[SciPy-User] NumPy and SciPy test failures

[Mads M. Hansen]

2012/2/1 Ralf Gommers <ralf.gommers at googlemail.com>:
>
>
> On Wed, Feb 1, 2012 at 5:17 PM, Mads M. Hansen <madsmh at gmail.com> wrote:
>>
>> 2012/2/1 Denis Laxalde <dlaxalde at gmail.com>:
>> > Mads M. Hansen wrote:
>> >> I have built NumPy 1.6.1 and SciPy 0.10.0 for Python 3.2 on a Fedora
>> >> 16 system and I used gfortran, but when I run the tests I get the
>> >> following failures and errors
>> >
>> > It's probably not scipy 0.10.0. Could you specify the exact versions of
>> > you have installed (e.g. from the header displayed by scipy tests)?
>> >
>> Here is the header
>>
>> >>> scipy.test('full')
>> Running unit tests for scipy
>> NumPy version 1.6.1
>> NumPy is installed in /usr/lib64/python3.2/site-packages/numpy
>> SciPy version 0.10.0
>> SciPy is installed in /usr/lib64/python3.2/site-packages/scipy
>> Python version 3.2.1 (default, Jul 11 2011, 18:54:42) [GCC 4.6.1
>> 20110627 (Red Hat 4.6.1-1)]
>> nose version 1.1.2
>>
>> I checked out the v.0.10.0 tag from the Git repository.
>>
> The reason you're seeing the _minimize_neldermead failure, and probably some
> others, is likely that you didn't clean the install dir before installing
> 0.10.0. That test was only added after 0.10.0 came out.
>
> Ralf

Hm that's odd, I delete /usr/lib64/python3.2/site-packages/scipy (and
numpy/) prior to each reinstall - is scipy installed in
other places by defaultr?