Daniel Arndt

For reference, https://dealii.org/9.4.0/doxygen/deal.II/structDataPostprocessorInputs_1_1CommonInputs.html#af00d5d79436ce5347e46d80575b97be8 shows that the overload taking a `DoFHandlerType` template argument was deprecated in 9.4.0.

Can you elaborate a bit on what you are trying to do? So far you are just changing the memory space the data is assumed to be allocated in which...

Why do we need a template argument? The mechanism for using scratch memory (assuming level 0) and global memory is fundamentally different. For unmanaged Views, the memory space is more...

> The implementation of the class doesn't depend on _where_ the shape data are allocated: The kernel evaluation logic works just as well when the shape data are allocated in...

You will have to construct the mesh yourself but the example in step-65 should be a good starting point. Just don't use the inner mesh for the ball and extent...

> I can compute precisely what all the different lengths should be so that the vertices match at the interface cube / triangle, but that would impose some restrictions about...

> [@masterleinad](https://github.com/masterleinad) Do you mean that it is the responsibility of the user to choose the right combination of lengths ? So, for example, the right approach for writing a...

CI is failing with ``` /__w/kokkos-tools/kokkos-tools/tests/vov-bug-finder/test_view_of_views_bug_finder.cpp(67): error: The enclosing parent function ("TestBody") for an extended __host__ __device__ lambda cannot have private or protected access within its class __attribute__((host)) __attribute__((device))(int) {...

> Like a "sanitizer" tool with a view of views check? I was thinking about not just covering kernels that have `Kokkos::View::initialization` in their label but always diagnose if the...

> That would produce lower quality error messages You could still special case for those particular kernels but my point is that we can diagnose more problems fairly easily.