Retry timeout does not work correctly
To reproduce: src/wget2 localhost:8090
All 20 retries occur immediately.
Solution #1 (quick & dirty): The thread who got the job just sleeps time.
Solution #2 (closed): The thread who got the job marks it with , deallocates the job, starts a timer (e.g. tell a 'timer' thread that it should wake the worker threads after ) and continues the work. On a simple connect error, we should immediately try the other IPs (if DNS gave us more than one). If none of the IPs works, we should mark URLs with this IP/port combination as 'waiting'.