pallet-revive-eth-rpc: Introduce SubstrateClientT trait abstraction and Asset Hub embedded RPC integration

[Nathy-bajo]

closes #11221

Removes the revive-dev-node standalone binary and integrates the pallet-revive ETH RPC server directly into the Asset Hub (Omni) node. Introduces a SubstrateClientT trait that abstracts all Substrate node interactions, allowing EthRpcServerImpl to be generic over its backend. A new NativeSubstrateClient implementation backed by sc_client_api traits provides a direct in-process path with no subxt dependency on the hot path. The SubxtClient path is preserved unchanged for the standalone eth-rpc binary. On the node side, BuildAssetHubRpcExtensions is added to polkadot-omni-node-lib behind an opt-in revive-rpc feature, registering the full eth_*/debug_*/net_*/web3_* surface on the standard RPC port alongside existing parachain RPCs.

[Nathy-bajo]

@Wizdave97 @seunlanlege Ready for review

[pgherveou]

Thanks for putting that together.

Before we can move forward, we need to check with the Node team
if they are ok adding these dependencies to the omninode cc @skunert

One of the reason we didn't do it initially was:

• 1 so we could move fast and deploy new version without touching the client.
• 2 so that it's not tied to a client implementation. But I guess since most collator should run the omni -ode and there are no other collator implementation, that doesn't really matter.

The eth server is more stable now so it make sense to remove the requirement for a second binary / host url, and your design allow to keep both the binary inside the client and to build it as a separate binary with subxt.

[pgherveou]

To simplify the review and not scare the reviewer, I would move the dev-node deletion to a follow up

[pgherveou]

Also tried to build it locally but seems not to compile yet.
Better to keep the PR in draft mode until CI / build pass and it's actually ready for reviews

[Nathy-bajo]

Also tried to build it locally but seems not to compile yet. Better to keep the PR in draft mode until CI / build pass and it's actually ready for reviews

Thanks for the review, fixing ci

[marian-radu]

@marian-radu @pgherveou for some unknown reasons the eth-rpc's cli::run keeps crashing immediately after connecting. Any suggestions please?

This could be caused by sync_backward panicking due to ChainMismatch. Please check #11153 for context.

Try one of:

# Use in-memory DB (no persistence, no mismatch possible):
./target/debug/eth-rpc --eth-pruning=256

# Or give it a persistent base-path so you can wipe it deliberately:
./target/debug/eth-rpc --base-path /tmp/eth-rpc-dev

[Nathy-bajo]

This could be caused by sync_backward panicking due to ChainMismatch. Please check #11153 for context.

Try one of:

# Use in-memory DB (no persistence, no mismatch possible):
./target/debug/eth-rpc --eth-pruning=256

# Or give it a persistent base-path so you can wipe it deliberately:
./target/debug/eth-rpc --base-path /tmp/eth-rpc-dev

Still no luck, can you review the codes please?

[athei]

This as any other project is bottlenecked by review time. Reviewing a giant PR that the author can't make to work is not a good use of that time. It needs to work first.

[Nathy-bajo]

@athei @marian-radu @pgherveou CI passes now so I believe it works, also tested internally.