This change adds a timeout to tools for use in the status updates. This makes

it so that if, for some reason, the tool doesn't execute, it will show up as

"timed out" after a period. In production this can happen if there are no

workers that have that tool available, or the broker has decided to stop

passing messages because the disk is filling up.

Enabled extension after switching to this branch and saw that the evolution

  was correctly applied (after a couple hours of banging my head against the

  wall because of some .pyc nonsense).
Refreshed the tool list and saw that the timeouts defined in the tool

  classes were properly plumbed through into the DB models.
Stopped my worker node and triggered some tools by publishing a review

  request. Saw that the new status updates were initially listed as "starting",

  and that after the configured timeout, were then listed as "timed out".
Restarted my worker node and saw the tools run and the status update

  change to "passed".