[NumPy-Tickets] [NumPy] #2049: Building 64-bit using VS2008 EE
NumPy Trac
numpy-tickets@scipy....
Sun Feb 12 18:50:52 CST 2012
#2049: Building 64-bit using VS2008 EE
---------------------------+------------------------------------------------
Reporter: jusic | Owner: pv
Type: enhancement | Status: new
Priority: normal | Milestone: Unscheduled
Component: Documentation | Version: devel
Keywords: |
---------------------------+------------------------------------------------
Comment(by mwiebe):
I've done this too, but without installing VS2008 EE, the compiler is
already fully included in the Windows SDK. The notes I kept about my
process are as follows:
1. Install the Windows SDK 7.0 (the one with .NET 3.5), available as a
free download from Microsoft’s website. This will install the C++
compilers used by Visual Studio 2008.
2. Fix a bug in the file C:\Program Files\Microsoft
SDKs\Windows\v7.0\Bin\SetEnv.cmd, by editing the file in a text editor.
Make a backup of the file first. You may have to run the text editor as
Administrator, then open this file, because it is protected by the OS.
a) Find a section labeled “Prevent path duplication if...” which is using
environment variables called ORIGINALPATH, ORIGINALINCLUDE, and
ORIGINALLIB. Replace every “!” with “%” here, for example
“!ORIGINALINCLUDE!” becomes “%ORIGINALINCLUDE%”.
b) Throughout the file, replace occurrences of “!Path!” with “%Path%”.
3. Create a batch file in the main numpy source directory, called
build_winsdk.bat, with the following commands in it. This is for 64-bit
Windows on Windows 7. On 32-bit windows, you should replace the part “/x64
/win7” with “/x86 /xp”.:
{{{
CALL "C:\Program Files\Microsoft SDKs\Windows\v7.0\Bin\SetEnv.cmd"
/release /x64 /win7
set DISTUTILS_USE_SDK=1
set MSSdk=1
"C:\Python27\Python.exe" setup.py install
pause
}}}
4. Run the batch file, which will build numpy and then install it to your
Python 2.7.
--
Ticket URL: <http://projects.scipy.org/numpy/ticket/2049#comment:2>
NumPy <http://projects.scipy.org/numpy>
My example project
More information about the NumPy-Tickets
mailing list