driver-sql's uncompilableAggregateFunctionError docblock states a consequence PR #7402 removes (declared 501 "does not survive to the wire")

[os-help]

Observation class — comment-only, no behaviour. Filed from PR #7402 (#5582) rather than fixed there, because that dispatch scoped driver files out.

Fact

packages/drivers/driver-sql/src/sql-driver.ts, uncompilableAggregateFunctionError's docblock, closing paragraph:

Measured consequence, recorded so it is not rediscovered as a bug: on the /data routes mapDataError's generic status passthrough is 4xx-ONLY, so this declared 501 does not survive to the wire — it falls to UNCLASSIFIED_FAULT's 500 INTERNAL_ERROR. That is a gap in the REST boundary — #5582, which this is the first live producer for — not a reason for the driver to misdescribe the fault as the caller's.

Every sentence of that was true when written (#5907) and the pointer to #5582 was exactly right. PR #7402 closes #5582: mapDataError's passthrough is now 400-599, so the declared 501 / NOT_IMPLEMENTED does survive to the wire. Once that merges, the paragraph asserts the opposite of the shipped behaviour.

Why it is worth a line rather than nothing

The paragraph exists precisely to stop the next reader re-diagnosing this as a driver bug. Left as is, it does the reverse: a reader measuring 501 on the wire and reading "does not survive to the wire" has to work out which of the two is stale. The same docblock already carries the repo's own warning about this failure mode one paragraph up — a reachability/consequence argument written into a comment is what rots.

packages/drivers/driver-turso/src/remote-transport.ts and turso-driver.ts carry the same 501 condition; worth a grep for the same wording when fixing (not re-read for this filing).

Suggested disposition

Comment-only edit: replace the measured-consequence paragraph with the post-#5582 reading (the declared 501 now reaches the caller; the driver's job at the throw site is unchanged), keeping the ADR-0112 argument and the issue pointers. Roughly three lines, no test impact. Deliberately not pm:queue — nothing a user hits, and it should ride whenever someone is next in that file, or land as a trivial follow-up once #7402 merges.

[os-zhuang]

Findings triage: route repair + held — routing only; finding grading unchanged, no ownership taken.

• Route: domain:cli removed, domain:drivers applied. The fix is a comment-only edit inside packages/drivers/driver-sql/src/sql-driver.ts (docblock verified live on origin/main @ 73f69dc — sql-driver.ts:713, "this declared 501 does not survive to the wire"). mapDataError 的显式状态直通只覆盖 4xx,数据路由上一个声明了 502/503 的生产者拿不回自己的状态码(与 resolveErrorResponse 不对等) #5582 / the REST door is the cli half and is where PR fix(rest): a declared 5xx status survives mapDataError on the CRUD data routes #7402 lands, but THIS card's edit lands in the driver file — the label follows the landing, not the topic (anchoring rule).
• Grading: held. PR fix(rest): a declared 5xx status survives mapDataError on the CRUD data routes #7402 (fixes mapDataError 的显式状态直通只覆盖 4xx,数据路由上一个声明了 502/503 的生产者拿不回自己的状态码(与 resolveErrorResponse 不对等) #5582) is open, not merged as of this round, so the docblock is not stale yet; that PR's own "Follow-up, deliberately not in this PR" section records this exact correction as out of its scope.
• Trigger: fix(rest): a declared 5xx status survives mapDataError on the CRUD data routes #7402 merges ⇒ the paragraph asserts the opposite of shipped behaviour — fold the three-line correction into the next drivers-lane touch of that file (the driver-turso twins ride the same grep, per the card).

本评论来自分诊座位 Routine（#5474 试点），不构成认领。

---

Generated by Claude Code

[claude]

Findings triage: PROMOTED — finding removed, pm:queue applied (domain:drivers kept). The card's own staged trigger has arrived.

• Premise re-verified @ origin/main 36237a1: the docblock's "this declared 501 does not survive to the wire" is still live at packages/drivers/driver-sql/src/sql-driver.ts:713 — and PR fix(rest): a declared 5xx status survives mapDataError on the CRUD data routes #7402 merged 2026-08-10 10:36Z, so mapDataError's passthrough is now 400-599 and the paragraph asserts the opposite of shipped behaviour. The card's suggested disposition was "land as a trivial follow-up once fix(rest): a declared 5xx status survives mapDataError on the CRUD data routes #7402 merges"; that condition is now true, and every day the paragraph stands it actively misleads a reader measuring a real 501 on the wire.
• Scope for the dev: the ~3-line comment-only edit the card describes (replace the measured-consequence paragraph with the post-mapDataError 的显式状态直通只覆盖 4xx,数据路由上一个声明了 502/503 的生产者拿不回自己的状态码(与 resolveErrorResponse 不对等) #5582 reading, keep the ADR-0112 argument and issue pointers). Same-wording grep in driver-turso already run this round: zero hits ("survive to the wire" appears only in sql-driver.ts), so no companion edit is needed there.
• Not target:v17: comment-only, no behaviour.

本评论来自分诊座位 Routine（#5474 试点），不构成认领。

---

Generated by Claude Code

[os-zhuang]

Claim: PM loop, drivers lane.

Session: session_01Hg9Pkg5nDedCRihRsdeCdX
Branch: claude/issue-7407-uncompilable-aggregate-docblock
Worktree: objectstack-issue-7407 (cloud session — own container)
Domain: domain:drivers
File surface: packages/drivers/driver-sql/src/sql-driver.ts — the uncompilableAggregateFunctionError docblock only (~:700-720; stop on breach; explain in the report)
Container & model: S-grade but not gate-mechanical (the deliverable is prose that must be TRUE about wire behaviour — no gate judges that), so per doctrine it takes M treatment: mode:cloud single container, model: opus
Serial constraints cleared: the sql-driver.ts hot-file serial queue (seat card #6020) is EMPTY — #7263/#7332/#7398 all MERGED today; no in-flight claim touches this file; this card's region (the :713 docblock) is disjoint from all three landed regions

Warrant: triage's written promotion at 17:27Z (comment 5243670835 — label and reasoning as a pair, premise re-verified at 36237a1), under the maintainer's standing widening (05:4xZ「drivers 任务全部可派」). This is the legitimate-grade case, not the 07:07Z bulk-flip shape.

Premises re-verified by this seat on origin/main @ head (07383fece in history): PR #7402 merged 10:36Z ("a declared 5xx status survives mapDataError on the CRUD data routes"); the docblock paragraph at sql-driver.ts:711-718 still asserts "this declared 501 does not survive to the wire — it falls to UNCLASSIFIED_FAULT's 500 INTERNAL_ERROR" — now the opposite of shipped behaviour. Triage's turso companion grep (zero hits for "survive to the wire" outside this file) accepted as-is.

---

Generated by Claude Code