In the TRS spec, describe ways to ensure clients can read base+change log after pruning/rebase

Describe ways to detect if a client is reading a base page that is no longer one from a current base, and hence no longer corresponds to the existing change log. Describe ways to implement an incremental base that does not have the above issue. Describe ways to keep older change events behind the previous cutoff to allow for clients that start reading the base+change log immediately before a rebase operation.