On 05/10/09 21:03, mabshoff wrote:
>>   1. Why isn't --enable-framework turned on by default for the sage
>> python build on the Mac?  If this isn't a hassle can you please consider
>> making this the default?
> 
> It caused some issues in the past and we would have to special case
> certain bits of Sage's build system to deal with OSX specific settings
> with frameworks. None of the components in Sage require Python as a
> framework and IIRC it caused relocation problems, but those might be
> fixed now.
> 
> I don't understand why VTK needs Python to be a framework. Googling
> around for 5 minutes did not reveal anything useful. Can you give me
> some pointers and/or documentation on this?

I really don't know if this is needed or not.  It could well be a VTK 
issue.  I need to build VTK with Carbon support and I'm not sure that 
requires a framework build of Python. Anyway, VTK tries to detect the 
Python library to link to when it builds the Python wrappers.  I think 
it finds the system Python by default which obviously will not work. 
When I point it to the libpython25.a file in 
SAGE_LOCAL/lib/python2.5/config, I get link errors since my guess is 
that file is not a proper universal library?  Upon some digging, 
fortunately, I seem to have an install.log lying around.  Here is the 
exact error message I got when linking to libpython25.a:

Linking CXX shared library ../bin/libvtkCommonPythonD.dylib

ld: common symbols not allowed with MH_DYLIB output format with the 
-multi_module option
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(object.o)
 
definition of common _Py_DivisionWarningFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(pythonrun.o)
 
definition of common _Py_VerboseFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(pythonrun.o)
 
definition of common _Py_DebugFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(pythonrun.o)
 
definition of common _Py_FrozenFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(pythonrun.o)
 
definition of common _Py_IgnoreEnvironmentFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(pythonrun.o)
 
definition of common _Py_InteractiveFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(pythonrun.o)
 
definition of common _Py_NoSiteFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(parsetok.o)
 
definition of common _Py_TabcheckFlag (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(myreadline.o)
 
definition of common _PyOS_ReadlineFunctionPointer (size 16)
/Users/prabhu/src/build/sage-3.4/local/lib/python2.5/config/libpython2.5.a(myreadline.o)
 
definition of common __PyOS_ReadlineTState (size 16)
/usr/libexec/gcc/i686-apple-darwin8/4.0.1/libtool: internal link edit 
command failed

If I leave the Python library unspecified it runs into linker errors 
with symbols being undefined at link time.  None of this happens with a 
framework build.

I guess I could retry with a non-framework build but it really takes a 
while to build on my machine and all previous attempts failed.

> Another problem is that if we changed the default for OSX we would
> need to force a rebuild of every other python component in Sage, so
> since we don't have any problems as it is seems plainly not worth it.

But isn't everything anyways built from source?  So how would this 
affect anything? In my case nothing else seems affected -- I rebuilt 
Python and installed that over the existing build. I didn't test too 
much apart from VTK related scripts though.

cheers,
prabhu




--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to 
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to