chore(net): use native happy eyeballs instead of the manual implementation

[Pavel Feldman (pavelfeldman)]

Summary

• Delete the manual RFC 8305 implementation (utils/happyEyeballs.ts); Node.js 20+ races connection attempts natively via autoSelectFamily, which is enabled by default.
• Keep the dual-family DNS lookup as a standard lookup option (dualStackLookup in utils/network.ts): a family-agnostic getaddrinfo can miss ::1 for localhost (AI_ADDRCONFIG filtering, /etc/hosts mapping ::1 to ip6-localhost only), breaking localhost servers bound to a single family.
• API request HAR timings now come from the standard socket lookup/connect/secureConnect events; __testHookLookup is adapted into the standard lookup option.

Fixes: #42193

[github-actions]

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

🔴 One failure is caused by this PR

mcp/cdp.spec.ts:75 › should throw connection error and allow re-connecting fails on every MCP browser/OS (chrome, chromium, msedge × ubuntu/macos/windows). Switching the CDP transport to dualStackLookup changes the shape of a refused-connection error, so the assertion no longer matches. The other two MCP failures are a pre-existing flake and one uncertain one-off.

Details

Overall: the latest tests 1 report has no real failures (22 flaky, all rescued on retry). In the latest MCP report, the cdp.spec.ts group is a real, PR-introduced regression; the other two are not.

Caused by this PR

• [chrome/chromium/msedge] › mcp/cdp.spec.ts:75 › should throw connection error and allow re-connecting (7 bots across ubuntu/macos/windows). The test asserts error: expect.stringContaining('connect ECONNREFUSED'). This PR points the CDP transport at dualStackLookup, which returns both::1 and 127.0.0.1 for localhost; with Node's default autoSelectFamily both attempts are raced. When the CDP endpoint is down, both refuse and Node raises an AggregateError — the received error is now "AggregateError: async initializeServer: …" instead of a string containing connect ECONNREFUSED, so the assertion fails on every platform. In the test-results DB this test has failed only on this PR (chore(net): use native happy eyeballs instead of the manual implementation #42240) — 2 runs, and passed the other 242, on no other SHA/PR, which is exactly what a freshly-introduced, deterministic regression looks like. To fix, match the aggregated message (assert on the nested .errors / a substring that survives aggregation) or preserve the single-address error shape for the refused case.

Pre-existing flake / infra

• [chrome] › mcp/cli-test.spec.ts:23 › debug test and snapshot (mcp-ubuntu-latest-chrome) — pre-existing flake. It fails with an empty snapshot (Expected substring: "button \"Submit\"", Received string: ""), and the same failure also hit unrelated PR feat(test runner): cancel operations on test end via default abort signal #41707 (mcp-macos-latest-firefox) — a place this PR can't be responsible for. It doesn't touch any networking path this PR changes.

Uncertain

• [firefox] › mcp/cli-config.spec.ts:130 › global config (mcp-windows-latest-firefox) — I can't prove this one either way. It timed out with an empty snapshot (Expected substring: "800x600", Received string: "") and has never failed before (0 of 753 prior runs), so I can't call it a flake by the same evidence I used above. But it exercises no networking path this PR changes, and its empty-snapshot signature matches the cli-test flake above, so infra noise is the likely story. A green re-run of that MCP job would confirm it's not the PR.

_{Triaged by the Playwright bot.}

_{Triaged by the Playwright bot - agent run}

[github-actions]

Test results for "MCP"

8101 passed, 1311 skipped

---

Merge workflow run.

[github-actions]

Test results for "tests 1"

20 flaky

⚠️ [chromium-library] › library/global-fetch.spec.ts:293 › should return security details from response `@frozen-time-library-chromium-linux`
⚠️ [chromium-library] › library/har.spec.ts:639 › should have security details `@frozen-time-library-chromium-linux`
⚠️ [chromium-library] › library/global-fetch.spec.ts:293 › should return security details from response `@chromium-ubuntu-22.04-arm-node20`
⚠️ [chromium-library] › library/har.spec.ts:639 › should have security details `@chromium-ubuntu-22.04-arm-node20`
⚠️ [chromium-library] › library/global-fetch.spec.ts:293 › should return security details from response `@realtime-time-library-chromium-linux`
⚠️ [chromium-library] › library/har.spec.ts:639 › should have security details `@realtime-time-library-chromium-linux`
⚠️ [chromium-library] › library/global-fetch.spec.ts:293 › should return security details from response `@chromium-ubuntu-22.04-node24`
⚠️ [chromium-library] › library/har.spec.ts:639 › should have security details `@chromium-ubuntu-22.04-node24`
⚠️ [chromium-library] › library/global-fetch.spec.ts:293 › should return security details from response `@chromium-ubuntu-22.04-node20`
⚠️ [chromium-library] › library/har.spec.ts:639 › should have security details `@chromium-ubuntu-22.04-node20`
⚠️ [chromium-library] › library/chromium/oopif.spec.ts:282 › should click `@chromium-ubuntu-22.04-node22`
⚠️ [chromium-library] › library/global-fetch.spec.ts:293 › should return security details from response `@chromium-ubuntu-22.04-node22`
⚠️ [chromium-library] › library/har.spec.ts:639 › should have security details `@chromium-ubuntu-22.04-node22`
⚠️ [chromium-library] › library/video.spec.ts:699 › screencast › should capture full viewport on hidpi `@chromium-ubuntu-22.04-node22`
⚠️ [firefox-library] › library/global-fetch.spec.ts:293 › should return security details from response `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-library] › library/har.spec.ts:639 › should have security details `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-page] › page/page-event-request.spec.ts:181 › should return response body when Cross-Origin-Opener-Policy is set `@firefox-ubuntu-22.04-node20`
⚠️ [webkit-library] › library/browsercontext-clearcookies.spec.ts:72 › should remove cookies by name regex `@webkit-ubuntu-22.04-node20`
⚠️ [webkit-library] › library/global-fetch.spec.ts:293 › should return security details from response `@webkit-ubuntu-22.04-node20`
⚠️ [webkit-library] › library/har.spec.ts:639 › should have security details `@webkit-ubuntu-22.04-node20`

51105 passed, 1208 skipped

---

Merge workflow run.