test: control lazy recovery timing

[KyleJune]

Summary

The macOS main run 34714681380 checked document recovery after a 10 ms wall-clock wait and saw no navigation yet. Advance a fake clock through import rejection and the scheduled handoff so the recovery assertions do not depend on runner timing.

Changes

• Use FakeTime throughout the eight lazy-route recovery tests.
• Exercise a 25 ms delayed import rejection in the route-preservation test.
• Preserve assertions for the destination, pending navigation, current page, error suppression, cancellation retries, and reload budget. Production code and sanitizer settings are unchanged.

Testing

• deno task check passed; deno task test --parallel --reporter=dot passed (34 suites, 388 steps).
• Red: deno task test --reporter=dot client.test.tsx, with only the delayed-import case selected and the original 10 ms helper, failed at assertEquals(assigned, ["/destination?q=1#top"]): actual [].
• Green: deno task test --reporter=dot --filter 'createLazyRoute failure recovery' client.test.tsx passed all eight cases with the clock fix.
• Mutation: resolving the recovery promise prematurely made the route-preservation test fail in assertPendingRecovery: actual resolved, expected pending. The mutation was restored before the full gates.

Closes

Nothing.

[github-actions]

🎉 This PR is included in version 0.11.6 🎉

The release is available on:

• JSR.io
• GitHub release

Your semantic-release bot 📦🚀