test: fix click tests on android

[Dmitry Gozman (dgozman)]

Summary

• Anchor the fixed header with bottom: 50px instead of calc(100vh - 50px): on Android 100vh exceeds innerHeight, so the header covered the whole viewport and the click could never succeed.
• Skip the iframe retention test on Android, where requestGC() is not enough to release the iframe.

[github-actions]

Test results for "tests 1"

1 failed
❌ [firefox-library] › library/browsercontext-storage-state.spec.ts:453 › should work when service worker is intefering and the origin is not open @firefox-ubuntu-22.04-node20

8 flaky

⚠️ [chromium-library] › library/chromium/chromium.spec.ts:436 › should produce network events, routing, and annotations for Service Worker (advanced) `@frozen-time-library-chromium-linux`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:179 › serviceWorker(), and fromServiceWorker() work `@chromium-ubuntu-22.04-arm-node20`
⚠️ [chromium-library] › library/beforeunload.spec.ts:130 › should support dismissing the dialog multiple times `@realtime-time-library-chromium-linux`
⚠️ [chromium-library] › library/video.spec.ts:736 › screencast › should work with video+trace `@realtime-time-library-chromium-linux`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:373 › should produce network events, routing, and annotations for Service Worker `@chromium-ubuntu-22.04-node20`
⚠️ [chromium-library] › library/video.spec.ts:664 › screencast › should capture full viewport `@chromium-ubuntu-22.04-node20`
⚠️ [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`

51719 passed, 1247 skipped

---

Merge workflow run.

[github-actions]

🟢 The one failure is pre-existing — it's broken on main too

browsercontext-storage-state.spec.ts:453 on Firefox fails on every run in the CI results db, including plain main pushes. Nothing here points at this PR.

Details

This PR touches two test files only — tests/page/page-click-scroll.spec.ts (CSS in an inline fixture) and tests/page/page-click.spec.ts (an Android skip). It changes no product code and no shared helper, so it can't reach a browsercontext-storage-state service-worker test.

Pre-existing flake / infra

• [firefox-library] › library/browsercontext-storage-state.spec.ts:453 › should work when service worker is intefering and the origin is not open — not a flake but a straight pre-existing breakage: 24 of 24 runs failed, 0 passed, spanning 16 distinct SHAs and 5 Firefox bots since 2026-09-10. That includes main pushes with no PR attached (e.g. 500c9c8, 249916f, 4dc79d4, a9da2db, d4ce80d), and other unrelated PRs (fix(server): end range responses when the file stream errors #42716, fix(reporter): do not hang perfetto writes after a stream error #42715, fix(trace-viewer): reject unsafe snapshot tag names #42712, fix(storage): preserve IndexedDB Map and Set values #42707, feat(ui-mode): update individual snapshots #42421). The error is identical everywhere — expect(received).toEqual(expected), 4 unexpected cookie entries where 0 were expected. Someone should fix it on main, but it isn't this PR's doing.

The 8 flaky results in the report (service worker network events, screencast video, beforeunload dialogs, partitioned cookies) all passed on retry and sit in library code this PR doesn't touch.

_{Triaged by the Playwright bot - agent run}

[github-actions]

Test results for "MCP"

2 failed
❌ [firefox] › mcp/cli-drag.spec.ts:19 › drag between elements @mcp-windows-latest-firefox
❌ [firefox] › mcp/cli-session.spec.ts:54 › idle timeout shuts the session down @mcp-windows-latest-firefox

8562 passed, 1446 skipped

---

Merge workflow run.

[github-actions]

🟢 Both failures are pre-existing Windows/Firefox flakes

cli-drag.spec.ts:19 and cli-session.spec.ts:54 both fail on plain main pushes with the exact same errors. This PR touches two test files and no product code, so it can't reach them.

Details

The PR changes only tests/page/page-click-scroll.spec.ts (inline CSS fixture) and tests/page/page-click.spec.ts (an Android skip). No product code, no shared helper — nothing that the MCP CLI suite exercises.

Pre-existing flake / infra

• [firefox] › mcp/cli-drag.spec.ts:19 › drag between elements — bimodal on mcp-windows-latest-firefox only: 9 failed / 685 passed over 694 runs, and green on all 12 other MCP bots (0 failures in ~8500 runs, including ubuntu and macOS Firefox). Same failure on the main push 654654f (run 34864024187) and on feat(mcp): add media emulation #42668, with the identical error: expect(received).toBe(expected), expected 0, received 1.

• [firefox] › mcp/cli-session.spec.ts:54 › idle timeout shuts the session down — same shape: 7 failed / 22 passed over 29 runs on mcp-windows-latest-firefox, 0 failures on every other bot. Fails on the main pushes d1ead3e, 654654f, 52a0e72 and on PRs feat(coverage): experimental istanbul coverage collection via tracing #42684, test: roll stable-test-runner to 1.64.0-alpha-2026-09-14 #42706, feat(mcp): add --file-paths=absolute for absolute paths in tool results #42673 — always with Expected substring: "(no browsers)" against a received ### Browsers listing, i.e. the session hadn't shut down yet when the assertion ran.

Both look like the Windows Firefox bot being slow enough that a shutdown/settle step loses its race. Worth someone fixing on main, but not this PR's doing.

_{Triaged by the Playwright bot - agent run}