Open tommling opened 6 years ago
For sawtooth event, it's able to do resolve fork based on block-commit events. But that's for the transaction processor layer and I'm afraid that's not low level enough for consensus algorithm. I mean when these events reached validator, they are already filtered and processed after consensus algorithm, are not raw event that suitable for using in consensus process.
Hey @ailisp good point. I am thinking the data will have to be stored in the chain. The affects are slightly delayed, its similar to a sign-up x block-info hybrid where we reference the information stored on chain to validate blocks and resolve forks.
So we add up the difficulty stored on the main chain and the difficulty stored in the trusses to come out with the total chain difficulty.
We need to consume the events from both chains and trigger transactions that store references.