hemelb::redblood::orientFacets gives inconsistent results when loading deformed mesh

[mobernabeu]

As part of #733, @Romain-Enjalbert found that hemelb::redblood::orientFacets will reorder 719 out 720 facets for a mesh that we know has consistent facet orientation throughout.

[mobernabeu]

3f7ea8e adds a failing test.

@Romain-Enjalbert, if you make all and run ./unittests_redblood hemelb::unittests::redblood::LoadingTimmMeshTests you should see 1 test out of 3 failing (see new code in previous commit for an explanation of what each of the tests does). The idea would be for us to provide a new implementation of hemelb::redblood::orientFacets that makes all three tests pass. This gives you a very fast an automated way of checking that the new code does what's expected from it in all three cases we have discussed this week.

[mobernabeu]

orientFacets method for meshes constructed from a .msh file has been deprecated. A more robust implementation based on VTK provided in a4764ed. We should use VTK input files from now on.

[rupertnash]

Hi @mobernabeu - you have added a failing test to the build. Can we remove this test as deprecated? Can we remove some code?

[mobernabeu]

Yes, we should remove the failing test and the support for .msh files altogether from the code as our implementation of orientFacets for them is subpar.