Comment 0 for bug 1517879

Trial has a "run until failure" feature which re-runs tests using `test.run(result)` until the test fails. This is very useful for debugging intermittently failing tests.

Currently, testtools tests cannot be run twice: the setUp & tearDown safety warnings will error out. testing-cabal/testtools#165 is in progress and solving that.

Even once this is fixed, the details attached to a test will be accumulated between runs.

Upstream: https://clusterhq.atlassian.net/browse/FLOC-3462