By default if a test is available and not blacklisted it will be run by
murdock.
This will move from a whitelisting everything that works, to
blacklisting what fails. This way, failing tests will be easier to
track instead of being silently just not run.
It could even allow introducing failing tests before a fix is available.