feat(rpc): implement debug_executionWitness API

[0x00101010]

Description

implement debug_executionWitness API

Tests

Tested offline with reth and op-program

Additional context

Metadata

[ajsutton]

Adding myself as reviewer so I don't forget this after the morning meeting run but no need to block on me if someone else gets to this first.

[ajsutton]

Looks good generally but would be worth adding a test for it. @refcell or @clabby does the returned format match what you'd expect?

There's also a comment saying the format might be simplified in the future - should be consider doing that now so we don't have to break compatibility? I'm not too sure what's actually likely to get standardised here.

[0x00101010]

Looks good generally but would be worth adding a test for it. @refcell or @clabby does the returned format match what you'd expect?

There's also a comment saying the format might be simplified in the future - should be consider doing that now so we don't have to break compatibility? I'm not too sure what's actually likely to get standardised here.

regarding format, I believe so, I compared returns from reth & geth, and also the queries from op-program, they match.

Currently trying to figure out return mismatch between reth & geth, working with reth teams on that

[0x00101010]

Looks good generally but would be worth adding a test for it. @refcell or @clabby does the returned format match what you'd expect?
There's also a comment saying the format might be simplified in the future - should be consider doing that now so we don't have to break compatibility? I'm not too sure what's actually likely to get standardised here.

regarding format, I believe so, I compared returns from reth & geth, and also the queries from op-program, they match.

Currently trying to figure out return mismatch between reth & geth, working with reth teams on that

@ajsutton, also regarding standardization, currently the conversation is around two things:

1. what exact data do we want to return here, geth returns headers in execution witness as well, while reth does not (currently not needed by op-program) => this is likely to be added into reth if we want the witness to be helpful to be used to execute blocks directly
2. how we want to return the data, from Peter's proposal, he suggests rlp everything to improve performance, currently since performance isn't a big concern, we opted for simplicify which is directly return json data, that can be revisited if needed

Added tests and ready for another review

[ajsutton]

LGTM. Given the benefit to fault proofs I think this is worth having in op-geth now, but would be good to submit this to upstream as well since it's a generically useful API.

[0x00101010]

LGTM. Given the benefit to fault proofs I think this is worth having in op-geth now, but would be good to submit this to upstream as well since it's a generically useful API.

Thanks @ajsutton, haven't done it now because there's still unanswered questions regarding return format and I sense resistance from geth side, but will try it to see how geth team reacts

[ajsutton]

Yeah, I'd say it's worth creating the PR to upstream to push the conversation forward.