Consider supporting standard debug rpc

[koushiro]

Motivation

Standardize debug namespace: ethereum/execution-apis#247

• debug_getRawHeader(hashOrNumber) -> bytes -- gets header RLP.
• debug_getRawBlock(hashOrNumber) -> bytes -- gets block RLP.
• debug_getRawTransaction(hash) -> bytes -- gets EIP-2718 encoded tx.
• debug_getRawReceipts(hashOrNumber) -> []bytes -- gets array of EIP-2718 encoded receipts.
• debug_getBadBlocks -> []BadBlock -- gets recent invalid blocks seen by the client.

We can also provide some client-specific debug rpc, which can be hidden behind some features.

Suggested Solution

Alternatives

Additional Information

[boundless-forest]

#768 is what's already being discussed.

[koushiro]

#768 is what's already being discussed.

this PR is about geth client-specific (non-standard) debug rpc

[koushiro]

@sorpaas what do you think?

[sorpaas]

I have no objection.

This is much simpler to implement than #768. Geth's debug does not include debug tracing, which means we can leave that unresolved "tracing in native / tracing in wasm" issue behind.

Those raw RLP APIs should be straightforward.

debug_getBadBlocks wouldn't really be useful in a Substrate context though. The rationale for that is for debugging multi-client consensus issues, which we'll never face (we may have multiple clients in the future, but for runtime it's only "multi-wasm-runtime", never "multi-EVM"). We can simply return empty array for that API.