[NumPy-Tickets] [NumPy] #2049: Building 64-bit using VS2008 EE

[NumPy Trac]

#2049: Building 64-bit using VS2008 EE
-------------------------+--------------------------------------------------
 Reporter:  jusic        |       Owner:  somebody   
     Type:  enhancement  |      Status:  new        
 Priority:  normal       |   Milestone:  Unscheduled
Component:  Other        |     Version:  devel      
 Keywords:               |  
-------------------------+--------------------------------------------------
 Hi,

 I just managed to compile a 64-bit version of latest numpy git-master on
 Windows 7 Prof. 64-bit and VS2008 Express Edition + PSDK 6.1. I would like
 to provide some input for "Microsoft Toolchain Support" wiki page
 (http://projects.scipy.org/numpy/wiki/MicrosoftToolchainSupport), since
 it's not too specific on how this is done.

 I installed the following packages:

 * "Windows SDK for Windows Server 2008 and .NET Framework 3.5":[[BR]]
 - download link:
 http://www.microsoft.com/download/en/details.aspx?id=24826 [[BR]]
 - file: 6.0.6001.18000.367-KRMSDK_EN.iso

 * "Microsoft Visual Studio 2008 Express Edition"[[BR]]
 - download link: http://www.microsoft.com/download/en/details.aspx?id=3092
 [[BR]]
 - file: VS2008ExpressENUX1397868.iso

 * Python 2.7.2 (amd64) + distutils + all that is needed. C:\Python27\ is
 in %PATH%.

 To make amd64 compilers work, the batch file "C:\Program Files
 (x86)\Microsoft Visual Studio 9.0\VC\vcvarsall.bat" needs to be fixed on
 lines 19 and 20 to:

 {{{
 if not exist "%~dp0bin\vcvars64.bat" goto missing
 call "%~dp0bin\vcvars64.bat"
 }}}

 Then, to build numpy install package, open command prompt and cd to numpy-
 git directory, and run:

 {{{
 > "C:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\vcvarsall.bat"
 amd64
 > python setup.py config
 > python setup.py build
 > python setup.py bdist_wininst
 }}}

 The package seems to work even if I didn't have a Fortran compiler. I had
 MinGW installed but not in %PATH%.

 Thanks.

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/2049>
NumPy <http://projects.scipy.org/numpy>
My example project