This is a hack.  The "blockstore" procedure is just returning a node
object, which causes skew in the resulting topology vs. what is generated
by the implicitly created node objects in the real testbed parser.  Harmless,
but not a good validity test.  Need to revisit.