Moreover, it's not a new problem, getNestMembers() has exactly the
same issue and nestmates have the same kind of lazy validation, the VM
doesn't check nestmate attributes untile an access orccurs.
So piggybacking on what getNestMembers() do seems a good idea.
This is a "for consistency" argument. Bu: ClassDesc did not exist at
the time we did nestmates! We were pretty unhappy about triggering
loading on that one too, but we didn't see a good alternative at the time.
So "let's do the same thing as something we hated doing in a different
situation, for consistency" is not the best of arguments :)
