fix(runner): do not force-kill worker while its teardown is in progress

[pavelfeldman]

Summary

• Worker sends heartbeats over IPC while gracefully closing, so a slow fixture teardown with timeout: 0 is not force-killed.
• A genuinely hung worker (blocked event loop, process that will not exit) sends no heartbeats and is still force-killed, preserving fix(runner): force-kill worker that hangs on stop #40637.

Fixes #42007

[github-actions]

Test results for "tests 1"

6 flaky

⚠️ [chromium-library] › library/video.spec.ts:495 › screencast › should capture static page in persistent context @smoke `@realtime-time-library-chromium-linux`
⚠️ [chromium-library] › library/video.spec.ts:356 › screencast › should work for popups `@chromium-ubuntu-22.04-node24`
⚠️ [chromium-library] › library/video.spec.ts:495 › screencast › should capture static page in persistent context @smoke `@chromium-ubuntu-22.04-node24`
⚠️ [chromium-page] › page/workers.spec.ts:191 › should attribute network activity for worker inside iframe to the iframe `@chromium-ubuntu-22.04-node20`
⚠️ [firefox-page] › page/page-emulate-media.spec.ts:144 › should keep reduced motion and color emulation after reload `@firefox-ubuntu-22.04-node20`
⚠️ [playwright-test] › ui-mode-trace.spec.ts:827 › should update state on subsequent run `@windows-latest-node22`

50739 passed, 1190 skipped

---

Merge workflow run.

[github-actions]

Test results for "MCP"

2 failed
❌ [firefox] › mcp/cli-core.spec.ts:189 › dialog-accept @mcp-windows-latest-firefox
❌ [firefox] › mcp/dashboard.spec.ts:108 › should activate session when show is called with -s @mcp-windows-latest-firefox

7806 passed, 1266 skipped

---

Merge workflow run.

[github-actions]

Hi, I'm the Playwright bot and I took a first look at the CI failures here.

🟢 Nothing looks caused by this PR

Both MCP failures are pre-existing Firefox flakes, on flows this PR doesn't touch. The "tests 1" report has only flaky tests, no real failures. You're clear.

Details

This PR changes worker heartbeat / force-kill logic in the test runner (process.ts, processHost.ts, exit-code.spec.ts). None of the failing tests exercise that path.

Pre-existing flake / infra

• [firefox] › mcp/cli-core.spec.ts:189 › dialog-accept <prompt> (@mcp-windows-latest-firefox) — Firefox-only flake. Across the test-results DB it flips verdict on unrelated SHAs: failed 7 of 488 firefox runs across 6 different PRs, passed the other 481; never failed on chromium/chrome/msedge/webkit.
• [firefox] › mcp/dashboard.spec.ts:108 › should activate session when show is called with -s (@mcp-windows-latest-firefox) — same story: failed 18 of 488 firefox runs across 13 different PRs, passed 470, and never failed on the other engines.

Both are Firefox-only and reproduce across many PRs that can't be responsible, and the MCP dialog/dashboard flows don't touch worker teardown or force-kill — so this PR isn't implicated.

Note

• The "tests 1" report lists 6 flaky (all rescued on retry) and 0 failures — nothing to triage there.

_{Triaged by the Playwright bot - agent run}