Preserve hotkey successor mappings until netuid dissolution

[UnArbosFour]

Summary

Stops clearing HotkeySuccessor when a neuron is replaced or re-registered, ensuring smart contracts can continue locating funds after a hotkey swap followed closely by neuron deregistration—even within the same batch.

Netuid dissolution already clears HotkeySuccessor and HotkeyRoot by prefix, so this PR retains that existing cleanup and documents why lineage must survive individual neuron deregistration.

Testing

• Adds regression coverage for same-block hotkey swap and neuron deregistration.
• Retains existing coverage confirming lineage is cleared when the netuid is dissolved.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
subtensor	Ready	Preview	Aug 28, 2026 8:16pm

[github-actions]

🛡️ AI Review — Skeptic (security review)

VERDICT: SAFE

VERY HIGH scrutiny (young account and one public repo), mitigated by repository write permission and substantive contributions; no Gittensor association found; fix/hotkey-successor → release-v451.

Static review found only a helper rename/documentation refinement in runtime code plus revised lineage tests. Bounded traversal and netuid-dissolution cleanup remain intact. No dependency, CI, or AI-review trust-boundary files changed.

Findings

No findings.

Conclusion

No malicious behavior or security vulnerability found in the current diff.

---

🔍 AI Review — Auditor (domain review)

VERDICT: 👎

Gittensor association UNKNOWN; the author is a recent but established repository contributor with write permission.

Description discrepancy

The PR claims to stop clearing successor mappings during neuron replacement/re-registration, but the production change only renames clear_stale_hotkey_successor; before and after this PR, replace_neuron and append_neuron remove the successor of new_hotkey. The described lifecycle behavior is therefore unchanged.

The overlapping PRs touch only individual shared files and address unrelated features.

Quick checks: git diff --check passed. cargo fmt --check --all was attempted but could not run because rustup cannot write to its locked environment under /home/runner/.rustup.

Findings

Sev	File	Finding
MEDIUM	pallets/subtensor/src/tests/hotkey_lineage.rs:427	Test does not reproduce the claimed regression	inline

Conclusion

The proposed regression test does not exercise the stated swap-then-deregistration scenario and would pass before this PR. The PR should demonstrate the actual failing behavior and include the production change that fixes it, or correct its stated purpose.

[github-actions]

🔄 AI review updated — Skeptic: SAFE Auditor: 👍

[github-actions]

🔄 AI review updated — Skeptic: SAFE Auditor: 👍

[github-actions]

AI review — see the sticky summary comment for the verdict and the inline comments below for specific findings.

[github-actions]

[MEDIUM] Test does not reproduce the claimed regression

This directly creates h0 → h1 while h0 remains registered, then replaces h0 with an unrelated key. That is not the stated swap-then-deregistration flow: after a real h0 → h1 swap, h1 is registered and h0 is not. Moreover, the pre-PR replace_neuron already cleared the successor of new_hotkey (replacement here), so this test passes without the PR. Exercise do_swap_hotkey followed by deregistration/replacement of the resulting live hotkey and first demonstrate that the test fails against the base branch.

[github-actions]

🔄 AI review updated — Skeptic: SAFE Auditor: 👎