Revert #2237 — forkid wire-format change is a breaking p2p change

[lucca30]

Summary

Reverts #2237 ("core/forkid: include polygon-specific forks in wire forkid").

This is a breaking p2p change that was not flagged as such when #2237 merged.

What went wrong

#2237 swapped gatherForks (lowercase, reflection-based, outer ChainConfig only) for GatherForks (uppercase, polygon-aware — also walks config.Bor.{Jaipur,Delhi,Indore,Ahmedabad,Bhilai,Rio,Madhugiri,MadhugiriPro,Dandeli,Lisovo,LisovoPro,Giugliano,Chicago}Block) in the wire-forkid path (core/forkid/forkid.go::NewID, NewStaticFilter).

This is correct in intent — bor's old forkid was lying about which forks were active and erigon ≥3.6.0 already includes these blocks — but changing what gets hashed in the wire forkid is, by definition, a coordinated-upgrade event. Any node running the new code computes a different forkid than nodes running the old code on the same chain, and the eth handshake fails with fork ID rejected: local incompatible or needs update.

Live evidence

Deployed v2.8.3-beta3 (which contains #2237) to an Amoy observation node (posdevnodes-amoy-2) on 2026-05-27. Symptoms within 5 minutes:

• Peer count oscillated 0 ↔ 1 (no stable handshakes).
• Looking for peers peercount=0 tried=10–35 static=13 every 10 s — static peers being dialed continuously, all dropped immediately.
• Chain stuck at the same block; no new imports.
• No errors at INFO level — silent rejection at the handshake layer.

Rolling the same node back to v2.8.0-beta2 (which does not contain #2237) restored peers to ~18 within seconds. Confirms the diagnosis: it's the forkid hash mismatch, not anything else in the beta3 delta.

This will affect mainnet identically the moment a v2.8.3 (with #2237) bor talks to any v2.8.2-or-older bor, because every Polygon-specific fork above is also activated on mainnet at non-zero blocks.

Why a plain revert, not a fix-forward

There is no algorithmic trick that makes a forkid hash change non-breaking on a chain where the affected forks have already activated. The only valid rollout strategies are:

1. Gate the new GatherForks behavior on a future fork block. Introduce a new BorConfig.ForkidFixBlock (or piggyback on the next planned consensus fork). Before that block: use legacy gatherForks for wire forkid computation. At and after: use GatherForks. Everyone deploys the binary well in advance; the protocol-level switchover is atomic at the block boundary. This is the canonical Ethereum pattern.
2. Coordinate a flag-day rollout with a config flag defaulting to legacy, flipped chain-wide on a known date.

Either approach should be a follow-up PR, designed and reviewed as a network upgrade. This PR just unblocks the existing release line.

What needs to follow

• Follow-up PR re-introducing the polygon-aware forkid behavior gated on a ForkidFixBlock (or equivalent) — must include chain-config entries for mainnet and amoy, an explicit announcement to operators, and a soak period on devnet first.
• Release notes for the upcoming v2.8.x lineage should call out that the forkid fix is deferred to a coordinated upgrade.

Test plan

• go build ./... passes after revert
• Same revert applied to v2.8.3-candidate and shipped as v2.8.3-beta4 — node recovered peers + chain progression on Amoy
• Reviewer to confirm core/forkid/forkid.go matches pre-core/forkid: include polygon-specific forks in wire forkid #2237 state (only the two GatherForks → gatherForks call sites)

[claude]

Claude Code Review

This repository is configured for manual code reviews. Comment @claude review to trigger a review and subscribe this PR to future pushes, or @claude review once for a one-time review.

_{Tip: disable this comment in your organization's Code Review settings.}

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 52.85%. Comparing base (d99ab21) to head (0f03fad).

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #2248      +/-   ##
===========================================
- Coverage    52.86%   52.85%   -0.02%     
===========================================
  Files          886      886              
  Lines       156811   156811              
===========================================
- Hits         82900    82883      -17     
- Misses       68663    68677      +14     
- Partials      5248     5251       +3     

Files with missing lines	Coverage Δ
core/forkid/forkid.go	76.69% <100.00%> (ø)

... and 22 files with indirect coverage changes

Files with missing lines	Coverage Δ
core/forkid/forkid.go	76.69% <100.00%> (ø)

... and 22 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.