wholmgren
commented
9 years ago Also struggling with conda's scipy 0.15 on Python 3.4 on linux. Edited history below:
$ rm -rf ~/miniconda3/pkgs/numpy-1.*
$ rm -rf ~/miniconda3/pkgs/scipy-0.1*
$ conda create -n scipytest3 python=3.4
$ source activate scipytest3
$ conda install -f scipy numpy
The following packages will be downloaded:
package | build
---------------------------|-----------------
numpy-1.9.2 | py34_0 7.7 MB
scipy-0.15.1 | np19py34_0 36.6 MB
------------------------------------------------------------
Total: 44.2 MB
$ python -c "from scipy.special import lambertw; print(lambertw(1))"
(nan+0j)
$ conda create -n scipytest2 python=2.7
$ source activate scipytest2
$ conda install -f scipy numpy
The following packages will be downloaded:
package | build
---------------------------|-----------------
numpy-1.9.2 | py27_0 7.8 MB
scipy-0.15.1 | np19py27_0 37.0 MB
------------------------------------------------------------
Total: 44.7 MB
$ python -c "from scipy.special import lambertw; print(lambertw(1))"
(0.56714329041+0j)Maybe related to #260 or conda #1331?
The scipy package was just updated to version 0.15.1 using Anaconda on a Linux server (CentOS 6.5) and a Linux workstation (Fedora 21). Seven tests failed (Python 3.4.2 x86_64). The same 7 tests failed for scipy version 0.14.0 and 0.15.0 with Python 3.4.2. The installation used is on both system is Miniconda.
For the same installation but using Python 2.7.9 instead, there were no failed test. No failed test for Python 3.4.2 on Windows either.
There are no failures when package is compile from source (i.e. installed using pip) on the same Fedora 21 workstation on Python 3.4.2 (compiled from source).
Here are the failures:
Thanks