Nyota EOF Breakout Summary

This meeting covered the following topics:

• Whether to include EOF in Pectra
• Spec updates
• Interoperability with other EIPs

Chatham House Rule

The meeting was conducted under the "Chatham House Rule" - anyone who comes to a meeting is free to use information from the discussion, but is not allowed to reveal who made any particular comment.

Whether to include EOF in Pectra

The first topic of discussion was wether or not to include EOF in Pectra. While the large majority of attendees who spoke about the issue were in favour of shipping EOF as soon as it's ready, this was not a universal opinion.

Objections discussed included stack validation, TX CREATE, testing complexities, and performance benchmarks. From these objections it was accepted that before EOF is activated:

• Performance impacts need to be quantified
• Testing methodologies need to be reiewed to ensure new surfaces are addressedd
• EEST tests need to be clear as to what they are testing
• Real world usage and metrics need to be simulated and validate there are no significant regressions

The specific timing of the Pectra fork may make inclusing in pectra infeasable (such as if we flip testnets in Q3)

A concern was raised that shipping EOF before Verkle could cause delays in Verkle, because of needed spec changes and/or other impacts.

Spec updates

• EIP-7692 is the merged "meta EIP" enumerating what EIPs make up EOF v1.
• TXCREATE and the Initcode transaciton have been removed. They may return in a future compatible version or an alternate facility may be used.

There was a discussion about some pending issues. Still unresolved are wether RETURNDATALOAD will zero pad, if RETURNDATACOPY will have different behavior in EOF to zero pad, and the final outcome of the inotcode mode validation rules.

We also spent some time discussing VLQ and a global code section, touching on how the removal of Simple Subroutines in Berlin (EIP-2315) makes devs hesitatnt to adopt a similar subroutine facility in EOF.

Interoperability with other EIPS

Two items were discussed.

• The idea of banning BLOCKHASH in EOF and requireing users to use the EIP-2935 contract was proposed. No commitements were made
• Interactions with EIP-7702 as a potential replacement for TXCREATE was discussed, however subsequent breakouts indicated the direction of 7702 would make it unworkable.

EOF testing was discussed in a different breakout.