Add free stream preservation tests for non-conforming P4estMesh and EC test

[efaulhaber]

Ninth point of #584.

In 3D, FSP only works in special cases for curved meshes (Kopriva et al., 2019, Section 6).
The three relevant cases (the other require different implementations) are:

1. conforming meshes,
2. two-dimensional extrusions,
3. meshes with N >= 2*Ngeo, where Ngeo is the degree of the geometry interpolant (kwarg polydeg of P4estMesh) and N is the polynomial degree of the solver (kwarg polydeg of DGSEM).

Case 1 is boring. I added tests for case 2 and case 3.

I also made the mesh in the 2D FSP example non-conforming.

Edit: As requested by @sloede, I now added an EC test as well.

[sloede]

I left two remarks regarding the error norms.

[sloede]

Congratulations on the devilish PR number by the way 👿

[codecov]

Codecov Report

Merging #666 (e926993) into main (c3e0903) will increase coverage by 0.00%.
The diff coverage is 97.05%.

@@           Coverage Diff           @@
##             main     #666   +/-   ##
=======================================
  Coverage   93.68%   93.69%           
=======================================
  Files         169      171    +2     
  Lines       16566    16600   +34     
=======================================
+ Hits        15520    15553   +33     
- Misses       1046     1047    +1     

Flag	Coverage Δ
unittests	93.69% <97.05%> (+<0.01%)	⬆️

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

Impacted Files	Coverage Δ
.../elixir_advection_amr_p4est_unstructured_curved.jl	100.00% <ø> (ø)
...s/3d/elixir_advection_p4est_unstructured_curved.jl	100.00% <ø> (ø)
src/solvers/dg_curved/dg_3d.jl	97.20% <ø> (ø)
src/solvers/dg_tree/dg_3d.jl	97.48% <ø> (ø)
src/equations/compressible_euler_3d.jl	94.86% <80.00%> (-0.13%)	⬇️
examples/2d/elixir_euler_free_stream_p4est.jl	100.00% <100.00%> (ø)
examples/3d/elixir_euler_ec_p4est.jl	100.00% <100.00%> (ø)
examples/3d/elixir_euler_free_stream_p4est.jl	100.00% <100.00%> (ø)
...ples/3d/elixir_euler_free_stream_p4est_extruded.jl	100.00% <100.00%> (ø)
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c3e0903...e926993. Read the comment docs.

[sloede]

LGTM! If someone else can confirm that the FSP numbers are OK, I think we're good to merge.

[andrewwinters5000]

Great work @efaulhaber ! Everything looks good from my end with respect to the errors / testing. Also, everything is documented for the FSP restriction of N>=2N_geo