gfortran
> and libatlas-base-dev.
>
> I would also strongly discourage you to use easy_install on 10.04. The
> version in ubuntu 10.04 uses distribute which is buggier than
> setuptools at this point (sic),
>
> cheers,
>
> David
>

Vishal, I would like to second David's sentiments.  easy_install for numpy
is not done very well and can cause a lot of issues.  You will need to make
sure you uninstall numpy using synaptic, and also see if you can find any
partial installs in the /usr/lib/python2.6/site-distribution folder, and
even the /usr/local/lib/python2.6/site-distribution folder and remove them
using rm -rf.

With a cleaned-out environment, and the previously mentioned packages (you
might need a couple more...) you can then build from source using 'sudo
python setup.py install' in the numpy directory.  Be sure to run the tests
after the install to ensure that everything works properly.  I had to redo
the build a few times before the tests finished successfully for me.

Ben Root


> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion@scipy.org
> http://mail.scipy.org/mailman/listinfo/numpy-discussion
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/numpy-discussion/attachments/20100614/ab35a853/attachment.html 


More information about the NumPy-Discussion mailing list