fix(health): surface a fetch error in TestsReachingList

[drakeo338]

Summary

TestsReachingList caught a rejected fetch into null, so a failed request rendered the same "no test reaches this file" sentence as a genuine empty answer. This change lets the fetch failure set an error state instead, so a 500 or a network error shows a distinct "could not load" message rather than the empty-graph answer.

There's a related spot worth a separate look: health_tests_reaching in coverage_routes.py:359 still folds a graph-walk exception into the same basis: "none" shape as a real empty answer, matching this file's own convention elsewhere, so the new client error state can't fire for that server-side case. I left it out of this PR since it would change an established convention the issue doesn't ask for, and the UI conflation this PR fixes holds regardless.

Fixes #2624.

Related Issues

Fixes #2624

Test Plan

Run from packages/ui: npx vitest run __tests__/health/inferred-tests.test.tsx

• Tests pass (pytest)
• Lint passes (ruff check .)
• Web build passes (npm run build) (if frontend changes)

Checklist

• My code follows the project's code style
• I have added tests for new functionality
• All existing tests still pass
• I have updated documentation if needed

[repowise-bot]

Reviewed the whole diff: nothing to fix.

⚠️ Health of changed files: 9.0 → 8.9 (-0.1)

👀 Suggested reviewers @RaghavChamadiya

---

📊 See the full report for this PR
_{Your repo map with this PR's blast radius lit up, every caller of the contracts it changes, and health before and after. No sign-in. · Reviewed with AI · Plain markdown for agents · ⭐ Star Repowise · 📥 Install bot · Silence on a single PR with [skip repowise] in the title · Per-repo toggle on repowise.dev/settings/bot · Updated 2026-09-28 08:52 UTC}

[RaghavChamadiya]

Thanks @drakeo338. Dropping the .catch(() => null) also brings the useSWR type back in line with what the fetcher prop actually promises, and the negative assertion that the empty-state text no longer renders is the check that matters. Agreed the server-side basis: "none" fold is a separate question.