Redesign fork choice abstractions

[zilm13]

PR Description

Spec: ethereum/consensus-specs#5249

Fixed Issue(s)

Closes #10754
Closes #10773

Documentation

• I thought about documentation and added the doc-change-required label to this PR if updates are required.

Changelog

• I thought about adding a changelog entry, and added one if I deemed necessary.

---

Note

High Risk
Changes core fork-choice ancestry, proposer-boost weighting, and Gloas late-reorg paths; incorrect payload-status handling could affect head selection and reorg safety.

Overview
Aligns fork-choice helpers with the Gloas/EIP-7732 spec by making ancestry walks payload-status-aware instead of defaulting every ancestor to a PENDING ForkChoiceNode.

getAncestorNode now takes a starting ForkChoiceNode (not just a block root) and is implemented on ForkChoiceStrategy by walking protoarray from that node identity. Call sites such as dependent-root resolution in ForkChoice pass the full node so parent-variant selection is preserved.

Adds isAncestor on ForkChoiceUtil (phase0: root match via getAncestorNode) with a Gloas override that also requires matching payload status (PENDING ancestors act as a wildcard). Gloas proposer-boost / attestation-weight logic uses isAncestor instead of ad-hoc ancestor-node comparisons.

Renames isShufflingStable to isNotEpochBoundary to mirror spec naming. Renames resolveVoteNode to getSupportedNode on ForkChoiceModel (protoarray score deltas).

specrefs drop several fork-choice items from the “not implemented” list and point is_ancestor, get_supported_node, is_epoch_boundary, and Gloas get_ancestor at the new Java implementations.

^{Reviewed by Cursor Bugbot for commit 461ef5a. Bugbot is set up for automated code reviews on this repo. Configure here.}

[tbenr]

LGTM