fix(test runner): forward the screen option from device descriptors

[Pavel Feldman (pavelfeldman)]

Summary

• Register screen as a test option so test.use({ ...devices[...] }) and use: { ...devices[...] } in config forward the descriptor's screen key, matching library mode.
• Add the TestOptions.screen type and docs entry, plus a regression test.

Fixes #42679

Breaking change, not for landing

Every config that spreads a device descriptor into use starts emulating screen on upgrade. That includes all desktop descriptors, so the default scaffolded project changes window.screen from 1280x720 to 1920x1080 and device-width media queries follow. A user-defined fixture named screen that depends on page now fails with a fixture dependency cycle, which affects @playwright-testing-library/test.

[github-actions]

Test results for "tests 1"

4 flaky

⚠️ [chromium-library] › library/video.spec.ts:356 › screencast › should work for popups `@chromium-ubuntu-22.04-node24`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@firefox-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`

51698 passed, 1247 skipped

---

Merge workflow run.

[github-actions]

Test results for "MCP"

1 failed
❌ [chrome] › mcp/webmcp.spec.ts:213 › browser_webmcp_call disambiguates same-name tools in identical same-origin frames @mcp-macos-latest-chrome

8495 passed, 1436 skipped

---

Merge workflow run.

[github-actions]

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

🟢 CI is clear — the one failure is a pre-existing flake

The only red test, mcp/webmcp.spec.ts:213, also fails identically on main (push, sha 6750deccd13b5703ab7c6d9aca9c723a7d1e5034) — a SHA this PR can't be responsible for. The four flaky results all passed on retry. This PR only wires up a screen test option; it doesn't touch MCP.

Details

Overall: nothing here points back at the diff. The PR adds a screen fixture in packages/playwright/src/index.ts plus docs/types and one playwright-test spec — it never reaches the MCP WebMCP tool-discovery path, and the one real failure reproduces off-PR.

Pre-existing flake / infra

• [chrome] › mcp/webmcp.spec.ts:213 › browser_webmcp_call disambiguates same-name tools in identical same-origin frames — pre-existing flake. The exact same failure lands on main at sha 6750deccd (run 34650183469, event=push), with the identical signature: the test expects Found 2 WebMCP tool(s) but gets Found 1 because only one of the two identical same-origin frames has registered its tools when the call runs. It's a registration/timing race, not something this PR changes. Chrome-only, and the WebMCP disambiguation code is untouched by the diff.

Already-retried flakes (passed on retry, not blocking)

• [chromium-library] › library/video.spec.ts:356 › screencast › should work for popups
• [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies
• [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe
• [firefox-page] › page/page-emulate-media.spec.ts:144 › should keep reduced motion and color emulation after reload

None of these exercise the screen option this PR adds, and all recovered on retry.

_{Triaged by the Playwright bot - agent run}