TOOLS-3203 I noticed a segfault among several apparently-spurious test failures. These appear to arise from our use of t.Error/Errorf when we should be using Skip/Skipf or Fatal/Fatalf; i.e., the intent is apparently that the test stop execution because setup failed. The fact that the test keeps going causes a segfault due to nil pointer dereference.
On my laptop, at least, all that's needed is to run the tests.
Whatever failures happen shouldn't trigger segfaults. I should see the test suite complete, even if it doesn't pass.
I see a segfault, which seems to cause a premature exit of the test suite. (If nothing else, it makes the test output harder to read.)