[Numpy-discussion] isnan and co: cleaning up

[David Cournapeau]

Andrew Dalke wrote:
> The following seems relevant in this thread.
>
> On a Mac, OS X 10.4, I just rebuilt from SVN and got
>
> compile options: '-Inumpy/core/src -Inumpy/core/include -I/Library/ 
> Frameworks/Python.framework/Versions/2.5/include/python2.5 -c'
> gcc: _configtest.c
> _configtest.c: In function 'main':
> _configtest.c:4: error: 'isnan' undeclared (first use in this function)
> _configtest.c:4: error: (Each undeclared identifier is reported only  
> once
> _configtest.c:4: error: for each function it appears in.)
> _configtest.c: In function 'main':
> _configtest.c:4: error: 'isnan' undeclared (first use in this function)
> _configtest.c:4: error: (Each undeclared identifier is reported only  
> once
> _configtest.c:4: error: for each function it appears in.)
> lipo: can't figure out the architecture type of: /var/tmp//ccTy1pzZ.out
> _configtest.c: In function 'main':
> _configtest.c:4: error: 'isnan' undeclared (first use in this function)
> _configtest.c:4: error: (Each undeclared identifier is reported only  
> once
> _configtest.c:4: error: for each function it appears in.)
> _configtest.c: In function 'main':
> _configtest.c:4: error: 'isnan' undeclared (first use in this function)
> _configtest.c:4: error: (Each undeclared identifier is reported only  
> once
> _configtest.c:4: error: for each function it appears in.)
> lipo: can't figure out the architecture type of: /var/tmp//ccTy1pzZ.out
> failure.
> removing: _configtest.c _configtest.o
>
> followed by a failure for "isinf".
>
> It kept on going and finished the compile cleanly, so I don't know if  
> this is an informational warning or an indication of a suppressed  
> error that shouldn't have been suppressed.  Should that error message  
> be there?  In my understanding, errors are bad.

It is not an error for the build, but an error at the configuration
stage. To get some informations about the platform, we do compile some
code snippets, and do our configuration depending on whether they fail
or not. We could log those in a different location than the actual
build, but I am not sure anyone wants to spend some time on it.

>
> I tried looking for where that test as being done, but it seems to be  
> somewhere in SCons (in numpy/core/SConscript) and I figured it would  
> be easier than ask than to root around.

It has nothing to do with scons, since the _configtest.c are generated
by the distutils build. With scons, the errors during configuration
stage logged in a separate file (which is one advantage of scons,
actually, since it does the work for us).

cheers,

David