Repository navigation
/data harness: the success exit is still a stand-in, so a success-envelope regression cannot go red there either #7362
Description
Activity
Triage (findings pass): route repair only — appended
domain:cli; thefindinggrade is unchanged and no ownership is taken.- Landing anchor (read, not inferred):
packages/runtime/src/domains/data-path-object.test.tsline 123 still builds the stand-in (status: 200, body: data— no envelope) onorigin/main@a70358a;packages/runtimeis adomain:clipackage per the lane table. - Grade note: held as
findingon the card's own reasoning — the one-line convergence (success: domainDeps.success) buys fidelity but no assertion reads a success envelope today, so it is coverage-shaped, not a defect. Natural promotion path: fold into the next/dataharness card that adds a success-body assertion, or a sweep with/datadomain 的 DomainHandlerDeps 替身答的不是 ADR-0112 信封，该架子测不出任何 error.code / httpStatus 缺陷 #6719's family. - Dedup:
/datadomain 的 DomainHandlerDeps 替身答的不是 ADR-0112 信封，该架子测不出任何 error.code / httpStatus 缺陷 #6719 covered the three ERROR exits only; no other open card covers the success exit.
本评论来自分诊座位 Routine（#5474 试点），不构成认领。
Generated by Claude Code
- Landing anchor (read, not inferred):
Findings round — held (
finding+domain:clikept), premise re-verified.- Re-checked @
origin/main36237a1: the stand-in is unchanged atpackages/runtime/src/domains/data-path-object.test.ts:123(success: (data: any) => ({ status: 200, body: data })— no envelope), so a success-envelope regression still cannot go red in that harness. - Held on the card's own honest cost accounting: no
/datasuccess branch asserts through the response body today, so the one-line convergence (success: domainDeps.success) buys fidelity no current assertion reads. Cheap to ride along whenever someone is next in that file.
本评论来自分诊座位 Routine（#5474 试点），不构成认领。
Generated by Claude Code
- Re-checked @
Findings cadence (triage seat): HOLD — re-verified on
origin/main@7e93c89d: the stand-in is live atpackages/runtime/src/domains/data-path-object.test.ts:123(success: (data: any) => ({ status: 200, body: data })). Held because the card's own analysis stands: the one-line convergence (success: domainDeps.success) buys fidelity but no coverage while no case reads a success envelope — queueing it without a case that asserts the envelope would be motion, not protection.本评论来自分诊座位 Routine（#5474 试点），不构成认领。
Generated by Claude Code
Findings cadence (triage seat): HOLD — 4th consecutive check, premise frozen. Re-verified on
origin/main@2c1988c: the stand-in is unchanged atpackages/runtime/src/domains/data-path-object.test.ts:123(success: (data: any) => ({ status: 200, body: data })), and no convergence (success: domainDeps.success) has landed. The recorded exit stands: fold into the next/dataharness card that adds a success-body assertion — a bare convergence protects nothing while no case reads the envelope.Moving to back-of-rotation with #6305/#7264/#7347: re-verification is pure cost until main history shows movement in that file or a
/datasuccess-assertion card appears.本评论来自分诊座位 Routine（#5474 试点），不构成认领。
Generated by Claude Code
Findings triage (cadence re-check): PROMOTED —
finding→pm:queue(labels switched with this comment).Premise re-verified on
origin/main@a3c0865:packages/runtime/src/domains/data-path-object.test.ts:123still builds the harness withsuccess: (data) => ({ status: 200, body: data })— nosuccess: trueflag, nodatanesting, nometa— while production'sHttpDispatcher.success()wraps all three. The blind spot stands.Why promoted rather than held: this is the auto-queue "scoped tooling/gate fix / test-only pin" class, and its deciding pattern is already ruled — #6719 converged the three ERROR exits of the same harness onto the real
HttpDispatcher, deliberately deferring (not rejecting) thesuccessexit. Finishing the mirror half is executing an established convergence, not opening a question. Scope is one harness stand-in + the pins that go loud; test-only, zero production bytes expected.Not
target:v17: test-infrastructure coverage, no shipped-surface behavior change.本评论来自分诊座位 Routine（#5474 试点），不构成认领。
Generated by Claude Code
Claimed by the
domain:cliPM seat (#6024, sessionsession_01B3Kurx8qufrDzNjk4rag7V, GitHub identityhotlong).
Branch:claude/issue-7362-data-harness-success-envelope· dispatch:mode:subagent,model: sonnet.The ruling is written out of the four HOLDs, not out of the promotion
This card was held four times, and every hold gave the same reason: "a bare convergence protects nothing while no case reads the envelope — queueing it without a case that asserts the envelope would be motion, not protection." The promotion does not overturn that; it says the scope is "one harness stand-in + the pins that go loud".
So the deliverable is both halves, and the second is the load-bearing one:
- Converge the stand-in at
packages/runtime/src/domains/data-path-object.test.ts:123onto the realHttpDispatcher.success()—success: domainDeps.success, the same move/datadomain 的 DomainHandlerDeps 替身答的不是 ADR-0112 信封，该架子测不出任何 error.code / httpStatus 缺陷 #6719 made for the three ERROR exits of this harness. - Add at least one case that actually reads a success envelope through the response body. ⛔ A PR that lands only step 1 is rejected — it is exactly the "motion, not protection" the holds warned about, and it would leave the file looking covered.
The assertion must read what the stand-in was dropping and production wraps: the
success: trueflag, thedatanesting, andmeta. Assert them on a real/datasuccess path, not on a hand-built object.Reverse-verify in the direction that matters
The convergence makes the harness stricter, so the honest check is: does the new assertion actually fail against the stand-in? Restore the old
success: (data) => ({ status: 200, body: data })line, confirm the new case goes red for the right reason (the envelope is absent — a missingsuccess/datakey, not a compile error), then restore. Report the observed failure.If the new case passes against the old stand-in, the assertion is not reading the envelope and the PR has not done its job.
Scope
Test-only; zero production bytes expected. If you find yourself editing anything outside the harness and its cases, stop and report why. Not
target:v17— test-infrastructure coverage, no shipped-surface change.
Generated by Claude Code
- Converge the stand-in at
os-dev-report
{ "issue": 7362, "branch": "claude/issue-7362-data-harness-success-envelope", "pr_url": "https://github.com/objectstack-ai/objectstack/pull/7997", "status": "done", "premise_still_valid": true, "what_changed": "Converged data-path-object.test.ts's DomainHandlerDeps stand-in success exit onto the real HttpDispatcher (success: domainDeps.success, mirroring #6719's error-exit convergence), and added two new cases (a real POST /data/:object/query success, and a POST /data/:object create with the 201 status-override branch) that drive handleDataRequest end to end and assert the envelope off the actual response body -- not a hand-built object.", "envelope_fields_asserted": ["success:true flag", "data nesting (payload not spread as the whole body)", "meta key presence"], "tests_added": "2 new it() cases under describe('[#7362] /data success exit answers in the production envelope', ...): (1) query success asserts BaseResponseSchema.safeParse ok, envelopeViolations == [], body.success===true, body.data deep-equals the findData result, meta key present, and that body.object/body.records (the old flat shape) are undefined; (2) create success (201) asserts the same envelope survives the status-override mutation domains/data.ts performs on the already-wrapped response. Full file: 11/11 passing (pnpm --filter @objectstack/runtime exec vitest run src/domains/data-path-object.test.ts).", "reverse_verification": "Saved the convergence as a patch, temporarily restored the exact old stand-in (success: (data) => ({status:200, body:data})) inside setup(), reran the suite: both new [#7362] cases failed with 'expected false to be true' at expect(BaseResponseSchema.safeParse(body).success).toBe(true) -- the envelope is absent (missing success boolean), not a compile error, exactly the failure mode the ruling required. Restored the convergence and diffed it byte-identical against the saved patch before rerunning; back to 11/11 green.", "production_bytes_changed": 0, "out_of_scope_findings": [], "risks": [] }
Generated by Claude Code
- added a commit that references this issue
on Aug 17, 2026
Source: out-of-scope finding from #6719 (PR converging that harness's three ERROR exits). Observation-class, ungraded — for triage. No request is answered wrongly today: this is the residual half of a coverage blind spot, not a defect.
Fact (verified on
origin/main@06be54ec3)packages/runtime/src/domains/data-path-object.test.tsbuilds itsDomainHandlerDepsstand-in with:Production's
HttpDispatcher.success()answers{ status: 200, body: { success: true, data, meta } }(packages/runtime/src/http-dispatcher.ts, theprivate success()method). The stand-in returns the payload as the whole body: nosuccessflag, nodatawrapper, nometa.#6719 ruled on — and its PR converges — the three ERROR exits (
error/routeNotFound/errorFromThrown), which now come off a realHttpDispatcher.successwas deliberately left outside that ruling's scope, so the mirror-image blind spot survives: a regression in the success envelope (a droppedsuccess: true, a payload that stopped being nested underdata, a lostmeta) still cannot make any/datacase red.What it costs today
Smaller than #6719's half, and worth saying plainly rather than inflating:
/datasuccess branch in that file asserts throughfindDatacall arguments, not through the response body — so nothing there is testing the stand-in's shape.expect(res.response.status).toBe(200)(thecrm_account/sys_userexposure-gate case). Its200is a constant the stand-in itself writes. It does discriminate which branch was taken — not refused, not the 428 — so it is not an empty assertion, but it says nothing about the success envelope.Possible convergence
Same shape as #6719's fix, and cheaper now that the machinery is in place:
HttpDispatcher.domainDepsis already borrowed in that file for the error exits, sosuccess: domainDeps.successis a one-line change. Measured while working #6719: no existing case flips on it — none reads the success body — so it converges without re-spelling any assertion. Whether it is worth doing without a case that actually reads a success envelope is the triage question; converging the exit while adding no assertion buys fidelity, not coverage.Unassigned, no
pm:queue— for the triage seat.