[issue4709] Mingw-w64 and python on windows x64

[Ralf Gommers]

Ralf Gommers added the comment:

> "We" is a lot of different companies and individuals. Anyone distributing prebuilt binaries is helping here, a few people are working on the licensing concerns for some components, other people are working on C BLAS libraries.

Note that we by default recommend to users to use a distribution like Anaconda/Canopy (for example at http://scipy.org/install.html). That's fine for many scientific users, but not for people that already have a Python stack installed or simply prefer to use pip for another reason.  So pre-built binaries like the ones in Anaconda/Canopy help, but don't solve the "make `pip install scipy` work" problem. And giving up on pip/PyPi would make no one happy...

> I see the issue approximately as "it's hard to install the scipy stack", which is broader than "Windows does not have a Free Fortran compiler" 

It's: "it's hard to install the scipy stack on Windows". On OS X and Linux it's really not that hard. On OS X, you can install all core packages with pip (there are binary wheels on PyPi). On Linux you can do that too after using your package manager to install a few things like BLAS/LAPACK and Python development headers.

And the lack of Windows wheels on PyPi is directly related to no free Fortran compiler.

> and allows for more solutions (apologies for putting words in your mouth, which is not my intent, though I have certainly seen a fixation on this one particular solution to the exclusion of other possibilities).

Much more effort has gone into pre-built binaries than into MinGW, as well as into other things that help but can't be a full solution like a C BLAS. And I haven't seen other solutions to "make the scipy stack pip-installable" that could work. So I have to disagree with "fixation".

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue4709>
_______________________________________