TestingWindow now explain why we need to call the pthread workaround method.

This commit is contained in:
Robert Maynard 2016-03-21 15:44:47 -04:00
parent d91424ce1c
commit a4148d4e6a

@ -59,14 +59,15 @@
// //
// //
// A: Specify 'no-as-needed' to the linker potentially causing over-linking // A: Specify 'no-as-needed' to the linker potentially causing over-linking
// and a slow down in link time // and a slow down in link time.
// //
// B: Use a method from pthread, making the linker realize that pthread is // B: Use a method from pthread, making the linker realize that pthread is
// needed. // needed. Note we have to actually call the method so that a linker with
// optimizations enabled doesn't remove the function and pthread requirement.
// //
// //
// So that is the explanation on why we have the following function which is // So that is the explanation on why we have the following function which is
// never used and seems very crazy. // used once, doesn't look to be useful and seems very crazy.
#include <pthread.h> #include <pthread.h>
#define VTKM_NVIDIA_PTHREAD_WORKAROUND 1 #define VTKM_NVIDIA_PTHREAD_WORKAROUND 1
static int vtkm_force_linking_to_pthread_to_fix_nvidia_libgl_bug() static int vtkm_force_linking_to_pthread_to_fix_nvidia_libgl_bug()