Testing and Visual Baselines

Testing and Visual Baselines

RFMC uses multiple test layers to protect behavior, visual stability, and release readiness.

Test Layers

Layer	Command	What It Covers
TypeScript	npm run typecheck:all	Compilation across shared/frontend/server
Unit tests	npm test -- --run	FMC state, page renderers, validation, training, display models
E2E smoke	npm run test:e2e:ci	Desktop Chromium: app loads, CDU renders, key flows
Visual baselines	npm run test:e2e:visual	Snapshot diffs for desktop, tablet, 3456x2234, Retina
High-res cockpit	Playwright project: desktop-3456x2234	Full-deck cockpit layout at physical pixel target
Visual fidelity gate	npm run measure:visual	Reference manifest, measurement profiles, baseline directories

Visual Baseline Workflow

npm run test:e2e:visual
npm run test:visual -- --project=desktop-chromium
npx playwright test e2e/visual/cockpit-highres.spec.ts --project=desktop-3456x2234
npx playwright test e2e/visual/cockpit-highres.spec.ts --project=retina-1728x1117-dsf2
npm run capture:baseline
npm run measure:visual

Visual Fidelity Classifications

npm run measure:visual generates docs/VISUAL_FIDELITY_REPORT.md that classifies each surface:

• snapshot-protected: Playwright visual baseline exists, prevents regressions
• measured: Geometry/color verified against an approved reference crop
• pilot-reviewed: Layout reviewed by a procedure-familiar user
• live-validated: Round-trip verified with real MSFS + PMDG hardware

Hardware accuracy remains "not measured" until rights-cleared hardware reference crops are approved.

What Visual Tests Do Not Prove

• Visual snapshots prove render stability, not hardware pixel accuracy.
• Hardware fidelity is not claimed until rights-cleared references are measured.
• npm run measure:visual does not imply measured real-hardware accuracy unless the report explicitly marks a surface as measured against approved references.

See docs/TESTING.md and docs/STATUS.md for current test counts and evidence.