fix: own the post-restart rerun, keep process.exit disabled in workers

[sheremet-va]

The rerun that startVitest schedules after a watch-mode server restart ran as a floating promise with no error handling. Anything failing inside it became an unhandled rejection in whatever process hosted the Vitest instance. This was the source of most of the recent CI flakes: our meta-tests run Vitest inside shared isolate: false forks workers, where the real process.exit was restored after the first test file, so a late rejection could escalate into the whole worker being killed mid-run ("Worker forks emitted error").

The restart now awaits its rerun and reports failures the same way the initial run does, and workers keep process.exit disabled until teardown, so a stray exit call fails loudly with the test file in the message instead of taking down the process.

[netlify]

✅ Deploy Preview for vitest-dev ready!

Name	Link
🔨 Latest commit	5fe1c01
🔍 Latest deploy log	https://app.netlify.com/projects/vitest-dev/deploys/6a81d14bc498d7000869a6b7
😎 Deploy Preview	https://deploy-preview-10963--vitest-dev.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.
🤖 Make changes	Run an agent on this branch

---

To edit notification comments on pull requests, go to your Netlify project configuration.