Open TheChymera opened 7 years ago
Hm, I can't really see why multilib on glibc should make any difference, especially if the errors appear in the header-files and not the shared library. Actually I don't see why we should build a multilib-profile in the first place, since we won't run either wine or pre-compiled 32bit-binaries.
By the way, the problem is there:
x86_64-pc-linux-gnu-gcc -O2 -pipe -O2 -pipe -fPIC -DREAD_WRITE_64 -DLINUX2 -Wcomment -Wformat -DUSE_TRACING -DHAVE_XDBE -DDONT_USE_XTDESTROY -c mri_cfft.c -I. -I/usr/include -Inifti/nifti2 -Inifti/niftilib -Inifti/nifticdf -Inifti/znzlib -I3DEdge/src -Irickr -DHAVE_ZLIB -DHAVE_GIFTI
Emphasis mine. This is hardcoded in many Makefiles.
As a quick fix sed'ing away all -I/usr/include
might work, since every sane compiler should add this implicitly anyways
Fix submitted upstream.
While the nipype issue has since been addressed on our RAP machines, it is now becoming apparent that I cannot emerge
sci-biology/afni
on them.AFNI can successfully be emerged on all our other machines, and the one difference that seems to overlap perfectly with AFNI not building is that glibc is built with
-multilib
. This seems consistent with the error I am getting:Is there any (easy) way to make a RAP install multilib? I don't see the point of keeping it no-multilib.
EDIT turns out this is entirely unrelated to multilib, but rather a prefix issue.