Arbitrum integration: Ensure that the espresso block justification is serialized with batch data

[nomaxg]

Currently, we include the block justification in L1IncomingMessageHeader, but this never actually makes it to the L1.

In EspressoSystems/nitro-espresso-integration#14, we introduce a new message type, L2MessageKind_EspressoTx, so we theoretically should be able to append the justification to the end of the l2 message and parse it accordingly.

[nomaxg]

@ImJeremyHe what do you think of this approach?

[sveitser]

@nomaxg I think that's fine. Once we no longer put the transactions in L1 do we still want to put the justification into L1? Do you recall what we did for the OP integration?

[nomaxg]

@nomaxg I think that's fine. Once we no longer put the transactions in L1 do we still want to put the justification into L1? Do you recall what we did for the OP integration?

Yeah we still include the justification at a minimum so that everything required for validation is available on the L1

[ImJeremyHe]

Currently, we include the block justification in L1IncomingMessageHeader, but this never actually makes it to the L1.

L1IncomingMessageHeader is a part of the message and now we get the justification here. IIUC, the justification is serialized with batch data already.

[ImJeremyHe]

I'm sorry, I used to think that all messages would be stored in L1 as they are. After debugging, I finally found that the header of the message is not completely made to L1. It seems that writing the justification into l2msg is an easy way. We can consider put the justification part to the first of L2Msg

[sveitser]

@ImJeremyHe @nomaxg we can close this right since

• Regard the BlockJustification as a part of L2Msg nitro-espresso-integration#20

has been merged?

[nomaxg]

yep 👍