[Pythonmac-SIG] Problems installing Numpy using Universal MacPython

[Robert Kern]

Ronald Oussoren wrote:
> On 27-mrt-2006, at 18:00, James Boyle wrote:
> 
> 
>>I posted this to the numpy mailing list - I am re-posting here  
>>since it is a more complete description of the situation than my  
>>previous post to this list.
>>
>>I am running OX 10.3.9. I installed python 2.4.2 using Universal  
>>MacPython 2.4.2.dmg.
>>I am on a PowerPC G4. The Universal build is designed to install on  
>>the Mac PPC and Intel machines.
>>
>>When I run the numpy 0.9.6 install it apparently fails on the  
>>configure. The whole trace back is at the end of this posting.
>>One thing that I noted  was the line:
>>
>>gcc options: '-arch ppc -arch i386 -isysroot /Developer/SDKs/ 
>>MacOSX10.4u.sdk -fno-strict-aliasing -Wno-long-double -no-cpp- 
>>precomp -mno-fused-madd -fno-common -dynamic -DNDEBUG -g'
>>
>>It appears that the configure routines think that I have two  
>>architectures - ppc and i386.
>>At my low level of understanding, all I can do is speculate that  
>>the Universal install generates some confusion as to the actual cpu.
>>Somewhere numpy is keying on information that is ambiguous in the  
>>universal install.
> 
> The problem is that gcc 3.3 on 10.3.9 doesn't support some of these  
> options. Specifcally it doesn't know '-isysroot' and also doesn't  
> support the i386 architecture. I'm pretty sure that I had implemented  
> support for stripping these arguments on a 10.3 system. Hopefully  
> I'll have time to boot into 10.3 tomorrow.

numpy extends distutils, so it may be overriding certain changes you've made to
distutils. We may have to duplicate those changes in numpy.distutils.

-- 
Robert Kern
robert.kern at gmail.com

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco