Write test checking whether effective peer targets are consistent with demanded ones

Churn and peer selection governors work together to establish the right peer targets for the node. This interplay is nuanced, considering that in Genesis consensus mode, there are two static targets bases to choose from, reflecting whether the node is caught up or still syncing. Churn has to pick the correct basis to modify and send to peer selection governor via a TVar, and OG has to make sure that it picks up these values from the TVar when churning is actively working, or whether to pick one of the targets bases when ledger state judgement changes. So the goal is to not lose information that churn provides, as well as to know when _not_ to use targets from churn in case they are out-of-date.