governance: ReScript frontend trips package anti-pattern policy on main

[hyperpolymath]

What

governance / Language / package anti-pattern policy job is failing on main (verified 2026-06-01):

```
❌ ReScript files detected - use AffineScript instead
'# hypatia:ignore cicd_rules/banned_language_file' pragma if intentional)
```

Source: estate language policy 2026-05-25 banned ReScript estate-wide in favour of AffineScript.

Impact

• All PRs land in BLOCKED state requiring `--admin` to merge. Today (2026-06-01) PRs test(quandle): KT-2 ext exact-value fixtures (Alexander + Conway) #41 (KT-2 ext exact-value fixtures) + feat(db-3): Phase A — 3 new indexes + strategy doc #42 (DB-3 Phase A indexes) were admin-merged for this reason.
• Auto-merge effectively non-functional.

Where the ReScript is

`frontend/src/` per `.claude/CLAUDE.md` — quandledb's UI shell.

Options

A. Migrate frontend/src → AffineScript (clean, estate-aligned; tracked separately under standards#252 estate-wide TS/ReScript→AffineScript campaign).

B. `.hypatia-ignore` pragma for `frontend/src/.res`** (escape hatch; keeps governance green until A lands).

C. Hand-port a minimal AffineScript shell that boots the same JSON API (smaller scope than full migration; aligned with estate STEP 8 first Mode-A batch pattern — see panll#72 / affinescript#505).

Out-of-scope today

The current krl/quandledb/tangle audit explicitly excludes estate-wide ReScript migration. This issue records the debt so that the next maintenance pass can take it.

Acceptance

Either `.hypatia-ignore` covers `frontend/`, or the ReScript files are gone and replaced by AffineScript equivalents, and the governance check is green on `main` for two consecutive pushes.

See also

• standards#252 (estate ReScript→AffineScript umbrella)
• panll#72 (Mode-A pattern: pure modules port first; user-shell defers)
• affinescript#508 (AS `pub` gap in user-modules ESM codegen — blocks Mode-B for now)

[hyperpolymath]

Status update — AffineScript stdlib codegen has landed (2026-06-01) 📌

Surfacing the AS-side baseline shift for the parking context here:

• affinescript#522 (merged 2026-06-01) — Sqlite.affine codegen foundation; first stdlib extern with full Deno-ESM lowering + tests end-to-end.
• affinescript#504 (merged 2026-06-01) — closes 6 of 10 STEP-3 Deno-ESM gaps; 6 new stdlib externs (stat helpers, byte ops, importMetaUrl) + wildcard-let codegen fix that was breaking strict-mode ESM.
• affinescript#463 (merged 2026-05-30) — non-ASCII string literals emit correct \uXXXX/\u{X} Unicode escapes in Deno-ESM (unblocks any source with emoji/CJK/accented chars).
• affinescript#421 — stdlib/json v0.3 rewired to hpm-json-rsr Zig FFI (11 externs).

What now works that didn't before: stdlib externs (Sqlite, Deno, json) lower to runnable Deno-ESM via lib/codegen_deno.ml; the lowering pattern (declare in stdlib/*.affine → wire in deno_builtins table + prelude → smoke-test in tests/codegen-deno/*) is repeatable and tested.

What is NOT yet resolved:

• User-module codegen — only stdlib has codegen so far; user-defined pub extern fn in app code is not yet wired.
• Stdlib primitive gaps still open: affinescript#160 (Http.fetch), #161 (portable Json — partially closed by stdlib/json v0.3), #162 (Dict/Map).
• Migration parser affinescript#57; module-qualified type paths affinescript#228; effect-row v1 affinescript#59.

Action for this ticket: No new sweep being kicked off here from this status post — surfacing so anyone landing here can see the current AS baseline. Re-evaluate when user-module codegen + the named stdlib primitives land.

🤖 Filed as part of a cross-estate AS-status sweep on 2026-06-01.

[hyperpolymath]

Re-triage: the cited Mode-B blocker affinescript#508 (pub gap in user-module ESM codegen) is CLOSED (verified 2026-06-11). Mode-B may be viable now; see hyperpolymath/affinescript#563 for current upstream state.