feat(webkit): roll to r2357

[microsoft-playwright-automation]

Browser build: https://github.com/microsoft/playwright-browsers/commit/bfe6a462f1d3c9403d856e685a2de9b11d74e3d2. Build has full platform coverage - check that all tests pass there.

[github-actions]

Test results for "tests 1"

1 failed
❌ [playwright-test] › playwright.trace.spec.ts:434 › should retain traces for interrupted tests @windows-latest-node22

7 flaky

⚠️ [chromium-library] › library/browsercontext-page-event.spec.ts:173 › should work with Ctrl-clicking `@chromium-ubuntu-22.04-arm-node20`
⚠️ [chromium-library] › library/video.spec.ts:664 › screencast › should capture full viewport `@chromium-ubuntu-22.04-node24`
⚠️ [chromium-library] › library/video.spec.ts:294 › screencast › should capture navigation `@chromium-ubuntu-22.04-node22`
⚠️ [chromium-library] › library/video.spec.ts:736 › screencast › should work with video+trace `@chromium-ubuntu-22.04-node22`
⚠️ [firefox-library] › library/heap.spec.ts:223 › should not leak workers `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-page] › page/page-goto.spec.ts:90 › should work with Cross-Origin-Opener-Policy `@firefox-ubuntu-22.04-node20`
⚠️ [playwright-test] › ui-mode-trace.spec.ts:715 › should indicate current test status `@windows-latest-node22`

51316 passed, 1239 skipped

---

Merge workflow run.

[github-actions]

🟢 The one failure is a pre-existing flake — this roll is clear

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

This is a WebKit roll (webkit r2355 → r2357), and the only failure is a playwright-test trace test on Windows — a suite that doesn't launch WebKit at all. It flakes with the same signature on main and unrelated PRs, so it isn't caused by this roll.

Details

The diff only bumps webkit revision 2355 → 2357 in browsers.json. Nothing in it touches the Playwright test runner or trace machinery.

Pre-existing flake / infra

• [playwright-test] › playwright.trace.spec.ts:434 › should retain traces for interrupted tests@windows-latest-node22 — pre-existing flake, not caused by this PR. In the aggregated CI results this test passes ~1000 runs and fails with the identical error expect(received).toBe(expected) // Expected: 1, Received: 0 on 5 distinct SHAs, including main (no PR) and unrelated PRs fix(chromium): keep closing the target so page.close() cannot hang on a navigation commit #41912 and feat(test-runner): annotate serial suites for custom sharding #42164. It's a playwright-test trace-retention test that never exercises WebKit, so a WebKit binary roll can't reach it. The failure clusters on windows-latest-node22, pointing at a timing flake in interrupted-test trace flushing on Windows.

The report also lists 7 flaky tests (rescued on retry) across chromium/firefox/playwright-test — none in WebKit code, all consistent with normal CI noise.

_{Triaged by the Playwright bot - agent run}

[github-actions]

Test results for "MCP"

1 failed
❌ [webkit] › mcp/storage.spec.ts:107 › browser_set_storage_state restores storage state from file @mcp-ubuntu-latest-webkit

8258 passed, 1361 skipped

---

Merge workflow run.

[github-actions]

🔴 One failure looks caused by this roll

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

This is a WebKit roll (webkit r2355 → r2357), and the MCP suite has a WebKit-only failure: browser_set_storage_state no longer restores localStorage on WebKit/Linux — the page reads back null instead of the saved value. It reproduces on two different SHAs of this PR and on no other run in the history, so it points at the new WebKit build, not a flake. The other failure (a playwright-test trace test on Windows) is a confirmed pre-existing flake.

Details

The diff only bumps webkit2355 → 2357 in packages/playwright-core/browsers.json, so anything that fails must be reachable through the new WebKit binary.

Caused by this PR

• [webkit] › mcp/storage.spec.ts:107 › browser_set_storage_state restores storage state from file@mcp-ubuntu-latest-webkit — real failure, tied to the roll. The tool restores storage state and the test expects the page to read back restoredValue; on the new build it gets null:

  - "result": StringContaining "restoredValue",
  + "result": "null",
  

  On mcp-ubuntu-latest-webkit this test is green in 719 of 720 runs in the aggregated CI results — the single red is this PR (run 33116054974, sha 0eaa6e1), and it failed again on the head sha ba38489 in the reported run. It's WebKit + Ubuntu specific: the same run passed WebKit on macOS and Windows. There is no instance of this test failing on main or any unrelated PR, so it can't be dismissed as a flake — the WebKit localStorage/storage-state behavior changed between r2355 and r2357. (The DB does hold 3 other reds for this title, but they're firefox on PR feat(storage): capture OPFS in storage state #42248 with a different signature — code/resultundefined — unrelated.)

Pre-existing flake / infra

• [playwright-test] › playwright.trace.spec.ts:434 › should retain traces for interrupted tests@windows-latest-node22 — pre-existing flake, not caused by this roll. It fails with the identical signature expect(received).toBe(expected) // Expected: 1, Received: 0 on 6 distinct SHAs, including a push to main (no PR) and unrelated PRs fix(chromium): keep closing the target so page.close() cannot hang on a navigation commit #41912, feat(test-runner): annotate serial suites for custom sharding #42164, feat(browserContext): add windowState option #42286. It's a playwright-test trace-retention test that never launches WebKit, so a WebKit binary roll can't reach it — a timing flake in interrupted-test trace flushing on Windows.

The report also lists 7 flaky tests (rescued on retry) across chromium/firefox/playwright-test — none in WebKit code, all normal CI noise.

_{Triaged by the Playwright bot - agent run}

[github-actions]

Test results for "tests 2"

1 failed
❌ [firefox-library] › library/inspector/cli-codegen-aria.spec.ts:87 › should update aria snapshot highlight @firefox-macos-15-large

20 flaky

⚠️ [chromium-library] › library/chromium/connect-over-cdp.spec.ts:523 › emulate media should not be affected by second connectOverCDP with noDefaults `@msedge-dev-windows-latest`
⚠️ [chromium-library] › library/defaultbrowsercontext-1.spec.ts:28 › context.cookies() should work Radoslav Kirilov (@smoke) `@chromium-windows-latest`
⚠️ [chromium-library] › library/video.spec.ts:736 › screencast › should work with video+trace `@chromium-windows-latest`
⚠️ [chromium-library] › library/trace-viewer.spec.ts:1893 › should not leak recorders `@chromium-macos-15-large`
⚠️ [chromium-library] › library/chromium/oopif.spec.ts:179 › should take screenshot `@chromium-macos-14-xlarge`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:213 › should intercept service worker requests (main and within) `@chromium-macos-15-xlarge`
⚠️ [chromium-library] › library/trace-viewer.spec.ts:1235 › should highlight target elements `@chrome-windows-latest`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:213 › should intercept service worker requests (main and within) `@chrome-macos-latest`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:301 › should report intercepted service worker requests in HAR `@chrome-macos-latest`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:436 › should produce network events, routing, and annotations for Service Worker (advanced) `@chrome-macos-latest`
⚠️ [chromium-library] › library/selector-generator.spec.ts:503 › selector generator › should use the name attributes for elements that can have it `@chrome-macos-latest`
⚠️ [chromium-library] › library/selector-generator.spec.ts:736 › selector generator › should not use icon fonts aria name `@chrome-macos-latest`
⚠️ [firefox-page] › page/page-event-request.spec.ts:181 › should return response body when Cross-Origin-Opener-Policy is set `@firefox-macos-15-large`
⚠️ [firefox-page] › page/page-goto.spec.ts:90 › should work with Cross-Origin-Opener-Policy `@firefox-windows-latest`
⚠️ [webkit-library] › library/screencast.spec.ts:57 › applies backpressure while async onFrame callback is pending `@webkit-macos-15-xlarge`
⚠️ [webkit-library] › library/trace-viewer.spec.ts:1673 › should highlight locator in iframe while typing `@webkit-macos-15-xlarge`
⚠️ [webkit-library] › library/trace-viewer.spec.ts:1975 › canvas clipping `@webkit-macos-15-xlarge`
⚠️ [webkit-library] › library/browsertype-connect.spec.ts:807 › launchServer › should upload a folder `@webkit-macos-15-large`
⚠️ [android-page] › page/page-request-fulfill.spec.ts:81 › should fulfill with unuassigned status codes
⚠️ [android-page] › page/selectors-frame.spec.ts:203 › should click in lazy iframe

110601 passed, 4712 skipped

---

Merge workflow run.

[github-actions]

🔴 One failure looks caused by this roll

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

This is a WebKit roll (webkit r2355 → r2357), and the MCP suite has a WebKit-only failure: browser_set_storage_state no longer restores localStorage on WebKit/Ubuntu — the page reads back null instead of the saved value. It's the only time this test has ever failed on WebKit in the aggregated CI results, and it landed on this PR, so it points at the new WebKit build, not a flake. The other two failures (a Firefox codegen test and a playwright-test trace test) are confirmed pre-existing flakes.

Details

The diff only bumps webkit2355 → 2357 in packages/playwright-core/browsers.json, so anything caused by it must be reachable through the new WebKit binary.

Caused by this PR

• [webkit] › mcp/storage.spec.ts:107 › browser_set_storage_state restores storage state from file@mcp-ubuntu-latest-webkit — real failure, tied to the roll. The tool restores storage state and the test reads it back, but on the new build localStorage.getItem("restoredKey") returns null:

  - "result": StringContaining "restoredValue",
  + "result": "null",
  

  In the test-results DB this test is green in 2169 of 2170 WebKit runs — the single red is this PR (sha 0eaa6e1), and it never fails on main or any other PR/sha on WebKit. It's WebKit + Ubuntu specific (WebKit on macOS and Windows stayed green). Per the flake rule there's nothing to point to elsewhere, so this can't be dismissed as noise — the WebKit storage-state / localStorage restore behavior changed between r2355 and r2357. (The DB holds 3 other reds for this title, but they're firefox on PR feat(storage): capture OPFS in storage state #42248 with a different signature — Expected -3 — unrelated.)

Pre-existing flake / infra

• [firefox-library] › inspector/cli-codegen-aria.spec.ts:87 › should update aria snapshot highlight@firefox-macos-15-large — pre-existing flake, not caused by this roll. It fails with the identical signature expect(locator).toBeVisible() failed — Locator: locator('x-pw-highlight') in 9 of 1885 runs; 8 of those are push events on main (no PR) between Aug 20–27, plus this PR once. It's a Firefox-only codegen highlight test, and a WebKit binary roll can't reach Firefox.

• [playwright-test] › playwright.trace.spec.ts:434 › should retain traces for interrupted tests@windows-latest-node22 — pre-existing flake. It fails with the same expect(received).toBe(expected) // Expected: 1, Received: 0 signature in 6 of 4551 runs across several SHAs, including a push to main. It's a playwright-test trace-retention test that never launches WebKit, so the roll can't reach it — a timing flake in interrupted-test trace flushing on Windows.

The reports also list flaky tests rescued on retry across chromium/firefox/webkit/playwright-test — normal CI noise, none pointing at the roll.

_{Triaged by the Playwright bot.}

_{Triaged by the Playwright bot - agent run}

[Devin Rousso (dcrousso)]

already handled by #42438