Remove ValidateMasternodeWinner in favour of IsTransactionValid

[kerryoncrypto]

Hello Telos-Team,

I know this PR will be a bit hard as you have spent much work for CMasternodePayments::ValidateMasternodeWinner() so far, but there is already such functionality on the source-code that just have to used the right way.

This PR removes CMasternodePayments::ValidateMasternodeWinner() from the code and re-arms the CMasternodePayments::IsTransactionValid()-Check with your SPORK_17_MASTERNODE_PAYMENT_CHECK-Spork.

Just like ValidateMasternodeWinner() assures IsTransactionValid() that there is a masternode-payment on the stake-transaction and has the right value spent. But in contrast to your implementation it is a bit more vague on the masternode that has to be paid. Your implementation requires that one specific masternode is paid which implies that the masternode-assets have to be the same on every node validating this transaction. As the number of masternodes grows and the synchronization-process is a bit laggy this is something that cannot be guaranteed and therefore should be avoided. IsTransactionValid adds a bit fuzz and allows a wider set of masternodes to be paid and therefore is likely to fail less. The function is taken from PIVX-upstream, used by most other masternode-coins and can be assumed to be safe - at least it's safe to assume that the payee is a valid masternode.

I'm looking forward for your feedback!

P.S.: It would be better to have this done via the already present but unused spork-key-mechanism (see SPORK_8_MASTERNODE_PAYMENT_ENFORCEMENT), but for the moment this PR is more like the update you intended first.

[DancingAxolotl]

Thank you for bringing this to attention.

Reusing existing code certainly would be better. But in this case IsTransactionValid performs checks based on a different masternode selection process. As we added several tiers of masternodes, it is likely that validation logic from PIVX will not work correctly.

The best approach would be a combination of both checking a wide set of masternodes and enforcing masternode tiers. We will have to check if IsTransactionValid can also support validating tiers of masternodes. It probably needs some modifications for that.

[kerryoncrypto]

Oh, you're right! I've missed this fact about the tiered MNs...
Whenever there is spare time, I'll look into it by myself.