[Swan-dev] should final.sh shut down pluto?

[D. Hugh Redelmeier]

| From: Paul Wouters <paul at nohats.ca>

| A swan-destroy type shutdown works for me. What I want to avoid is
| triggering all tests as failures when we have a leak-detective report
| in shutdown.

A leak-detective difference could be reported as something other than
a normal diff / normal failure.  Kind of like white-space changes are
reported differently by kvmresults.py.

But not all failures during a swan-destroy should be considered special 
(ignorable by user).