You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm trying to clean up issues and it looks that there are many issues that are specific to complex and/or batch relayers. Since we are already using standalone relayers for Rococo <> Westend and soon we will be using standalone relayers for Kusama <> Polkadot, let's remove complex and batch relayers code instead of trying to fix remaining issues and maintain it without having possibility to properly test it.
We should be able to use standalone relayers in all bridges, including Bulletin chain bridge. There's one inconvenience with standalone relayers - since they are totally independent of each other, it will take at least 3 blocks to relay e.g. a message: block N would include GRANDPA finality transaction, N+1 would include parachain finality transaction and N+2 would include message delivery transaction. We can't include e.g. GRANDPA and parachain transactions into a single block, because parachain tx will be rejected by our tx extension, because it relies on the GRANDPA proof, missing from the runtime storage. So we can't even e.g. use some signals (or watch tx pool) to connect standalone relayers.
It can be solved later by having all relayers running in a single process and a specific fn submit_* implementation, which would connect all transactions together. But I do not see it as a big issue, so better start with removing old code and if that'd be a problem, we could have some solution later.
I'll be using this issue as a reason to close existing issues with complex, on-demand and batch relayers.
The text was updated successfully, but these errors were encountered:
I'm trying to clean up issues and it looks that there are many issues that are specific to complex and/or batch relayers. Since we are already using standalone relayers for Rococo <> Westend and soon we will be using standalone relayers for Kusama <> Polkadot, let's remove complex and batch relayers code instead of trying to fix remaining issues and maintain it without having possibility to properly test it.
We should be able to use standalone relayers in all bridges, including Bulletin chain bridge. There's one inconvenience with standalone relayers - since they are totally independent of each other, it will take at least
3
blocks to relay e.g. a message: blockN
would include GRANDPA finality transaction,N+1
would include parachain finality transaction andN+2
would include message delivery transaction. We can't include e.g. GRANDPA and parachain transactions into a single block, because parachain tx will be rejected by our tx extension, because it relies on the GRANDPA proof, missing from the runtime storage. So we can't even e.g. use some signals (or watch tx pool) to connect standalone relayers.It can be solved later by having all relayers running in a single process and a specific
fn submit_*
implementation, which would connect all transactions together. But I do not see it as a big issue, so better start with removing old code and if that'd be a problem, we could have some solution later.I'll be using this issue as a reason to close existing issues with complex, on-demand and batch relayers.
The text was updated successfully, but these errors were encountered: