fix(webkit): allow service worker navigation while offline

[Devin Rousso (dcrousso)]

offline emulation rejects navigation even when a service worker supplies the response

apply configured offline state when page targets are initialized

cover service worker responses, context isolation, and reopening pages

fixes #42775

[Yury Semikhatsky (yury-s)]

let's drop this tests and keep coverage only for the functionality that is being fixed

[github-actions]

Test results for "tests 1"

1 failed
❌ [playwright-test] › playwright.spec.ts:1042 › page.pause() should disable test timeout @ubuntu-latest-node24

4 flaky

⚠️ [chromium-library] › library/popup.spec.ts:260 › should not throw when click closes popup `@chromium-ubuntu-22.04-arm-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`
⚠️ [playwright-test] › ui-mode-trace.spec.ts:448 › should work behind reverse proxy `@macos-latest-node22`

52516 passed, 1270 skipped

---

Merge workflow run.

[github-actions]

🟢 The failure is pre-existing CI infra noise

Hi, I'm the Playwright bot and I took a first look. The PR is clear: page.pause() should disable test timeout failed because its headed Chromium subprocess could not connect to an X server, outside the WebKit offline/service-worker code changed here.

Details

Pre-existing flake / infra

• [playwright-test] › playwright.spec.ts:1042 › page.pause() should disable test timeout (@ubuntu-latest-node24) — the subprocess exited with Missing X server or $DISPLAY, so the outer test received exit code 1 instead of 0. This PR only changes WebKit page initialization and adds service-worker policy tests; it does not touch Playwright Test, Chromium launch, or CI display setup.
• The same test failed on unrelated #38599, whose next CI report had no real failures. It has also failed amid the same headed-browser infrastructure pattern on unrelated #39350 and #39476.

The four retry-rescued tests in the report are flaky rather than failed, so they do not affect this verdict.

_{Triaged by the Playwright bot - agent run}

[github-actions]

Test results for "MCP"

2 failed
❌ [firefox] › mcp/cli-core.spec.ts:57 › click link @mcp-windows-latest-firefox
❌ [msedge] › mcp/annotate.spec.ts:291 › should enter annotate mode on fresh dashboard.tsx mount with -s --annotate @mcp-windows-latest-msedge

9013 passed, 1489 skipped

---

Merge workflow run.

[github-actions]

🟢 Both failures are pre-existing flakes — the PR is clear

Hi, I'm the Playwright bot and I took a first look at the latest MCP failures. Both tests have failed on unrelated SHAs on the same Windows/browser bots, while this PR only changes WebKit offline initialization and its library coverage.

Details

The latest MCP report has two failures. Neither test exercises WebKit, offline emulation, service workers, or a helper changed by this PR.

Pre-existing flake / infra

• [firefox] › mcp/cli-core.spec.ts:57 › click link (@mcp-windows-latest-firefox) — the click timed out and the snapshot contained that timeout instead of the navigated page. This Windows/Firefox-only test failed in 14 of 442 retained Firefox executions in a recent triage, including unrelated main sha da087a92 (run); it also failed on unrelated chore(locators): rewrite locator parser with per-language parsers #42895 and chore(server): simplify action helper signatures #42904.
• [msedge] › mcp/annotate.spec.ts:291 › should enter annotate mode on fresh dashboard.tsx mount with -s --annotate (@mcp-windows-latest-msedge) — the dashboard never exposed main[name="Dashboard: annotate"] within 15 seconds. The same test on this bot failed 4 of 716 runs in the aggregated results, on unrelated chore(coverage): istanbul instrumentation for the trace viewer and html reporter #42785, chore(deps): bump the github-actions group with 2 updates #42824, fix(net): apply socketTimeout from socket creation in httpRequest #42660, and a main push. The same missing-dashboard failure also occurs on Windows Chrome.

The required results-database refresh was attempted twice, but GitHub's repository-artifacts API returned HTTP 500. The historical counts above come from recent Playwright bot triages backed by that database; the linked unrelated runs establish the same-test flake history required for this verdict.

_{Triaged by the Playwright bot - agent run}