For these tests, the date returned by clock_gettime() will be compared
with that returned by the gettimeofday() function.

Assumptions:
- Machines running these test cases also have access to the gettimeofday
  function.
- Ideally, the gettimeofday() function has different internals than
  clock_gettime().  [Hopefully, clock_gettime() doesn't just _call_
  gettimeofday(), for instance.]
  However, even if not, it's still a valid test to make sure clock_gettime()
  doesn't have an error somewhere else.
==> See
http://sourceforge.net/mailarchive/forum.php?thread_id=591583&forum_id=2469
It turns out these two functions _should_ return the same value but do
have different internals, so this is a good test case.


-----
Another note:  At one point, I considered using ntpd to do date checking.
That idea ended up getting thrown out, but one thing I want to document is
that, if I did chose that idea, the tester would need to install RPMs:
  ntp
  libcap
for it to run.