test errors should be reported differently from test failures

[Nate Finch]

Right now there's very little visual different between a test failure (result == false) and a test error (error != ""). The only difference I see is an additional line of text that indicates the error, which is not significantly different from the lines simply showing the values passed into the check.

Test errors indicate a problem with the test, whereas test failures indicate a problem with tested function (at least, that appears to be the intent). It would be useful if test errors were counted separately from failures, and given an indicator that's different than simple failure, so it's clear to the test writer that the test needs to be fixed, not the tested function.