test(#565): correct two stale widen specs (widen is one-dimensional by design)

[BorisTyshkevich]

What & why

Closes #565 — and corrects its premise. I opened that issue calling this a user-visible regression ("widen no longer grows a grid tile's height"). That was wrong. Widen behaves exactly as specified; the two specs were stale.

Measured live in Chromium against the real fixture:

initial   h=208px w= 867px   t-sales={"grid":{"span":8}}
press 1   h=208px w=1305px   t-sales={"grid":{"span":12,"height":2}}
press 2   h=208px w= 101px   t-sales={"grid":{"span":1,"height":2}}

Span doubles 8 → 12 and wraps to 1 at the maximum; the height never moves. src/dashboard/application/panel-widen.ts says so three times:

• the module reads "only, and never anything two-dimensional: the corner drag (Add a rowless Grafana-style tile-grid Dashboard layout #291) stays the free-form resize";
• for grid: "doubles span on each press and wraps to 1 at the maximum. Height is always resent unchanged";
• nextPanelPlacement: "Height is preserved."

t-sales authors no height, so it resolves to DEFAULT_GRID_HEIGHT_UNITS (2) → 32 + 88*2 = 208px, before and after. The failing assertion demanded > 208 and got exactly 208.

What was actually stale

Both defects come from the grafana-grid@2 style-keyed placement shape, not from the widen step:

	was	now
rendered height	toBeGreaterThan(before.height) — demanded the one thing widen must not do	toBe(before.height) — the real invariant; fails if widen ever becomes two-dimensional
persisted placement	{ span: 12, height: 4 } — flat v1 shape, and a height that never existed	{ grid: { span: 12, height: 2 } }
narrow-tile poll	items['t-sales']?.span — v1 path, resolves to undefined, so ?? null turned a shape mismatch into an opaque 5s timeout	items['t-sales']?.grid?.span

The narrow-tile spec's expected value (2) was always correct and is unchanged: the two presses wrap span to 1, and the ⋯ row it then clicks steps it to 2. I initially misread that as wrong too — the menu click is what makes 2 right.

Note the direction of the height fix: asserting equality is strictly stronger here than the old toBeGreaterThan, because it now pins the documented one-dimensional contract rather than leaving it unconstrained.

Quarantine removed

The test.fail() markers I added in #566 were based on the misdiagnosis. They're gone — these specs pass on their own merits now, no expected-failure bookkeeping.

Verification

npx playwright test tests/e2e/tile-open-workbench.spec.js → chromium 21 passed, webkit 20 passed, quarantine removed. Nothing under src/ changes; this is a test-only correction.

Worth recording why this rotted unnoticed: the placement shape changed under these specs and the browser suite never ran on a pull request until #564. This is the first thing that gate caught.

Checklist

• npm test passes (the per-file coverage gate is non-negotiable) — unaffected, no src/ or unit change
• Tests added/updated in the same change as the code
• npm run build succeeds (single-file dist/sql.html) — unaffected
• Layers kept honest: pure logic in src/core/, network in src/net/ (injected fetch), DOM in src/ui/
• No new runtime dependency
• README / CHANGELOG.md ([Unreleased]) updated if behavior or the deployed surface changed — n/a, no behaviour change
• Reconciled affected tracked work (roadmap Roadmap to 1.0.0 #68, the issue body, ADR/CHANGELOG) if this change reshaped it — Two tile-widen e2e specs are stale against the grafana-grid@2 placement shape #565 retitled and its premise corrected

🤖 Generated with Claude Code

https://claude.ai/code/session_01GiubaoqEuBzAyo5C4P8Vqr