Allows pushing multiple protobufs per queue item, needed in the case of
listing retrieve queue jobs, which can have several VIDs in the same
job.

When the object is move constructed, everything is moved except
list::end(), which doesn't point to an actual object in memory and is
recreated with a new address. So in the case where the iterator points
to rhs.list::end(), it needs to be explicitly reset.

Default move constructor did not work in case of an empty list, as in
this case the list is not moved, a new empty list is instantiated. So
the copied iterators are invalid.

New move constructor is the same as default with an additional check to
reset the iterators if the lists to be moved are empty.

Conflicts:
	scheduler/OStoreDB/OStoreDB.cpp

This allows helgrind to be aware of succession of events.
Also reduced the default queue depth for unit tests.

Adapted the in-RPM script generation.

The delay will initially slow down the user and then the semaphore will block.
This prevent the queue from going too big and the agent object from growing
so big all becomes slow.

...before moving to first item of next job queue queue, which requires
one additional call to operator++

Summary now includes all tapepools, including those with no queued jobs
(previously we only showed a summary of queues containing jobs)