Helping with testnet forking issues?
All forking issues have been resolved! Next I'm going to start working on stabilizing the masternode payments ghost masternodes, then I'll set a date for the fork (prob about two weeks).
** Please update your client to 10.9.2 or 9.5.2 **
** If you updated and are running either of those versions from a few hours ago, please update as well, the keys were changed in the code **
(If you're not helping with testnet, you can ignore this. I'll be having everyone else update soon)
Quick question about the forking issues being resolved. If Evan or maybe someone else who understands auto-checkpointing as it was implemented in DarkCoin (Hiro?
) could answer I would be really grateful.
As I understand it, and I mean absolutely no offence, the underlying issue of the voting system causing forks remains unfixed and now autocheckpointing will be used to keep all clients on the correct chain. Please correct me if I am wrong.
Assuming I am not... then what is the plan if the "checkpoint" node (this a central node according to the code that was posted by chapelin) has some downtime or even worse is attacked? If the node is down then it won't send checkpoints. If it has been compromised then the individual clients will not sign. That seems ok at a glance, but wouldn't the blockchain just fork once again during the time that no checkpoints are being included in the blocks?
Has this been considered and DarkCoin is using a modified version of autocheckpointing? Or am I misunderstanding something fundamental related to how autocheckpointing actually works?
Thanks.
