Ensure that refshapes matches when creating FEValues

[KnutAM]

This bit me giving silent wrong results

julia> ip = Lagrange{RefTriangle,1}()
Lagrange{RefTriangle, 1}()

julia> qr = QuadratureRule{RefQuadrilateral}(2);

julia> CellValues(qr, ip)
CellValues(scalar, rdim=2, and sdim=2): 4 quadrature points
 Function interpolation: Lagrange{RefTriangle, 1}()
Geometric interpolation: Lagrange{RefTriangle, 1}()^2

This PR throws an ArgumentError in this case.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.77%. Comparing base (0c0f9b5) to head (153ee3e).
Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1185      +/-   ##
==========================================
- Coverage   93.78%   93.77%   -0.02%     
==========================================
  Files          39       39              
  Lines        6256     6262       +6     
==========================================
+ Hits         5867     5872       +5     
- Misses        389      390       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[termi-official]

I think we can add something analogue for InterfaceValues.

[KnutAM]

I think we can add something analogue for InterfaceValues.

This is already included. I implemented it on the level of FunctionValues and GeometryMapping to ensure that anything that uses this building blocks will give the error.

[termi-official]

Good point. Should we still add path coverage in test_interfacevalues.jl for this?

[termi-official]

LGTM. That should help tracking down issues in more complicated code bases.

Edit: CI failure is unrelated to PR.