docs(qa/dogfood): retire stale "older client prepends code" rationale

[os-trump]

Fixes #11589

What

packages/qa/dogfood/test/hook-error-format.dogfood.test.ts carried the retired
"older bundled @objectstack/client would prepend code as [ObjectStack] CODE: …"
rationale at two comment sites — header (line 12) and the assertion site (line 118,
now spans 118-119 after the edit). #10345 measured both halves of that rationale
false and retired the same sentence from packages/rest/src/error-response.ts:

• the shipping client keeps .message to the server's human-readable message by
  explicit rule, attaching the code programmatically as error.code;
• "older bundled client, current server" is not a supported pairing — Decide when the client's three-location code probe retires — ADR-0112 D5 says "after batch 3", the code now says "legacy-server fallback, NOT debt" #4007 retired
  the compat read, since SDK and server ship on one release train.

This PR is a comment-only reword in that one file, replacing the stale rationale
with the real one: expect(body.code).toBeUndefined() is correct because the hook
throws new Error(BUSINESS_MSG) with no code declared — ADR-0112 invents
nothing for a producer that declared none. No executable line changed; both
assertions are byte-identical to before.

Verified line numbers (re-derived against this tree, no drift from the issue)

• Header site: line 12
• Assertion-comment site: line 118 (now spans 118-119 after rewording)

Scope fence honored

packages/rest/src/** was not touched — its four sibling comments carrying the same
sentence were already rewritten in #10345's PR, and a grep for the retired phrasing
(older bundled, for old clients, CODE: …) across the repo found no further live
site: the remaining packages/rest hits (rest.test.ts:2695,
rest-sandbox-declared-status.test.ts:30, rest-hook-script-fault-envelope.test.ts:331)
already narrate the retirement in the past tense, not as live rationale.

Tests

Ran at 4663677c (this branch's only commit, current head):

• pnpm --filter '@objectstack/dogfood^...' build — full dependency closure, clean,
  no errors (via the shared verify lock).

• pnpm --filter @objectstack/dogfood exec vitest run --maxWorkers=2 test/hook-error-format.dogfood.test.ts
  — 4 passed (4), via the shared verify lock (VERDICT command-exit 0). No
  red-before state exists for a comment-only change, as expected; this is the
  file's suite green after the edit, not evidence the edit itself is correct — the
  anchored greps below are that evidence.

• Anchored grep proving the retired sentence is gone from both sites and the new
  reason is present:

  $ grep -n "older bundled\|for old clients\|CODE: …" packages/qa/dogfood/test/hook-error-format.dogfood.test.ts
  (no matches)
  
  $ grep -n "declared no code\|threw no code\|declared none" packages/qa/dogfood/test/hook-error-format.dogfood.test.ts
  12:// and not a `code` field: `throw new Error(...)` declares no code, and
  118:    // No sandbox debug wrapper, no branding, and no `code` — the hook threw
  119:    // no code, and ADR-0112 invents nothing for a producer that declared none.
  

Gate family (re-derived with node scripts/pm/dispatch-gates.mjs against this diff,
not reconciled to the dispatch prompt's list) — all green:

• pnpm --filter @objectstack/spec run check:empty-state
• pnpm --filter @objectstack/spec run check:liveness
• pnpm --filter @objectstack/spec run check:strictness-ledger
• pnpm --filter @objectstack/spec run check:variant-docs
• pnpm check:published-files
• pnpm check:slot-lookup
• pnpm check:test-source-alias
• pnpm check:type-source-resolution
• node scripts/check-plugin-teardown-shape.mjs
• node scripts/docs-audit/check-affected-docs.mjs
• node scripts/docs-audit/check-drift-comment.mjs
• Convention-triggered (this diff edits a *.test.ts file): pnpm check:query-options-erasure,
  pnpm check:engine-double-contract, pnpm check:cross-package-test-inputs,
  pnpm check:where-matcher, pnpm check:type-check-coverage — all green.
• pnpm check:type-check-debt (--re-measure) refuses locally: it needs the FULL
  78-package workspace closure built (turbo run build --filter='./packages/*' --filter='./packages/*/*'), not just @objectstack/dogfood's own dependency
  closure — it errored naming @objectstack/service-knowledge unbuilt, a package
  outside dogfood's graph. Deferred as disproportionate to a comment-only diff:
  the DEBT/TEST_DEBT counters are raw tsc diagnostic counts, and this diff
  changes zero non-comment bytes in an already-tracked (not newly added) test
  file, so tsc's AST — and therefore every ledger count this gate re-measures —
  is unaffected by construction. CI's lint.yml builds the full closure first and
  will re-measure for real.
• pnpm lint (full-repo, per this lane's standing instruction) — 5429 files
  scanned, 0 errors, 0 warnings (VERDICT command-exit 0, held the shared
  verify lock 154s after waiting 435s behind three other agents' jobs). A direct
  single-file eslint run on the target file alone was also 0/0, confirmed before
  the full run landed.

skip-changeset

Comment-only test-file change; no changeset, per triage. skip-changeset label
applied via the additive labels endpoint and read back: size/xs, tests,
skip-changeset.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

Nothing in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from 0 changed package(s)), so this run has no opinion about the docs.

What this run could not see

• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.

Coarse fallback — 0 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json e7480a80676dc70c1a21555d6a11a86c8f518209 → packageMentionDocs.