Among the changes done for this implementation we have the StorageWeightReclaim feature which is introduced in Polkadot v1.9.0 and no issues reported about it since then. However my investigation shows the polkadot system parachains have not adopted it yet. This feature would allow the differences between benchmarked proof sizes and the actual proof sizes to be accounted for plus a facility for manual weight metering inside on_idle. This needs both the client and the runtime to support it. I noticed we had brought a part of this feature to our runtime. In this PR our client and runtime will become compatible about this check. My own assessment is this should be a pretty safe adoption especially if no problems are discovered during testing and try-runtime. The good thing about keeping it is that our upgrade will become closer to the upstream template/suggestion for parachains.
[x] Remove xToken
[x] Update Dependencies
[x] Fix XCM build while sticking to v3
[x] Fix Runtime
[x] Fix Node (Update jsonrpsee here)
[x] Fix tests
[x] Fix benchmarks
[x] Fix and test try-runtime
[x] Bump up all our crate versions
[x] Fix on-chain-release-build feature for eden-runtime
[x] Update CI to produce the runtime with on-chain-release-build feature
[x] Add and enable the rest of the benchmark command
[x] Bake eden-dev chainspec into runtime which is useful for benchmarking
[x] Remove redundant code in chain-spec
[x] Verify new frame_system::Config
[x] Check if async backing is properly supported
[x] Verify block duration for parachain (6s according to polkadot parachain templates or 12s as before and according to parachain-commons
Among the changes done for this implementation we have the StorageWeightReclaim feature which is introduced in Polkadot v1.9.0 and no issues reported about it since then. However my investigation shows the polkadot system parachains have not adopted it yet. This feature would allow the differences between benchmarked proof sizes and the actual proof sizes to be accounted for plus a facility for manual weight metering inside on_idle. This needs both the client and the runtime to support it. I noticed we had brought a part of this feature to our runtime. In this PR our client and runtime will become compatible about this check. My own assessment is this should be a pretty safe adoption especially if no problems are discovered during testing and try-runtime. The good thing about keeping it is that our upgrade will become closer to the upstream template/suggestion for parachains.
on-chain-release-build
feature for eden-runtimeon-chain-release-build
featureStorageWeightReclaim
: https://github.com/paritytech/polkadot-sdk/pull/3002[cfg(feature = "runtime-benchmarks")]
type XcmpQueue = (); => Result: it doesn't seem so and it's not fixing xcm benchmarking