I'm reworking some internal APIs and have a few questions about how much you would like to expose to CLM.
The internal reworks are not the goal of this discussion, but whether I expose these to CLM via the LedgerFactory is.

Sync module moves away from authorizationHandler and registerReceiveHandler, instead it now exposes handleMessage(message: table, sender: string, trusted: bool).

Approach one:
I can fully hide this in the factory; by calling the callback on every message I received after registering via registerReceiveHandler.

Approach two:

• LedgerFactory.createLedger changes signature and removes registerReceiveHandler.
• The resulting ledger object gets a new function handleMessage(message: table, sender: string, trusted: bool) that CLM calls after receiving a message and deserializing it.

Which do you prefer?