Infra-board — tests¶
Frontend: 17 files · 188 tests. E2E: none yet.
The infra-board is well covered at the unit level — src/pages/infra-board is
the single best-covered area in the frontend at 99.64% statements — and has
no end-to-end coverage at all.
Frontend¶
| Area | Files | Tests |
|---|---|---|
pages/infra-board (data and pure logic) |
6 | 88 |
components/InfraBoardDashboard |
10 | 86 |
pages/InfraBoardDashboard.test.tsx |
1 | 14 |
| File | Tests | Covers |
|---|---|---|
components/InfraBoardDashboard/PhaseDetail.test.tsx |
26 | Phase detail view |
pages/infra-board/infra-board.data.test.ts |
26 | The board's data module |
pages/infra-board/rip-model.test.ts |
20 | The RIP phase model |
pages/infra-board/rail-stats.test.ts |
15 | Rail statistics |
pages/InfraBoardDashboard.test.tsx |
14 | The page container |
components/InfraBoardDashboard/InfraSectionRouter.test.tsx |
14 | Section routing |
pages/infra-board/rip-phases.catalog.test.ts |
11 | The twelve-phase catalogue |
components/InfraBoardDashboard/FaseladderOverview.test.tsx |
9 | Faseladder overview |
components/InfraBoardDashboard/Portfolio.test.tsx |
9 | Portfolio view |
pages/infra-board/modes.config.test.ts |
9 | Mode configuration |
components/InfraBoardDashboard/InfraCommandPalette.test.tsx |
8 | Command palette |
pages/infra-board/rip-phase-counts.test.ts |
7 | Phase counts |
components/InfraBoardDashboard/ProjectDetail.test.tsx |
6 | Project detail |
components/InfraBoardDashboard/MijnDag.test.tsx |
5 | The "Mijn Dag" section |
Coverage: pages/infra-board 99.64 / 97.2 / 100 / 100 —
the highest in the frontend — and components/InfraBoardDashboard
90.96 / 84.54 / 87.87 / 91.68.
The gap between those two rows is the usual one: the pure data and model modules are exhaustively covered, while the components carry the "critical interactions only" scoping used across the frontend.
E2E¶
None. There is no Playwright spec that drives this board.
This is a gap rather than a decision. The PA cockpit work in the 21–22 August
window demonstrated concretely what unit tests cannot see on a board like this
one: a saved write that was a bare return;, a confirm that built an object and
discarded it, a panel hardcoded to empty, a resource fetched once at mount and
never again. Every one of those passed its component tests, because a component
test mocks the very seam that was broken.
The infra-board has the same shape — a section router, a command palette, data panels reading through a provider — so it is exposed to the same class of defect, and nothing currently would catch it.
If a spec is added, the PA cockpit's structure is the model to copy: drive mock mode against the real store with no mocking, which is affordable precisely because the fixtures are deterministic. See PA cockpit → Why mock mode is worth an E2E suite.