fix(orb): invalidate durable CI cache on coalesced fork-PR completions

[bitfathers94]

fix(orb): invalidate durable CI cache on coalesced fork-PR completions

maybeReReviewOnCiCompletion returns early for fork PRs (empty
pull_requests[]) once a completion event in a burst is head-SHA
coalesced, before ever reaching the invalidation loop whose own comment
promises it runs "for EVERY resolved PR, regardless of whether the
re-review below actually fires". Only the first completion per 60s
window invalidated the durable CI-state cache, so a reader could observe
a stale pre-completion aggregate for up to the cache TTL — for fork PRs
only (same-repo PRs invalidate before their coalesce check).

Invalidate the durable cache in the coalesced branch too, resolving via
the fast stored-DB head-SHA lookup only (no live fork fallback — the
round-trip the coalesce exists to avoid; an untracked fork the DB misses
has no cache entry to clear, mirroring maybeInvalidateCiCacheOnLegacyCiEvent).
The coalescing itself is unchanged: it still suppresses the duplicate
re-review dispatch.

Closes #8684

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.19%. Comparing base (6735ef1) to head (cbbfe50).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8737      +/-   ##
==========================================
+ Coverage   90.56%   91.19%   +0.63%     
==========================================
  Files          96       97       +1     
  Lines       22490    25677    +3187     
  Branches     3884     5016    +1132     
==========================================
+ Hits        20367    23417    +3050     
- Misses       1945     1993      +48     
- Partials      178      267      +89     

Flag	Coverage Δ
backend	95.70% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/queue/processors.ts	95.70% <100.00%> (ø)

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-26 05:32:27 UTC}

2 files · 1 AI reviewer · no blockers · readiness 90/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This PR closes a real gap: the coalesced-completion early-return for fork PRs (empty pull_requests[]) skipped the durable CI-state cache invalidation that the non-coalesced path in the invalidation loop below always runs, so a reader could see a stale CI aggregate for the cache TTL on fork PRs specifically. The fix resolves the coalesced PR via the fast stored-DB lookup (listOpenPullRequests filtered by headSha) rather than the live commits/pulls round-trip the coalescing exists to avoid, mirroring the stated maybeInvalidateCiCacheOnLegacyCiEvent pattern, and both catch()-wraps keep the resolve/invalidate best-effort so a DB or cache hiccup doesn't fail the webhook. The two added tests are well-targeted: one drives two real check_suite events through processJob and asserts the durable sync-state row is actually nulled on both the non-coalesced and coalesced event (with a throwing fetch stub proving no live round-trip occurred), and the other confirms the coalesced second event still suppresses the duplicate re-review audit — this is a genuine reachable-path fix, not a can't-occur guard.

Nits — 3 non-blocking

• src/queue/processors.ts: the new invalidation branch calls listOpenPullRequests(env, repoFullName) and filters in-memory by headSha rather than a headSha-scoped DB query — fine at typical open-PR counts but worth confirming this matches the existing invalidation loop's approach rather than adding a second, differently-shaped query path.
• The PR description states the invalidation loop's own comment already promises it runs 'for EVERY resolved PR, regardless of whether the re-review actually fires' — worth double-checking that comment is still accurate post-fix or updating it to reference the new coalesced branch as well.
• Consider whether invalidateCiStateCache could be batched/parallelized (e.g. Promise.all) instead of sequential awaits in the for-loop if a repo can have many open fork PRs sharing edge-case head SHAs, though this is a minor perf nit not a blocker.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#8684
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (1 linked issue).
Validation posture	⚠️ 20/25	Preflight is ready, but the PR body does not name the validation run.
Contributor workload	✅ 10/10	Author activity: 44 registered-repo PR(s), 29 merged, 0 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	bitfathers94; Gittensor profile; 44 PR(s), 0 issue(s).
Improvement	✅ Minor	risk: clean · value: minor

Linked issue satisfaction

Addressed
The diff moves cache invalidation into the coalesced fork-PR branch (resolving via the stored DB head-SHA lookup before the early return) so it now runs unconditionally, and adds both a test proving invalidation happens on both coalesced events and a regression test proving the re-review dispatch is still deduplicated.

Review context

• Author: bitfathers94
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: TypeScript
• Official Gittensor activity: 44 PR(s), 0 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Start here: Add validation command/output.
• Then work through the remaining 1 step in the Signals table above.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by LoopOver, a quiet PR intelligence layer for OSS maintainers.

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.