op-node: Handle malformed/incomplete receipt data from L1

[protolambda]

Bug-reports specify that Besu as L1 RPC returns receipt data with invalid log-index attributes. Since these can be inferred, rather than logging that the data is wrong, we can correct the data (if the receipts-root still matches the consensus data). This data "hydration" is also already happening for the raw receipts of the debug-geth RPC kind. And then, on the mismatch, just make a verbose log of what's wrong about the response exactly, so the L1 client can be fixed.

[ajsutton]

Do we have a link to the bug reports? I know a while back besu was starting log index from 0 for each transaction rather than each block but I thought that had been fixed.

Fixing the data also sounds like a good idea so long as anything important to us is covered by the checks against the block hash.

[protolambda]

Regarding Besu log index: see the errors in the screenshot in ethereum-optimism/developers#59 (comment)

[ajsutton]

Ah yeah that very much looks like hyperledger/besu#4114 which should be fixed in the latest besu (but may have regressed).

[protolambda]

@ajsutton that issue is a year old, it must have regressed, or the user wouldn't be able to sync the latest L1 chain with it.

[ajsutton]

oh good point, I missed the 2022 part of the date. The error message would fit perfectly with that bug though...

[ajsutton]

Ah that fix was only for eth_getLogs. Almost certain the eth_getReceipt will have the same bug - just seeing if I can find someone with an existing besu node to confirm.

[ajsutton]

Confirmed, so I've logged hyperledger/besu#6204

[protolambda]

Looks like this was fixed in Besu. Can we close this issue? @ajsutton