Ok, thanks for the advice on splice, didn't know it was that expensive.
On the other hand the routine is rarely called and I have no control over what users submit / priority some jobs have ( the files itself contain that info ),
and so the parser decides what file gets processed / computed
( these are long running job descriptions > +1 hour / job / CPU / compute node)
I just mean to say that this snippet will only process a couple of files / hour ( not thousands / minute )

I tested the Thread::Queue module and I'll keep it in mind as future reference, as I don't see a real use for it right now in the current code, (both the main / thread needs access to the same array > preventing jobs from getting processed multiple times )