Consider: should a call to `alterValidator` that sets the same validator be a no-op?

[wildmolasses]

Worried that our test suite doesn't account for this call path by often assuming validator != newValidator