#53 Timestamp of versions seem wrong in some cases

[Kenneth C. Schalk]

The timestamp (shown through the NFS interface with ls
-l, or returned by the VestaSource::timestamp method
through the SRPC interface) on versions often seems to
have no correspondence to the time of the last change
or the time of checkin. Often a sequence of versions
will have identical timestamps, in spite of the fact
that changes have been made between them.

I don't have a good theory on why this is the case, and
I haven't really investigated it. I've noticed it
before, and just now it's causing me some difficult in
trying to sort the output of vlatest in chronological
order.

It seems as though this may only be the case at the
repository where each version is checked in. In one
package I've been looking at, the timestamps of the
versions at non-master replicas seem to correspond to
the last modification of the contents of the version.