Until now, quarantining a flaky test in Tuist meant skipping it entirely. That solved the immediate problem of your CI failing randomly, but it created a new one: once a test was quarantined, there was no signal to know when it was fixed. Teams had to manually unquarantine tests and hope for the best, or let the quarantine list grow indefinitely.
We changed how quarantine works. Quarantined tests now run normally, but their failures no longer affect your CI exit code. If only quarantined tests fail, your build passes. If any non-quarantined test fails, you still get a non-zero exit.
What this unlocks
Because quarantined tests keep running, Tuist can now observe whether they are still failing or have been fixed. This is the foundation for automatic unquarantine: once a test consistently passes, Tuist can remove it from the quarantine list and unmark it as flaky without anyone having to remember to do it manually.
Learn more
As always, feedback is very much welcome!