Say why the unconfigured-node assertion cannot pass vacuously

[Babissimo]

The three minor notes the review carried over from #338. All were non-blocking, and production was a release behind at the time, so they were left rather than holding that deploy.

The one that matters

not.toHaveProperty("detection_area") was not gated on the response status. An absence assertion holds against an error body as readily as against a correct one, so a failed request would have taken the test green for the wrong reason — the exact shape of vacuous pass that is worth guarding explicitly.

The metrics and trust assertions beside it did already catch that case, which is why the reviewer rated it clarity rather than a gap. But they caught it as a side effect rather than by intent, and a future edit could remove them without anyone noticing what else they were holding up. The status check now says it outright.

The two smaller ones

Symmetry. The placed node's status test now also asserts the body is a non-null object, matching its REAL_NODE_ID sibling. The two are a deliberate pair; they read better making the same claims.

Wording. "legacy" leaves the test title. POST /api/radar/detections is the current route for a node registering without a config, not a deprecated one, and the word implied a dead code path a reader would go looking for and not find.

Verification

No behaviour change: three assertions added, one word removed. The Playwright job does not run on pull requests, only on deploy, so CI here cannot exercise this file — the same limitation noted on #338. The assertions themselves were checked against staging while preparing that PR: an unconfigured node returns metrics, trust, reputation and coverage_map with no detection_area, and the endpoint answers 200 for both nodes in the pair.

Follows #338. Ticket: 86cbauhxd (closed).

🤖 Generated with Claude Code

[claude]

Claude finished @Babissimo's task in 1m 15s —— View job

---

Review complete

• Gather context (PR diff, related file)
• Run code review
• Post findings

Findings

None. This is a tiny, self-contained e2e-test hardening change (8/-1 lines, no production code) and it does what the description says:

• expect(analyticsStatus).toBe(200) (frontend/e2e/nodes.spec.ts:~638) now gates the not.toHaveProperty("detection_area") absence check, so a failed request can no longer make the assertion pass vacuously — the metrics/trust checks right below it already caught that case as a side effect, but this makes the intent explicit.
• expect(typeof placedBody).toBe("object") / expect(placedBody).not.toBeNull() added to the placed-node test mirror the existing pair used on its REAL_NODE_ID sibling — correct pairing (typeof null === "object" in JS, hence the separate not.toBeNull()), consistent with the pattern already in use nearby.
• "legacy" dropped from the test title — string-only change, no assertions affected.

No correctness issues, no removed behavior, nothing touching /v1/nodes contracts or other CLAUDE.md-gated concerns (this is Playwright e2e only, which per the PR description doesn't run in CI on pull requests). Diff matches the PR description precisely.