The original RIP-7560 previously required a separation of a Native Account Abstraction transaction flow into non-atomic validation frame and execution frame.
We now believe that this separation may not be necessary for some L2 implementations, so this separation of transactions into non-atomic parts has been extracted into a separate document here:
Hey, I have a question related to the motivation for this RIP. It mentions:
On “single sequencer” Layer 2 chains that do not have a “transaction mempool” in a traditional sense, this proposal provides no benefit compared to the original RIP-7560.
This claim seems to have an assumption that the sequencer will serve transactions in a first-come first-served order, so it won’t be bothered by state conflicts among native aa transactions.
However, if the sequencer picks transactions from a public mempool based on factors like effective gas tip, this RIP could still help mitigate state conflicts among native AA transactions. Am I understanding this correctly?