Needless unlock when job's completed or existed?

Is it necessary to unlock the current job when checking it’s completed, exited or cancelled? Here is the redash/ at master · getredash/redash · GitHub .

If we unlock the current job, WatchError will be raised and then it’ll come back to the second loop. While if we don’t unlock the current job, the _job_lock_id will be set during current while loop. I think it’ll also work well.

Am I right? Looks forward for comments here, Thanks!