SpatialUnderstanding returns vertices with strange (near-zero) normals

[ForrestTrepte]

1. Open the SpatialUnderstanding solution. Set a breakpoint in RunTest_RealTimeScan_DynamicInputData and enable the code (currently disabled with #if 0) that checks normLength.
2. Run TestApp in SpatialUnderstanding in the UniversalRelease x86 config.

RESULT: Notice that normLength is zero for each of the vertices. The normals returned by GeneratePlayspace_ExtractMesh_Extract have tiny coordinates such as 5.69660380e-27.

EXPECTED: Normals have a length of 1.

Either the meaning of normals in the context of SpatialUnderstanding isn't what I would expect it to be based on the concept of a normal in computer graphics or something is wrong with the process of computing normals or the test app.

[ForrestTrepte]

By the way, running in UniversalDebug x86 config I was getting vary large values for normals. It seems like something is very wrong there. For now, I've disabled that part of the test and added a comment about it to the test code.

I'm done working on spatial understanding, at least for now. So I'll leave this issue for someone else to investigate if they want.