[bokelley]

Naming slice for #3504 (T4 in the approved DX plan, section 3.1): the naming freeze before beta.3. It stacks on #3545 (vectors) → #3544 → #3543 → #3542 → #3534 and the rest of the #3504 stack. Review only feat(reporting): freeze reporting-subscription names before beta.3. It targets main because CI runs only for PRs into main.

Old → new

Every old name stays as a @deprecated alias for one beta.

Old	New	Where
PrincipalStoreScope, ReportingLedgerSubscriptionPrincipalV1	PrincipalRef { tenantId, principalId }	root, /server, /reporting/advanced, /reporting/ledger; used by every reporting-subscription signature (engine, engine store, service hooks, ledger admission, principal runtime)
expectedSubscriptionPeriods	owedSubscriptionPeriodsV1	/reporting/advanced (also fixes the module's one V1 outlier)
groupExpectedSubscriptionPeriodsV1	groupOwedSubscriptionPeriodsV1	/reporting/advanced
ExpectedSubscriptionPeriodV1 / …ObligationV1 / …PeriodsInputV1	OwedSubscriptionPeriodV1 / OwedSubscriptionObligationV1 / OwedSubscriptionPeriodsInputV1	/reporting/advanced
subscriptionExpectedReportingPeriodsV1	toExpectedReportingPeriodsV1 (keeps "expected": it feeds the consumer's ExpectedReportingPeriod)	/reporting/advanced
lifecycle.recordReportingSubscriptionCoverage	recordSubscriptionReadback (it writes readback lists, not subscription_coverage)	principal runtime
engine.listPrincipals	engine.listSubscriptionPrincipals (one name with the service hook it feeds)	engine
createReliableReportingService({ reportingSubscriptions }), createPostgresReliableReportingProductionService({ reportingSubscriptions })	{ subscriptions }. Passing both is a PlatformConfigError	service, @adcp/reporting-storage/service

After this PR, every function and type on /reporting/advanced carries V1, apart from the deprecated alias, the error classes, a constant and PrincipalRef, which is the facade identity type.

Deviations from the plan

• Only one of the three option keys collapses here. The service and production-service blocks become subscriptions. createPrincipalLifecycle({ reportingSubscriptions }) doesn't fold into that block yet. That would make the service build and install the lifecycle's subscription section, which is the T5 managed path ("the server installs the lifecycle section and capabilities"). Today the lifecycle must exist before the engine and the service can be built. So the lifecycle key stays as rung-3 wiring, and accountConfigurations is still set on both. T5 removes that duplication.
• The block keeps today's fields. subscriptions still carries engine, resolvePrincipal, listSubscriptionPrincipals and resolveSubscriptionConfiguration, with the type ReliableReportingSubscriptionsOptionsV1 from /reporting/advanced. The rung-1 shape (maxAccounts, backfillOnRegrant, override hooks) and its suffix-free type are T5. Python calls that type SubscriptionOptions, so T5 should use the same name.
• The Owed* types are renamed too, not only the three functions the plan lists, so the vocabulary stays consistent. The engine store port keeps listPrincipals, as section 2.4 specifies.

Names checked against Python

The Python SDK hasn't done its naming pass yet. These names follow the section 3.1 table: Python owed_subscription_periods (no suffix) and PrincipalRef(tenant_id, principal_id).

Checks

npm run typecheck, lint:boundaries, check:package-size, Prettier, and 1,277 tests across reporting, principal, server-decisioning and packages/reporting-storage (PostgreSQL suites run in CI). Tests cover each deprecated alias and the both-keys refusal.

Refs #3504

🤖 Generated with Claude Code

[aao-secretariat]

Ladon verdict: Approve

Approve — no changes to review.

This is a subsequent pass. The head commit (147c5fa) and the prior-review commit (e27f6ed) share an identical tree (334891da...); the stack was rebased forward 23 commits but the PR content is byte-for-byte unchanged. The diff delta is empty (0 files changed), and the prior pass was clean with no Critical/High/Medium findings. Nothing new to re-surface.

Decision table walkthrough:

• Row 1 (critical/high): no findings — skip.
• Row 2 (gated_paths + not approved): gated_paths is false — skip.
• Row 3 (deleted high-risk): high_risk false — skip.
• Rows 4–5 (medium + category/high-risk): no findings — skip.
• Row 6 (sticky escalate): prior decision was approve, not escalate — skip.
• Row 7 (no-auto-approve team): no team match — skip.
• Row 8 (≥3 medium): zero medium findings — skip.
• Row 9: approve.

No author-team gate and no gated-path gate apply. Clean carry-forward of the prior clean approval.

[aao-secretariat]

Ladon verdict: Approve

Approve — subsequent pass, clean diff with no findings.

The delta since the prior approve is a single documentation comment in scripts/check-package-size.mjs recording #3504's measured packed size (26,370,041 bytes) against the unchanged MAX_PACKED_TARBALL_BYTES cap of 26,420,000 (~40 KB headroom). The budget number is correctly left untouched — no behavior change, no changeset impact.

No critical/high/medium findings. Checked: no author no-auto-approve team match, high_risk false, gated_paths false, no deletions. Although review_decision is REVIEW_REQUIRED, gated_paths is false so row 2 does not apply. No escalation trigger fires; falls through to row 9.

[aao-secretariat]

Ladon verdict: Approve

Approve — clean pass on the 6-file delta (3 source, 3 test), no blocking or medium findings.

Checked:

• calendar.ts:272 — refuses sub-daily billing_cycle alignment per adcp#8622 (fail-closed, correct).
• engine.ts:808-815 — refreshes open action_required exclusion on host pending-action change (JCS-canonical compare, from preserved, mutates live target.exclusions via open, sets changed); latestPeriodEnd uses a max-scan instead of withheld.at(-1) since frozen periods aren't end-ordered — non-empty guaranteed by the withheld.length guard. Correct.
• reporting-subscriptions.ts:357 — narrows one-shot reset from !== true to === false, fixing a durable-state bug where an omitted flag wiped a committed-but-failed replace retry; covered by new regression test.

Witness-not-translator, version-pathway, official-transport, and changeset coverage all satisfied (changeset present in the stack). No new I/O. No no-auto-approve team match; gated_paths false; high_risk false. Prior decision was also approve with no findings. None of decision rows 1–8 fire → row 9 approve.