Also update the VTK-m based filters to work properly with the changes

Previously all the CMake variables would not be created

The free standing function allows for us to change the CastAndCall
implementation details in the future inside of VTK-m without having to
also update VTK at the same time.

The property isn't used anymore by the wrapping tools.

There are static asserts which trigger when building from the wrapper
sources.

Now that Tcl wrapping is gone, all wrapping tools now use
`WRAP_EXCLUDE_PYTHON` instead.

The counting array in these 2 occurrences would be 32bit integers and we need
them to be 64bit integers ( when vtkIdType == 64bit ).

The direct use of VTK-m was to switch the device being used.

TODO: add a means of specifying the device a VTK-m filter in
vtkAcceleratorsVTKm should use without directly interfacing with VTK-m.

For better encapsulation (and simplification of the build), do not
expose VTK-m code itself from the vtkAcceleratorsVTKm package.

All the VTK-m include and libraries should be hidden from code outward
facing of the vtkAcceleratorsVTKm library. Besides, the VTK-m
accelerator tests are not called as a subdirectory from the
vtkAcceleratorsVTKm code, so the CMake variables specifying the include
directories and libraries are not set.

If it is later decided that code using the accelerators library also
needs to directly compile VTK-m code, it should be specified as PUBLIC
includes and libraries to the vtkAcceleratorsVTKm library.

Typos fixes for documentation and source code comments. This commit is a follow up to:   !3424

Also bump VTK-m to bring in the new VertexClustering impl.

Previously we didn't build when tbb was not installed in a system include,
and the CUDA code paths would never be entered.

Updated baseline due to unrelated marching cube changes while
bumping vtkm.

Now that VTK requires a C++11 compiler we can ditch using custom typedefs to
wrap =delete;

Previously vtkmlib headers would be installed twice, and source files
where accidentally being installed too.

Does not make sense for OpenGL2 and is not implemented
Also deprecate PainterDelegate

Now that VTK requires a C++11 compiler we can ditch using custom typedefs

Adding in functionality to allow inserting vtkNew objects into
stl containers of vtkSmartPointer objects as well as comparisons
between vtkNew objects and raw pointers. Also removing Get() and
GetPointer() calls from vtkNew objects since in most instances we
can just pass in the vtkNew<> object instead of having to use the
Get() or GetPointer() methods to get the raw pointer.

just a minor warning fix

1) Support mapping cell fields
2) Support more types of datasets