Thanks @unclebilly for your pull request. Resque-loner now supports a maximum time for which a job should be unique. Just define @loner_ttl in your job (or leave it at -1 to never expire) and after @loner_ttl seconds your job can be enqueued again, even if an older one is still marked as running.
Merged in @ryansch's pull requests to clean up things a bit.
This removed the
dequeue_from methods from
resque-loner because it caused troubles with resque's own
Pulled in #8 and #9 so that removing empty queues does not fail