Detecting numpy vs micronumpy
Dear all, I tried asking this on the NumPy mailing list, but realise here is more likely: http://mail.scipy.org/pipermail/numpy-discussion/2011-September/058439.html How should a python script (e.g. setup.py) distinguish between real numpy and micronumpy? Or should I instead be looking to distinguish PyPy versus another Python implementation? Thanks, Peter
I think, for the time being, the appropriate solution is to just check the Python version, the original NumPy doesn't run on PyPy so it should be fine. Alex On Thu, Sep 15, 2011 at 5:00 PM, Peter Cock <p.j.a.cock@googlemail.com>wrote:
Dear all,
I tried asking this on the NumPy mailing list, but realise here is more likely: http://mail.scipy.org/pipermail/numpy-discussion/2011-September/058439.html
How should a python script (e.g. setup.py) distinguish between real numpy and micronumpy? Or should I instead be looking to distinguish PyPy versus another Python implementation?
Thanks,
Peter _______________________________________________ pypy-dev mailing list pypy-dev@python.org http://mail.python.org/mailman/listinfo/pypy-dev
-- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero
On Thu, Sep 15, 2011 at 10:02 PM, Alex Gaynor <alex.gaynor@gmail.com> wrote:
I think, for the time being, the appropriate solution is to just check the Python version, the original NumPy doesn't run on PyPy so it should be fine. Alex
How precisely? The problem I am running into is that "import numpy" appears to work under PyPy 1.6 (you get micronumpy) but later things like numpy.get_include() don't work (AttributeError). Should I just treat that exception itself as meaning it is micronumpy not real numpy? Thanks, Peter
On Thu, Sep 15, 2011 at 5:55 PM, Peter Cock <p.j.a.cock@googlemail.com>wrote:
I think, for the time being, the appropriate solution is to just check
On Thu, Sep 15, 2011 at 10:02 PM, Alex Gaynor <alex.gaynor@gmail.com> wrote: the
Python version, the original NumPy doesn't run on PyPy so it should be fine. Alex
How precisely?
The problem I am running into is that "import numpy" appears to work under PyPy 1.6 (you get micronumpy) but later things like numpy.get_include() don't work (AttributeError). Should I just treat that exception itself as meaning it is micronumpy not real numpy?
Thanks,
Peter
Well, until we implement it anyways :) That's why I think something like "import platform; platform.python_implementation == 'PyPy'" is a godo way to check. Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero
On Thu, Sep 15, 2011 at 11:04 PM, Alex Gaynor <alex.gaynor@gmail.com> wrote:
The problem I am running into is that "import numpy" appears to work under PyPy 1.6 (you get micronumpy) but later things like numpy.get_include() don't work (AttributeError). Should I just treat that exception itself as meaning it is micronumpy not real numpy?
Thanks,
Peter
Well, until we implement it anyways :) That's why I think something like "import platform; platform.python_implementation == 'PyPy'" is a godo way to check. Alex
Thanks, I'll use that. Its a shame that wasn't in Python 2.5 though, my copy of Jython doesn't support it either. Peter
Hi, On Fri, Sep 16, 2011 at 11:48 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
Thanks, I'll use that. Its a shame that wasn't in Python 2.5 though, my copy of Jython doesn't support it either.
The older and more robust way to check this is: "__pypy__" in sys.builtin_module_names Armin
participants (3)
-
Alex Gaynor -
Armin Rigo -
Peter Cock