fix(metadata-protocol): say what the predicate covers, not what the repo runs, in driver-text-disclosure test

[hotlong]

Fixes#8822

What

packages/metadata-protocol/src/protocol.driver-text-disclosure.test.ts claimed, in two
places named on the card plus a third found while in the file, that this repo runs only
two SQL dialects. That is false — MySQL is measurably run here (driver-sql branches on
mysql/mysql2, CI stands up a live mysql:8.0 for the required
Temporal Conformance (live PG + MySQL) check, and live MySQL 8.0.46 measurements landed
merged driver fixes in #8621 / #8622). The file's substance was already correct — this is
a wording-only correction, landing the same shape #8739 already landed in
packages/types/src/error-leak.ts: say what the shared looksLikeInternalErrorLeak
predicate covers, not what the repo runs.

Three sites corrected, all in this one file:

1. Module note (was ~line 35): "the two dialects this repo runs" to "the two dialects it
   COVERS".
2. Test name (was ~line 276): recognises the two engines this repo runs, and none of the three it does not to recognises the two engines the predicate covers, and none of the three it does not.
3. A third instance found while in the file, in the DIALECTS matrix's doc comment (not
   named on the card, same false claim, same file, folded in per the card's instruction):
   "the list of engines nobody here has run is unbounded" to "the list of dialects this
   predicate does not cover is unbounded".

What this explicitly does NOT do

• No behaviour change and no assertion change every expect(...) is untouched.
• No DIALECTS matrix change. The matrix deliberately includes dialects the predicate does
  NOT recognise, asserts looksLikeInternalErrorLeak returns false for them, and then
  asserts the text is withheld anyway so correctness never depends on having enumerated
  the world's SQL engines. That design is preserved exactly as-is.
• No position on whether MySQL is a supported deployment target. That question is [finding] error-leak.ts asserts "nobody here runs" MySQL — but CI stands up a live mysql:8.0 for a required check, and the claim is load-bearing for security reasoning #8739's
  and stays open; the correction is written to be true regardless of how it resolves.

Tests

At 9ca3831f2 (HEAD):

• pnpm --filter '@objectstack/metadata-protocol^...' build — 0 (full dependency closure
  built before testing).
• pnpm --filter '@objectstack/metadata-protocol' test -- --maxWorkers=2 — 1444/1444
  passed, 98/98 files, including the edited file's 16/16 tests (same count as before,
  wording only).
• pnpm check:cross-package-test-inputs — OK, 9 packages, all declared.
• pnpm check:durability-log-level — OK, 25 seams, all loud/declared.
• node scripts/check-cross-package-test-inputs.mjs — same as above (single script backing
  both).
• pnpm check:query-options-erasure — OK, ratchet holds, none new, baseline unchanged
  against fd6bdf8.
• pnpm check:type-check-coverage — OK, structural coverage unchanged.
• pnpm check:type-check-debt --re-measure (on a full turbo run build of the workspace,
  70/70 tasks) — OK, 33 ledger entries re-measured, 1926 raw tsc errors total, none above
  its recorded ceiling (no regression from this change).

Out of scope

None filed no defects found beyond the wording this card targets.

---

Generated by Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
objectstack	Ignored		Aug 15, 2026 11:30am

[github-actions]

📓 Docs Drift Check

No hand-written docs reference the 0 changed package(s). ✅