Nonces missmatch between pool and historyDB

[arnaubennassar]

When a tx is sent to the pool, it uses the current nonce (starting by 0), however in the HistoryDB the next nonce is used (starting by 1). This causes a missmatch of TxID, that defeats the purpose of the ID (being able to track the tx from the moment it's sent to the pool to the moment it's forged, and therefore added to HistoryDB)

[ed255]

From an online discussion; the correct spec is this:

• The account is created with nonce = 0
• The nonce in the account states the next valid nonce for that account

[ed255]

This also affects the purger methods:

var checkNoncesQuery = fmt.Sprintf(`
		UPDATE tx_pool SET
			state = '%s',
			batch_num = %%d
		FROM (VALUES
			(NULL::::BIGINT, NULL::::BIGINT),
			(:idx, :nonce)
		) as updated_acc (idx, nonce)
		WHERE tx_pool.from_idx = updated_acc.idx AND tx_pool.nonce <= updated_acc.nonce;
	`, common.PoolL2TxStateInvalid)

tx_pool.nonce <= updated_acc.nonce should be tx_pool.nonce < updated_acc.nonce

[ed255]

This should have been fixed after #412
@arnaubennassar if you are able to test this and verify that it's fixed, please write back to this issue and close it!

[arnaubennassar]

I've run the integration-testing env and forged a L2 tx, and I can confirm that the nonces now start with 0