Skip to content

docs: record code-mode delivery and qualify Desktop model switching - #3884

Merged
lidge-jun merged 3 commits into
devfrom
codex/260907-e-code-mode-docs
Sep 7, 2026
Merged

docs: record code-mode delivery and qualify Desktop model switching#3884
lidge-jun merged 3 commits into
devfrom
codex/260907-e-code-mode-docs

Conversation

@lidge-jun

@lidge-jun lidge-jun commented Sep 7, 2026

Copy link
Copy Markdown
Owner

Summary

Verification

  • Current manual chain: docs: complete Raycast status and model-name recovery guidance #3883 head e0a1a8836e447ece3010ba98813e6feb3161eefddocs: record code-mode delivery and qualify Desktop model switching #3884 top f8c8f6dd15b1518f25a3b1cd7b15c9b6c8b24f62, based on dev@f4a4b468feb04094b90c909d82e04606caf92663.
  • Central Cross-platform CI on the top: run 34121907231, lane=all, all 16 required per-chain jobs SUCCESS at the exact top (attempt 1). The lower layer has no separate suite dispatch. Prior run 34120282456 was cancelled after scheduling instructions crossed and is not evidence.
  • Required per-chain checks: Linux 4 shards, macOS 2 shards, gates/storage/api/keyring/npm/docker. Windows 6 shards and macOS control are excluded under the owner-approved per-chain policy and reserved for final train validation.
  • Independent Astra source audit (Tesla): PASS, no corrections, at the exact base/lower/top above. Layer patches and six Desktop locale replacements match reviewed content. C's seven code-mode translations are already present and semantically consistent in the pinned base.
  • Per main instruction after D landed, merge-tree against dev@da707ccb6 is conflict-free, expected tree c415b6abd77d5f05148fa2a114116d4b492177e0. It includes D changes and therefore differs from tested top tree 986ae11d785934bb4e118a869d38532c1365863d; no combined-tree test claim is made.
  • Local suite, typecheck, build, install and privacy scan: NOT RUN by instruction. Source diff/applicability checks passed; they do not establish a documentation build or runtime pass. Required-head CI passed. CodeRabbit review after ready-for-review and current policy refresh remain to be checked.

Manual chain (stack: null; merge bottom-up):

Layer PR / branch Base Scope
1 #3883 / codex/260907-e-docs-followups dev Raycast and locale followups
2 this PR / codex/260907-e-code-mode-docs codex/260907-e-docs-followups Delivery record and Desktop qualification; CI top

Depends on #3883. Review this layer's diff only. Every commit uses [skip ci]; final centrally dispatched lane=all CI and policy gates are required before LANDABLE.

Checklist

  • Scope stays focused and avoids unrelated cleanup.
  • Docs or release notes were updated when needed.
  • Security-sensitive changes were reviewed for secrets, auth, and unsafe defaults (documentation only; findings cited are already public).

@coderabbitai

coderabbitai Bot commented Sep 7, 2026

Copy link
Copy Markdown
Contributor

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

  • Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@github-actions

github-actions Bot commented Sep 7, 2026

Copy link
Copy Markdown
Contributor

Deterministic PR hygiene checks passed.

@github-actions github-actions Bot added the documentation Improvements or additions to documentation label Sep 7, 2026
@lidge-jun

Copy link
Copy Markdown
Owner Author

리뷰 · 우선순위 46 / 80

이 PR은 이미 dev에 합쳐진 #3854(code-mode 호스트 계약·실패 주석)의 인도 기록과, Claude Desktop 모델 전환 문서의 표현을 약하게 고치는 문서 레이어입니다. 베이스가 dev가 아니라 #3883 브랜치(codex/260907-e-docs-followups)라서, 아래층 Raycast/로케일 후속이 먼저 있어야 깨끗이 올라갑니다. 지금 HEAD d00615d56에는 #3854 본문 코드가 이미 있고, 플랜 폴더 devlog/_plan/260907_code_mode_host_contract/에는 000·001·010·020·030까지 있습니다. 이 PR이 새로 넣는 040_delivery_record.md는 “무엇을 고쳤다”가 아니라 “어느 커밋·어느 CI 런에서 무엇이 성공/스킵이었는지”를 남기는 감사 스냅샷입니다.

기록 내용은 솔직합니다. PR 헤드 6bdcba5bff와 머지 ece556a6ed, 런 34090946313·34091933836, 잡별 성공/스킵 표를 붙입니다. Windows full-suite 매트릭스가 두 런 모두 SKIPPED였다는 점, ci 집계가 스킵을 통과로 바꾸지 않는다는 점, merge-head의 gates에서 GUI lint/build·dashboard-preview가 스킵이었다는 점을 분명히 씁니다. Limits 절에서는 Anthropic/Google/OpenAI-chat/command-code에 exec-result 주석 시암이 없다는 잔여와, 공개 리뷰 스레드 두 개(discussion_r3947178410, discussion_r3947178418)가 감사 시점에도 unresolved였다는 사실을 고쳤다고 말하지 않고 남깁니다. 이건 릴리스 노트가 아니라 인도 감사 문서라서, 과장 없이 남는 쪽이 dev의 유지보수 습관과 맞습니다.

두 번째 파일은 docs-site/.../guides/claude-code.md입니다. 예전 문장은 Desktop footer 피커가 안 바뀌면 /model <id>를 쓰라고 단정에 가깝게 읽혔습니다. 열려 있는 #3782(Windows Claude Desktop 1.46388.4에서 피커와 /model 모두 초기 모델을 유지)를 인용해, 그 우회도 실패할 수 있다고 낮춥니다. 프로필 기본 모델 선택·재적용·새 대화는 보증된 수정이 아니라 점검 단계라고 적고, OpenCodex는 피커 상태를 보지 못하며 요청에 실려 온 모델 id만 따른다고 다시 말합니다. #3782는 이 PR로 닫지 않습니다. code-mode 가이드 7개 번역은 lane C 소유라 이 레이어가 만지지 않습니다.

범위는 문서 두 파일이고 런타임·라우팅을 바꾸지 않습니다. types/config 대분할과도 무관합니다. 다만 드래프트·[skip ci]·#3883 의존이라, 단독 머지 점수는 낮고 “체인으로 정리할 문서 부채” 점수입니다. #3879#3880#3881 코드 캐리 스택보다는 급하지 않습니다.

devlog/_plan/260907_code_mode_host_contract/040_delivery_record.md - #3854 헤드/머지 SHA·CI 잡표·Windows full-suite 스킵·미해결 리뷰 스레드를 역사 스냅샷으로 고정합니다. 수정 주장 없음.
docs-site/.../guides/claude-code.md Desktop /model 절 - #3782를 인용해 우회 실패 가능·프로필/새 대화는 점검 단계라고 낮춥니다. #3782는 열린 채 유지.
베이스 codex/260907-e-docs-followups (#3883) - dev 직베이스가 아님. 아래층 없이 머지하면 충돌·리베이스 비용이 납니다.
lane C 번역 7개 - 이 PR 범위 밖. 인도 기록과 Desktop 문구만 이 레이어 책임입니다.

메인테이너의 판단이 필요한 지점

  • #3883이 먼저 ready·머지된 뒤에만 이 PR을 ready로 올릴지
  • 040의 unresolved 리뷰 두 건을 후속 코드 PR로 잡을지, 문서로만 남길지
  • #3782를 help wanted/needs-design 등으로 라벨만 손볼지(이 PR 범위 밖)
  • 문서 체인 top에서 중앙 lane=all을 누가 돌릴지(본문에 “main session”으로만 적혀 있음)

너의 추천
#3883 머지 후 이 레이어를 리베이스·ready로 올리고, 체인 top CI가 초록이면 머지하세요. 내용은 과장 없는 인도 기록과 Desktop 문서 정직화라 가치는 있으나, #3882나 #3879#3881 코드 캐리보다 급하지 않습니다. 닫지 말고 드래프트 체인으로 유지하세요.

이 댓글은 grok-bot이 작성했습니다

@lidge-jun
lidge-jun force-pushed the codex/260907-e-docs-followups branch from bdd9956 to 3ee30e8 Compare September 7, 2026 11:04
@lidge-jun
lidge-jun force-pushed the codex/260907-e-code-mode-docs branch from 124fc03 to b201307 Compare September 7, 2026 11:04
@lidge-jun
lidge-jun force-pushed the codex/260907-e-docs-followups branch from 3ee30e8 to 403acd2 Compare September 7, 2026 11:55
@lidge-jun
lidge-jun force-pushed the codex/260907-e-code-mode-docs branch from 1ae0a59 to 8d4edc7 Compare September 7, 2026 11:55
@lidge-jun
lidge-jun force-pushed the codex/260907-e-docs-followups branch from 403acd2 to e0a1a88 Compare September 7, 2026 12:08
@lidge-jun
lidge-jun force-pushed the codex/260907-e-code-mode-docs branch from 8d4edc7 to f8c8f6d Compare September 7, 2026 12:08
@lidge-jun
lidge-jun marked this pull request as ready for review September 7, 2026 12:37
@lidge-jun
lidge-jun requested a review from Ingwannu as a code owner September 7, 2026 12:37
@lidge-jun
lidge-jun changed the base branch from codex/260907-e-docs-followups to dev September 7, 2026 12:37
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Sep 7, 2026

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review Completed 2026-09-07T12:39:29.109933Z f8c8f6d Draft marked ready
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@lidge-jun
lidge-jun merged commit 74089fd into dev Sep 7, 2026
29 of 56 checks passed
@lidge-jun
lidge-jun deleted the codex/260907-e-code-mode-docs branch September 7, 2026 12:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

documentation Improvements or additions to documentation

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant