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
{{ message }}
This repository has been archived by the owner on Nov 15, 2023. It is now read-only.
Now that there is a dedicated client-side gadget with easy access to offchain db (#12753), move all offchain-specific functionality including proof generation from the pallet to the gadget.
This will greatly simplify the pallet which will no longer need both Storage<RuntimeStorage> and Storage<OffchainStorage> abstractions - it will simply have Storage (and will only work with runtime storage).
We should also move the MMR RPC code from frame/merkle-mountain-range/rpc to client/merkle-mountain-range/rpc; RPC is definitely not FRAME, but client-side.
The text was updated successfully, but these errors were encountered:
After giving this a try, I'm not convinced it's actually better. We still need to call into runtime to get MMR size when generating proof (so it's not really more efficient), and we ultimately end up with proof generation and verification split across multiple components. I now think keeping proof generation in the pallet is better. The gadget is then responsible for only offchain canonicalization and pruning.
acatangiu
changed the title
client/mmr: move all offchain functionality including proof generation to mmr gadget
mmr: move RPC from frame to client
Nov 30, 2022
Now that there is a dedicated client-side gadget with easy access to offchain db (#12753), move all offchain-specific functionality including proof generation from the pallet to the gadget.
This will greatly simplify the pallet which will no longer need both
Storage<RuntimeStorage>
andStorage<OffchainStorage>
abstractions - it will simply haveStorage
(and will only work with runtime storage).We should also move the MMR RPC code from
frame/merkle-mountain-range/rpc
toclient/merkle-mountain-range/rpc
; RPC is definitely not FRAME, but client-side.The text was updated successfully, but these errors were encountered: