CUDA 6.0, clang, libstdc++: something (cudafe?) seems to run clang in c++11 mode

When compiling which checks for clang features, like boost or opencv, it seems that something is invoking clang in c++11 mode, where it should be c++03 mode. Subsequent compile errors are the result. I’m not sure what exactly is going on.

See nvcc compile errors with with CUDA 6.0, clang, OS X 10.9 when including Boost.Function or opencv - CUDA Programming and Performance - NVIDIA Developer Forums for resulting errors.

Does anyone have any insights on this?

Hm… I’m also happy to recieve comments on where I might find more of a response for this issue. Where are the Mac OS X CUDA devs hinding?