Fix current_state_events table to not lie

If we try and persist two state events that have the same ancestor we
calculate the wrong current state when persisting those events.

Implement delete_devices API

(But this doesn't implement the same for deleting access tokens or e2e keys.

Also respond to code review.

Add helpful texts to logger config options

This implements the proposal here https://docs.google.com/document/d/1C-25Gqz3TXy2jIAoeOKxpNtmme0jI4g3yFGqv5GlAAk for deleting multiple devices at once in a single request.

Refactor FederationServer._handle_new_pdu

Add an option to disable stdio redirect

Reread log config on SIGHUP

This makes it tractable to run synapse under pdb.

- to make it easier to add more config options.

When we are using a log_config file, reread it on SIGHUP.

Noop repated delete device inbox calls from /sync

Unfortunately this significantly increases the size of the already-rather-big
FederationHandler, but the code fits more naturally here, and it paves the way
for the tighter integration that I need between handling incoming PDUs and
doing the join dance.

Other than renaming the existing `FederationHandler.on_receive_pdu` to
`_process_received_pdu` to make way for it, this just consists of the move, and
replacing `self.handler` with `self` and `self` with `self.replication_layer`.

When we receive PDUs via `get_missing_events`, we have already checked their
sigs, so there is no need to do it again.

This should be functionally identical: it just seeks to improve readability by
reducing indentation.

Support registration & login with phone number

This file post-dates OM

looks like the upsert function was accidentally sending sql logging to the
general logger. We already log the sql in `txn.execute`.