fix(ci): coalesce Bandit pull request scans - #1606
Conversation
Keep push and manual scans independent while cancelling only superseded scans for the same pull request and repository. Signed-off-by: Seongho Bae <me@seonghobae.me> Co-authored-by: Codex <noreply@openai.com>
|
Important Draft PR not reviewedDraft PRs are not automatically reviewed by default.
To automatically review draft PRs, update your CodeRabbit configuration: reviews:
auto_review:
drafts: trueThanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
현재 head acb374e9b68faf1911a4298155aead3e4144b22e에서 Application CI, Security, Docker image validation, CodeQL 및 중앙 review 관련 검사가 실제로 접수됐습니다. 확인 시 실행 job은 queued, publish/deploy는 pull_request 조건상 skipped입니다. queued를 성공으로 세지 않으며, exact-head terminal checks와 독립 리뷰를 기다립니다. 38개 로컬 회귀검사와 diff check는 별도 분모입니다. |
|
재확인: 현재 exact head acb374e9b68faf1911a4298155aead3e4144b22e의 12개 Actions run이 모두 queued이며 완료된 실행은 없습니다. Bandit run 34194670652도 queued입니다. 동시성 그룹 API는 |
|
최신 exact head acb374e9의 hosted 상태가 변했습니다. Bandit |
|
현재 head 재검증: backend, security, CodeQL, Bandit 및 정적 분석은 SUCCESS로 완료됐습니다. Docker image validation·frontend는 IN_PROGRESS이고 중앙 OpenCode/Strix/Noema 및 admission 단계는 QUEUED입니다. 따라서 부분 성공만 기록하며 전체 required checks 통과나 merge readiness로 해석하지 않습니다. |
|
재확인: backend/frontend 및 정적 보안 검사는 계속 완료됐고, Docker 이미지 검증은 IN_PROGRESS입니다. admission·required workflow bootstrap·변경 범위 검사는 QUEUED입니다. 현재 head는 변경되지 않았으므로 기존 SUCCESS를 정확히 이 head의 부분 결과로만 사용합니다. |
|
Exact-head audit: head |
|
현재 head 이 결함은 Naruon에서 복제·우회하지 않고 canonical owner 선행 PR ContextualWisdomLab/.github#2043 및 exact successor #2044에서 수리 중입니다. 두 PR의 보호 Checks와 인증된 terminal verdict가 완료되기 전에는 이 PR을 Ready/승인/병합하지 않습니다. |
Current authority — 2026-09-08
develop@042b0c70531b229af3acbd0421a2f23098d848b3c1286c1631365862e2b576fb045388796d3bd246develop: 0 files, +0/-0, behind 0Finding and owner repair
The original
acb374e9b68faf1911a4298155aead3e4144b22echange created a second Bandit concurrency owner and grouped every pull-request run by repository + PR number while cancelling all PR runs. That conflicts with the existing Bandit concurrency lineage #1554 and its successor #1562, which preserve a separate generation for intentional manual reruns so a later PR event cannot cancel a user-requested rerun of an older head.This was treated as an ownership/contract repair, not as disposable race output. Ordinary child
55d601cdc494921cb2e98e04612709e43e4f6a69restores the protected product/workflow tree while retaining the generated commit in ancestry. Fresh comparison then showed that the branch ancestry itself was stale/diverged from protecteddevelop, so two-parent ordinary mergec1286c1631365862e2b576fb045388796d3bd246adopts exact protecteddevelopas the second parent without force-push or destructive rebase. The resulting compare has the protected head as merge base, is ahead-only, and has no effective files.Succession boundary
The useful intent — coalescing superseded Bandit PR scans without cancelling manual reruns — remains owned by #1554/#1562. This PR stays open/Draft as provenance until the canonical owner/successor normally integrates into protected
developand a post-merge comparison proves the intended Bandit concurrency/test contract is present there. Only then is zero-delta closure justified as complete succession.Do not merge this zero-delta PR as an alternate Bandit owner. No force-push, destructive rebase, self-approval, dummy/no-op requeue, review fabrication/dismissal, synthetic status, central-workflow copy, or gate weakening.