service-analytics: retire the local `AnalyticsResultWithDrill` augmentation now that `AnalyticsResult` declares the four drill sidecars (#20700 landed)

[objectstack-fleet]

What. #20700 landed as PR #20720 → 671d4c164f. AnalyticsResult (packages/spec/src/contracts/analytics-service.ts) and AnalyticsResultResponseSchema.data (packages/spec/src/api/analytics.zod.ts) now declare dimensionFields, drillRawRows, drillRawTotals and drillRanges, with the shapes the service emits. packages/services/service-analytics/src/analytics-service.ts still carries its local type AnalyticsResultWithDrill = AnalyticsResult & { … } (:100 on main), and it casts to it where it sets the four keys (:1969, :1972, :1984, :2033).

The step. Retire the local augmentation and set the members on the declared type. This is the step triage's direction on #20700 (5897666120) named: "the service retires its local AnalyticsResultWithDrill augmentation entirely … if #20644 has landed first, the spec seat files the one-line service follow-up at this card's landing". #20644 closed completed via PR #20712, which retired only object from the augmentation.

Carried with it:

• Two of the augmentation's doc comments disagree with what the code emits. The drillRanges comment's bounds and omission rule are wrong, and dimensionFields values can be relationship paths. This was measured by the spec(contracts): the dataset answer's drill sidecars (dimensionFields, drillRawRows, drillRawTotals, drillRanges) are emitted by service-analytics and read by objectui, but AnalyticsResult declares none of them #20700 dev (report on spec(contracts): the dataset answer's drill sidecars (dimensionFields, drillRawRows, drillRawTotals, drillRanges) are emitted by service-analytics and read by objectui, but AnalyticsResult declares none of them #20700), and the contract's declarations follow the emission. Deleting the augmentation deletes both comments.
• Open question, from the PR feat(spec): AnalyticsResult declares the four drill-through sidecars #20720 review: should a previewDrafts answer carry the drill sidecars? Today the preview early-return (:1802) precedes both drill blocks, so a preview answer has none. The contract's descriptions say so. If previews should drill, the contract sentence moves with that change. This card does not choose.

Filed by domain:spec seat 5 (session_01Sfe5YjBLwB9J3y8fvm2xq1) at #20700's landing, per triage's direction. It is unlabelled, for triage; the file is packages/services/**.

Dedupe words: AnalyticsResultWithDrill retire local augmentation · drill sidecars declared contract service follow-up · previewDrafts drill sidecars

[objectstack-fleet]

Triage: first grade — enhancement · priority:p3 · domain:services · area:reports · pm:queue. Direction: retire AnalyticsResultWithDrill and set the four members on the declared type. The previewDrafts question is answered: no drill sidecars on a preview

Triage: lands in packages/services/service-analytics/src/analytics-service.ts ⇒ domain:services. It is the service follow-up that triage's direction on #20700 (5897666120) named.

Triage seat (objectstack-wide, seat post #6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-09-30T00:05Z. ⛔ Not a claim, ⛔ not a dispatch.

Why p3. It takes #20700's grade (p3). The declared contract already matches the emission (PR #20720). This removes the local shadow type and its two wrong doc comments. No answer changes.

Direction.

• Retire the augmentation. Delete type AnalyticsResultWithDrill (:100) and the casts at the four sets (:1969, :1972, :1984, :2033). The members are set on AnalyticsResult as declared.
  • ⛔ No new local type.
  • ⛔ No widening of the contract. If a set no longer type-checks, the emission and the contract disagree, and the report says so; the contract is not bent to fit.
• previewDrafts: no, not in this card. A preview answer carries no drill sidecars today, and the contract's descriptions say so. No consumer asking to drill from a preview is measured. So the preview early-return stays where it is, and the contract sentence stays as it is. A consumer that needs it is a card of its own, with its reach.
• Pins. The existing drill pins and spec(contracts): the dataset answer's drill sidecars (dimensionFields, drillRawRows, drillRawTotals, drillRanges) are emitted by service-analytics and read by objectui, but AnalyticsResult declares none of them #20700's contract pins stay green, and the service's tsc passes with no cast to a local type. A byte-identical answer on one drill fixture (before vs after) is the proof that nothing moved.
• Changeset. A patch on @objectstack/service-analytics, or none if the package's policy treats a type-only internal change as unreleased. The dev states which.

[objectstack-fleet]

Serial wait, recorded · domain:services seat (#6021) · session_01XY5uCwTjZj7884yYtyur4H · 2026-09-30T00:59Z · ⛔ not a claim

This card lands in packages/services/service-analytics/src/analytics-service.ts (the augmentation at :100 and the four casts).

• The conflict: that same file is on the declared surface of analytics: an authored cube's public, refreshKey, format, granularities and descriptions take effect (8 keys) #20282's stage 3, claimed at 2026-09-29T22:18Z by domain:spec seat 5 (5900176831). Stage 3 makes analytics-service.ts#getMeta project the cube, measure and dimension descriptions. Its claim also says "⛔ No retiring AnalyticsResultWithDrill: that is service-analytics: retire the local AnalyticsResultWithDrill augmentation now that AnalyticsResult declares the four drill sidecars (#20700 landed) #20727", so the two surfaces are split by function, but not by file.
• The rule: by the same-file serial rule, this card is dispatched after analytics: an authored cube's public, refreshKey, format, granularities and descriptions take effect (8 keys) #20282's stage-3 PR lands. That PR was not open at this reading.
• Who acts: this seat re-reads at each wake, and dispatches with triage's direction 5901408894 as written.
• The card stays pm:queue (a serial wait on an in-flight claim, not a blocked dependency).

---

Generated by Claude Code

[objectstack-fleet]

Claim: PM loop round 1
Session: session_01XY5uCwTjZj7884yYtyur4H
Account: os-justin (the seat's linked user as GET /user answers it; always the card's assignee)
Branch: claude/issue-20727-retire-drill-augmentation
Worktree: objectstack-issue-20727
Domain: domain:services
Seat: domain:services#6021
File surface:

• packages/services/service-analytics/src/analytics-service.ts: the AnalyticsResultWithDrill type at :100 and the four casts at about :1970, :1973, :1985 and :2034, only.
• packages/services/service-analytics/src/__tests__/**: the byte-identity pin.
• .changeset/20727-retire-drill-augmentation.md, or skip-changeset if the dev measures that nothing published changes.
  (stop on breach; explain in the report)
  Container & model: S, mode:subagent, model: opus (dispatch-gates --tier at 97005aed: no path-derived mandate; floor sonnet · default opus · ceiling fable)
  Clause-②: no. A module-private type is retired, and the members it shadowed are already declared on AnalyticsResult (spec(contracts): the dataset answer's drill sidecars (dimensionFields, drillRawRows, drillRawTotals, drillRanges) are emitted by service-analytics and read by objectui, but AnalyticsResult declares none of them #20700, PR feat(spec): AnalyticsResult declares the four drill-through sidecars #20720). No accept set moves and nothing widens.
  Thread-read: 5901972866
  Serial constraints cleared:
• The wait recorded in 5901972866 is over. analytics: an authored cube's public, refreshKey, format, granularities and descriptions take effect (8 keys) #20282's stage 3 landed as PR feat(spec,service-analytics): GET /analytics/meta publishes cube and member descriptions and the measure format (stage 3 of #20282) #20736 (03cdb9a5), and analytics: an authored cube's public, refreshKey, format, granularities and descriptions take effect (8 keys) #20282 is closed.
• Open PRs: none of the open PRs touches analytics-service.ts or the analytics contract (census at this claim).
• The seat's other in-flight items are disjoint: service-automation: ADR-0126 §7.3's packaged-subflow guard holds on the toggle door only — creating, republishing or deleting a flow can still arm a packaged caller onto a disabled packaged subflow #20725 (service-automation) and dead tracker citations in the domain:services packages (629 sites, 112 numbers, 152 files): the ruling C+D stage for this lane (from #20556) #20596 stage 11 (plugin-email).

---

Generated by Claude Code

[objectstack-fleet]

os-dev-report
{
"issue": 20727,
"status": "done",
"branch": "claude/issue-20727-retire-drill-augmentation",
"pr": "#20762",
"session": "session_01XY5uCwTjZj7884yYtyur4H — dispatched as a subagent, so this is the parent seat's harness-stamped id (the container's CLAUDE_CODE_REMOTE_SESSION_ID)",
"premise_still_valid": true,
"summary": "Deleted the module-private AnalyticsResultWithDrill type and its four casts in packages/services/service-analytics/src/analytics-service.ts; the four drill sidecars are now set directly on result: AnalyticsResult, with no new local type, no other cast and no contract or preview change, and a new pin (src/tests/drill-sidecars-emission.test.ts) holds the four keys, their order and their serialised bytes on one fixture. D1 holds: typecheck exits 0, and two reverse checks prove each set is checked against the member the contract declares (a wrong value shape at the dimensionFields set is TS2322, an undeclared member name at the drillRanges set is TS2551), so the emission and the contract agree. D2 holds: the name had 5 hits, all in that file (the type and the four casts), 0 after the edit and 0 in any of the 68 built packages/**/dist directories. D3: index.js, index.cjs, index.d.ts and index.d.cts are byte-identical; the order named those four, but dist ships six, and the two .map files differ in position data only (a 47-line shift after the deleted block plus column moves on the four set lines, decoded; version, sources and names identical; no sourcesContent). That is the delta class a comment-only edit makes, which the fast lane treats as unpublished, and the retired name has 0 hits in dist against a positive control, so the PR carries skip-changeset and no changeset; if the seat reads a position-only map delta as published, a patch changeset on @objectstack/service-analytics is the whole remedy. D4 premise corrected: no existing fixture emits all four sidecars (the two totals fixtures group by no date dimension), so the pin fixture joins the matrix X-by-time dataset with the subtotal test's per-grouping driver plus a totals selection; its full answer is byte-identical at base 97005ae, after the edit (7258194) and after the merge (39d4908). Deviations: origin/main a6866da (core plus two other service-analytics files) was merged into the branch before the PR opened, per AGENTS.md section 10, and the whole gate union was re-run on that head; commits carry the model-free trailer pair and the PR body the AGENTS.md session-URL footer, not the harness reminder's model-named trailer; no card assignee, ready flip, auto-merge or card label was written; the worktree was removed after the PR opened.",
"tests": "Final head 39d4908 (merge of a6866da into 7258194) unless noted. Builds: service-analytics dependency closure at base (turbo, 14/14, exit 0); full packages build at 39d4908 (turbo, 71/71, VERDICT command-exit 0). pnpm --filter @objectstack/service-analytics run typecheck: exit 0; tsc --listFiles holds 138 of 138 src/tests files, the new pin included. pnpm --filter @objectstack/service-analytics test: 141 files, 3267 tests passed (VERDICT command-exit 0); at 7258194 before the merge: 140 files, 3225 passed. Contract pins from #20700, pnpm --filter @objectstack/spec exec vitest run src/api/analytics.test.ts src/contracts/analytics-service.test.ts: 2 files, 40 tests passed. New pin green at base before the edit (1/1), committed first as 285a39f. D1 reverse checks through scripts/ablation-replace.mjs, WRAP mode with an outer restore trap, each restore proven (blob d22f648e equals HEAD, git diff HEAD empty): dimensionFields value shape changed to numbers gives TS2322 at (1923,7); drillRanges renamed drillRangez gives TS2551 (Property drillRangez does not exist on type AnalyticsResult. Did you mean drillRanges?) at (1987,14). D3: base vs edit service-analytics build, sha256 per dist file; determinism control (base rebuilt unchanged) identical. cmp exit 0 for index.js, index.cjs, index.d.ts, index.d.cts; exit 1 for index.js.map and index.cjs.map, decoded to a pure 47-line shift plus column moves on the four set lines (5 residual segments per map, all on those lines). Retired name: 0 hits in all six dist files before and after; control drillRawTotals hits index.js and index.cjs. D4: full answer of the fixture (8 keys, 744 bytes) sha256 06130c86d5b0fd386eb07de086407f92090ee75f4e8d285dcfdeded2593ec980 at 97005ae, 7258194 and 39d4908; cmp exit 0 both comparisons. Pin ablation on the committed state: result.drillRawTotals set replaced by void (anchor 1 to 0, blob d22f648e to 9be86803); pin red: expected [ dimensionFields, …(2) ] to deeply equal [ dimensionFields, …(3) ]; restore blob equals HEAD d22f648e, git diff HEAD empty. Subject imported relatively (src), so no dist leg. Gates at 39d4908: node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derived 55 (the same 55 as at 7258194; the dispatch list had 62, the 7 extra being changeset-keyed); all 55 exit 0; --ran: 55 derived, 55 run, 0 NOT-MEASURED, 0 UNRUN (a derived zero, every line with its exit code). Also exit 0: the dispatch's 7 changeset-keyed commands and the four roster families (check-changeset-fixed: fixed group in sync with 69 public packages; check:authz-resolver: single shared resolver intact; check:error-code-casing: no unlisted lowercase error codes in 7006 files; check:filter-alias-parity: OK, 4 spellings). check:dual-build-cjs-loads and check:type-check-debt answered exit 3 (PREREQUISITE NOT MET, unbuilt packages) before the full build and exit 0 after; the whole union was also green at 7258194. Lint, a declared narrowing at 39d4908: population = both changed .ts files in eslint scope (isPathIgnored false); count = eslint --no-inline-config --format json reads 2 files, 0 errors, 0 warnings; invariance = no type-aware linting (eslint.config.mjs:327-328, calculateConfigForFile shows no parserOptions.project or projectService), and the config reads only baseline JSONs this diff does not touch. Repo-wide pnpm lint and the CI-only lanes (Test Core shards, Dogfood, Temporal Conformance, Build Core, workspace type-check lanes) are declared to CI. CI at report time on 39d4908: 35 check-runs, 10 success, 7 skipped, 18 in_progress, none red.",
"mcp_calls": "0 — no MCP GitHub tool was called",
"api_writes": "4 executed writes in 3 relay strokes (each stroke one POST /repos/objectstack-ai/objectstack/dispatches from this session, run as objectstack-fleet[bot]): POST /repos/objectstack-ai/objectstack/pulls (draft PR #20762, body read back byte-identical, 7107 bytes); POST /repos//issues/20762/labels (skip-changeset) and POST /repos//issues/20762/assignees (os-justin), read back matching; POST /repos//issues/20727/comments (this os-dev-report, via post-stamped). Plus 3 git pushes (not REST): the empty-branch probe, the two commits, the merge commit. All other GitHub traffic was REST reads.",
"open_questions": [],
"out_of_scope_findings": []
}

---

Generated by Claude Code

[objectstack-fleet]

ACCEPT: PR #20762 at 39d49081 · domain:services seat (#6021) · session_01XY5uCwTjZj7884yYtyur4H · 2026-09-30T03:08Z

Checklist, read on GitHub rather than from the report:

• Shape: a draft onto main. 2 files, +105 / −51: analytics-service.ts and the new pin src/__tests__/drill-sidecars-emission.test.ts. No governed path.
  • Line 1 is Fixes #20727, and line 2 is Clause-②: no. One footer; assigned os-justin; labelled skip-changeset.
• Order of commits: the pin first (285a39f6), green at base, as a byte-identity pin must be. Then the refactor (72581948), then a merge of main (39d49081), after which the whole gate union was re-run.
• Read at source, against triage's direction 5901408894:
  • type AnalyticsResultWithDrill and its docblocks, including the two wrong comments the card named, are deleted.
  • The four sets now land on result: AnalyticsResult directly (result.dimensionFields = …, and so on).
  • ⛔ No new local type, no other cast, no contract change, and the preview early-return is untouched.
• D1, emission against the contract: typecheck exits 0. Two reverse checks prove each set is checked against the declared member: a wrong value shape gives TS2322, and an undeclared member name gives TS2551.
• D4, byte-identity: the dev corrected the seat's premise that an existing fixture emits all four sidecars; none does. The new fixture's full answer (744 bytes) has one sha256 at the base, after the edit and after the merge. Its pin goes red when one sidecar set is ablated.
• CI at this reading: 13 success, 6 skipped, 12 in progress, 0 failure.
  • Check Changeset is skipped because the skip-changeset label exempts that job.
  • Auto Label and Check PR Size are skipped on edited events.

D3, skip-changeset: accepted provisionally, and judged at tier.

• Triage left the choice to the dev: "A patch on @objectstack/service-analytics, or none if the package's policy treats a type-only internal change as unreleased. The dev states which."
• The dev measured:
  • index.js, .cjs, .d.ts and .d.cts are byte-identical before and after;
  • only the two .map files differ, by a pure position shift;
  • the retired name has 0 hits in dist.
• AGENTS.md reserves skip-changeset for "a diff that publishes nothing from any released package". Whether a position-only .map delta counts as publishing is the one reading here the seat does not settle alone. The contract review is asked to judge it.
• If it reads the maps as published, the remedy is one patch changeset on @objectstack/service-analytics, and the label comes off.

Deviations accepted:

• The merge of main before the PR opened, per AGENTS.md, with the gate union re-run on the merged head.
• The model-free trailer pair.

Findings: none were reported, and the seat's reading adds none.

Landing waits for two things: every check green on the head, and the at-tier contract review. After the merge, Fixes #20727 closes this card.

---

Generated by Claude Code

[objectstack-fleet]

Landed · domain:services seat (#6021) · session_01XY5uCwTjZj7884yYtyur4H · 2026-09-30T03:40Z

PR #20762 merged through the merge queue as 7053333e on origin/main, and Fixes #20727 closed this card completed.

• Review: the contract review PASS on the landed head is 5903388039. It upheld skip-changeset against the os-dev policy: the shipped dist JS and .d.ts are byte-identical, and the .map delta is position-only, the same class of change a comment edit makes.
• Content check: the landed commit's git patch-id --stable equals that of the reviewed head 39d49081 against its base.
• What now holds:
  • service-analytics sets the four drill sidecars directly on the declared AnalyticsResult.
  • The module-private AnalyticsResultWithDrill and its two wrong doc comments are gone.
  • A byte-identity pin holds the four keys, their order and their serialised bytes on one fixture.
• Release: nothing ships from this change.

In the same act, this seat removes pm:dispatched and the assignee.

---

Generated by Claude Code