server: make sure that the various replication loggers use consistent logging

[rboyer]

This generic replicator code is used for 3 places: config entries, federation states, and enterprise stuff. This gets all 3 variants to actually log all of their components with the appropriate prefixes:

server.replication.config_entry: finished fetching config entries: amount=0
server.replication.config_entry: Config Entry replication: local=0 remote=0
server.replication.config_entry: Config Entry replication: deletions=0 updates=0
server.replication.config_entry: replication completed through remote index: index=1

server.replication.federation_state: finished fetching remote objects: amount=1
server.replication.federation_state: diffing replication state: local_amount=0 remote_amount=1
server.replication.federation_state: diffed replication state: deletions=0 updates=1
server.replication.federation_state: performing updates: updates=1
server.replication.federation_state: finished updates
server.replication.federation_state: replication completed through remote index: index=14
server.replication.federation_state: finished fetching remote objects: amount=2
server.replication.federation_state: diffing replication state: local_amount=1 remote_amount=2
server.replication.federation_state: diffed replication state: deletions=0 updates=1
server.replication.federation_state: performing updates: updates=1

[rboyer]

A replicator instance takes the Logger on the ReplicatorConfig and does this Named+Named expansion before using it for all internal logging.

When using the function replicator style, the function accepts a logger as an argument, and that logger is the derived one with the Named+Named prefix above (this is what config entries do).

The federation states version uses the delegate style, and so has it's logger configured separately here. Since this was just using s.logger before it was not getting the Named+Named prefix.

[dnephin]

LGTM