[boost] [speedsnail] Re: [python] cygwin-gcc config problem

From: Anthony Williams (anthony_w.geo_at_[hidden])
Date: 2007-02-20 12:15:51

Peter Bartlett <pete_at_[hidden]> writes:

> Douglas Gregor wrote:
>> |python|
>>  andreas_beyer:  gcc-cygwin-3.4.4
>>  [..]: gcc-cygwin-3.4.4
>
> All the gcc-cygwin python failures occur to a configuration problem  
> that occurs before Boost code is reached. (see e.g.  
> http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/output/speedsnail-gcc-d-boost-bin-v2-libs-python-test-andreas_beyer-test-gcc-cygwin-3-4-4-debug_release.html)

That's a speedsnail result, not a JustSoftwareSolutions one.

Roland's been running the cygwin regressions.

> PY_LONG_LONG is resolving as __int64 which is not a type on gcc/cygwin.
>
> Anthony, this is probably because you are using a Windows-specific  
> pyconfig.h (this would happen by default for example if you point gcc  
> at the Windows MSI version of Python for the Python includes). If your  
> "C:/Programme/Python25/include/pyconfig.h" has a message about being  
> Windows-specific at the top, then this is the case.
>
> As a hack to get you further you can add -D__int64="long long" to your  
> cxxflags, however there are likely to be other issues. If my  
> hypothesis as to the cause is correct, then a better route would be to  
> make sure the cygwin python package is installed and make sure cygwin  
> is pointing at /path_to_cygwin/usr/include/python2.4/ for the includes.
>
> If my hypothesis is not correct, then I'm sorry for the noise.

Anthony
-- 
Anthony Williams
Just Software Solutions Ltd - http://www.justsoftwaresolutions.co.uk
Registered in England, Company Number 5478976.
Registered Office: 15 Carrallack Mews, St Just, Cornwall, TR19 7UL