There may be leftover state from a previous failed test run in the file system. If we don't clean this up before running the test, the test can fail perpetually until the machine is rebooted. This results in wasted engineering effort to diagnose this. So, we'll ensure the state is clean before starting the test.
<rdar://problem/98950007>
Pull request: https://github.com/WebKit/WebKit/pull/3514
Committed 253633@main (92c64e3b4c25): <https://commits.webkit.org/253633@main> Reviewed commits have been landed. Closing PR #3514 and removing active labels.