Skip to content

feat(benchmark): add frozen P1 offline assessment - #560

Open
zengchuang-hw wants to merge 6 commits into
mainfrom
codex/benchmark-434-offline-assess
Open

zengchuang-hw wants to merge 6 commits into
mainfrom
codex/benchmark-434-offline-assess

Conversation

@zengchuang-hw

@zengchuang-hw zengchuang-hw commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

Refs #434
Refs #371
Refs #430

Apps can now assess a fixed retained P1 record through the production Benchmark port instead of an injected result fixture. The new offline service verifies protocol/record binding, actual retained candidate bytes and trustworthy check proofs, then returns separate format (F), correctness (C) and measurement-validity (M) claims. Apps retains primary-key selection, archiving and publication, including exits 0/8/9 and archive-failure exit 6 with the in-memory result preserved.

Authority: #434 limited authorship delegation and final interface record. Dong1017 retains bounded delivery accountability and port ownership; zengchuang-hw / Codex authors the implementation. This does not assign S3 module ownership. The implementation baseline is main 9466bc8d910c9fef55d4a9e23ee1a82950280c30, including merged #530/#537/#542; current integration base is main 10ba7a5500bcd95949219b138f9ada5033b922b0 (including the #556 recovery-test correction), merged without rewriting branch history. Head: 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2.

Frozen task and public ports

The same accepted D023 / t108-v4 family drives preparation and assessment: p1-fixture@1 / p1-rules@1 / evaluate / p1, explicitly linked in the merged prepare contract. Public ports are benchmark.assessment.assess(prepared, evidence, *, artifacts), pure encode / decode / validate_assessment, and AssessmentService(clock=...) for reproducible assessment timestamps. See production-assess.md.

  • Bind the complete canonical record payload to RecordRef hash/size/attempt/revision, task, operation, scope, semantic prepared ID and protocol digest. Relocated outer references remain valid; contradictory declared candidate expectations are binding errors. No EvaluationResult is returned for a top-level mismatch.
  • Verify every opened artifact's bytes through the supplied read-only capability. The three portable open-error codes become per-claim gaps; unrelated host programming errors propagate. Raw proofs may have ordinary JSON whitespace, but duplicate keys, float/nonfinite tokens, invalid Unicode and unsupported fields are refused without rewriting retained bytes.
  • Match frozen check keys, revision, phase, schema, producer kind/domain, actual input capture and actual runtime entry/revision/code pin. A proof's self-declared candidate ID/hash is not evidence by itself. Unknown clocks remain unknown; confirmed execution is not made conditional on fabricating timestamps.
  • Aggregate conflicting trustworthy proofs once (V21). Unsupported same-family grader/claim-rule revisions are refused (V22). P1 emits no invented metrics or comparisons because the frozen protocol defines none. Eligibility is conservative: ineligible on failed/conflicting conditions, unknown on gaps, eligible only when all frozen conditions pass; there is no overall verdict field.
  • Compute V23 identity using only the declared typed-reference projection. Supporting content participates; self ID/time and typed ArtifactRef locations are excluded. Fixed-clock canonical bytes match in an isolated process.

Benchmark imports no Apps/Harness/Knowledge module, launches no worker/process, uses no network/device, and writes no archives. The installed-wheel check now permanently verifies the public offline port can import independently in a new -I interpreter. The three new canonical Apps test consumers are registered in the migration inventory.

Validation and limits

  • Current-head production prepare/assess + Apps assess group: 141 passed, 0 skipped. Includes the review regression for a good proof plus an unusable second declared proof, both within one check and across matching checks, both orders, all three portable read errors, bad bytes, invalid JSON and empty evidence. Verified refs remain retained, affected claims stay INCONCLUSIVE and other claims remain independent. Public identity projection now reports malformed nested values through a declared AssessmentError.
  • Expanded current-head validation: 435 passed, 332 subtests passed, 0 skipped across core ports, the corrected project-explanation recovery tests, migration inventory, import boundaries, first-wave isolation, contracts and package-root inventory.
  • Fresh sdist→wheel→new environment, outside-checkout installed checker (including the independent public-port import probe): passed on this head.
  • Entire-tree Ruff and diff whitespace passed. No dependency, frozen fixture/evidence, threshold or CI-lane change.
  • Current-head Ready CI passed (all eight jobs). The prior head failed only the known Fix the retired digest import in the project-explanation recovery test on main #555 recovery test; Use the Harness digest in the interrupted recovery test #556 is now included. The old failure is retained, not reported as a passed run. Required current-head CI remains a merge condition.

The CPU cross-tests run the real prepare/assess ports and Apps archive/read/publish code over retained software-test record/proof fixtures. They do not claim a production Harness evaluation worker execution, a service/model gain, or completion of #429/#371. Production shared CPU acceptance must still consume #429's actual evaluation target output. Keep #434 open for that joint acceptance after port integration.

Review and integration

Please inspect frozen P1 trust/eligibility rules, the complete-record binding, per-item failures versus top-level errors, pure identity projection, and Apps T35/T39 ordering. Dong1017's owner technical review is requested but cannot count as an independent approval, because he is the accountable parent for this slice. Obtain two current eligible independent write+ approvals (requested from Amy/David; first non-empty line LGTM), passing required checks and no blockers before squash integration. Author/self-review does not count.

Agent-assisted-by: Codex (author)
Signed-off-by: zengchuang <zengchuang3@huawei.com>
Agent-assisted-by: Codex (author)
Signed-off-by: zengchuang <zengchuang3@huawei.com>
Agent-assisted-by: Codex (author)
Signed-off-by: zengchuang <zengchuang3@huawei.com>
Agent-assisted-by: Codex (author)
Signed-off-by: zengchuang <zengchuang3@huawei.com>
Dong1017
Dong1017 previously approved these changes Nov 21, 2026

@Dong1017 Dong1017 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

PR review verdict — PASS
Reviewed base 7a7ab84c3b5f26b20c098f439470d5bcfa972809 and head 4258c76a84f3c9cf7086a0b7965a7ec17cd445f9 (full diff, diffComplete: true). prior-reviews.md shows no formal review and no outstanding CHANGES_REQUESTED on this head, so this is the first independent review. This is Dong-bot (authorized delegate) acting for Dong1017.

What I verified by reading the diff and the base-tree code it depends on

  • Module boundary. benchmark/assessment.py imports only contracts.json_v1 and the sibling benchmark.preparation. It has no Apps, Harness or Knowledge import, which fits the import gate in docs/module-layout.md. The new tests/check_installed.py probe covers this in a fresh -I -B interpreter. It checks that the module ships in the wheel and that no inferrsi.apps, inferrsi.harness or inferrsi.knowledge module gets loaded.
  • Helpers reused from preparation. The private helpers it calls exist at the head and have the expected shapes: _artifact, _id, _SHA, _DECIMAL, _closure, PreparationError (in preparation.py), validate_prepared, protocol_digest and prepared_id.
  • Migration inventory. The docs/apps-module-migrations.json edits add tests/test_benchmark_assessment.py to the consumers of apps.storage.archive, apps.storage.artifact_reader and apps.handoff.bridge. These are the canonical modules the new test imports. test_active_consumers_use_canonical_paths_and_are_registered requires exactly this registration. It only checks that each consumer is unique and is a file, so the append order is fine.
  • Binding order. _binding checks the full record hash and size, the attempt and revision, the task, operation and scope, prepared_id and protocol_digest. It does this before any artifact read, and test_binding_mismatch_is_refused_before_any_artifact_open asserts it. The inputs are snapshotted through a canonical round-trip, which the mutating-reader test covers.
  • Byte authority. _RetainedBytes.read reads at most size+1 bytes, so an oversize artifact is caught. It checks size and SHA-256, closes any stream it opened, and caches the outcome. Only the portable code values UnsafeArtifactPath, ArtifactUnavailable and ArtifactIntegrityError become claim-level gaps. Other host exceptions propagate, and the tests cover TypeError and CanonicalError. A failure to read the protocol itself is a port error, not a guessed assessment.
  • Trust checks. _trusted checks the check, producer, domain, runtime pin (entrypoint, revision and code hash) and a started step bound to the candidate input. The tests cover wrong-code, foreign-check, uncaptured-input and not-started cases.
  • Claim logic.
    • F and C are independent, so one claim's missing evidence does not erase another's result.
    • M additionally requires a completed measurement, confirmed cleanup and lease release.
    • Contradictory trustworthy proofs, including conflicting sample values, collapse into one INCONCLUSIVE claim with conflicting_evidence.
    • validate_assessment enforces exactly three claims, empty metrics and comparison, calendar-checked timestamps and the semantic ID. The ID excludes only the clock, the self-ID and typed ArtifactRef locations.
  • Docs. docs/contracts/production-assess.md and the new docs/operations.md row match the code, including the explicit statement that the Harness evaluation worker and T37/T40 (#429) are not claimed.

Non-blocking follow-ups (not gating)

  1. _observations drops problems whenever at least one other proof for the same check_id parsed. Two cases follow, and both end up PASS:

    • A trusted check lists a valid passed proof plus a second evidence ref that is unreadable or fails its hash.
    • A second same-check_id check has unreadable evidence.

    Neither case is surfaced as a gap or as INCONCLUSIVE. A corrupted failed proof could therefore be hidden by a good passed one. The doc says an affected proof becomes a claim-level gap, so I suggest either marking the claim INCONCLUSIVE when any trusted declared ref is unusable, or stating the current rule in the contract doc. A test with two refs on one check would pin whichever behavior you choose.

  2. assessment_id() is public but only checks the top-level fields. A malformed nested result can raise KeyError or TypeError there instead of a declared AssessmentError. validate_assessment runs the shape checks first, so the supported paths are fine.

Not verified here

  • I did not run the tests, CI or the wheel/installed check, and I did not see the CI status.
  • I did not diff the code against the t108-v4 golden fixtures or the Apps assess wiring beyond what the new tests exercise.
  • I did not audit preparation.validate_prepared internals.
  • Joint production CPU acceptance (#371/#429) remains outside this PR, as the PR states.
  • The merge ruleset still needs the second eligible approval and green required checks.

Posted by Dong-bot, an automated agent acting on behalf of @Dong1017 (always-on-bot, model claude-sonnet-5-5).

@Dong1017 Dong1017 added the kind:task Bounded delivery task with a delegated slice label Nov 21, 2026

Copy link
Copy Markdown
Contributor Author

@Dong1017 请按 7034454814 的已收口检查点 技术检视当前 head 4258c76a84f3c9cf7086a0b7965a7ec17cd445f9:同一 D023 / t108-v4 / p1-fixture@1 / p1-rules@1、按语义键绑定、逐项实际字节完整性、F/C/M 独立规则与 V21 冲突、V22 同族未知 revision 拒绝、V23 跨进程固定时钟字节、T35/T39 Apps 交叉路径。当前 CPU 输入是明确的留存记录/证明夹具,不宣称 #429 真实 evaluation worker 已执行。

@amy-why-3459 @david6666666 请作为独立评审检视同一 head。当前源码对应核心交叉组 114 passed;边界/登记组与隔离安装的实际受测版本关系见正文(无隐藏跳过或弱化检查)。Ready CI 44898147291 已启动,合入前须通过。

角色明确:作者为 zengchuang-hw / Codex,accountable parent 是 Dong1017;Dong 的 owner 技术意见不计独立批准。两位有效独立 write+ 账户通过后,请各自以首个非空行独立为 LGTM 的正式 Approve 记录结论;有意见先处理。这里不修改 S3 整体 ownership,也不将本 PR 或目录任务关闭算作 #371 总验收。

Copy link
Copy Markdown
Collaborator

Reviewed 9466bc8d910c9fef55d4a9e23ee1a82950280c30 → 4258c76a84f3c9cf7086a0b7965a7ec17cd445f9 (full diff). The PR is also two commits behind current main (91addd96 #544, 7a7ab84c #547); those changes do not appear to overlap the assessment port itself.

The module boundary, binding-before-read, portable artifact error handling, F/C/M independence, V21/V22/V23 coverage, and the Apps T35/T39 tests look consistent with the accepted #434 interface and docs/contracts/production-assess.md. Accuracy/performance comparison evidence is not applicable: this is an offline assessment of retained software-test fixtures, and the PR does not claim a production worker run or a service gain. Architecture also looks appropriate: benchmark.assessment imports only contracts.json_v1 and sibling benchmark.preparation, with no Apps/Harness/Knowledge imports.

One issue I would like resolved before approving:

_observations in src/inferrsi/benchmark/assessment.py keeps problems only when proofs is empty, so _claim can still PASS if a trusted check has one parsable proof and another declared evidence_refs entry that is unreadable or fails its hash. The unusable ref is dropped. The contract says "An affected proof becomes a claim-level evidence gap," and a corrupted failed proof could therefore be hidden by a good passed one. test_each_read_failure_is_only_a_claim_gap only covers the single-ref case.

Please either treat a trusted check as incomplete/INCONCLUSIVE when any declared ref is unusable (and add a two-ref test), or state the current "any one valid proof is enough" rule in production-assess.md and pin it with that test.

Non-blocking notes: please rebase onto current main before merge. Required CI on this head was still running when I reviewed, and the known test_recovered_attempt_does_not_trust_a_rewritable_declaration failure on main (#555 / fix #556, not yet merged) may also affect full-check here. As the PR already states, Dong1017's owner review should not count as one of the two independent approvals.

Keep an affected claim inconclusive when any matching declared proof is unusable, even if another proof is valid. Return declared errors for malformed identity projections.

Agent-assisted-by: Codex (author)
Signed-off-by: zengchuang <zengchuang3@huawei.com>
Agent-assisted-by: Codex (author)
Signed-off-by: zengchuang <zengchuang3@huawei.com>
cursor[bot]
cursor Bot previously approved these changes Nov 21, 2026

@cursor cursor Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

PR review verdict — PASS
Reviewed base 7a7ab84c3b5f26b20c098f439470d5bcfa972809 and head 4258c76a84f3c9cf7086a0b7965a7ec17cd445f9.

Technical / third-party verify only (cursor App collaborator permission none — not write+ eligible / dual-tip / merge-ready under docs/agent-collaboration.md).

Checked

  • New inferrsi.benchmark.assessment offline P1 port: no Apps/Harness/Knowledge imports; binding + retained-byte proof path; independent F/C/M claims; AssessmentError portable codes.
  • Tests cover codec identity, claim gaps vs host errors, binding refusal before artifact open, unsupported rule revision, clock/location projection, no worker/network, Apps exits via fixture path.
  • Docs: production-assess.md + operations.md row; check_installed wheel isolation probe; migrations JSON lists the new test.

Ready note (non-code)

  • Tip Ready full-check still FAIL on tests/test_project_explanation.py::test_recovered_attempt_does_not_trust_a_rewritable_declaration (inferrsi.infra.artifacts.digest) — same mainline residue fixed by merged #556. Rebase/update onto main that includes #556 and re-run Ready before steward merge.

@cursor cursor Bot left a comment •

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

PR review verdict — PASS
Reviewed base 10ba7a5500bcd95949219b138f9ada5033b922b0 and head 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2 (re-review after tip move dismissed prior Approve on 4258c76a84f3…).

Technical / third-party verify only (cursor App collaborator permission none — not write+ eligible / dual-tip / merge-ready).

Delta since prior tip

  • Evidence-gap integrity: empty evidence_refs and any unusable declared proof keep the claim INCONCLUSIVE even when another proof is valid (_observations / _claim); docs note order independence.
  • assessment_id maps malformed nested projection to declared InvalidRequest.
  • New tests: test_good_proof_cannot_hide_another_declared_evidence_gap (order x gap kinds) and malformed identity projection.

Also re-checked offline port boundaries (no Apps/Harness/Knowledge imports) and installed-package probe from earlier tip.

Ready full-check in progress on post-#556 base at review time — steward merge after green + write+ independent Approve (Dong tip Approve was dismissed by the push).

Copy link
Copy Markdown
Contributor Author

@amy-why-3459 @Dong1017 已采纳多份证明缺口的意见:head 4887f1e9 中,匹配检查的任一声明证明不可读、完整性失败、字段非法或无证明时,该 claim 保持 INCONCLUSIVE;有效证明引用和缺口原因保留,顺序不影响结果,其余 claim 独立。已先复现两种错误 PASS,再补齐同一检查多 ref / 同 check_id 多检查、正反顺序、三种 portable error、错误字节/非法 JSON/空列表的矩阵。也处理了 assessment_id() 嵌套结构泄露 KeyError/TypeError 的建议。

核心交叉组 141 passed、0 skipped,Ruff/whitespace 通过。按主线同步建议,以普通 merge 纳入已合入的 #556 和当前 main 10ba7a55(保留公开提交历史,无强推);旧 CI 的唯一 full-check 失败正是 #555,两版 installed-package 和 Dashboard browser 均已通过。正在补跑新 head 的扩大回归与 fresh wheel 检查,并等待新 CI;尚不宣称 checks 通过。

@amy-why-3459 @david6666666 请复审新 head。Dong 是 accountable parent,他的技术批准仍不计两份独立审批。

@Dong1017 Dong1017 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

PR review verdict — PASS
Reviewed base 10ba7a5500bcd95949219b138f9ada5033b922b0 and head 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2 (full diff, diffComplete: true). This is a Dong-bot (authorized delegate) technical owner review. It is not one of the two independent write+ approvals the PR requires, because Dong1017 is the accountable parent. I found no outstanding CHANGES_REQUESTED in prior-reviews.md. My earlier Approve on 4258c76a was dismissed by the push, so this is the re-review on the new head.

Earlier blocker (amy-why-3459): a good proof hiding an unreadable or invalid second proof. Resolved on this head.

  • _observations now records problems for every matching check, whether or not other proofs parsed:

    • an untrusted or incomplete check;
    • an empty evidence_refs;
    • an unreadable ref;
    • an integrity failure;
    • an invalid proof.
  • _claim marks the check missing when not proofs or problems (previously, only when proofs was empty). Valid refs and gap reasons are kept in the claim, and the other claims are not affected.

  • test_good_proof_cannot_hide_another_declared_evidence_gap covers:

    • the three portable error codes, plus wrong bytes and invalid JSON;
    • the same check with two refs, and two checks with the same check_id;
    • an empty evidence_refs;
    • both orders.

    It asserts INCONCLUSIVE, the missing_evidence == ["format"] list, retention of the good refs, and unknown eligibility.

  • production-assess.md now says "Every declared proof for a matching check must be usable" and that the result does not depend on order. The contract and the code agree.

assessment_id hardening. assessment_id wraps _projection and catches KeyError, TypeError, CanonicalError and PreparationError. It re-raises them as AssessmentError("InvalidRequest"). test_public_identity_reports_malformed_nested_projection_as_declared_error covers five malformed shapes. Host adapter exceptions on open still propagate; that is pinned by test_host_programming_errors_are_not_disguised_as_missing_evidence.

Re-checked from the earlier review.

  • Module boundary: benchmark.assessment imports only contracts.json_v1 and benchmark.preparation. The -I -B probe in tests/check_installed.py asserts that no apps, harness or knowledge modules load.
  • Binding runs before any artifact read: test_binding_mismatch_is_refused_before_any_artifact_open.
  • The three claims are independent, and V21, V22 and V23 are covered.
  • The T35/T39 Apps exit-code paths (0/8/9, exit 2, exit 6 with the in-memory result kept) are covered.
  • The migration inventory lists tests/test_benchmark_assessment.py for the canonical modules it imports.

Not verified here (no execution).

  • I did not run the 141-test core group, the Ruff and whitespace checks, or the fresh-wheel check. Those figures come from the author.
  • Ready CI on this head was still running. The author does not claim it passes. It must be green, including full-check after the #556 merge, before squash.
  • The merge also needs two eligible independent write+ approvals, from Amy and David. cursor[bot]'s Approve does not count toward that, since its collaborator permission is none.
  • This does not change the stated scope: it is an offline assessment of retained fixtures. It does not accept #429, #371 or #434 jointly.

Non-blocking: if the production protocol ever needs "any one valid proof is enough" semantics, that would be a deliberate contract change. It should come with its own acceptance, not a silent relaxation of the current strict rule.


Posted by Dong-bot, an automated agent acting on behalf of @Dong1017 (always-on-bot, model claude-sonnet-5-5).

@amy-why-3459 amy-why-3459 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

PR review verdict — PASS
Reviewed 4258c76a84f3c9cf7086a0b7965a7ec17cd445f9 → 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2 (incremental, including the merge of current main 10ba7a5500bcd95949219b138f9ada5033b922b0).

The earlier blocker is resolved on this head:

  • _observations now records a problem for an empty evidence_refs list, and still records unreadable, integrity, and invalid-proof gaps even when another proof parsed.
  • _claim treats a matching check as missing evidence when problems is non-empty, not only when proofs is empty. Verified refs are kept; other claims stay independent.
  • test_good_proof_cannot_hide_another_declared_evidence_gap covers two refs on one check, two checks with the same check_id, both orders, the three portable error codes, wrong bytes, invalid JSON, and empty refs, and asserts INCONCLUSIVE / missing_evidence == ["format"] / retained good refs.
  • production-assess.md now states that every declared proof for a matching check must be usable and that order does not change the result.

I also checked the assessment_id() InvalidRequest mapping for malformed nested projections, and the merge with main (#544, #547, #556). The PR-versus-main diff remains the six assessment-port files; the previously unmerged test_recovered_attempt_does_not_trust_a_rewritable_declaration fix is in this head via #556.

Accuracy/performance comparison evidence is still not applicable (offline assessment of retained fixtures; no service-gain claim). Dong1017's owner review should not count as one of the two independent approvals.

Note: Ready full-check on this head was still running when I reviewed (run 44900119820); lint-and-smoke, installed-package, dashboard-browser, and declarations had passed. Merge should wait until the Ready lane is green.

Copy link
Copy Markdown
Contributor Author

@david6666666 请补充当前 head 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2 的独立检视。Amy 已正式批准该 head(6005015737),原多份证明缺口已修复;仍需要第二份有效独立 write+ 批准。Dong 的 owner 技术批准和 Cursor 技术批准均不计此处独立票。

当前源码扩大回归 435 passed、332 subtests passed、0 skipped,fresh sdist→wheel 隔离安装通过;Ready CI 等两版 full-check 完成,其余安装/浏览器检查已通过。请重点看每个声明 ref 的缺口是否保留、F/C/M 独立性、V21/V22/V23,以及接口没有执行副作用;有意见我先处理。

Copy link
Copy Markdown
Contributor Author

@david6666666 当前 head 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2 的 Ready CI 44900119820 已全部通过,knowledge-model-policy 也为 success。Amy 的同 head 正式批准已核对;现在缺第二个独立 write+ 账户的正式复审结论。请检视后独立记录 PASS / DO NOT PASS;若有意见我先处理,不将 Dong(parent)或 Cursor(无 write)的技术批准凑成两票。

Copy link
Copy Markdown
Contributor Author

@jiangkuaixue123 补请你独立检视当前 head 4887f1e94f39c1d4b675e965cfcc0a6e932d50a2。冻结 P1 离线判定:固定记录绑定、每个声明证明的完整性与缺口、F/C/M 独立性、V21/V22/V23、无执行副作用。

当前 head Ready CI 已全部通过,细项本地回归与 fresh wheel 结果见正文。Amy 已正式批准;原 David 请求仍保留。Dong 是本切片的 accountable parent / owner,他的批准不计独立票,作者也不自审。请按实际 diff 和接受范围给出 PASS / DO NOT PASS;符合时正式 Approve 的首个非空行请独立为 LGTM。若参与过本切片的实质材料或实现,请注明,以便准确记录角色。

Copy link
Copy Markdown
Contributor Author

@hsliuustc0106 补请你独立检视本 PR。当前 head 的 Ready CI 全部通过(3.10 / 3.11 各 5980 passed、31 skipped、674 subtests),核心/边界回归与 fresh wheel 结果见正文;Amy 已同 head 正式批准。现有 David / jiangkuaixue123 的请求仍保留。请按实际差异给出结论,有意见会先处理。

本切片 parent 是 Dong1017,他的技术批准不计独立审批;作者不自审。需要另一位独立 write+ 评审正式批准后才可按既定顺序合入。

Copy link
Copy Markdown
Contributor Author

当前 4887f1e9 Ready CI 仍全绿,已确认只有 Amy 是一份有效独立审批。@david6666666 @jiangkuaixue123 @hsliuustc0106 请复审并在判断通过后给当前头 formal Approve(首个非空行 LGTM);若有意见先列出,我会处理。Dong 是 accountable parent,Cursor 没有 write,二者不能作为第二份。

本 PR 还卡住 #566 的同步集成;新的 H capture 已是 Ready #567,真实证据闭环的剩余任务被明确跟踪为 #565。没有把后续实现或 fixture 测试偷算成 #434/#371 完成,也不会绕过两份独立批准合入。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

kind:task Bounded delivery task with a delegated slice

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants