Integral on `Tetrahedron` not fully supported

[mikeingold]

Integrals on Tetrahedron are only currently supported for the GaussKronrod algorithm, which uses a triply-nested quadgk formulation (sub-optimal but it works).

Much like Triangle, the default integral methods for GaussLegendre and HAdaptiveCubature will not work because they require the integration limits to be rectangular on $(0,1)^n$ for an $n$-dimensional space. However, to extend the derivation that I used to implement these methods for Triangle seems like a non-trivial job.