[Python-modules-team] Bug#527707: Bug#527707: python-numpy: FTBFS: _configtest.c:5: undefined reference to `exp'

[Sandro Tosi]

Hi all,

On Fri, Jul 3, 2009 at 11:35, Cyril Brulebois <kibi at debian.org> wrote:
> severity 527707 important
> retitle  527707 python-numpy: Error during build: _configtest.c:5: undefined reference to `exp'
> thanks
>
> Lucas Nussbaum <lucas at lucas-nussbaum.net> (08/05/2009):
>> Justification: FTBFS on amd64
>
> Yes, but not because of the following:
>> Relevant part:
>> > gcc -pthread _configtest.o -o _configtest
>> > _configtest.o: In function `main':
>> > /build/user-python-numpy_1.2.1-1-amd64-f6BuY2/python-numpy-1.2.1/_configtest.c:5: undefined reference to `exp'
>> > collect2: ld returned 1 exit status
>
> The real cause of the FTBFS is #521525, and that bug doesn't make the
> package FTBFS. Keeping it as important bug, though, because errors
> during buildings aren't a good idea.
>
> The fix is pretty trivial, I'd guess, like adding -lm to the command
> line. I didn't fix it in my NMU though, since that was already happening
> in the previous (successful) builds.

I'm attaching the upstream reply (from [1]):

 The configtest are, as the name suggests, attempts to compile some simple
 C code with different sizes for built-in types, as I understand it. Maybe
 some other checks, not sure about that. In the end a configuration should
 be found that compiles without problems. So it's allowed to fail, but not
 to throw an error that stops the build process.

[1] http://projects.scipy.org/numpy/ticket/1526

so the purpose of _configtest.c is to actually fail, in order to
identify the config options set to use: the failures seen in the build
log are "under control". That said, I'm closing this bug.

Regards.
-- 
Sandro Tosi (aka morph, morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi