abb9f7d443
Full-deployment QA against a copy of the real DB surfaced launch/require
references the server .cts migration didn't update. All are non-source
tooling/edge paths, which is why typecheck/check/tests stayed green:
- e2e/setup/prepare-db.js: require('../../db/database') and
require('../../scripts/seedDemoData') → .cts. This broke `npm run
test:e2e*` entirely (harness couldn't load). [material]
- playwright.config.js + scripts/prod-smoke.sh: webServer/boot command
`node server.js` → `node server.cts`. Also blocked the e2e + prod-smoke
suites. [material]
- setup/firstRun.js → .cts, require('../services/auditService') → .cts,
and server.cts call site → require('./setup/firstRun.cts'). Latent:
server.cts only reaches firstRun when userCount===0, but database.cts
auto-seeds a default admin during getDb() first, so the path isn't hit
in normal boot — fixed defensively so it can't crash if ever reached.
- .env.example: doc reference `node server.js` → server.cts.
Verified: typecheck:server + check:server clean; server test suite
226/226; e2e probe 17/17 (was 0 — suite couldn't boot before); the
production Docker image builds and boots on a copy of the real
production DB (44 bills, 1159 payments) with /api/health 200 and all 64
page endpoints returning 2xx.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
||
|---|---|---|
| .. | ||
| setup | ||
| smoke.spec.js-snapshots | ||
| README.md | ||
| a11y.authed.spec.js | ||
| a11y.spec.js | ||
| api.probe.spec.js | ||
| auth.setup.js | ||
| b-ui.spec.js | ||
| categories.spec.js | ||
| constants.js | ||
| critical-path.spec.js | ||
| smoke.spec.js | ||
README.md
E2E harness (Playwright)
The automated regression net for the QA plan — real clicks in a real browser,
plus visual-regression and axe-a11y that re-run every cycle. See
docs/QA_PLAN.md §4.4 and the B-UI / B14 / B15 batches.
Setup (one-time)
npm install
npx playwright install chromium # add: firefox webkit for cross-browser
Run
npm run test:e2e # seed scratch DB, then run headless
npm run test:e2e:ui # seed scratch DB, then Playwright UI mode (watch/debug)
npm run test:e2e:update # re-baseline visual-regression screenshots
Safe by construction
Each command runs e2e/setup/prepare-db.js first, which builds a fresh scratch
DB (db/e2e.db, git-ignored) with a seeded regular user — it refuses to touch
db/bills.db. playwright.config.js then boots the
app against that scratch DB on dedicated ports (UI 5199, API 3099) with
reuseExistingServer: false, so E2E never collides with or reuses your normal
dev server on 5173/3000. You can keep npm run dev running while E2E runs.
Override defaults via env: E2E_DB_PATH, E2E_USER, E2E_PASS, E2E_UI_PORT,
E2E_API_PORT (see e2e/constants.js).
What's here
| File | Covers |
|---|---|
setup/prepare-db.js |
fresh scratch DB + seeded regular user (run automatically) |
auth.setup.js |
logs in via the UI once, saves .auth/user.json for reuse |
smoke.spec.js |
(logged-out) login renders, empty-submit doesn't crash, routing, login visual baseline |
a11y.spec.js |
axe-core scan of public pages — fails on critical/serious WCAG violations |
critical-path.spec.js |
(logged-in) tracker renders seeded data, quick-pay + undo, all authed pages render error-free, sidebar nav |
Growing it (next steps)
- Numeric reconciliation — assert Summary's paid/unpaid/remaining totals
match the Tracker for the same month (QA_PLAN B5); see the TODO in
critical-path.spec.js. - Create-bill flow — drive BillModal (create → appears on Tracker → delete).
- Primitive state matrix — one spec per
client/components/ui/*(B-UI). - Fault injection —
page.route()to force 401/429/500/timeout per page and assert a recoverable error, never a white screen (B14). - Visual coverage — a
toHaveScreenshotper page × {desktop,mobile}. - Cross-browser — enable the firefox/webkit projects in the config.
Screenshot baselines live in e2e/**/*-snapshots/ and are committed (that's
the point of visual regression). test-results/, playwright-report/, and
e2e/.auth/ are not.