// release-only crash
Worked on web, crashed on phones
- found
- A club card crashed the release app. Web and unit tests passed.
- cause
- Hermes lacks some date formatting APIs.
- changed
- Release check runs a real Hermes build. A lint rule flags the APIs.
> riverbound / quality engineering
Riverbound is my book tracker and book club app for iPhone and Android. AI writes much of the code, so the tests are what I trust.
As of November 2026.
01 — ./strategy
Many fast tests, a few slow ones on a device. The database gets its own layer because it decides who sees what.
02 — ./quality-gates
One pipeline runs on every pull request. The other runs on the exact commit that ships.
npm ci from the lockfilerelease, one retryA flow tagged release runs in the release check. No pipeline edits to forget.
Each attempt keeps its own report, so a test that passed on retry still shows as flaky.
03 — ./layers
Excerpts from the real suites.
Estimated page counts, pages past the end, 100% shown too early. Each edge case gets a test named after its rule.
test('percent labels round down, so only the end shows 100%', () => { assert.equal(formatProgressPercent(99.7, 'en'), '99%'); assert.equal(formatProgressPercent(100, 'en'), '100%'); }); test('a page beyond an estimated length never shows more than 100%', () => { assert.equal(getProgressPercent({ page: 420, totalPages: 400 }), 100); }); test('pages must be whole numbers within the limit', () => { for (const bad of ['', 'abc', '12.5', '-3', String(MAX_PAGES + 1)]) { assert.deepEqual(parseProgressInput('page', bad, '400'), { ok: false, error: 'invalidPage' }, bad); } });
Tests find controls by role and name, like a screen reader. A double tap can't submit twice.
// Two taps on a saving button would submit twice; // the button, not each caller, has to stop that. it('ignores presses while loading, and says it is busy', async () => { const handlePress = jest.fn(); await render(<Button label="Save" onPress={handlePress} loading />); const button = screen.getByRole('button', { name: 'Save' }); await fireEvent.press(button); expect(handlePress).not.toHaveBeenCalled(); expect(button).toBeBusy(); }); // The Maestro flows find this button by id, so the prop must reach the control. it('passes its test id through to the pressable', async () => { … });
All migrations run on PGlite, a real Postgres inside Node. The suite then acts as each user and checks what they can and can't do.
// Run a query as a given user, the way the API would. async function as(userId, sql, params = []) { await db.query(`select set_config('request.jwt.claim.sub', $1, false)`, [userId ?? '']); await db.exec(userId ? 'set role authenticated' : 'set role anon'); try { return await db.query(sql, params); } finally { await db.exec('reset role'); } } check('reviews need a verdict or thoughts', !!(await fails(users.bea, `insert into public.reviews (user_id, work_id) values ($1, $2)`, [users.bea, work]))); check('anonymous users cannot call excluded library stats', !!(await fails(null, `select * from public.get_library_stats()`)));
Flows tap by test ID and clean up after failed runs, so they're safe to rerun. Here a reader reports a review and an admin resolves it.
name: Reports - report a review, then resolve it as an admin tags: [release, safety] --- # 1. Admin: start clean (one open report per target). - runFlow: file: subflows/sign-in.yml env: { IDENTIFIER: ${MAESTRO_E2E_ADMIN_IDENTIFIER} } - runFlow: file: subflows/resolve-reader-reports.yml env: { TARGET_KIND: review } # 2. Bad user: make sure there's a review to report. - extendedWaitUntil: visible: { id: book-status-button-.* } timeout: 30000 # 3. Reader reports it. 4. Admin marks it reviewed. …
04 — ./data-quality
Three places move data. Each is tested for accuracy and partial failure.
05 — ./test-data
Real testers use production, so tests never do.
PR checks never touch a hosted database. The release check uses dev only. The import tool refuses production.
The test password is scrubbed from every result file. If any copy remains, nothing is uploaded and the job fails.
Shared subflows for sign-in, deep links and the keyboard. New tests are built from them, not copied.
Flows use test IDs, not labels. Component tests confirm the IDs reach the screen.
06 — ./bugs-caught
Each one changed how Riverbound is tested.
// release-only crash
// environment gap
where. Local allows it.// flaky infrastructure
A book tracker and book club with spoiler-free buddy reads. Coming to iPhone and Android.