Implement Aura+Grandpa warp syncing

bin/wasm-node/CHANGELOG.md

@@ -5,6 +5,7 @@
 ### Changed
 
 - The GRANDPA warp sync algorithm now downloads Merkle proofs of all the necessary storage items at once, rather than one by one sequentially. This removes approximately 11 networking round-trips and thus significantly reduces the time the warp syncing takes. ([#2578](https://github.com/paritytech/smoldot/pull/2578))
+- The GRANDPA warp sync algorithm now works on AURA-based chains. It previously only worked for chains that are using BABE. Note that GRANDPA warp sync is irrelevant for parachains. ([#2581](https://github.com/paritytech/smoldot/pull/2581))
 - The GRANDPA warp sync implementation has been considerably refactored. It is possible that unintended changes in behaviour have accidentally been introduced. ([#2578](https://github.com/paritytech/smoldot/pull/2578))
 
 ## 0.6.27 - 2022-07-29

src/sync/all.rs

@@ -154,7 +154,11 @@ impl<TRq, TSrc, TBl> AllSync<TRq, TSrc, TBl> {
                     requests_capacity: config.sources_capacity, // TODO: ?! add as config?
                 }) {
                     Ok(inner) => AllSyncInner::GrandpaWarpSync { inner },
-                    Err((chain_information, warp_sync::WarpSyncInitError::NotGrandpa)) => {
+                    Err((
+                        chain_information,
+                        warp_sync::WarpSyncInitError::NotGrandpa
+                        | warp_sync::WarpSyncInitError::UnknownConsensus,
+                    )) => {
                         // On error, `warp_sync` returns back the chain information that was
                         // provided in its configuration.
                         AllSyncInner::Optimistic {