Anastasios Kichidis
Anastasios Kichidis
> I’ll be back fully back tomorrow and will be able to give a complete review. In the meantime, have you considered re-using the “reconfiguration” watch channel to propagate the...
Let's reopen this in https://github.com/MystenLabs/sui if it's still relevant
Closing the ticket as @aschran @bmwill already working on this and tracking via other issues
Let's re-open this in https://github.com/MystenLabs/sui if it's still relevant
Closing this as we have now moved Narwhal to https://github.com/MystenLabs/sui . @gdanezis please re-open it there if it's still relevant
> I think this should work now that ConsensusHandler is atomic - it either will process all transactions in commit or none. > > But lets add a big warning...
> The tests you switched from #[test] to #[tokio::test] were not simtests anyways so we don't lose any additional testing that simtest provides running them or am I missing something?...
> > Right now the checkpoint fork message does not give very clear indication of which consensus subdag has the fork. We actually just used this log message to debug...
> I know mutating CommittedSubDag just to set reputation scores is a bit weird. But sending reputation scores to linearizer just to construct CommittedSubDag does not seem the most ideal...
> I like the safety provided by not mutating the subdag but I would still vote for keeping the setter method especially because I left a todo so that we...