Avoid allocations in boundary flux for parabolic RHS

[DanielDoehring]

The previous implementation caused allocations due to the slicing.

[jlchan]

Thanks!

[andrewwinters5000]

LGTM! Nice catch to avoid allocations.

[DanielDoehring]

Credits for pin-pointing the allocations go to @jlchan, I just noticed this while working on the mortars & amr for parabolic terms.

[codecov]

Codecov Report

Merging #1594 (d1e4dad) into main (6c97c48) will increase coverage by 6.73%.
The diff coverage is 75.00%.

@@            Coverage Diff             @@
##             main    #1594      +/-   ##
==========================================
+ Coverage   89.40%   96.13%   +6.73%     
==========================================
  Files         401      401              
  Lines       33041    33072      +31     
==========================================
+ Hits        29539    31791    +2252     
+ Misses       3502     1281    -2221     

Flag	Coverage Δ
unittests	96.13% <75.00%> (+6.73%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
...s/tree_2d_dgsem/elixir_navierstokes_shear_layer.jl	0.00% <0.00%> (ø)
...s/tree_2d_dgsem/elixir_navierstokes_convergence.jl	100.00% <100.00%> (ø)
...s/tree_3d_dgsem/elixir_navierstokes_convergence.jl	100.00% <100.00%> (ø)
src/solvers/dgsem_tree/containers_2d.jl	97.73% <100.00%> (+45.10%)	⬆️
src/solvers/dgsem_tree/containers_3d.jl	98.53% <100.00%> (ø)

... and 37 files with indirect coverage changes