cfgmgr32.h is a windows-specific header file preventing compilation in Linux about vk_raytrace

[chrisdonlan]

In order to get around this, I deleted the following function entirely (there are no uses in the project):

  void* LoadNmvlLibrary()

And I commented out the content here:

  float getCpuLoad()
  {
      /*
    static uint64_t _previousTotalTicks = 0;
    static uint64_t _previousIdleTicks  = 0;

    FILETIME idleTime, kernelTime, userTime;
    GetSystemTimes(&idleTime, &kernelTime, &userTime);

    auto FileTimeToInt64 = [](const FILETIME& ft) {
      return (((uint64_t)(ft.dwHighDateTime)) << 32) | ((uint64_t)ft.dwLowDateTime);
    };

    auto totalTicks = FileTimeToInt64(kernelTime) + FileTimeToInt64(userTime);
    auto idleTicks  = FileTimeToInt64(idleTime);

    uint64_t totalTicksSinceLastTime = totalTicks - _previousTotalTicks;
    uint64_t idleTicksSinceLastTime  = idleTicks - _previousIdleTicks;

    float result = 1.0f - ((totalTicksSinceLastTime > 0) ? ((float)idleTicksSinceLastTime) / totalTicksSinceLastTime : 0);

    _previousTotalTicks = totalTicks;
    _previousIdleTicks  = idleTicks;
       */

    // return result * 100.f;
    return -1 * 100.f;
  }

---

The CPU function seems approachable, so I am working on a linux work-around to get the CPU load in either a cross platform manner, or a switchable manner with macros.

But, I would have no idea where to start in LoadNmvlLibrary().

[mklefrancois]

We should have it working under Linux. Thanks for reporting.