[vtkusers] Problem building Universal Binary version of VTK on Mac OSX

Jens Thomas j.m.h.thomas at dl.ac.uk
Thu May 10 10:42:35 EDT 2007


Sean McBride wrote:
> On 2007-05-10 13:30, Jens Thomas said:
>
>   
>> Many thanks for the reply - I really appreciate it.
>>
>> Actually, it was rather remiss of me not to mention that I was building 
>> against python 2.5 which I'd built myself (configured with: 
>> --enable-universalsdk), not the Mac framework python which is 2.3.
>>
>> I've just tried building vtk against the framework python (using the 
>> same VTK source code and CMake binary as for the failing build) and it's 
>> worked fine - so it looks like it's something to do with the interaction 
>> between CMake and Python 2.5.
>>
>> I've attached a diff of the CMakeCache.txt files in case it sheds light 
>> on what's gone wrong (the first file on the command line was the 
>> framework vtk, the second the python 2.5).
>>     
>
> I'm afraid I don't have any advice then... I don't know python at all. 
> My knowledge is limited to being able to flip the VTK_WRAP_PYTHON
> boolean on and off. :)
>
> Have you searched the archives?  Does Python 2.5 work with VTK on other OSs?
>
>   
It does. I've built it many times on various flavours of Linux and even 
Solaris! I've got it to build fine under Darwin 7.9.0 on PPC too - it's 
just peculiar to the combination of python2.5 and the new Intel Mac that 
causes the problem.

Unfortunately, great as CMake can be, when things do go wrong, it's a 
complete nightmare trying to work out why: the Makefile where the build 
goes wrong (Common/CMakeFiles/vtkCommonPython.dir/build.make) is 
identical as far as I can make out between the build where it works and 
the one where it fails, so it looks like it's something CMake-specific 
that's foobar.

I guess I'll try asking on the CMake list.

Thanks again for your help.

Best wishes,

Jens



More information about the vtkusers mailing list