Fix the issues found in the current logic of catch-up

Size: S (PoA)

In scope of this ticket fix the following issues in the current logic of catch-up:

• Execution of 3PC-batch after catch-up in case it was applied before catch-up and then reverted at catch-up start.
• Absence of removal of PrePrepares from Replica.stashingWhileCatchingUp.
• If catch-up is in progress and some ledger catch-up has not been started yet then the node does not respond older LedgerStatuses for this ledger with ConsistencyProofs until it starts catch-up of this ledger.
• Different quorums for not newer LedgerStatuses and none-proofs which actually mean the same.
• Mode.syncing is used instead of Mode.discovering.
• Wrong log message on reaching the quorum of none-proofs in LedgerManager.canProcessConsistencyProof.