Update EIP-7892: harden BPO forks spec.

[raulk]

This is a minimal changeset on top of the existing BPO hardforks spec to solve some loose ends.

Concretely, this adjusts the following at the consensus layer:

• Assigns a fork_version to pure BPO forks.
• Updates compute_fork_version() to take BPO forks into account.
• Specifies ENR field mapping to account for upcoming BPO forks.
• Notes that p2p topics will roll over.

Furthermore, this makes a broader adjustment to avoid future ambiguity:

• We now require that all future blob parameter changes be performed through blob schedule configuration—even those introduced by regular (named) hardforks.

---

While thinking through this, I realized there may be value in formalizing the broader concept of "parametric forks" by defining how they interact with regular hard forks, establishing a process for bringing additional parameters under this management mechanism, and iterating on client config schemas (https://github.com/eth-clients) to introduce time-based triggers as a top-level organizational axis. Perhaps the subject of a future proposal.

[eth-bot]

File EIPS/eip-7892.md

Requires 1 more reviewers from @ethDreamer

[raulk]

Following further discussion on Discord, I posted a synthesis of current consensus/opinions and recommendation here: ethereum/consensus-specs#4331 (comment)

[rolfyone]

why are we renaming the blob schedule? we can just do this under blob_schedule still unless we're worried about old parsers breaking or something... The blob_schedule already resulted in objects and we're just adding an attribute imo..

[raulk]

This is a leftover from a previous version that backported Deneb and Electra. Fixing.

[raulk]

Superseded by #9840.