Comment 12 for bug 1130146

The place where this fails in testr is testsuite.py L100

    def _run_test(self, test, process_result, queue):
        try:
            test.run(process_result)
        finally:
            queue.put(test)

It seems that under these fail conditions the test.run(process_result) leaves us with a test which is invalid enough that it just gets silently dropped later. That's hugely dangerous.

If test isn't queueable after all is said and done, we need to insert an explicit fail so we at least know that things went horribly wrong and we don't pass when there should be failures.