Compilation errors with `dim==1`

[mschreter]

Compilation errors with `dim==1`

[mschreter]

@kronbichler It looks now much better, using dealii/dealii#16776 adaflo (and MeltPoolDG) compiles now without any compatibility changes :-). This should be a good sign since we consider many 1D cases. Thank you very much for the quick fix.

[kronbichler]

I see the point; it seems that the tests in deal.II are less comprehensive than I thought they are, so anyway this indicates we should work towards additional 1D tests. My apologies for not testing here, this is an oversight.
In my later PR dealii/dealii#16776, I indeed enabled get_divergence for 1D due to the benefits. I think we should aim to avoid the constexpr cases as you introduced them here, and I think we can get all the features into the now single call function. I think we should start from the latter PR because, as you state correctly, it again changes some of the edge cases. We should discuss it there and try to edge out possible incompatibilities for the 1D case. Essentially, the problem there can be traced back to the missing compatibility layer we had for 1D/vector-valued variables. I added two functions now to dealii/dealii#16776, can you check again?