chore(i18n): string sweep, shared intl test helper, toast lint guard

[innolope-dev]

Stacked on #2554 — review that first; this PR's diff is against it, and the base retargets to dev once #2554 merges.

Why

Follow-up to #2447. react/jsx-no-literals only inspects JSX children, so hardcoded copy kept reaching users through toasts, placeholders, aria-labels and alt text.

The ticket was right about QRScanner

It named QRScanner as a known offender with "3–4 toasts". It had four, on dev, and three of them had keys that already existed and simply weren't being used. (An earlier pass of mine looked at a divergent branch where they'd been extracted and wrongly concluded the ticket was stale.)

The lint rule

A no-restricted-syntax selector banning string literals as the first argument to toast.error|success|info|warning|loading.

Appended to the existing array, not a new config block — ESLint flat config replaces rule options when a later block names the same rule, so a second no-restricted-syntax entry would have silently disabled the four router.back selectors. That array already sits under files: ['src/**/*.{ts,tsx}'], which is exactly the coverage needed, since every offender lives in src/hooks or src/context — outside the jsx-no-literals globs.

Verified it fires by planting a synthetic violation, not just by observing a green run.

Scoped to toast.*, not throw new Error('…'). Every throw site flagged in the ticket was traced to its catch handler: they all funnel through toFriendlyError to errors.genericSupport, or into an already-localized toast. Those strings are Sentry breadcrumbs and issue-grouping keys — translating them would fragment triage for zero user-facing gain.

Note the rule can't actually block a merge: the eslint job is still continue-on-error and absent from ci-success.needs. Left as-is deliberately; promoting it is a separate team decision.

Shared test helper

src/test-utils/intl.tsx replaces 39 hand-rolled NextIntlClientProvider wrappers (the ticket estimated ~15). Exports the three shapes call sites actually need: IntlWrapper (composable, for suites nesting QueryClientProvider), renderWithIntl, and renderHookWithIntl.

Two suites had omitted timeZone, so next-intl fell back to the runner's system zone and their date formatting was machine-dependent — the shared wrapper pins UTC. renderHookWithIntl exists so useSumsubKycFlow.test.ts (a .ts file, no JSX) can drop its React.createElement workaround and the eslint-disable react/no-children-prop that came with it.

ClientProviders.test.tsx deliberately keeps its own reference — it asserts on provider ordering, so the provider is the subject under test.

Extractions

Four QRScanner toasts, both logout/login toasts, two card-approval toasts, the waitlist email placeholder, two "Peanut Logo" aria-labels (reusing the existing navigation.peanutLogoAlt), two descriptive alt strings. The QRScanner payment-method row keeps its brand names verbatim — proper nouns — and only the EVM row, whose label is prose, moves to a key.

NavHeader gains an optional titleKey: /receipt/[entryId] is an async server component and this app has no server-side next-intl setup (locale is resolved entirely client-side by AppIntlProvider), so useTranslations is unavailable there.

Two structural fixes

Clipboard. The match on "There is no data on the clipboard" was not a locale bug — that string is a hardcoded English literal inside @capacitor/clipboard's Java (ClipboardPlugin.java:53), not an OS message. It was a dependency coupling: a plugin patch bump silently flips the branch and an empty clipboard starts reporting as a failure. readClipboard() now classifies the rejection and returns a reason.

'Enjoy Peanut!' is consolidated behind one isTestTransaction() predicate. It compares backend data, so it's locale-safe, but it was repeated in four places and one of them rendered the raw literal instead of the transaction.enjoyPeanut key that already existed and was already translated. The backend should own this flag — worth a ticket.

Verification

pnpm typecheck clean · 174 suites / 2311 tests pass · eslint 0 errors · prettier --check . clean

Note: peanut-ui's pull_request CI trigger only fires for main/dev/develop bases, so this PR's checks come from the push trigger.

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
peanut-wallet	Ready	Preview, Comment	Jul 28, 2026 4:35pm

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: c589de41-9b30-4ab8-804a-ccd6addcb7bc

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch chore/i18n-sweep

---

_{Comment @coderabbitai help to get the list of available commands.}