I have been investigating this issue, and have filed binutils bug #16452 (http://sourceware.org/bugzilla/show_bug.cgi?id=16452) to track this. This can be worked around by explicitly linking the executable against pthreads, or explicitly specifying the path to NVIDIA's libGL using gcc's -L flag, e.g.:
Hi Alberto,
I have been investigating this issue, and have filed binutils bug #16452 (http:// sourceware. org/bugzilla/ show_bug. cgi?id= 16452) to track this. This can be worked around by explicitly linking the executable against pthreads, or explicitly specifying the path to NVIDIA's libGL using gcc's -L flag, e.g.:
$ g++ ./test.cpp -pthread -lGL
or
$ g++ ./test.cpp -L/usr/ lib/nvidia- 319-updates -lGL