[SciPy-user] installation problems on SUSE 10.2

[Eike Welk]

On Monday 29 June 2009, John Schulman wrote:
> I'm going to post my solution, because it's a really easy general
> solution to these installation problems, and it isn't mentioned on
> the scipy webpages.
>
> I downloaded and built sage from source. All it requires is a c/c++
> compiler. http://www.sagemath.org/download-source.html
>
> This took about 3 hours.
>
> This gives me a totally self-contained local python installation
> with scipy in it.
> It also has ipython, easy_install, and all that other nice stuff in
> the bin directory.
>
> When I log on, I go to the sage directory and type in the command
> ./sage -sh
> This sets all the shell variables so the executables in the sage
> bin directory becomes default, so I just have to type 'ipython' and
> I've got my sage ipython.

Thanks for telling us an alternative, and painless way to install 
Numpy/Scipy/Matplotlib on Suse Linux. Could you put a description of 
the installation process into Scipy's Wiki, please?

The "science" repository mentioned in the Wiki worked when I installed 
openSuse 11.0 on my computer, about a year ago. (Both repositories 
that you talked about, 'science' and 'ScientificLinux', don't contain 
packages for your operating system, Suse 10.2, anymore. Therefore 
these repositories are of little help for you.)

That there are now two repositories for Numpy/Scipy and also  
David Cournapeau's repository for Atlas is a bit confusing. Some time 
ago I have E-mailed some of the maintainers and proposed that they 
merge these repositories, but unfortunately it has not happend.

I wrote the section about installing Numpy/Scipy/Matplotlib on Suse 
when I had trouble installing them myself. The wording has to be 
updated somehow, since there are no longer packages for Suse 10.2 at 
the mentioned locations.

Kind regards,
Eike.