|
31103
|
7
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31104
|
9
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31105
|
1
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31106
|
3
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31107
|
4
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31108
|
8
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31109
|
10
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31110
|
11
|
5
|
7
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"3a794261c {"Commits":[{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"}],"HeadCommit":{"Sha1":"3a794261c6bc270385e7eb7ed870e8ec38b8902d","Message":"1\n","AuthorEmail":"11833999+littlemaidi@user.noreply.gitee.com","AuthorName":"LITTLEMAIDI","CommitterEmail":"11833999+littlemaidi@user.noreply.gitee.com","CommitterName":"LITTLEMAIDI","Timestamp":"2026-09-18T17:33:15+08:00"},"CompareURL":"vodtest/pc/compare/f0a3ea0e395022ad8ffd30ff8eeb1bba490c9fa5...3a794261c6bc270385e7eb7ed870e8ec38b8902d","Len":1}...
|
1789724002
|
Edit
Delete
|
|
31111
|
5
|
5
|
5
|
76
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"3deb4ea37 {"Commits":[{"Sha1":"3deb4ea37b83c2e2977c9c2b8e8166c101002860","Message":"chore(reports): PEER_UNSATISFIED 判据的二十八份静态证据 回绑 @ 8e5c633，刷新快照锚点\n\n本批次最后一个提交，按 C241 同时刷新快照锚点到 8e5c633（声明落后 0，本提交自身使实测\n落后 1，在 +1 自指窗内）。\n\n二十八份 reports/*.latest.json 全部 clean 绑定（gitSha=8e5c633、worktreeDirty:false）。\n整轮 28 步全执行，前 27 步全绿，末步 check:kernel-admission 仍是那 8 处已知诚实红。\n\nplatform-dependency-provenance 报告新增一处 registered：PEER_UNSATISFIED（client-fact\nrc.3 的 peer 0.16.1 ↔ 本仓已解析 0.22.0），期限 2026-12-18。门禁整体仍绿——登记的作用是\n把「pnpm 只打 WARN、门禁零覆盖」变成「可见、有期限、到期转红」，不是让它消失。\n\n工作区内三份 reports/ui-*.latest.json 的改动早于本批次，未提交也未还原。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:38:50-07:00"},{"Sha1":"8e5c633cbfbc16afc61c95b8cfcecb90a2f1aa88","Message":"feat(governance): PEER_UNSATISFIED 判据——把 pnpm 只打 WARN 的跨包版本不一致变成可见可登记\n\nB1 pin @juhai/kernel@0.22.0 后 pnpm 报了一行 unmet peer 就继续装，**本仓原有门禁一条都\n抓不到**。按内核退回 pin 立项 §6 三条出路裁为 ③（登记偏差 + 加判据，不停等发布列车）。\n\n新判据 PEER_UNSATISFIED：lock 里 @juhai/* 之间声明的 peer 必须被已解析版本满足；\n未满足即红，可按既有五字段口径登记降级为 registered，过期转红。peer 指向本仓根本没装的\n包时不判（防过度判红）。自检从 7 条扩到 12 条，新增 5 条全覆盖上述分支。\n\n实测命中一条真问题，并顺带拆出更深的一层：\n\n- 已发布的 @juhai/client-fact@1.0.0-rc.3 声明 peer @juhai/kernel: 0.16.1（发布物自证），\n 而控制面仓里**同版本的源码**声明 0.22.0——2026-09-17 框架同步只改了 package.json 没重发\n- 把 registry 上的 rc.3 拆开看：dist 里**没有** Authorization 回查鉴权，而源码 reader.ts:31\n 发 Bearer ${token}。**同一个版本号，registry 与源码是两份不同的代码**\n- 该发布物 provenance.json 为 runner:\"local\" / reports:[] / **gitSha:\"\"**——空字符串。\n 任何消费者都无法验证装到的包出自哪次提交，上一条「同版本不同内容」因此无法被机器发现\n\n不兼容风险经 delta 分析可论证为极低：内核 v0.16.1 → v0.22.0 只改 health.ts(+126)、\nalert.ts(+53)、tenant.ts(+12) 三处纯增，client-fact 一个都不消费。故登记而非阻断，\n期限 2026-12-18，blockedBy 指向平台发布列车（G-12 计费阻塞）。\n\n为什么不选另两条路（立项 §6 已记）：② 退回 pin 0.16.1 的当前代价几乎为零——三处加固\n（replay cursor-unknown、WS 载荷上限、订阅数上限）**实测都已在 v0.16.1 里**，我上一轮说\n「退回等于放弃 6 个版本加固」是错的——但它是伪解，B2 要迁 alert/tenant、B3 要迁 health 时\npeer 冲突原样回来；① 等平台重发会把本仓进度挂在不可控的跨仓列车上，且若仍走本地旁路，\nprovenance 依旧无 gitSha，只解决 peer 不解决可对账性。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:37:17-07:00"}],"HeadCommit":{"Sha1":"3deb4ea37b83c2e2977c9c2b8e8166c101002860","Message":"chore(reports): PEER_UNSATISFIED 判据的二十八份静态证据 回绑 @ 8e5c633，刷新快照锚点\n\n本批次最后一个提交，按 C241 同时刷新快照锚点到 8e5c633（声明落后 0，本提交自身使实测\n落后 1，在 +1 自指窗内）。\n\n二十八份 reports/*.latest.json 全部 clean 绑定（gitSha=8e5c633、worktreeDirty:false）。\n整轮 28 步全执行，前 27 步全绿，末步 check:kernel-admission 仍是那 8 处已知诚实红。\n\nplatform-dependency-provenance 报告新增一处 registered：PEER_UNSATISFIED（client-fact\nrc.3 的 peer 0.16.1 ↔ 本仓已解析 0.22.0），期限 2026-12-18。门禁整体仍绿——登记的作用是\n把「pnpm 只打 WARN、门禁零覆盖」变成「可见、有期限、到期转红」，不是让它消失。\n\n工作区内三份 reports/ui-*.latest.json 的改动早于本批次，未提交也未还原。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:38:50-07:00"},"CompareURL":"luoanwu/digital-employee-os/compare/fd3ba87ed111d9245012f0578178cccb4743ab20...3deb4ea37b83c2e2977c9c2b8e8166c101002860","Len":2}...
|
1789724335
|
Edit
Delete
|
|
31112
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"7433af77e {"Commits":[{"Sha1":"7433af77e643c41f2ccb34fb8aebe363ec76e306","Message":"test(治理): L6 的判定搬进 core 并补负向用例——它此前是唯一零覆盖的一条\n\n一致性门禁 L1—L5 / L7—L13 的判定都在 lib/workspace-consistency-core.mjs 里、各有负向\n用例，**唯独 L6（worktree 两侧 gitdir）与文件系统读取缠在 check-workspace-consistency.mjs\n里**，全量扫描确认 tests/ 里没有任何一条用例触达它。它「能工作」的唯一证据是恰好没在\n实数据上报错——而「没报错」与「判不出来」在外部看来一模一样。\n\n本轮只搬不改：IO 函数改为只收集事实，判定交给新的纯函数 evaluateWorktreeGitdirs。\n**判词逐字未变，改前改后门禁 --json 输出实测完全相同**（status / counts / issues 三者\n逐字相等，prunableWorktrees 仍为 5、L9 仍为 1）。\n\n补 8 条负向用例，覆盖两个方向的全部分支：\n 方向一 .git 为空 / 指向不存在的登记 / 登记反指不回来（含反指读不出时须写「(空)」）；\n 三种坏法互斥，前一种命中不再叠加后一种\n 方向二 gitdir 为空即红；目标还在就跳过（反指归方向一，不重复报）；\n **目标没了且旁边没有同名工作树 = prunable，只计数不报**——那是 git worktree\n prune 的日常；目标没了但工作树就在旁边 = 失联，判红且不计入 prunable\n\n顺带核实两条初测假阳性，记在这里免得再查一遍：L9 的判定其实在 core 里且有用例\n（composeProjectName / composeProjectCollisions），只是测试里不出现「L9」这个字面量；\nL0 与 L28 是正则误匹配（一个字符类、一处注释里的行号）。\n\n一致性门禁测试 34 → 42，治理层全量 389 → 401（395 过 / 6 跳 / 0 失败）；\n统一入口阻断性门禁 11/11。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:39:16-07:00"},{"Sha1":"1e6a813a72945400129128947654df4ecdc8849d","Message":"docs(回灌): U-39 / U-40——发布物与源码同版本不同内容，且 provenance 无 gitSha\n\n新增 §2E 第 5 轮条目（基础基座 → 基础设施）。采集自 OS 执行「内核退回 pin 立项」B1\n（contracts exact pin @juhai/kernel@0.22.0）时对接已发布客户端包的实测。\n\nU-39 发布物与其源码同版本不同内容，peer 只是露出来的一角。已发布的\n @juhai/client-fact@1.0.0-rc.3 声明 peer @juhai/kernel: 0.16.1（发布物自证），而控制面仓\n 同版本源码声明 0.22.0——2026-09-17 框架同步只改了 runtime/clients/fact/package.json 未重发。\n 进一步拆开发布物：dist 里没有 Authorization 回查鉴权，源码 reader.ts:31 却发\n Bearer ${token}。消费侧对此完全不可见：pnpm 对未满足 peer 只打一行 WARN 就继续装。\n → 平台侧动作：重发 rc.4 携带正确 peer，发布前确认 dist 与源码同一提交。\n 基座侧已自保：新增 PEER_UNSATISFIED 判据 + 五字段登记（期限 2026-12-18）。\n\nU-40 本地旁路发布的发布物不可对账到提交。registry 上 rc.3 的 provenance.json 为\n runner:\"local\" / reports:[] / gitSha:\"\"——空字符串。任何消费者都无法验证装到的包出自哪次\n 提交，U-39 那种「同版本不同内容」因此无法被机器发现。这比 peer 错一版严重：peer 至少还有\n 一个可比的值，gitSha 为空是判据缺席。\n → 平台侧动作：发布物 provenance.json 必须带真实 gitSha，本地旁路也不得留空——旁路可以解除\n CI 身份检查，但不应连「这是哪次提交构建的」都放弃。\n\n两条均标 🔴 下层缺口，阻塞于平台发布列车（G-12 计费）。基座侧不停等：按立项 §6 三条出路\n裁为 ③（登记偏差 + 加判据），B2 / B3′ 继续推进。\n\n一致性门禁复跑：合计 1 条，仅 L9（8 仓共用 compose project api-nestjs），与本次改动无关——\n它正是 OS 在 framework-criteria.json 里登记为 missing 的框架 F14。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:39:13-07:00"}],"HeadCommit":{"Sha1":"7433af77e643c41f2ccb34fb8aebe363ec76e306","Message":"test(治理): L6 的判定搬进 core 并补负向用例——它此前是唯一零覆盖的一条\n\n一致性门禁 L1—L5 / L7—L13 的判定都在 lib/workspace-consistency-core.mjs 里、各有负向\n用例，**唯独 L6（worktree 两侧 gitdir）与文件系统读取缠在 check-workspace-consistency.mjs\n里**，全量扫描确认 tests/ 里没有任何一条用例触达它。它「能工作」的唯一证据是恰好没在\n实数据上报错——而「没报错」与「判不出来」在外部看来一模一样。\n\n本轮只搬不改：IO 函数改为只收集事实，判定交给新的纯函数 evaluateWorktreeGitdirs。\n**判词逐字未变，改前改后门禁 --json 输出实测完全相同**（status / counts / issues 三者\n逐字相等，prunableWorktrees 仍为 5、L9 仍为 1）。\n\n补 8 条负向用例，覆盖两个方向的全部分支：\n 方向一 .git 为空 / 指向不存在的登记 / 登记反指不回来（含反指读不出时须写「(空)」）；\n 三种坏法互斥，前一种命中不再叠加后一种\n 方向二 gitdir 为空即红；目标还在就跳过（反指归方向一，不重复报）；\n **目标没了且旁边没有同名工作树 = prunable，只计数不报**——那是 git worktree\n prune 的日常；目标没了但工作树就在旁边 = 失联，判红且不计入 prunable\n\n顺带核实两条初测假阳性，记在这里免得再查一遍：L9 的判定其实在 core 里且有用例\n（composeProjectName / composeProjectCollisions），只是测试里不出现「L9」这个字面量；\nL0 与 L28 是正则误匹配（一个字符类、一处注释里的行号）。\n\n一致性门禁测试 34 → 42，治理层全量 389 → 401（395 过 / 6 跳 / 0 失败）；\n统一入口阻断性门禁 11/11。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:39:16-07:00"},"CompareURL":"luoanwu/platform-governance/compare/6b045590e6ba93dd703453bc478b3b6c879491f3...7433af77e643c41f2ccb34fb8aebe363ec76e306","Len":2}...
|
1789724358
|
Edit
Delete
|
|
31113
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"72c053ccb {"Commits":[{"Sha1":"72c053ccbb2b4effb1451b6b61e69faf9c313fd5","Message":"docs(治理): README 订正——「判定逻辑在 core」这句话此前对 L6 并不成立\n\n第 49 行写着「判定逻辑在 workspace-consistency-core.mjs，夹具用例 19/19」。前半句对\nL6 是假的（它的判定与文件系统读取缠在 IO 脚本里、测试零覆盖），后半句的 19 也早已陈旧。\n\nL6 的判定已于同日搬进 core 并补 8 条用例，这句话现在成立了；数字改为当前的 42，\n并写明这条例外曾经存在——免得下一个人以为一直如此。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:39:56-07:00"}],"HeadCommit":{"Sha1":"72c053ccbb2b4effb1451b6b61e69faf9c313fd5","Message":"docs(治理): README 订正——「判定逻辑在 core」这句话此前对 L6 并不成立\n\n第 49 行写着「判定逻辑在 workspace-consistency-core.mjs，夹具用例 19/19」。前半句对\nL6 是假的（它的判定与文件系统读取缠在 IO 脚本里、测试零覆盖），后半句的 19 也早已陈旧。\n\nL6 的判定已于同日搬进 core 并补 8 条用例，这句话现在成立了；数字改为当前的 42，\n并写明这条例外曾经存在——免得下一个人以为一直如此。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:39:56-07:00"},"CompareURL":"luoanwu/platform-governance/compare/7433af77e643c41f2ccb34fb8aebe363ec76e306...72c053ccbb2b4effb1451b6b61e69faf9c313fd5","Len":1}...
|
1789724399
|
Edit
Delete
|
|
31114
|
11
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31115
|
9
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31116
|
1
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31117
|
3
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31118
|
4
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31119
|
7
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31120
|
8
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31121
|
10
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"f897e9c04 {"Commits":[{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"}],"HeadCommit":{"Sha1":"f897e9c044c0085de57de1a598692f9a5a54936c","Message":"团购管理包厢设置按钮显示\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:45:09+08:00"},"CompareURL":"vodtest/pc/compare/3a794261c6bc270385e7eb7ed870e8ec38b8902d...f897e9c044c0085de57de1a598692f9a5a54936c","Len":1}...
|
1789724731
|
Edit
Delete
|
|
31122
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"54738a32d {"Commits":[{"Sha1":"54738a32d101f97cc5dd3a5bca7ea426dceea9c6","Message":"feat(治理): L9 的完整仓名单带进 JSON——裁决输入不该每次重新推导\n\nL9 目前是全工作区唯一在报的那条，挂在 N-11 的裁决上，而裁决要回答的正是「哪些仓各自\n补 compose `name:`」。判定本来就算出了完整名单（composeProjectCollisions 返回\nrepos / files），但建 issue 时被丢掉：消息里只写前三个加「等 N 个」，JSON 里一个都没有。\n于是每次想知道是哪八个，都得重新推导一遍。\n\n`issue()` 增一个可选 extra 参数，只用于把**判定已算出、消息里放不下**的结构化数据带进\nJSON；没有 extra 的检查一个字段都不多（用例钉住 L13 的键仍是四个）。\n\n实测：门禁现在报出的 8 个仓与我按同一口径（排除 29 个工作树、按 .git 目录归仓）独立\n重推的结果完全一致——数字员工OS、知识云、AI项目与工单、AI人事系统、设备云、巨嗨智服、\n巨嗨报价系统、嗨赞造意/image-generation。counts 未变（total 1 / L9 1）。\n\n补两条用例：L9 的发现必须携带 repos 且数量与消息声称的一致、不得重复、files 不少于仓数；\n以及 issue() 的默认形状不变。前者守的是「有冲突时名单不得丢」，不是「必须有冲突」——\nL9 清零后它自然为空。\n\n一致性门禁测试 42 → 44，治理层全量 397 过 / 6 跳 / 0 失败。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:46:51-07:00"}],"HeadCommit":{"Sha1":"54738a32d101f97cc5dd3a5bca7ea426dceea9c6","Message":"feat(治理): L9 的完整仓名单带进 JSON——裁决输入不该每次重新推导\n\nL9 目前是全工作区唯一在报的那条，挂在 N-11 的裁决上，而裁决要回答的正是「哪些仓各自\n补 compose `name:`」。判定本来就算出了完整名单（composeProjectCollisions 返回\nrepos / files），但建 issue 时被丢掉：消息里只写前三个加「等 N 个」，JSON 里一个都没有。\n于是每次想知道是哪八个，都得重新推导一遍。\n\n`issue()` 增一个可选 extra 参数，只用于把**判定已算出、消息里放不下**的结构化数据带进\nJSON；没有 extra 的检查一个字段都不多（用例钉住 L13 的键仍是四个）。\n\n实测：门禁现在报出的 8 个仓与我按同一口径（排除 29 个工作树、按 .git 目录归仓）独立\n重推的结果完全一致——数字员工OS、知识云、AI项目与工单、AI人事系统、设备云、巨嗨智服、\n巨嗨报价系统、嗨赞造意/image-generation。counts 未变（total 1 / L9 1）。\n\n补两条用例：L9 的发现必须携带 repos 且数量与消息声称的一致、不得重复、files 不少于仓数；\n以及 issue() 的默认形状不变。前者守的是「有冲突时名单不得丢」，不是「必须有冲突」——\nL9 清零后它自然为空。\n\n一致性门禁测试 42 → 44，治理层全量 397 过 / 6 跳 / 0 失败。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:46:51-07:00"},"CompareURL":"luoanwu/platform-governance/compare/72c053ccbb2b4effb1451b6b61e69faf9c313fd5...54738a32d101f97cc5dd3a5bca7ea426dceea9c6","Len":1}...
|
1789724816
|
Edit
Delete
|
|
31123
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"a7075e40e {"Commits":[{"Sha1":"a7075e40e9aed8fbe69edb7fb7f0ebc05e477214","Message":"docs(计划): 登记 L6 / L9 两处补强，并订正自己提交信息里的一句过头话\n\nL6 此前是唯一判定与 IO 缠在一起、测试零覆盖的规则，已搬进 core 并补 8 条用例，\n只搬不改、输出逐字相同；README 那句「判定逻辑在 core」此前对它是假的，已订正。\n\nL9 的完整仓名单现在进 JSON。但我在那条提交信息里写的「每次都得重新推导」过头了：\nN-11 正文 2026-09-17 就手写了这 8 个仓名。真实缺口是名单只有手抄的一份、门禁不产\n机器形态，两者可以无声分叉。本轮独立重推的 8 个与 N-11 手抄的完全一致。\n\n另记三次初测假阳性与方法教训：这个仓一贯把判定拆进另起名字的核心库、测试用自动发现，\n按名字扫必然假阳性，要判一条规则有没有被测得找它的判定函数。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:47:33-07:00"}],"HeadCommit":{"Sha1":"a7075e40e9aed8fbe69edb7fb7f0ebc05e477214","Message":"docs(计划): 登记 L6 / L9 两处补强，并订正自己提交信息里的一句过头话\n\nL6 此前是唯一判定与 IO 缠在一起、测试零覆盖的规则，已搬进 core 并补 8 条用例，\n只搬不改、输出逐字相同；README 那句「判定逻辑在 core」此前对它是假的，已订正。\n\nL9 的完整仓名单现在进 JSON。但我在那条提交信息里写的「每次都得重新推导」过头了：\nN-11 正文 2026-09-17 就手写了这 8 个仓名。真实缺口是名单只有手抄的一份、门禁不产\n机器形态，两者可以无声分叉。本轮独立重推的 8 个与 N-11 手抄的完全一致。\n\n另记三次初测假阳性与方法教训：这个仓一贯把判定拆进另起名字的核心库、测试用自动发现，\n按名字扫必然假阳性，要判一条规则有没有被测得找它的判定函数。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:47:33-07:00"},"CompareURL":"luoanwu/platform-governance/compare/54738a32d101f97cc5dd3a5bca7ea426dceea9c6...a7075e40e9aed8fbe69edb7fb7f0ebc05e477214","Len":1}...
|
1789724855
|
Edit
Delete
|
|
31124
|
11
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31125
|
9
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31126
|
1
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31127
|
3
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31128
|
4
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31129
|
7
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31130
|
8
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31131
|
10
|
5
|
11
|
18
|
0
|
0
|
refs/heads/pc-260915
|
0
|
{"Commits":[{"Sha1":"ab955d708 {"Commits":[{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"}],"HeadCommit":{"Sha1":"ab955d7085bb1c39efeb02b307b86043375aace6","Message":"小程序内容设置\n","AuthorEmail":"1091045324@qq.com","AuthorName":"caihongyu","CommitterEmail":"1091045324@qq.com","CommitterName":"caihongyu","Timestamp":"2026-09-18T17:52:12+08:00"},"CompareURL":"vodtest/pc/compare/f897e9c044c0085de57de1a598692f9a5a54936c...ab955d7085bb1c39efeb02b307b86043375aace6","Len":1}...
|
1789725138
|
Edit
Delete
|
|
31132
|
1
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31133
|
9
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31134
|
3
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31135
|
4
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31136
|
7
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31137
|
8
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31138
|
10
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31139
|
11
|
5
|
1
|
22
|
0
|
0
|
refs/heads/app-260915
|
0
|
{"Commits":[{"Sha1":"0be2474e1 {"Commits":[{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"}],"HeadCommit":{"Sha1":"0be2474e1e34ad93ff9ac4b30d694c52573501af","Message":"修改包厢mac的时候增加判断。\n","AuthorEmail":"121158035@qq.com","AuthorName":"zhangjunnan","CommitterEmail":"121158035@qq.com","CommitterName":"zhangjunnan","Timestamp":"2026-09-18T17:52:59+08:00"},"CompareURL":"vodtest/app/compare/b95bc8ca61a1eee32824ca00934dcd4561260235...0be2474e1e34ad93ff9ac4b30d694c52573501af","Len":1}...
|
1789725188
|
Edit
Delete
|
|
31140
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"05e3826e9 {"Commits":[{"Sha1":"05e3826e9e9e4dbbdd753252c2134b8494b08f2b","Message":"chore(reports): 工作台两份快照回绑 @ dfd4832\n\ndocs/runbook.md 在 workbench-ops-snapshot 的作用域里且是 error 级——改 §7 那段就会当场\n把 check:evidence 打红。经 reports:rebind 在 HEAD 的干净检出里重跑，两份同源快照一并带回。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:11:36-07:00"},{"Sha1":"dfd483287cee48db9791915cf33e037c4f758f6f","Message":"chore(reports): :68 修复波及的四份验收重跑回绑 @ 08c3788\n\n作用域含 runtime/test 的三份（runtime-acceptance / mainline-acceptance / revocation-sla）\n加上连带重写的 conformance-differential，在 HEAD 的干净检出 + 隔离栈 enterprise-platform-ms23\n重跑，全部 worktreeDirty=false：\n\n- runtime-acceptance passed（7 步全绿）\n- conformance-differential passed\n- mainline-acceptance passed（fact-read-http 11 / fact-persistence 17 / audit-tamper 19 /\n mainline-revocation-chaos 18）\n- revocation-sla partial（p95 5036 ms，20 样本）\n\n结果与修复前一致，说明这次动的只是写报告的时机，没有动判据。\nui-acceptance / identity-ui 作用域不含 runtime/test，未受影响，保持原绑定。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:11:12-07:00"},{"Sha1":"925eee8195821074b0e2a699acc87b4af2ea35c4","Message":"docs(部署,runbook): 记录补 §10——:68 已修、两个方向的负向实测与隔离栈会被清空的教训\n\n- runbook §7 第四处那段的收尾从「代码修复尚未做」改成已落 08c3788，并写清改的是调用点不是\n 删掉占位写；那句「跑之前先确认这份报告已提交」现在只对「跑到一半失败」成立\n- 部署记录补 §10：负向实测两个方向（旧代码毁报告 / 新代码 sha256 不变）、正向确认 t+85s\n 仍写 RUN_NOT_COMPLETED、四份重跑结果，以及中途踩的一脚——enterprise-platform-ms23 是\n 具名共享验收栈且无卷，库和角色会被别的会话清掉，每轮开跑前重新供给比事后排查便宜\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:11:12-07:00"},{"Sha1":"08c3788da42c5e8a4949daa8d5c7f23d30016205","Message":"fix(runtime): revocation-sla 的「已开跑」标记挪到前置校验之后——先毁证据再校验是反的\n\nadda76a 把三个验收 runner 改成「前置失败即退出、不碰报告」，但 revocation-sla 的报告不是\nrunner 写的，是用例自己写的，所以那次没覆盖到它——这是同病的第四处（runbook §7，2026-09-18 补注）。\n\n原形态：beforeAll 的**第一句**就无条件写 status=failed / RUN_NOT_COMPLETED，**下一句**才做\nisolated() 白名单校验。于是「环境没配对、这一轮根本没开始」也会先把上一轮的好报告毁掉。\n2026-09-18 重跑六份验收时第一次就踩到：模块库用了不在白名单里的名字，绑 765300b 的 partial\n当场没了（只因为跑在独立检出里才幸存）。\n\n改法不是删掉这个标记——「跑起来但没跑完」确实该在盘上留 RUN_NOT_COMPLETED 而不是上一轮的\npartial，这是用例自己写报告带来的责任。改的是**调用点**：挪到全部前置都满足之后，即模块库白名单\n通过、真实连库建好动作、API 子进程起来且 /health 就绪的那一刻。\n\n负向实测（两个方向都做了，确认这条改动真的有牙）：\n- 同一条件下跑旧代码：报告被改写成 RUN_NOT_COMPLETED（sha256 变化，好报告已毁）\n- 同一条件下跑新代码：测试照样以 MAINLINE_ISOLATED_DATABASE_REQUIRED 失败，报告 sha256 逐字节不变\n两次都用 DATABASE_URL_PERMISSION=…/platform_permission_notallowed 触发，5 例 skipped。\n\ntypecheck 通过（platform-tests 无 lint 配置，只有 test / typecheck 两个脚本）。\n受影响报告（runtime-acceptance / mainline-acceptance / revocation-sla，作用域含 runtime/test）\n随后在干净检出重跑回绑。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:51:53-07:00"},{"Sha1":"1f6fa2521e4d947ecd0ee43b18e691f836522049","Message":"docs(部署): 记录补 §9——六份验收在隔离栈重跑的结果与三个未写全的前置\n\n六份全部真跑回来（4 passed / 1 partial 设计内 / 1 passed），check:evidence 36 新鲜 / 4 过期。\n补上 runbook §7 没写全的三条：模块库后缀白名单、必须非超级用户角色、模块迁移要先 deploy。\n并登记 revocation-sla.test.ts:68 先毁证据再校验前置的缺陷与不在本轮修的理由。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:47:01-07:00"}],"HeadCommit":{"Sha1":"05e3826e9e9e4dbbdd753252c2134b8494b08f2b","Message":"chore(reports): 工作台两份快照回绑 @ dfd4832\n\ndocs/runbook.md 在 workbench-ops-snapshot 的作用域里且是 error 级——改 §7 那段就会当场\n把 check:evidence 打红。经 reports:rebind 在 HEAD 的干净检出里重跑，两份同源快照一并带回。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:11:36-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/1b2328540134f0118d2e3ed9abe10095e2ae4679...05e3826e9e9e4dbbdd753252c2134b8494b08f2b","Len":6}...
|
1789740912
|
Edit
Delete
|
|
31141
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/claude/zealous-tu-6f0fca
|
0
|
|
1789740931
|
Edit
Delete
|
|
31142
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/claude/zealous-tu-6f0fca
|
0
|
{"Commits":[{"Sha1":"031878b2a {"Commits":[{"Sha1":"031878b2a7dbd2e2021b6dd71b0197e50cec3c96","Message":"docs(计划): 第四处同病改判已修（08c3788），并订正本条原先给错的一个修法\n\n登记时给了两个修法，取的是「把写报告挪到前置校验之后」；另一个「直接删掉占位写」被接手\n会话否掉，理由成立，本条一并订正：「跑起来但没跑完」确实该在盘上留 RUN_NOT_COMPLETED 而\n不是上一轮的 partial——这是用例自己写报告带来的责任，和「上轮结论已被后续提交推翻」不是\n同一个事实，不能用 check:evidence 判过期替代。\n\n实现是把那句抽成 markRunStarted()，调用点挪到模块库白名单通过、真实连库、API 子进程\n/health 就绪之后。两个方向都做了负向实测：同一条件下旧代码毁报告（sha256 变化），新代码\n测试照样失败而报告逐字节不变；正向确认 t+85s 兜底仍写 RUN_NOT_COMPLETED。\n\n受影响四份已重跑回绑 dfd4832（全绑 08c3788 / dirty=false，结果与修复前一致，说明动的只是\n写报告的时机）；runbook §7 收尾同步改成已修（925eee8）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:13:53-07:00"},{"Sha1":"f7a6b8eb62651f612a8a0e9666b22a5413b4fb1a","Message":"docs(计划): 登记第四处「前置失败销毁好证据」——revocation-sla 由用例自己在 beforeAll 写 failed\n\nadda76a 把三个 runner（runtime / UI / 双后端差分）统一改成前置失败即退出、不碰报告，\n本表上一条据此记「三处同病」。复核接手会话的重跑计划时发现同族还有第四处，且不在那\n三个 runner 里：revocation-sla.latest.json 的写入者是用例自己，不是 runner。\n\n要害是写入时机。runtime/test/e2e/revocation-sla.test.ts:68 在 beforeAll 的第一句就\n无条件写下 status:failed / reason:RUN_NOT_COMPLETED，下一句才做 isolated() 库名校验\n（mainline-support.ts:11 要求 /platform_\u003ckind\u003e_(ms23|test|ci)）。先毁证据再校验前置，\n恰是 adda76a 所确立口径的反面。\n\n触发面比「跑主线验收」宽：runtime/test 是工作区包 platform-tests，pnpm\nruntime:check:runtime 第 5 步的 turbo run test 就会跑到它。runbook §7 那句「前置守卫\n段已修复，但跑到一半失败仍会写 failed」对本处不成立——连跑到一半都不需要。\n\n当日由接手会话实测复现：七个模块库建成 platform_\u003cm\u003e_acc0918 后 e2e/chaos 整片失败，\n其独立检出里那份报告已被清成 RUN_NOT_COMPLETED 绑 46265a6；主工作区绑 765300b 的\npartial 因其在独立检出里跑而幸存。连带结论是七个模块库后缀只能是 ms23 / test / ci，\n且 check:runtime 与 mainline:check 本就该共用同一批库。\n\n本轮只登记不修：六份验收报告的重跑已移交主会话并在其独立检出里进行，改同一批文件必\n撞车。真源仓侧的代码修正与 runbook §7 的措辞订正留给接手会话。全程只读，未跑门禁、\n未起底座、未改任何报告。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T02:04:54-07:00"}],"HeadCommit":{"Sha1":"031878b2a7dbd2e2021b6dd71b0197e50cec3c96","Message":"docs(计划): 第四处同病改判已修（08c3788），并订正本条原先给错的一个修法\n\n登记时给了两个修法，取的是「把写报告挪到前置校验之后」；另一个「直接删掉占位写」被接手\n会话否掉，理由成立，本条一并订正：「跑起来但没跑完」确实该在盘上留 RUN_NOT_COMPLETED 而\n不是上一轮的 partial——这是用例自己写报告带来的责任，和「上轮结论已被后续提交推翻」不是\n同一个事实，不能用 check:evidence 判过期替代。\n\n实现是把那句抽成 markRunStarted()，调用点挪到模块库白名单通过、真实连库、API 子进程\n/health 就绪之后。两个方向都做了负向实测：同一条件下旧代码毁报告（sha256 变化），新代码\n测试照样失败而报告逐字节不变；正向确认 t+85s 兜底仍写 RUN_NOT_COMPLETED。\n\n受影响四份已重跑回绑 dfd4832（全绑 08c3788 / dirty=false，结果与修复前一致，说明动的只是\n写报告的时机）；runbook §7 收尾同步改成已修（925eee8）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:13:53-07:00"},"CompareURL":"luoanwu/platform-governance/compare/58a896f8a4030c0d72fa3b631d8f823f144d1650...031878b2a7dbd2e2021b6dd71b0197e50cec3c96","Len":2}...
|
1789740931
|
Edit
Delete
|
|
31143
|
5
|
5
|
5
|
76
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"91b525700 {"Commits":[{"Sha1":"91b5257007ab9d588dcada5c5ca8d0c41a024bb9","Message":"chore(reports): B2 的四十一份三级证据 回绑 @ d1eb586，刷新快照锚点\n\n本批次最后一个提交，按 C241 同时刷新快照锚点到 d1eb586（声明落后 0，本提交自身使实测\n落后 1，在 +1 自指窗内）。\n\n四十一份 reports/*.latest.json 全部 clean 绑定（gitSha=d1eb586、worktreeDirty:false）：\n静态 28 + 运行态 13。\n\n运行态实测与基线逐项持平，**tenantId 收敛在真实写链上行为中性**：\n testsPassed 733（基线 733）· behaviorParityCases 204（基线 204）· testFailures 0\n 87 项指标零违规\n静态整轮 28 步全执行、前 27 绿，末步 check:kernel-admission 仍是那 8 处已知诚实红。\n\n本批次跑了三轮 runtime，前两轮红，都不是 B2 的代码问题，如实记录：\n\n一、第一轮 outcomes.integration.test.ts「outcome not opened in time」（15 秒 E2E 超时）。\n 判为偶发，依据三条：该测试用 tenant-a-nestjs-outcome，在新 schema 下合法；整份日志里\n 新校验的拒绝消息 / ZodError / invalid_string **0 次命中**；同一份代码隔离重跑该文件\n 14/14 通过。第三轮整轮亦一次过。\n 顺带记一个行为：**失败的 runtime 会删掉 8 份运行态报告**（跑到一半中止、未重新生成）。\n\n二、第二轮测试全过（733 / 204-204 / 0 失败）但整轮记 failed，红在末尾的文档复核\n governance-kernel-boundary-number-consistency。根因是本会话的一次操作失误：\n `git restore reports/` 用了目录级范围，把已重新生成的 kernel-boundaries.latest.json\n 退回 HEAD 版（143/143），而 CLAUDE.md 已是 144/144。重跑静态后边界报告回到 144/144，\n 又撞上 governance-number-evidence-shape（要求一份 passed 的 runtime），于是必须再跑\n 第三轮才解环——这正是 CLAUDE.md 记载过的那个环。\n\n 同一次失误还抹掉了三份 reports/ui-*.latest.json 的未提交改动（不属本会话）。\n 已用 git fsck --lost-found 从游离 blob 全数找回并逐字核对：它们是 2026-09-15 的 UI 验收，\n 绑 c433d57 且 dirty=false。本提交仍未包含它们。\n 教训具体化：reports/ 下同时存在「本会话产出的」与「别人的」两类文件，任何范围操作都会\n 同时命中两类，**还原必须逐文件点名**。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:31:09-07:00"},{"Sha1":"d1eb5860c4337cf457a03c76670d2551dc848cee","Message":"feat(contracts): B2 迁入内核租户原语，收敛 tenantId 的三套口径为单源\n\n立项 B2 原写「净迁入六个」。开工时改了做法并写回立项：**净迁入的模块不该先暴露后接线**。\napi-error / replay-buffer / request-context / platform-ports / alert 只加到 barrel 而不接线，\n就是新增零消费者的 CORE 面——正是本仓 check:kernel-admission 的 K5 在盯、K4 判「声明无回执」\n的同型。改为**接线时才迁入**。六个里只有 tenant 是「迁入即能用」的，本批只做它。\n\n一、packages/contracts/src/tenant.ts 转发 @juhai/kernel 的 11 个租户解析原语。\n 这个名字是上一步 tenant.ts → authorization.ts 改名腾出来的；本仓此前**完全没有这一组**。\n\n二、用它收敛 tenantId 的三套口径（8 处 / 6 个文件）：\n z.string().min(1) × 6（无上限、无字符约束） → tenantIdSchema\n .max(128) × 2（有长度无形状） → tenantIdSchema\n ^ten_.+$（事实投影，自成一套） → **叠加**在内核形状之上，不再另起\n\n 判据直接委托 normalizeTenantId，不在仓内重写正则——重写就又是一套口径。\n 用 normalizeTenantId(v) === v 而非 !== undefined：后者放行带首尾空白的值\n （normalize 先 trim 再判），于是「校验通过的值」与「真正该存的值」不是同一个。\n\n收紧实测咬住 11/11。B2 之前 z.string().min(1) 会放行全部这些：\"\"、\" t1 \"、\"t/1\"、\n**\"t\\n1\"**、\"_t\"、129 字符。含换行符那条是框架注释点名的伪造日志行向量（NestJS 侧有字符串\n拼接日志）；129 字符那条对应 users 表 @@unique([tenantId, email]) 撞 Postgres btree 约\n2704 字节行上限打 500。也就是说本仓在此之前**没有任何 tenantId 形状与长度校验**。\n\n三处门禁如实红过再修，没有一处先改门禁：\n check:kernel-boundaries B3 新内核面未归类 → tenant.ts 归 CORE（contractModules 32）\n check:kernel-packages P7 公开面变更须伴随版本上调 → @repo/contracts 1.18.0 → 1.19.0\n check:kernel-admission K1 导出清单漂移 → 重签（40 模块 / 756 导出）\n连带同步接入手册下游锁行与 CLAUDE.md 边界数字 143/143 → 144/144。\n\n验证：contracts 312/312 · typecheck 13/13 · 整轮静态 28 步前 27 绿（末步仍是那 8 处已知红）。\n运行态另行重跑后回绑——本提交不含任何运行态报告。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:04:17-07:00"}],"HeadCommit":{"Sha1":"91b5257007ab9d588dcada5c5ca8d0c41a024bb9","Message":"chore(reports): B2 的四十一份三级证据 回绑 @ d1eb586，刷新快照锚点\n\n本批次最后一个提交，按 C241 同时刷新快照锚点到 d1eb586（声明落后 0，本提交自身使实测\n落后 1，在 +1 自指窗内）。\n\n四十一份 reports/*.latest.json 全部 clean 绑定（gitSha=d1eb586、worktreeDirty:false）：\n静态 28 + 运行态 13。\n\n运行态实测与基线逐项持平，**tenantId 收敛在真实写链上行为中性**：\n testsPassed 733（基线 733）· behaviorParityCases 204（基线 204）· testFailures 0\n 87 项指标零违规\n静态整轮 28 步全执行、前 27 绿，末步 check:kernel-admission 仍是那 8 处已知诚实红。\n\n本批次跑了三轮 runtime，前两轮红，都不是 B2 的代码问题，如实记录：\n\n一、第一轮 outcomes.integration.test.ts「outcome not opened in time」（15 秒 E2E 超时）。\n 判为偶发，依据三条：该测试用 tenant-a-nestjs-outcome，在新 schema 下合法；整份日志里\n 新校验的拒绝消息 / ZodError / invalid_string **0 次命中**；同一份代码隔离重跑该文件\n 14/14 通过。第三轮整轮亦一次过。\n 顺带记一个行为：**失败的 runtime 会删掉 8 份运行态报告**（跑到一半中止、未重新生成）。\n\n二、第二轮测试全过（733 / 204-204 / 0 失败）但整轮记 failed，红在末尾的文档复核\n governance-kernel-boundary-number-consistency。根因是本会话的一次操作失误：\n `git restore reports/` 用了目录级范围，把已重新生成的 kernel-boundaries.latest.json\n 退回 HEAD 版（143/143），而 CLAUDE.md 已是 144/144。重跑静态后边界报告回到 144/144，\n 又撞上 governance-number-evidence-shape（要求一份 passed 的 runtime），于是必须再跑\n 第三轮才解环——这正是 CLAUDE.md 记载过的那个环。\n\n 同一次失误还抹掉了三份 reports/ui-*.latest.json 的未提交改动（不属本会话）。\n 已用 git fsck --lost-found 从游离 blob 全数找回并逐字核对：它们是 2026-09-15 的 UI 验收，\n 绑 c433d57 且 dirty=false。本提交仍未包含它们。\n 教训具体化：reports/ 下同时存在「本会话产出的」与「别人的」两类文件，任何范围操作都会\n 同时命中两类，**还原必须逐文件点名**。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:31:09-07:00"},"CompareURL":"luoanwu/digital-employee-os/compare/3deb4ea37b83c2e2977c9c2b8e8166c101002860...91b5257007ab9d588dcada5c5ca8d0c41a024bb9","Len":2}...
|
1789741873
|
Edit
Delete
|
|
31144
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/feat/registry-rule-hardening
|
0
|
|
1789742570
|
Edit
Delete
|
|
31145
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/feat/registry-rule-hardening
|
0
|
{"Commits":[{"Sha1":"92a6b8e64 {"Commits":[{"Sha1":"92a6b8e647c2c4d764c1fcdb8448bfc48a4dc4ed","Message":"fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键\n\n蓝图 docs/domain/registry-snapshot-blueprint.md 从迁入起就写着「失败关闭」与\n「review key 仅用于人工复核幂等」两条不变量，实测三处不成立。本次按蓝图自己的口径补齐，\n改动只在 contracts/src/domain/application-contract-registry/：不新增写入点、不新增宿主、\n不动 Catalog 与 Schema，每条输出仍固定 activationAllowed=false / registryWriteAllowed=false /\nhumanApprovalRequired=true。\n\n1. 失败关闭：assessRegistrySnapshot / assessRegistryTransition /\n reviewCandidateSnapshotUpgrade 对畸形输入抛 TypeError（快照缺字段、now 不是 Date、\n decisions 整个缺失都会抛），四个入口里只有 LocalManifestAdmissionPlanner.plan 是\n 失败关闭的。现四个入口一致返回 REJECT + 原因码。这不是洁癖：check-fixtures 把抛出当\n 配置错误处理，实测一次 TypeError 会让整套 15 例逐例判定塌成一条 configuration 失败。\n\n2. 十六进制引用规范化：digest / commit 的正则带 /i 收大小写两种写法，比较与入键却\n 大小写敏感——同一摘要写成大写会得到另一个 reviewIdempotencyKey，同版本大小写翻转被判成\n SAME_VERSION_DIGEST_CHANGED，大写回滚 pin 被判成 ROLLBACK_PIN_MUST_MATCH_CURRENT。\n 现一律规范化为小写后再比较、再入键，ACCEPT 回吐规范化后的值。\n\n3. 版本文法收敛：一个模块三套（快照流与清单准入用 \\d+\\.\\d+\\.\\d+，升级复核用严格 SemVer），\n 新增 registry-primitives.ts 作单一来源，含预发布序；DEC-009—012 此前在两份文件各写一份，\n 一并收敛。snapshot-upgrade-review.ts 因此净减约 60 行重复实现。\n\n两处语义变化（蓝图新增一节已如实登记）：\n- 收紧：1.02.0 这类前导零版本此前被放行，现在拒；assessRegistrySnapshot 另开始拒 latest\n 这类可变标签——此前它能过体检却必然过不了迁移规则。\n- 放宽：1.0.0-rc.1 这类预发布版本此前被快照流整条挡在外面，现在按 SemVer 序参与迁移。\n 平台自己发的就是 1.0.0-rc.N，升级复核那一侧本来就这么判。风险面有限：本模块所有出口\n 都不激活、不写 Registry，多进人工复核队列不等于多放行制品。\n\n证据（本分支隔离工作树，clean）：注册中心单测 44/44（原 34 + 加固 10）；\ncontracts check:local 302/302；check:fixtures 九个契约域套件 100 例 0 失败，\n其中本套件 15 例（正 4 / 反 11，原 8 例）。篡改必红实测两次：sameHexRef 退回大小写敏感\n→ 单测 1 红 + 夹具 P03 红；去掉快照入口守卫 → 整套塌成 0 例 1 配置失败。\n\n仍为仓内候选 sdk_e2，不代表正式 Registry、Snapshot 发布或跨仓 Required Check 已上线。\n清单登记的 C01.04 / C01.06 是「接线」型缺口且前置 Q01 运行宿主未裁，本次未动。\nreports/fixtures.latest.json 未回绑：它是 17 套件的聚合报告，只跑 9 个套件去覆盖会缩小结论面，\n应在整批（含并行会话的 public-file / IM / 跨域流程改动）落定后统一重跑回绑。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:40:10-07:00"},{"Sha1":"793d0a106f07c2b9974991a2a5953fdebc5be1b3","Message":"fix(contracts): 公共文件规则补入参体检——plan 遇畸形请求此前抛 TypeError，非布尔授权位此前被判成放行\n\n迁入版本在「入参可信」的前提下是对的，但本能力的实际入口是 JSON：夹具评估器、策略快照、\n三个消费仓的载荷都从反序列化来，TS 类型只在编译期成立。判定面没有扩张，改的只是拿到脏值时往哪边倒。\n\n三个缺口（均有实测，非读代码推断）：\n\n1. `LocalPublicFileAccessPlanner.plan()` 遇畸形请求抛 TypeError 而不是拒绝。补强前 dist 实测\n `plan({})` → `TypeError: Cannot read properties of undefined (reading 'trim')`，`plan(null)`\n 与缺 `file` 的请求同理。蓝图写的「失败闭合」只覆盖快照畸形（bundled deny-all），没覆盖请求畸形——\n 抛异常落到消费侧是 500，不是拒绝。现按同包 `webhook-delivery-planner.plan()` 的既有先例先体检请求：\n 非对象 → FILE_ACCESS_REQUEST_INVALID，归属三件套缺失或非字符串 → FILE_ACCESS_ATTRIBUTION_REQUIRED，\n `file` 不是对象 → FILE_REFERENCE_INVALID。\n\n2. 非布尔的授权位被判成放行。`if (!context.permissionGranted)` 对 \"false\" / 1 / {} 全部放行；\n `authorizationVersion \u003c requiredAuthorizationVersion` 与 `expiresAt \u003c= now` 在另一端缺失时都是 false，\n 即「比不出来」被当成「比过了」。现在授权位只认字面 true，两个比较的两端都必须是整数。\n\n3. `PublicFileRef.version` 声明为 \"v1\" 却从来没被校验，换版本的引用会被 v1 判定顺带放行。\n 加 FILE_REF_VERSION_UNSUPPORTED。\n\n覆盖：夹具评估器不再先把缺字段补成 \"\" / NaN（那层补齐等于夹具永远喂不进脏值，而脏值正是这次要固定的\n那一侧），payload 原样交给规则；新增 refresh 规则——planner 第三个公开方法此前在门禁上零覆盖。\n夹具 13 → 29 例（正 4 / 反 25），补齐六条此前只有单测、门禁上没有的策略白名单原因，以及\nFILE_ACCESS_ATTRIBUTION_REQUIRED / FILE_TENANT_NOT_ALLOWED 两条此前任何地方都没覆盖的原因；\n定向测试 15 → 22 例，迁入的 15 例一条未改。\n\n未做（需 CHG，已在 index.ts 抬头与 SOURCE.md 写明，不再沿用旧措辞）：\nschemas/public-file-ref.v1.schema.json（version=1 数字 + ownerEntity*/storageRef）与本目录的\nPublicFileRef（version=\"v1\" 字符串 + sha256/size/mimeType/scanStatus）不是同一个对象——一份通得过\nSchema 的引用喂不进 assessPublicFileRef，反之亦然。合并属契约变更，不在本次范围。Catalog 登记未动\n（public-file 仍 module_e2、code_truth 指向工单仓 packages/attachments），三个消费仓的副本未同步本次补强。\n\n证据：在仅含本提交改动的导出树上 tsc 绿、定向 22/22、contract-public-file 夹具 29/29；\n工作树另有并行会话的 WIP，故本次不回绑任何 reports/。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:38:28-07:00"},{"Sha1":"0c6a1918aa223b1779ce4653ba2c4e4609aea336","Message":"chore(reports): dev 常驻运行时切换后的部署门禁回绑 @ 06c3d4c\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:34:13-07:00"},{"Sha1":"06c3d4cb7fd09b2b44e621a56d7b14653bb8b445","Message":"chore(reports): runtime 镜像重建与冒烟证据回绑 @ 05e3826\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:33:44-07:00"}],"HeadCommit":{"Sha1":"92a6b8e647c2c4d764c1fcdb8448bfc48a4dc4ed","Message":"fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键\n\n蓝图 docs/domain/registry-snapshot-blueprint.md 从迁入起就写着「失败关闭」与\n「review key 仅用于人工复核幂等」两条不变量，实测三处不成立。本次按蓝图自己的口径补齐，\n改动只在 contracts/src/domain/application-contract-registry/：不新增写入点、不新增宿主、\n不动 Catalog 与 Schema，每条输出仍固定 activationAllowed=false / registryWriteAllowed=false /\nhumanApprovalRequired=true。\n\n1. 失败关闭：assessRegistrySnapshot / assessRegistryTransition /\n reviewCandidateSnapshotUpgrade 对畸形输入抛 TypeError（快照缺字段、now 不是 Date、\n decisions 整个缺失都会抛），四个入口里只有 LocalManifestAdmissionPlanner.plan 是\n 失败关闭的。现四个入口一致返回 REJECT + 原因码。这不是洁癖：check-fixtures 把抛出当\n 配置错误处理，实测一次 TypeError 会让整套 15 例逐例判定塌成一条 configuration 失败。\n\n2. 十六进制引用规范化：digest / commit 的正则带 /i 收大小写两种写法，比较与入键却\n 大小写敏感——同一摘要写成大写会得到另一个 reviewIdempotencyKey，同版本大小写翻转被判成\n SAME_VERSION_DIGEST_CHANGED，大写回滚 pin 被判成 ROLLBACK_PIN_MUST_MATCH_CURRENT。\n 现一律规范化为小写后再比较、再入键，ACCEPT 回吐规范化后的值。\n\n3. 版本文法收敛：一个模块三套（快照流与清单准入用 \\d+\\.\\d+\\.\\d+，升级复核用严格 SemVer），\n 新增 registry-primitives.ts 作单一来源，含预发布序；DEC-009—012 此前在两份文件各写一份，\n 一并收敛。snapshot-upgrade-review.ts 因此净减约 60 行重复实现。\n\n两处语义变化（蓝图新增一节已如实登记）：\n- 收紧：1.02.0 这类前导零版本此前被放行，现在拒；assessRegistrySnapshot 另开始拒 latest\n 这类可变标签——此前它能过体检却必然过不了迁移规则。\n- 放宽：1.0.0-rc.1 这类预发布版本此前被快照流整条挡在外面，现在按 SemVer 序参与迁移。\n 平台自己发的就是 1.0.0-rc.N，升级复核那一侧本来就这么判。风险面有限：本模块所有出口\n 都不激活、不写 Registry，多进人工复核队列不等于多放行制品。\n\n证据（本分支隔离工作树，clean）：注册中心单测 44/44（原 34 + 加固 10）；\ncontracts check:local 302/302；check:fixtures 九个契约域套件 100 例 0 失败，\n其中本套件 15 例（正 4 / 反 11，原 8 例）。篡改必红实测两次：sameHexRef 退回大小写敏感\n→ 单测 1 红 + 夹具 P03 红；去掉快照入口守卫 → 整套塌成 0 例 1 配置失败。\n\n仍为仓内候选 sdk_e2，不代表正式 Registry、Snapshot 发布或跨仓 Required Check 已上线。\n清单登记的 C01.04 / C01.06 是「接线」型缺口且前置 Q01 运行宿主未裁，本次未动。\nreports/fixtures.latest.json 未回绑：它是 17 套件的聚合报告，只跑 9 个套件去覆盖会缩小结论面，\n应在整批（含并行会话的 public-file / IM / 跨域流程改动）落定后统一重跑回绑。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:40:10-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/05e3826e9e9e4dbbdd753252c2134b8494b08f2b...92a6b8e647c2c4d764c1fcdb8448bfc48a4dc4ed","Len":4}...
|
1789742570
|
Edit
Delete
|
|
31146
|
5
|
5
|
5
|
76
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"4da2c656a {"Commits":[{"Sha1":"4da2c656ae691428be3f177563952cba66c37b88","Message":"chore(reports): 静态链首次 28/28 全通过 回绑 @ 801d916，刷新快照锚点\n\n本批次最后一个提交，按 C241 同时刷新快照锚点到 801d916（声明落后 0，本提交自身使实测\n落后 1，在 +1 自指窗内）。\n\n二十八份 reports/*.latest.json 全部 clean 绑定（gitSha=801d916、worktreeDirty:false）。\n**pnpm check 退出码 0：28/28 个步骤通过**——自 check:kernel-admission 引入以来首次整链绿，\n也是本仓静态轴扩到 28 步后的首次。\n\n绿的成色写在实现提交里，此处只记边界：唯一没有机器回执的 CORE.crossProductRequired 走的是\nevidenceDeferred 五字段登记（期限 2026-12-18、触发条件＝第二个产品真实装载），它在报告里是\nobserved 一行，**不是绿，是登记在案的缺证**。到期不补即转红。\n\n运行态未受本批影响（只动门禁、注册表、夹具与文档，未动 apps/packages 运行代码），\nruntime 报告仍绑在上一批的 d1eb586，13 份运行态证据未重跑也未失效。\n\n工作区内三份 reports/ui-*.latest.json 的改动早于本批次，未提交也未还原。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:42:54-07:00"},{"Sha1":"801d916f36af0ee32c77b2138ec9afcf14b34db7","Message":"feat(governance): K3/K4 裁决——八处准入声明逐条补上真回执，静态链首次 28/28\n\ncheck:kernel-admission 引入时诚实红 8 处（K3 一处 + K4 七处）。本批逐条裁决并清零，\n**没有一处是把判据改松**——三条是补实体，一条是承认无法证明。\n\n一、K3（跨产品证据只有命名空间导入）→ 补实体\n 下游夹具此前只有 10 个 import * as，按构造在契约模块全空时同样通过。改为携带 5 个\n **真实被产品消费**的符号（ProductManifestInput / Composition / RuntimeContractCatalog /\n DataContractImplementation / EventContractImplementation，实测自唯一装载的 juhai.hr）的\n 具名类型导入并用于类型位置——编译通过才第一次构成「下游能按名字用上这些契约」的证据。\n\n二、K4 的结构性问题：两条门禁要求互相冲突，拆字段解决\n check-kernel-boundaries 的 B2 明令 evidenceSources **不得指 reports/**（运行产物随时重跑、\n 可被删除——本会话实测过：失败的 runtime 会删掉 8 份报告），而 K4 要求必须有机器回执。\n 两条各自成立，是同一件事的两半。改为两个字段：\n evidenceSources 持久判据源（脚本 / 词典 / 锁）——B2 判它\n evidenceReceipts 该判据源产出的机器回执——K4 判它\n K4 同时加强：回执须真实存在、带机器结论（status 或 violations）、带 provenance.gitSha、\n 且自身不是红盘。另加两条配套判据：\n · 禁止把**本门禁自己产出的报告**当回执（结构性自指：它只要有发现就是红的）\n · 确实无法被证据支撑时走 evidenceDeferred 五字段登记，降级 observed、过期转红\n\n三、逐条落点\n industryIndependent → kernel-boundaries（逐文件扫 CORE 源码不得内嵌已登记 Product\n 命名空间，check-kernel-boundaries.mjs:181）+ naming（业务词典）\n dualBackendGoverned → dual-backend-parity\n stableContract → kernel-packages\n crossProductRequired → **无回执，登记为「证据待补」**：本仓只装载一个产品，跨产品消费面\n 实测仅 5 个符号 / 34 个模块为零，该证据客观上无从产生。期限\n 2026-12-18，触发条件＝第二个产品真实装载；到期转红。\n 硬指一份回执等于把自证换个地方做，不做。\n\n四、订正一处我自己的误判\n 先前称「parity 对 employment 零断言」——**查错了地方**：断言在 governance.registry.json\n 的 employment.parityAssertions 里，由 check-dual-backend-parity.mjs:507 通用执行，本就有\n 10 条（消费者标识分区、new FactInbox、FACT_READ_NOT_CONFIGURED fail-closed、\n autoCommit:false、未配 broker 返回 null）。\n 据此我一度把新断言**硬编码**进脚本，被 check:fork-readiness 的 F2 当场抓住\n （「门禁不得硬编码已登记业务标识」）——F2 抓的是我的代码，抓得对。已撤回硬编码，\n 把两条新不变式（事件载荷与行映射必须走契约单源）按注册表形式补入，10 → 14 条。\n\nkernel.boundaries.json 新增 evidenceNotes：逐份回执记录**它到底判了什么**，含上述订正与\n「dual-backend-behavior 的 63 个矩阵用例不含 llm/employment，故两个 pack 不以它为据」这类边界。\n指针只说明有机器判过，不代表覆盖该词全部含义。\n\n自检 10 → 17 条（新增回执质量、自指、证据待补三组，含 1 条基线正例、2 条防过度判红）。\n整轮 pnpm check：**28/28 步骤全部通过**，退出码 0——自本门禁引入以来首次。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:41:16-07:00"}],"HeadCommit":{"Sha1":"4da2c656ae691428be3f177563952cba66c37b88","Message":"chore(reports): 静态链首次 28/28 全通过 回绑 @ 801d916，刷新快照锚点\n\n本批次最后一个提交，按 C241 同时刷新快照锚点到 801d916（声明落后 0，本提交自身使实测\n落后 1，在 +1 自指窗内）。\n\n二十八份 reports/*.latest.json 全部 clean 绑定（gitSha=801d916、worktreeDirty:false）。\n**pnpm check 退出码 0：28/28 个步骤通过**——自 check:kernel-admission 引入以来首次整链绿，\n也是本仓静态轴扩到 28 步后的首次。\n\n绿的成色写在实现提交里，此处只记边界：唯一没有机器回执的 CORE.crossProductRequired 走的是\nevidenceDeferred 五字段登记（期限 2026-12-18、触发条件＝第二个产品真实装载），它在报告里是\nobserved 一行，**不是绿，是登记在案的缺证**。到期不补即转红。\n\n运行态未受本批影响（只动门禁、注册表、夹具与文档，未动 apps/packages 运行代码），\nruntime 报告仍绑在上一批的 d1eb586，13 份运行态证据未重跑也未失效。\n\n工作区内三份 reports/ui-*.latest.json 的改动早于本批次，未提交也未还原。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:42:54-07:00"},"CompareURL":"luoanwu/digital-employee-os/compare/91b5257007ab9d588dcada5c5ca8d0c41a024bb9...4da2c656ae691428be3f177563952cba66c37b88","Len":2}...
|
1789742577
|
Edit
Delete
|
|
31147
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"0c6a1918a {"Commits":[{"Sha1":"0c6a1918aa223b1779ce4653ba2c4e4609aea336","Message":"chore(reports): dev 常驻运行时切换后的部署门禁回绑 @ 06c3d4c\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:34:13-07:00"},{"Sha1":"06c3d4cb7fd09b2b44e621a56d7b14653bb8b445","Message":"chore(reports): runtime 镜像重建与冒烟证据回绑 @ 05e3826\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:33:44-07:00"}],"HeadCommit":{"Sha1":"0c6a1918aa223b1779ce4653ba2c4e4609aea336","Message":"chore(reports): dev 常驻运行时切换后的部署门禁回绑 @ 06c3d4c\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:34:13-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/05e3826e9e9e4dbbdd753252c2134b8494b08f2b...0c6a1918aa223b1779ce4653ba2c4e4609aea336","Len":2}...
|
1789742609
|
Edit
Delete
|
|
31148
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"c190a6ab5 {"Commits":[{"Sha1":"c190a6ab5d2c50e4c90f027bbcca72f1f3050005","Message":"fix(contracts): 通知与 Webhook 规则补入参体检——一处倒向放行、两处崩溃面与目标地址的同义写法\n\n迁入版把策略快照一侧校验到底（normalizeSnapshot 逐字段抛），请求一侧、目标地址一侧与夹具评估器三处漏风。\n补强前对 dist 逐条实测：\n\n- 放行面（唯一一处方向倒向放行）：三个状态位按真值判，`replayAuthorized: \"false\"` 在 `!value` 下为 false，\n 未授权重放被判成已授权 → DELIVER；planner 原样透传这三位，计划面同样通。现在只认布尔字面量。\n- 崩溃面：classifyNotificationEffect(undefined) / classifyProviderOutcome(undefined) 抛 TypeError 而不是拒绝，\n 夹具的 effect / provider 两条规则都能直接触发——门禁那时拿到的是异常不是判定，落到消费侧是 500 不是拒绝。\n- 目标地址（C08.05 / 蓝图「私网地址失败关闭」）：isPrivateLiteral 只认四段十进制，2130706433 / 0x7f000001 /\n 0177.0.0.1 / 127.1 全判成公网；0.0.0.0/8、CGNAT 100.64/10、组播与保留段、.internal、末尾点均未覆盖。\n 现在数字形主机一律按 IP 字面量处理，只放行严格四段十进制的公网地址；allowedHosts 自己先过这道地板\n （WEBHOOK_ALLOWED_HOST_NOT_PUBLIC），小写归一后再查一次重名。\n- 本次唯一放宽的一处，据实登记：迁入版按字符串前缀 fc / fd / fe80 判私网 IPv6，fc-portal.example.com 这类\n 正常域名被误判成私网。改判后它可用，但仍须先进运营批准的 allowlist 且为 HTTPS，不构成新的可达面。\n- 死代码：WEBHOOK_TENANT_WILDCARD_FORBIDDEN 写在字符集校验之后，`*` 永远先报 WEBHOOK_ROUTE_TENANTS_INVALID，\n 蓝图写明的「tenant allowlist 禁止 *」从未以自己的名义拦下过。\n- 退避：decideProviderOutcome 原取 min(maximumBackoff, retryAfter)，供应商回一个 Retry-After: 0 即可把批准过的\n 退避清零。现在外部回执只能把退避拉长到策略上界，上界语义与既有用例逐字不变。\n- 夹具评估器把两种非放行当成放行：SKIP_DUPLICATE 与 DLQ 都没有 reasons 字段，被 `\"reasons\" in r` 判成空数组\n = ACCEPT，与该文件自己第一句「只有 DELIVER / DELIVERED 算放行」正相反。另补 refresh 规则。\n- effect_id 改为强校验：五段必须是规范标识符（因此不含 :），否则 a:b+c 与 a+b:c 会拼出同一个 id；\n 请求一侧的标识符不再先 trim 再判定。\n\n合法快照与合法请求的判定结果逐字不变：迁入的 14 例断言一条未改。单测 14 → 24 例；夹具 8 → 17 例，并修好\n名不副实的 N06-plan-without-policy-denies-all——它此前只喂了两个字段的请求，压根走不到「无策略即拒」那一步。\n\n核验（工作树 dirty，报告未回绑，只绑本地）：contracts build / typecheck 绿、本域单测 24/24、\ncontracts/test/*.test.mjs 330/330、check:fixtures 17 套件 236 例 0 失败 0 不可用（本域 17/17；用临时报告名跑，\n跑完删除，reports/ 未动）。负向实测两次：① 把 HEAD 版三支规则单独编译后喂新夹具，8 例新增负例全部不通过\n（N07/N10/N11 直接放行、N08/N09 抛 TypeError、N12 换原因、N13/N14 旧评估器不识别 refresh）；\n② 篡改 N10 的声明原因 → 门禁 exit 1 并点名该例，翻 expect → 被 id/expect 一致性先挡下，还原后 exit 0。\n\n未做：缺失模块完善清单 C12 的模板审批 / 多渠道 Provider / 签名防重放 / 渠道切换本仓仍一件没有（该行类型是\n「裁决」，前置是运行宿主 Q01），本轮仍是契约包形态，不代表通知服务上线；Catalog 登记未动（仍 module_e2，\ncode_truth 指工单仓 packages/notification-delivery）；三个消费仓的 packages/notification-delivery 未同步；\n入站回调验签（C12.05）与 UNKNOWN 回执（C12.04）属判定面扩张，只登记未实现。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:43:48-07:00"},{"Sha1":"9d13fb058d2a0356797f26d1d17a4b2149b6ab97","Message":"docs(runbook): 订正 CORS 一句——服务端代理不解 CORS，工作台吃的是 :3101 那个进程的白名单\n\n原文「控制面 Web 与 identity dev 走服务端代理，不吃这条」把两件事混了。代理只是把上游的\nAccess-Control-Allow-Origin 原样回传，吃不吃只看浏览器所在页面的 origin 与被请求 origin\n是否同源：控制面 Web 的页面与它的 /api 代理同源，不吃；工作台 dev（:3010）经 :3098 代理\n打到 :3101，跨源，放行的必须是 :3101 那个进程的 CORS_ORIGIN，而不是容器的\nPLATFORM_CORS_ORIGIN——后者 2026-09-18 补过、前者一直缺，于是总览两块读面全被浏览器拦成\n「依赖故障」，而服务端日志里是一次正常 200。今天按这句话排查，方向被带偏过一次。\n\n同轮登记两处仓外事实：identity dev 的 env 文件位置（企业控制面/.secrets/\nenterprise-platform-identity-dev.env，已含 CORS_ORIGIN=http://localhost:3010），\n以及工作台 API 基地址的单一出处（apps/workbench/.env.local，launch.json 不再覆盖）。\n\n作用域 = 文档订正，不含代码与门禁证据。本机 dev 实测：预检 204 / 三个读面 200 /\nidentity 探针 database:up redis:up，证据只绑本地运行态。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:42:44-07:00"},{"Sha1":"13c7eb7271af67df9d0b4e5223bda3eaa4ab7b10","Message":"fix(contracts): 跨域流程规则补入参体检，并补上 FLOW-01 与 effect 幂等两条自己没做到的红线\n\n迁入版的判定在「入参可信」时是对的，但本能力的真实入口是 JSON：夹具评估器、approved 快照、\n未来的 Owner Adapter 载荷都从反序列化来，TS 类型只在编译期成立。四类问题均有 dist 实测，非读代码推断。\n\n1. 崩溃面八处抛 TypeError 而不是拒绝。assessWorkflowDefinition(undefined) →\n Cannot read properties of undefined (reading 'domains')；{}、domains 非数组、steps 非数组、\n 步骤缺 dependsOn、effectIdempotencyKey 是数字同理；LocalWorkflowPlanner.plan({}) →\n Cannot read properties of undefined (reading 'trim')。夹具 definition 与 plan 两条规则都能直接触发——\n 门禁那时拿到的是崩溃不是判定，落到消费侧是 500 不是拒绝。现在两侧都先按 unknown 体检：\n 定义非对象 → WORKFLOW_DEFINITION_INVALID，请求非对象或归属三件套非空串 → WORKFLOW_ATTRIBUTION_REQUIRED。\n\n2. 放行面三处脏值倒向合格。timeoutSeconds: \"abc\" 因 \"abc\" \u003c= 0 为 false 而合格；\n irreversibleEffect: 0 被当成可逆，整条不可逆效果的幂等键要求随之失效；steps: [] 的「流程」也合格，\n approved 快照喂进来就是一份零命令的 PLAN。现在 timeout 必须是正整数、irreversibleEffect 必须是布尔、\n 零步骤 → WORKFLOW_STEPS_REQUIRED。\n\n3. 判定面在两处确有扩张，都是本规则自己没做到自己的红线。FLOW-01 Eligibility 是开发计划写死的\n 第一道闸（「拒绝把域内定时任务和单域状态机迁入平台」），但迁入版只数 domains 声明了几个域，\n 不数步骤实际落在几个域——把单域状态机的 domains 写成两个即可整条绕过，实测\n {domains:[\"hr\",\"identity\"], steps:[两步都 ownerDomain:\"hr\"]} 判 ELIGIBLE。现在 Owner 面体检通过后\n 再数步骤真实跨度，沿用 NOT_CROSS_DOMAIN 不新增原因码。另一条红线写「不可逆外部效果使用独立\n effect idempotency，平台重放不得重复付款、通知」，但两个不可逆步骤共用同一把键时无人拦，\n 一次效果的幂等记录会压掉另一次 → 新增 EFFECT_IDEMPOTENCY_NOT_UNIQUE。\n\n4. 第三条扩张在 governed 步骤一侧：不可逆步骤既无同 Owner 补偿又不开人工接管时此前照常编进波次，\n 而退出门要求「每步唯一 Owner、timeout 和恢复责任明确」，这种步骤失败后没有任何恢复责任人 →\n STEP_RECOVERY_RESPONSIBILITY_REQUIRED。补偿仍是可选项：人工接管或补偿有其一即可，\n 可逆步骤不受约束，两条都有正向用例钉住。\n\n另修一处诊断损失：normalizeSnapshot 把定义体检的多条原因只抛第一条，改为逐条抛出。\n合法定义与合法请求的判定结果逐字不变，迁入的 11 例单测断言一条未改。\n\n覆盖：定向测试 11 → 22 例，补强用例另起两节；夹具 7 → 16 例（正 2 / 反 14），并给原 N01—N05\n补 reasons 声明——此前只断言「拒绝」，换一条原因拒绝也算过。\n\n未做（越界，已在 SOURCE.md 与蓝图增量节写明）：流程实例、持久调度、补偿执行器、Timer、\nOwner Adapter 与 M5 分工一件未建，按缺失模块完善清单 C15 卡在运行宿主 Q01 与 G4，本轮仍是契约包形态，\n不代表编排平台上线；Catalog 登记一律未动（仍 candidate_e0 / E0）；步骤 Permission 的资源归属仍不校验\n（ownerDomain \"hr\" 的步骤声明 identity 资源照样编进计划，而资源名与域的绑定在本包里没有可判定的真源，\n纯规则读不到 catalogs/permissions.json）；commandIdempotencyKey 含定义版本号，同一实例升版重放会换键，\n不可逆效果由不含版本的 effectIdempotencyKey 兜底，改它属运行宿主裁决时一并定的事。\n\n证据：在仅含本提交改动的导出树上 tsc 与 typecheck 绿、定向 22/22、contract-cross-domain-workflow\n夹具 16/16；负向实测把 N07-declared-domains-without-real-owner-span 的声明原因改成 WORKFLOW_CYCLE，\n门禁 exit 1 并点名该例（reasons=[\"NOT_CROSS_DOMAIN\"]），还原后转绿。导出树里 governance-linter.test.mjs\n因 /tmp 下工作区根不可解析而失败，pristine HEAD 导出树同样失败，与本次改动无关；真实工作树\ncontracts:check:local 330/330、check:fixtures 17 套件 227 例 0 失败。工作树另有并行会话的 WIP，\n故本次不回绑任何 reports/。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:42:22-07:00"},{"Sha1":"5720165763b35035e6778fe9a802493a16eb5aca","Message":"fix(contracts): 会话准入规则补入参体检——四条 fail-open 与一处「兜底自己会抛」收口\n\n九项契约包自 2026-09-14 迁入以来一直是「原样迁入、未改语义」，本条是第一次动判定。\n起因是对 collaboration-messaging 逐条实测而非读代码：迁入版把策略快照一侧校验到底\n（normalizeSnapshot 逐字段抛并被工厂捕获），请求一侧整个信任声明类型，而声明类型\n约束不了 HTTP 与夹具送进来的 JSON。五条缺陷均可复现，且都打在该能力自己 README 的\n「最小验收」与蓝图的「失败闭合」上：\n\n1. 抛而不拒。三个导出在畸形 JSON 下抛 TypeError 而不是返回 REJECT。兜底的 bundled\n deny-all 也不例外——createRuntimeConversationAdmissionPlanner().plan({}) 抛异常，\n 即「什么都不放行」的最后一道在畸形入参下等于没有。\n\n2. 未知 senderType 放行，且 \"AI_AGENT\"（只改大小写）同样放行。AI_ATTRIBUTION_REQUIRED\n 只在严格等于 ai_agent 时触发，于是「AI 不得冒充人员」有一条改大小写就能走的绕过路径。\n 不做大小写归一化——归一化等于替冒充者补齐形状。\n\n3. 游标不绑定会话。ConversationAppendState 有 expectedTenantId 却没有会话标识，拿 conv-1\n 的游标判 conv-9 的消息得到 APPEND，planner 直接 ADMIT 并按 conv-9 生成幂等键；重复与\n 序列 Gap 判定全程可能在读另一个会话的账本。\n\n4. processedMessageIds 传成字符串时 includes 退化为子串判断，幂等两个方向都错：\n \"msg-1\" 不含 \"msg-2\" 则漏判重复，\"msg-1,msg-2\" 含则误判重复。\n\n5. memberActive 传成真值字符串即算在籍（原判据 !state.memberActive，\"false\" 是真值）。\n\n附带：hasAttachment 非布尔即拒。此前缺字段时 attachmentAuthorizationRequired 是\nundefined，下游 if (plan.attachmentAuthorizationRequired) 就跳过公共文件平台重新授权，\n等于「缺字段 = 不需要授权」。\n\n改法只收紧失败一侧，合法输入的判定结果逐字不变：迁入的 16 例断言一条未改，只给游标补了\n新的必填字段。expectedConversationId 做成必填而不是「在场才查」——可选只能拦住已经想起来\n要绑定的调用方，拦不住忘了的那个，而忘了的那个正是第 3 条的全部成因；破坏面可枚举且全在\n仓内（本包测试与夹具各一处），Catalog 上该项 candidate_e0 / code_truth 指向本目录 /\n上层真实消费者为零。ConversationAppendDecision 新增 CONVERSATION_BINDING_REQUIRED /\nCONVERSATION_MISMATCH / STATE_INVALID 三档（原先「序列非整数」复用 REJECT，与「消息体检\n未过」同码，改后 REJECT 专指后者）；planner 新增 REQUEST_SHAPE_INVALID /\nATTACHMENT_FLAG_INVALID 两条原因与两条映射。新增原因一律排在既有原因之后判定，既有原因码\n的优先级未动。\n\n顺带修一条证据登记缺口：reports/fixtures.latest.json 的作用域只写到 runtime/modules，\n九个契约包套件的两类输入整个在作用域外——夹具用例在 contracts/test/fixtures/，判定用的\nevaluateFixturePayload 在 contracts/src/domain/。作用域写于只有运行时模块有夹具的时候，\n2026-09-14 迁入时没随迁，后果是改夹具或改规则都不会让这份报告过期，而它正是这九项在\nCatalog 上登记的证据之一。实据不用构造：本轮改动与并行会话的 public-file 改动同时在树上，\ncheck:evidence 仍不点它。补两条后该报告立刻转 EVIDENCE_WORKTREE_CHANGED，check:evidence\n由 4 条过期变 6 条——多出来的不是新问题，是此前看不见的问题。\n\n核验：contracts test 285 → 292（我的增量 +7，既有断言未改；树上现读 330，多出的是并行会话\n在途改动）；check:fixtures 本套件 8 → 14 例（正 2 / 反 12）0 失败；负向实做：把 N11 的声明\n原因改成不会产生的串，门禁 exit 1 并点名该例，还原后 exit 0；governance 测试 340/340；\n工作区一致性 L4 / L10 / L13 均 0。\n\n未做（越界）：会话 / 成员 / 消息账本、长连接服务、Fact、独立运行服务一件未建——按缺失模块\n完善清单 C16 卡在「裁决 · 实现」，前置是 Owner ADR、三消费者与 G4。Catalog 的\nimplementation_state / evidence_level / code_truth / evidence 一律未动（改登记须 CHG）。\n蓝图迁入原文逐字未改，硬化增量另起一节附在其后。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:41:37-07:00"},{"Sha1":"793d0a106f07c2b9974991a2a5953fdebc5be1b3","Message":"fix(contracts): 公共文件规则补入参体检——plan 遇畸形请求此前抛 TypeError，非布尔授权位此前被判成放行\n\n迁入版本在「入参可信」的前提下是对的，但本能力的实际入口是 JSON：夹具评估器、策略快照、\n三个消费仓的载荷都从反序列化来，TS 类型只在编译期成立。判定面没有扩张，改的只是拿到脏值时往哪边倒。\n\n三个缺口（均有实测，非读代码推断）：\n\n1. `LocalPublicFileAccessPlanner.plan()` 遇畸形请求抛 TypeError 而不是拒绝。补强前 dist 实测\n `plan({})` → `TypeError: Cannot read properties of undefined (reading 'trim')`，`plan(null)`\n 与缺 `file` 的请求同理。蓝图写的「失败闭合」只覆盖快照畸形（bundled deny-all），没覆盖请求畸形——\n 抛异常落到消费侧是 500，不是拒绝。现按同包 `webhook-delivery-planner.plan()` 的既有先例先体检请求：\n 非对象 → FILE_ACCESS_REQUEST_INVALID，归属三件套缺失或非字符串 → FILE_ACCESS_ATTRIBUTION_REQUIRED，\n `file` 不是对象 → FILE_REFERENCE_INVALID。\n\n2. 非布尔的授权位被判成放行。`if (!context.permissionGranted)` 对 \"false\" / 1 / {} 全部放行；\n `authorizationVersion \u003c requiredAuthorizationVersion` 与 `expiresAt \u003c= now` 在另一端缺失时都是 false，\n 即「比不出来」被当成「比过了」。现在授权位只认字面 true，两个比较的两端都必须是整数。\n\n3. `PublicFileRef.version` 声明为 \"v1\" 却从来没被校验，换版本的引用会被 v1 判定顺带放行。\n 加 FILE_REF_VERSION_UNSUPPORTED。\n\n覆盖：夹具评估器不再先把缺字段补成 \"\" / NaN（那层补齐等于夹具永远喂不进脏值，而脏值正是这次要固定的\n那一侧），payload 原样交给规则；新增 refresh 规则——planner 第三个公开方法此前在门禁上零覆盖。\n夹具 13 → 29 例（正 4 / 反 25），补齐六条此前只有单测、门禁上没有的策略白名单原因，以及\nFILE_ACCESS_ATTRIBUTION_REQUIRED / FILE_TENANT_NOT_ALLOWED 两条此前任何地方都没覆盖的原因；\n定向测试 15 → 22 例，迁入的 15 例一条未改。\n\n未做（需 CHG，已在 index.ts 抬头与 SOURCE.md 写明，不再沿用旧措辞）：\nschemas/public-file-ref.v1.schema.json（version=1 数字 + ownerEntity*/storageRef）与本目录的\nPublicFileRef（version=\"v1\" 字符串 + sha256/size/mimeType/scanStatus）不是同一个对象——一份通得过\nSchema 的引用喂不进 assessPublicFileRef，反之亦然。合并属契约变更，不在本次范围。Catalog 登记未动\n（public-file 仍 module_e2、code_truth 指向工单仓 packages/attachments），三个消费仓的副本未同步本次补强。\n\n证据：在仅含本提交改动的导出树上 tsc 绿、定向 22/22、contract-public-file 夹具 29/29；\n工作树另有并行会话的 WIP，故本次不回绑任何 reports/。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:38:28-07:00"}],"HeadCommit":{"Sha1":"c190a6ab5d2c50e4c90f027bbcca72f1f3050005","Message":"fix(contracts): 通知与 Webhook 规则补入参体检——一处倒向放行、两处崩溃面与目标地址的同义写法\n\n迁入版把策略快照一侧校验到底（normalizeSnapshot 逐字段抛），请求一侧、目标地址一侧与夹具评估器三处漏风。\n补强前对 dist 逐条实测：\n\n- 放行面（唯一一处方向倒向放行）：三个状态位按真值判，`replayAuthorized: \"false\"` 在 `!value` 下为 false，\n 未授权重放被判成已授权 → DELIVER；planner 原样透传这三位，计划面同样通。现在只认布尔字面量。\n- 崩溃面：classifyNotificationEffect(undefined) / classifyProviderOutcome(undefined) 抛 TypeError 而不是拒绝，\n 夹具的 effect / provider 两条规则都能直接触发——门禁那时拿到的是异常不是判定，落到消费侧是 500 不是拒绝。\n- 目标地址（C08.05 / 蓝图「私网地址失败关闭」）：isPrivateLiteral 只认四段十进制，2130706433 / 0x7f000001 /\n 0177.0.0.1 / 127.1 全判成公网；0.0.0.0/8、CGNAT 100.64/10、组播与保留段、.internal、末尾点均未覆盖。\n 现在数字形主机一律按 IP 字面量处理，只放行严格四段十进制的公网地址；allowedHosts 自己先过这道地板\n （WEBHOOK_ALLOWED_HOST_NOT_PUBLIC），小写归一后再查一次重名。\n- 本次唯一放宽的一处，据实登记：迁入版按字符串前缀 fc / fd / fe80 判私网 IPv6，fc-portal.example.com 这类\n 正常域名被误判成私网。改判后它可用，但仍须先进运营批准的 allowlist 且为 HTTPS，不构成新的可达面。\n- 死代码：WEBHOOK_TENANT_WILDCARD_FORBIDDEN 写在字符集校验之后，`*` 永远先报 WEBHOOK_ROUTE_TENANTS_INVALID，\n 蓝图写明的「tenant allowlist 禁止 *」从未以自己的名义拦下过。\n- 退避：decideProviderOutcome 原取 min(maximumBackoff, retryAfter)，供应商回一个 Retry-After: 0 即可把批准过的\n 退避清零。现在外部回执只能把退避拉长到策略上界，上界语义与既有用例逐字不变。\n- 夹具评估器把两种非放行当成放行：SKIP_DUPLICATE 与 DLQ 都没有 reasons 字段，被 `\"reasons\" in r` 判成空数组\n = ACCEPT，与该文件自己第一句「只有 DELIVER / DELIVERED 算放行」正相反。另补 refresh 规则。\n- effect_id 改为强校验：五段必须是规范标识符（因此不含 :），否则 a:b+c 与 a+b:c 会拼出同一个 id；\n 请求一侧的标识符不再先 trim 再判定。\n\n合法快照与合法请求的判定结果逐字不变：迁入的 14 例断言一条未改。单测 14 → 24 例；夹具 8 → 17 例，并修好\n名不副实的 N06-plan-without-policy-denies-all——它此前只喂了两个字段的请求，压根走不到「无策略即拒」那一步。\n\n核验（工作树 dirty，报告未回绑，只绑本地）：contracts build / typecheck 绿、本域单测 24/24、\ncontracts/test/*.test.mjs 330/330、check:fixtures 17 套件 236 例 0 失败 0 不可用（本域 17/17；用临时报告名跑，\n跑完删除，reports/ 未动）。负向实测两次：① 把 HEAD 版三支规则单独编译后喂新夹具，8 例新增负例全部不通过\n（N07/N10/N11 直接放行、N08/N09 抛 TypeError、N12 换原因、N13/N14 旧评估器不识别 refresh）；\n② 篡改 N10 的声明原因 → 门禁 exit 1 并点名该例，翻 expect → 被 id/expect 一致性先挡下，还原后 exit 0。\n\n未做：缺失模块完善清单 C12 的模板审批 / 多渠道 Provider / 签名防重放 / 渠道切换本仓仍一件没有（该行类型是\n「裁决」，前置是运行宿主 Q01），本轮仍是契约包形态，不代表通知服务上线；Catalog 登记未动（仍 module_e2，\ncode_truth 指工单仓 packages/notification-delivery）；三个消费仓的 packages/notification-delivery 未同步；\n入站回调验签（C12.05）与 UNKNOWN 回执（C12.04）属判定面扩张，只登记未实现。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:43:48-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/0c6a1918aa223b1779ce4653ba2c4e4609aea336...c190a6ab5d2c50e4c90f027bbcca72f1f3050005","Len":5}...
|
1789742672
|
Edit
Delete
|
|
31149
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"e8fda74e2 {"Commits":[{"Sha1":"e8fda74e279a5884440d5efaac2265e513c814b6","Message":"chore(reports): 入口门禁与工作台两份快照回绑 @ a9abcbe\n\n在 a9abcbe 的干净 detached 检出里重跑（governance/rebind-reports.mjs --gates caddy,workbench-snapshots，\n带 CHECK_CADDY_DOCKER=1），三份均 provenance.gitSha=a9abcbe / worktreeDirty=false。\n\n- reports/caddy.latest.json：0 违规。12 个 HTTP 契约前缀 × 2 profile 全覆盖；新增的 routing 段逐条记下\n 模拟出来的块链，13 条登记前缀（11 个去重契约前缀 + 2 个长连接前缀）全部到达 {$PLATFORM_API_UPSTREAM}；\n workbenchGuard.reached=true / conditionalBefore=0；语法校验 dev / staging 均 passed，via docker，\n 镜像 caddy@sha256:5f5c8640…d58648（compose 锁定的那一个，不再是浮动 tag）。\n- reports/workbench-catalog-snapshot.latest.json：只有 provenance 变。caddy-routes.json 进了它的作用域\n （新增 validate / mount 段），被 check:evidence 判过期，但快照只吃 routes.contracts，内容逐字未变——\n 重建后与旧版按字段比对一致，已实测。\n- reports/workbench-ops-snapshot.latest.json：与 catalog 同一脚本产出，必须一起带回（分开会绑不同提交）。\n 它有一处内容变化不由本轮改动引起：image 段从 d8f09205 / sha256:fe418667… 更新到 05e3826e /\n sha256:c0869934…，来源是 06c3d4c 已提交的 reports/image-digest.json——此前那份快照相对该提交是旧的。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:46:25-07:00"},{"Sha1":"a9abcbe392de8a903512ab4b25529a6e277113ee","Message":"feat(gateway): 入口门禁改为解析块树后模拟路由——「文件里配了」不等于「这条路径上生效」\n\nC08 应用群网关是采纳形态，判定在 Caddyfile 这份配置里，由 check-caddy 守。它此前是逐行正则：\n某条指令在文件里出现过就算配了。本轮不动入口运行行为（两份 Caddyfile 一个字没改），只把它\n自己在注释里写明判不了的那一类失效变成静态判定。\n\n逐行正则判不出来的三件事：\n- 有个更靠前的 handle 块把 /api/platform/* 截走——入口鉴权配得好好的，请求根本走不到它。\n caddy-routes.json 的 workbench.note 原文是「这种绕过只能靠运行态探针发现」。\n- 请求体上限 / 上游超时 / request-id 注入写在别的块里，门禁全绿，平台路由上一层保护都没有。\n BODY_LIMIT_MISSING 这类规则问的是「文件里有没有」，不是「这条路由上有没有」。\n- 契约前缀有匹配器、但请求先被别人服务了：ROUTE_UNCOVERED 绿，流量到不了 API。\n\n新增块树解析（tokenizeLine / parseCaddyfile：`{` 只有作为行末独立 token 才开块，占位符\n{$VAR} 与 {http.request.uuid} 是 token 内部的花括号）与路由模拟（按 Caddy 的 handle 互斥语义\n与指令序，逐层并入命名匹配器），据此新增：\n- ROUTE_SHADOWED 前缀有匹配器但被更靠前的 handler 截走，到不了 {$PLATFORM_API_UPSTREAM}\n- WORKBENCH_AUTH_BYPASS 被守护前缀在到达 forward_auth 所在块之前就被别的 handler 服务了；\n 更靠前的 handler 若条件不是纯路径（header / method / file），按「可能截走」同样判红\n- INGRESS_GUARD_OUT_OF_SCOPE 入口约定在文件里有，却不在服务该前缀的那条块链上\n- STREAMING_ROUTE_UNCOVERED 登记的长连接前缀没被路由到上游（与「禁 read_timeout」凑成一对）\n- CADDYFILE_UNPARSED 花括号不配平 / 没有站点块——判不了就报，不静默通过\n\n同轮补三条登记面的：\n- PROFILE_MOUNT_UNBOUND profile 的 Caddyfile 没被登记的 compose 挂进入口容器。staging 那份此前\n 只被校验、没人核对它有没有被 overlay 加载；校验一份没人加载的配置等于没校验。\n- CADDY_IMAGE_UNPINNED compose 的入口镜像不是 digest 锁定。docker 校验原先写死浮动 tag\n caddy:2-alpine，而跑的是锁定镜像——拿另一个版本的 Caddy 校验，过了也\n 不代表运行的那个认它。现在从 compose 取镜像，取不到即 skipped，不退回浮动 tag。\n- ROUTE_MAPPING_ORPHANED 登记表映射了 modules.json 里已不存在的契约（CONTRACT_UNMAPPED 的反向死登记）\n\nstripComments 按 Caddy 口径重写：引号内的 # 是字面量，# 只有在 token 开头才起注释——原先一刀切到\n第一个 #，会把 respond \"a#b\" 这类值截半，反过来让门禁看不见后面真实生效的指令。\nevidence-scopes 的 caddy 作用域补进 stack/compose.yaml 与 stack/overlays/staging.yaml：结论现在\n也取决于「挂没挂进容器」和「校验用的是不是那个镜像」，这两个输入一改结论就可能不再成立。\n\n核验（工作树 dirty，报告未回绑，只绑本地）：\n- CHECK_CADDY_DOCKER=1 check:caddy exit 0：12 个 HTTP 契约前缀 × 2 profile 全覆盖；模拟路由 13 条\n 前缀（11 个去重契约前缀 + 2 个长连接前缀）全部到达上游且入口约定就在各自链上；\n workbenchGuard.reached=true；两份 Caddyfile 语法均 passed，镜像 caddy@sha256:5f5c8640…d58648\n （= compose 锁定、dev 常驻容器正在跑的那一个）。\n- check-caddy.test.mjs 19 → 37 例，每条新规则都有篡改夹具且夹具断言「确实改到了」。两例专门对照\n 旧规则：插入抢先的 handle 后 evaluateCaddyfile / evaluateWorkbench 仍全绿，新规则判红。\n- 与真实 Caddy 对照（同一 digest 镜像，两个一次性容器，上游指向死端口）：真实基线\n GET /api/platform/probe → 502（forward_auth 去校验，没放行）；插入抢先 handle 的那份 → 200\n 「bypassed」，门禁判红。同一份里 /api/v1/decisions 仍 502，截走范围与模拟一致。\n- 一处自测抓到的实现缺陷：scopeNodes 顶层那一跳没挡 handler 分支，\"别的块里配了也算数\"——\n 正是要判的那种失效；已修并由「请求体上限挪到别的块」一例钉住。另一例钉住命名匹配器声明在\n 里层块时必须判得动（撤掉修复即落在 reverse_proxy 而非 respond，实测过）。\n- 回归：governance 全套 340/340；工作区一致性门禁只余既有 L9 compose 重名 1 条，与本轮无关。\n\n未做：G-3 限流要 caddy-ratelimit 插件即自建镜像，会打破 digest 锁定的官方镜像，是 stack 决策；\nHTTP 方法白名单归属 2026-09-16 已裁为不阻塞退役、转 stack 待办；C08.04 East-West 身份、C08.05\nWebhook ingress 与 Egress SSRF、C08.06 外发 / WAF / HA 全无落点（Q03）。静态判定的边界据实登记：\n非路径条件只能判「可能」，route 只作近似，每个前缀只探一条代表路径（/t/\u003ctenant\u003e/login 这类\n「有意分流一部分」不在判定范围），入口鉴权仍挡不住文档导航——运行态探针仍是必要补充。\n\n报告回绑另起 chore(reports) 提交（仓纪律 5）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:45:36-07:00"}],"HeadCommit":{"Sha1":"e8fda74e279a5884440d5efaac2265e513c814b6","Message":"chore(reports): 入口门禁与工作台两份快照回绑 @ a9abcbe\n\n在 a9abcbe 的干净 detached 检出里重跑（governance/rebind-reports.mjs --gates caddy,workbench-snapshots，\n带 CHECK_CADDY_DOCKER=1），三份均 provenance.gitSha=a9abcbe / worktreeDirty=false。\n\n- reports/caddy.latest.json：0 违规。12 个 HTTP 契约前缀 × 2 profile 全覆盖；新增的 routing 段逐条记下\n 模拟出来的块链，13 条登记前缀（11 个去重契约前缀 + 2 个长连接前缀）全部到达 {$PLATFORM_API_UPSTREAM}；\n workbenchGuard.reached=true / conditionalBefore=0；语法校验 dev / staging 均 passed，via docker，\n 镜像 caddy@sha256:5f5c8640…d58648（compose 锁定的那一个，不再是浮动 tag）。\n- reports/workbench-catalog-snapshot.latest.json：只有 provenance 变。caddy-routes.json 进了它的作用域\n （新增 validate / mount 段），被 check:evidence 判过期，但快照只吃 routes.contracts，内容逐字未变——\n 重建后与旧版按字段比对一致，已实测。\n- reports/workbench-ops-snapshot.latest.json：与 catalog 同一脚本产出，必须一起带回（分开会绑不同提交）。\n 它有一处内容变化不由本轮改动引起：image 段从 d8f09205 / sha256:fe418667… 更新到 05e3826e /\n sha256:c0869934…，来源是 06c3d4c 已提交的 reports/image-digest.json——此前那份快照相对该提交是旧的。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:46:25-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/c190a6ab5d2c50e4c90f027bbcca72f1f3050005...e8fda74e279a5884440d5efaac2265e513c814b6","Len":2}...
|
1789742873
|
Edit
Delete
|
|
31150
|
5
|
7
|
5
|
116
|
0
|
0
|
|
0
|
1|fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeE 1|fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键...
|
1789742918
|
Edit
Delete
|
|
31151
|
5
|
11
|
5
|
116
|
0
|
0
|
|
0
|
1|fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeE 1|fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键...
|
1789742925
|
Edit
Delete
|
|
31152
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"b62945890 {"Commits":[{"Sha1":"b62945890987b341b65684046c6e4d8ae612c529","Message":"Merge pull request 'fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键' (#1) from feat/registry-rule-hardening into main\n\nReviewed-on: https://gitea.g-hi.com/luoanwu/enterprise-platform/pulls/1\n","AuthorEmail":"law@g-hi.com","AuthorName":"luoanwu","CommitterEmail":"law@g-hi.com","CommitterName":"luoanwu","Timestamp":"2026-09-18T22:48:44+08:00"},{"Sha1":"92a6b8e647c2c4d764c1fcdb8448bfc48a4dc4ed","Message":"fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键\n\n蓝图 docs/domain/registry-snapshot-blueprint.md 从迁入起就写着「失败关闭」与\n「review key 仅用于人工复核幂等」两条不变量，实测三处不成立。本次按蓝图自己的口径补齐，\n改动只在 contracts/src/domain/application-contract-registry/：不新增写入点、不新增宿主、\n不动 Catalog 与 Schema，每条输出仍固定 activationAllowed=false / registryWriteAllowed=false /\nhumanApprovalRequired=true。\n\n1. 失败关闭：assessRegistrySnapshot / assessRegistryTransition /\n reviewCandidateSnapshotUpgrade 对畸形输入抛 TypeError（快照缺字段、now 不是 Date、\n decisions 整个缺失都会抛），四个入口里只有 LocalManifestAdmissionPlanner.plan 是\n 失败关闭的。现四个入口一致返回 REJECT + 原因码。这不是洁癖：check-fixtures 把抛出当\n 配置错误处理，实测一次 TypeError 会让整套 15 例逐例判定塌成一条 configuration 失败。\n\n2. 十六进制引用规范化：digest / commit 的正则带 /i 收大小写两种写法，比较与入键却\n 大小写敏感——同一摘要写成大写会得到另一个 reviewIdempotencyKey，同版本大小写翻转被判成\n SAME_VERSION_DIGEST_CHANGED，大写回滚 pin 被判成 ROLLBACK_PIN_MUST_MATCH_CURRENT。\n 现一律规范化为小写后再比较、再入键，ACCEPT 回吐规范化后的值。\n\n3. 版本文法收敛：一个模块三套（快照流与清单准入用 \\d+\\.\\d+\\.\\d+，升级复核用严格 SemVer），\n 新增 registry-primitives.ts 作单一来源，含预发布序；DEC-009—012 此前在两份文件各写一份，\n 一并收敛。snapshot-upgrade-review.ts 因此净减约 60 行重复实现。\n\n两处语义变化（蓝图新增一节已如实登记）：\n- 收紧：1.02.0 这类前导零版本此前被放行，现在拒；assessRegistrySnapshot 另开始拒 latest\n 这类可变标签——此前它能过体检却必然过不了迁移规则。\n- 放宽：1.0.0-rc.1 这类预发布版本此前被快照流整条挡在外面，现在按 SemVer 序参与迁移。\n 平台自己发的就是 1.0.0-rc.N，升级复核那一侧本来就这么判。风险面有限：本模块所有出口\n 都不激活、不写 Registry，多进人工复核队列不等于多放行制品。\n\n证据（本分支隔离工作树，clean）：注册中心单测 44/44（原 34 + 加固 10）；\ncontracts check:local 302/302；check:fixtures 九个契约域套件 100 例 0 失败，\n其中本套件 15 例（正 4 / 反 11，原 8 例）。篡改必红实测两次：sameHexRef 退回大小写敏感\n→ 单测 1 红 + 夹具 P03 红；去掉快照入口守卫 → 整套塌成 0 例 1 配置失败。\n\n仍为仓内候选 sdk_e2，不代表正式 Registry、Snapshot 发布或跨仓 Required Check 已上线。\n清单登记的 C01.04 / C01.06 是「接线」型缺口且前置 Q01 运行宿主未裁，本次未动。\nreports/fixtures.latest.json 未回绑：它是 17 套件的聚合报告，只跑 9 个套件去覆盖会缩小结论面，\n应在整批（含并行会话的 public-file / IM / 跨域流程改动）落定后统一重跑回绑。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T07:40:10-07:00"}],"HeadCommit":{"Sha1":"b62945890987b341b65684046c6e4d8ae612c529","Message":"Merge pull request 'fix(contracts): 注册中心规则补失败关闭与摘要规范化——三个入口此前抛 TypeError，同一摘要的两种写法此前分叉成两个复核键' (#1) from feat/registry-rule-hardening into main\n\nReviewed-on: https://gitea.g-hi.com/luoanwu/enterprise-platform/pulls/1\n","AuthorEmail":"law@g-hi.com","AuthorName":"luoanwu","CommitterEmail":"law@g-hi.com","CommitterName":"luoanwu","Timestamp":"2026-09-18T22:48:44+08:00"},"CompareURL":"luoanwu/enterprise-platform/compare/e8fda74e279a5884440d5efaac2265e513c814b6...b62945890987b341b65684046c6e4d8ae612c529","Len":2}...
|
1789742926
|
Edit
Delete
|