Skip to content

feat(firefox): roll to r1541 - #42441

Open
microsoft-playwright-automation[bot] wants to merge 1 commit into
mainfrom
roll-into-pw-firefox/1541
Open

feat(firefox): roll to r1541#42441
microsoft-playwright-automation[bot] wants to merge 1 commit into
mainfrom
roll-into-pw-firefox/1541

Conversation

@microsoft-playwright-automation

Copy link
Copy Markdown
Contributor

Browser build: https://github.com/microsoft/playwright-browsers/commit/f81945fd6dce01624d8286736d47c4d014ab3a50. Build has full platform coverage - check that all tests pass there.

@github-actions

Copy link
Copy Markdown
Contributor

Test results for "tests 1"

5 failed
❌ [firefox-library] › library/har.spec.ts:629 › should have connection details @firefox-ubuntu-22.04-node20
❌ [firefox-library] › library/har.spec.ts:639 › should have security details @firefox-ubuntu-22.04-node20
❌ [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects @firefox-ubuntu-22.04-node20
❌ [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests @firefox-ubuntu-22.04-node20
❌ [firefox-library] › library/har.spec.ts:699 › should return server address directly from response @firefox-ubuntu-22.04-node20

7 flaky⚠️ [chromium-library] › library/chromium/chromium.spec.ts:301 › should report intercepted service worker requests in HAR `@frozen-time-library-chromium-linux`
⚠️ [chromium-library] › library/video.spec.ts:495 › screencast › should capture static page in persistent context Radoslav Kirilov (@smoke) `@realtime-time-library-chromium-linux`
⚠️ [chromium-library] › library/video.spec.ts:736 › screencast › should work with video+trace `@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`
⚠️ [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`

51241 passed, 1240 skipped


Merge workflow run.

@github-actions

Copy link
Copy Markdown
Contributor

🔴 The Firefox roll breaks 5 HAR serverIPAddress tests

Hi, I'm the Playwright bot — I took a first look at the failing CI.

This is a real regression from the roll, not flake. Firefox r1541 now reports the HAR serverIPAddress as the bare ::1 instead of the bracketed [::1], so library/har.spec.ts's toMatch(/^127\.0\.0\.1|\[::1\]/) no longer matches. All 5 failures share that exact signature, and they reproduce on every retry — 10/10 attempts red on this PR, 0 green.

Details

Caused by this PR

  • [firefox-library] › library/har.spec.ts:629 › should have connection details

  • [firefox-library] › library/har.spec.ts:639 › should have security details

  • [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects

  • [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests

  • [firefox-library] › library/har.spec.ts:699 › should return server address directly from response

    All five fail identically:

    Expected pattern: /^127\.0\.0\.1|\[::1\]/
    Received string: "::1"
    

    This is a Firefox-only browser roll (roll-into-pw-firefox/1541), and serverIPAddress is populated by the rolled juggler's networking layer — squarely in the diff's blast radius. The new build emits the loopback IPv6 address unbracketed. The fix is either to restore the bracketed [::1] form in the juggler's HAR reporting, or to relax the test to accept a bare ::1. Chromium and WebKit are unaffected, which is exactly why only the firefox-library project went red.

    Test-results DB confirms it's the roll, not a flake: across the recent window these tests failed on one branch only — this PR (10 failed / 0 passed). On every other PR and branch that ran them they're green (0 failures over hundreds of runs), so the flake hard-rule is not met — there's no run outside this PR where they misbehave.

Pre-existing flake / infra

  • The time test runner - frozen job also shows red, but its blob report lists no failing tests — only the 7 flaky tests below, all rescued on retry. That's the frozen-time job's usual flake noise, unrelated to the roll.

  • Reported flaky (passed on retry), not caused by this PR: chromium.spec.ts:301, video.spec.ts:495, video.spec.ts:736, browsercontext-cookies-third-party.spec.ts:257 & :470, heap.spec.ts:223, page-goto.spec.ts:90.

Triaged by the Playwright bot - agent run

@github-actions

Copy link
Copy Markdown
Contributor

Test results for "MCP"

4 failed
❌ [chromium] › mcp/clipboard.spec.ts:19 › clipboard write without permission dialog @mcp-macos-latest-chromium
❌ [chromium] › mcp/clipboard.spec.ts:19 › clipboard write without permission dialog @mcp-windows-latest-chromium
❌ [chromium] › mcp/clipboard.spec.ts:19 › clipboard write without permission dialog @mcp-ubuntu-latest-chromium
❌ [firefox] › mcp/cli-core.spec.ts:394 › --raw on command without output @mcp-windows-latest-firefox

8255 passed, 1361 skipped


Merge workflow run.

@github-actions

Copy link
Copy Markdown
Contributor

🔴 The Firefox roll breaks 5 HAR serverIPAddress tests

Hi, I'm the Playwright bot — I took a first look at the failing CI.

One real regression: Firefox r1541 reports the HAR serverIPAddress for the loopback as the bare ::1 instead of the bracketed [::1], so library/har.spec.ts's /^127\.0\.0\.1|\[::1\]/ no longer matches. All 5 firefox-library failures share that signature and fail on every attempt (2/2 red on this PR, 0 red across ~hundreds of other runs). The MCP clipboard and cli-core reds are pre-existing flakes a firefox roll can't be responsible for.

Details

The diff is firefox-only — browser_patches/firefox/juggler/* plus browsers.json. serverIPAddress is populated by juggler's networking layer, squarely in the blast radius, which is why only firefox-library went red while Chromium and WebKit stayed green.

Caused by this PR

  • [firefox-library] › library/har.spec.ts:629 › should have connection details

  • [firefox-library] › library/har.spec.ts:639 › should have security details

  • [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects

  • [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests

  • [firefox-library] › library/har.spec.ts:699 › should return server address directly from response

    All five fail identically:

    Expected pattern: /^127\.0\.0\.1|\[::1\]/
    Received string: "::1"
    

    Test-results DB confirms the roll, not a flake: each failed 2/2 on this PR and 0 times across 356–983 other runs on unrelated SHAs. Fix by restoring the bracketed [::1] form in juggler's HAR reporting, or relaxing the test to accept a bare ::1.

Pre-existing flake / infra

  • [chromium] › mcp/clipboard.spec.ts:19 › clipboard write without permission dialog (macOS, Windows, Ubuntu) — the clipboard read returns "" instead of "Hello from Playwright!". This is a Chromium-only test (Firefox is skipped), so a Firefox-only roll can't cause it. It flakes across the DB independent of this PR — 24 failures over 5824 other runs, including on main. The three OSes failing together points at a clipboard-access hiccup in that one MCP run.

  • [firefox] › mcp/cli-core.spec.ts:394 › --raw on command without output (Windows only) — TimeoutError: Timeout 5000ms exceeded waiting for aria-ref=e2. Firefox on macOS and Ubuntu passed in the same run; only mcp-windows-latest-firefox failed. That exact bot has a standing flake history — 4 failures over 262 runs on main, plus one-off reds on PRs test: roll stable-test-runner to 1.63.0-alpha-2026-08-10 #42186, chore(tracing): reference trace and har blobs via relative file paths #42192, chore: remove experimental component testing #42168 and [DO NOT LAND] roll ffmpeg to 1012 #42067. Meets the flake bar; this is one more instance, not the roll.

  • The 7 tests reported flaky (rescued on retry) — chromium.spec.ts:301, video.spec.ts:495/:736, browsercontext-cookies-third-party.spec.ts:257/:470, heap.spec.ts:223, page-goto.spec.ts:90 — are the usual noise, not caused by the roll.

Triaged by the Playwright bot - agent run

@github-actions

Copy link
Copy Markdown
Contributor

Test results for "tests 2"

25 failed
❌ [playwright-test] › playwright.artifacts.spec.ts:127 › should work with screenshot: on @frozen-time-runner-chromium-linux
❌ [playwright-test] › reporter-html.spec.ts:307 › created › should not include image diff with non-images @frozen-time-runner-chromium-linux
❌ [firefox-library] › library/har.spec.ts:629 › should have connection details @tracing-firefox
❌ [firefox-library] › library/har.spec.ts:639 › should have security details @tracing-firefox
❌ [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects @tracing-firefox
❌ [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests @tracing-firefox
❌ [firefox-library] › library/har.spec.ts:699 › should return server address directly from response @tracing-firefox
❌ [firefox-library] › library/har.spec.ts:629 › should have connection details @firefox-macos-15-large
❌ [firefox-library] › library/har.spec.ts:639 › should have security details @firefox-macos-15-large
❌ [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects @firefox-macos-15-large
❌ [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests @firefox-macos-15-large
❌ [firefox-library] › library/har.spec.ts:699 › should return server address directly from response @firefox-macos-15-large
❌ [firefox-library] › library/har.spec.ts:629 › should have connection details @firefox-windows-latest
❌ [firefox-library] › library/har.spec.ts:639 › should have security details @firefox-windows-latest
❌ [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects @firefox-windows-latest
❌ [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests @firefox-windows-latest
❌ [firefox-library] › library/har.spec.ts:699 › should return server address directly from response @firefox-windows-latest
❌ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies @firefox-macos-15-xlarge
❌ [firefox-library] › library/har.spec.ts:629 › should have connection details @firefox-macos-15-xlarge
❌ [firefox-library] › library/har.spec.ts:639 › should have security details @firefox-macos-15-xlarge
❌ [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects @firefox-macos-15-xlarge
❌ [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests @firefox-macos-15-xlarge
❌ [firefox-library] › library/har.spec.ts:699 › should return server address directly from response @firefox-macos-15-xlarge
❌ [android-native] › android/device.spec.ts:56 › androidDevice.fill
❌ [android-native] › android/webview.spec.ts:44 › should navigate page externally

23 flaky⚠️ [chromium-library] › library/chromium/chromium.spec.ts:436 › should produce network events, routing, and annotations for Service Worker (advanced) `@tracing-chromium`
⚠️ [chromium-library] › library/chromium/launcher.spec.ts:37 › should not create pages automatically `@msedge-dev-windows-latest`
⚠️ [chromium-library] › library/inspector/cli-codegen-2.spec.ts:202 › cli codegen › should upload a file via hidden input triggered by button `@chrome-ubuntu-22.04`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:213 › should intercept service worker requests (main and within) `@chromium-macos-15-large`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:373 › should produce network events, routing, and annotations for Service Worker `@chromium-macos-15-large`
⚠️ [chromium-library] › library/chromium/chromium.spec.ts:436 › should produce network events, routing, and annotations for Service Worker (advanced) `@chromium-macos-15-large`
⚠️ [chromium-library] › library/video.spec.ts:390 › screencast › should scale frames down to the requested size `@chromium-macos-15-large`
⚠️ [chromium-library] › library/chromium/oopif.spec.ts:179 › should take screenshot `@chromium-macos-14-xlarge`
⚠️ [chromium-library] › library/video.spec.ts:736 › screencast › should work with video+trace `@chromium-macos-15-xlarge`
⚠️ [chromium-library] › library/beforeunload.spec.ts:130 › should support dismissing the dialog multiple times `@chrome-macos-latest`
⚠️ [chromium-library] › library/selector-generator.spec.ts:343 › selector generator › should trim long text `@chrome-macos-latest`
⚠️ [chromium-library] › library/selector-generator.spec.ts:673 › selector generator › should generate noText: no text engine `@chrome-macos-latest`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies `@tracing-firefox`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@tracing-firefox`
⚠️ [firefox-page] › page/page-goto.spec.ts:90 › should work with Cross-Origin-Opener-Policy `@tracing-firefox`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies `@firefox-macos-15-large`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@firefox-macos-15-large`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies `@firefox-windows-latest`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@firefox-windows-latest`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@firefox-macos-15-xlarge`
⚠️ [webkit-library] › library/playwright-client.spec.ts:26 › should connect from a page and drive the same browser `@tracing-webkit`
⚠️ [webkit-library] › library/har.spec.ts:472 › should return receive time `@webkit-windows-latest`
⚠️ [webkit-library] › library/tracing.spec.ts:494 › should produce screencast frames fit `@webkit-macos-15-xlarge`

110414 passed, 4714 skipped


Merge workflow run.

@github-actions

Copy link
Copy Markdown
Contributor

🔴 The Firefox roll breaks the HAR serverIPAddress tests

Hi, I'm the Playwright bot — I took a look at the failing CI.

One real regression from the roll: Firefox r1541 now reports the loopback HAR serverIPAddress as the bare ::1 instead of the bracketed [::1], so library/har.spec.ts's /^127\.0\.0\.1|\[::1\]/ no longer matches. The five affected tests fail on every Firefox platform in this run (ubuntu, macOS, windows, tracing) and on every retry. The test-results DB is unambiguous: 10 failures, all on this PR, 0 across 3034 runs on other SHAs. Everything else that's red is a pre-existing flake a Firefox-only browser roll can't be responsible for — with one Firefox cookie test I'd give a second glance.

Details

The diff is Firefox-only — browser_patches/firefox/juggler/*, preferences/playwright.cfg, and browsers.json. serverIPAddress is populated by juggler's networking layer, squarely in the blast radius, which is why only the firefox-library project went red while Chromium and WebKit stayed green.

Caused by this PR

  • [firefox-library] › library/har.spec.ts:629 › should have connection details

  • [firefox-library] › library/har.spec.ts:639 › should have security details

  • [firefox-library] › library/har.spec.ts:664 › should have connection details for redirects

  • [firefox-library] › library/har.spec.ts:686 › should have connection details for failed requests

  • [firefox-library] › library/har.spec.ts:699 › should return server address directly from response

    All five fail identically:

    Expected pattern: /^127\.0\.0\.1|\[::1\]/
    Received string: "::1"
    

    Across the recent window these five failed 10/10 on this PR and 0 times over 3034 runs on unrelated SHAs — the flake hard-rule isn't met anywhere outside this PR. Fix by restoring the bracketed [::1] form in juggler's HAR reporting, or by relaxing the test to accept a bare ::1.

Uncertain — worth a second look

  • [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies — hard-failed on @firefox-macos-15-xlarge and was rescued-on-retry on several other Firefox bots in the same run. It has 0 failures over 985 runs outside this PR, so I can't call it a flake, and partitioned-cookie behavior is the kind of thing a Firefox version bump shifts — so it's within the roll's blast radius. It mostly self-rescued and didn't block, but I can't prove it's unrelated. The COOP page-goto.spec.ts:90 flake is a milder instance of the same "only-on-this-PR Firefox flakiness" pattern. Neither is confirmed, but both are worth a glance alongside the HAR fix.

Pre-existing flake / infra

  • [chromium] › mcp/clipboard.spec.ts:19 › clipboard write without permission dialog (macOS, Windows, Ubuntu) — a Chromium-only test (Firefox is skipped), so a Firefox roll can't touch it. Flakes independently of this PR: 8 failures across 5 other PRs over 731 runs.
  • [firefox] › mcp/cli-core.spec.ts:394 › --raw on command without output (Windows only) — TimeoutError waiting for aria-ref=e2; Firefox on macOS and Ubuntu passed in the same run. Standing flake: 5 failures across 4 other PRs over 722 runs.
  • [android-native] › android/device.spec.ts:56 › androidDevice.fill and [android-native] › android/webview.spec.ts:44 › should navigate page externally — Android runs don't exercise juggler's Firefox networking. Both flake across the DB: 12 failures on 2 other PRs over 348 runs.
  • [playwright-test] › playwright.artifacts.spec.ts:127 and [playwright-test] › reporter-html.spec.ts:307 (both @frozen-time-runner-chromium-linux) — Chromium test-runner tests; a Firefox browser roll can't change Chromium screenshot or HTML-reporter output. These have 0 failures anywhere in the DB and are the frozen-time job's usual harness noise.
  • The remaining tests reported flaky (rescued on retry) — chromium.spec.ts:301/:213/:373/:436, video.spec.ts:390/:495/:736, launcher.spec.ts:37, cli-codegen-2.spec.ts:202, oopif.spec.ts:179, beforeunload.spec.ts:130, selector-generator.spec.ts:343/:673, heap.spec.ts:223, playwright-client.spec.ts:26, har.spec.ts:472, tracing.spec.ts:494 — are the usual cross-engine noise, not caused by the roll.

Triaged by the Playwright bot - agent run

Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants