experimental - deposit queue

[tynes]

Description

Adds an experimental spec for a deposit queue.

This feature is meant to solve:

• Complexity of ResourceMetering
• No deposit transaction frontrunning grief
• Consistent amount of gas used on L1 when depositing

An additional consideration is to be able to handle L3s with equal or higher block time than the L2. (issue currently is that those chains can’t catch up their L1BlockInfo to the latest block). It isn't clear if this should be included as part of this or be its own spec yet

[kaleb-keny]

appreciate pushing for a change on this @tynes 🙏🏼

[tynes]

It would be good to also include a nonce with each deposit event so that we don't need to observe all logs that create deposit txs during the fault proof program to be sure that we have not skipped a deposit. It would also be ideal to have a nonce on L2 that increments for each deposit consumed, so that we can look at the L1 nonce and know how many not yet pulled in deposits there are on L2.

We would also want to do the same nonce strategy for the system config ConfigUpdate event, its possible that given the static config change in #122 we can deprecate ConfigUpdate in favor of simply making calls to L1Block as a special identity, similar to how custom gas token works

[mdehoog]

+1 on including a nonce with each deposit event, and also with config updates. This would allow us to skip processing L1 blocks that don't have any L2-relevant events in them, speeding up derivation and proving.

[BlocksOnAChain]

@benjaminion I believe we can move this for Isthmus, since we aren't going to ship it with Holocene.