fix(webkit): do not override Accept-Language header set by the page

[yury-s]

Fixes #23732

[github-actions]

Test results for "MCP"

7760 passed, 1249 skipped

---

Merge workflow run.

[github-actions]

Test results for "tests 1"

1 failed
❌ [webkit-page] › page/page-leaks.spec.ts:164 › waitFor should not leak @webkit-ubuntu-22.04-node20

4 flaky

⚠️ [chromium-library] › library/video.spec.ts:699 › screencast › should capture full viewport on hidpi `@realtime-time-library-chromium-linux`
⚠️ [firefox-page] › page/page-emulate-media.spec.ts:144 › should keep reduced motion and color emulation after reload `@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`
⚠️ [playwright-test] › ui-mode-trace.spec.ts:827 › should update state on subsequent run `@windows-latest-node22`

50214 passed, 1190 skipped

---

Merge workflow run.

[github-actions]

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

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

[webkit-page] › page/page-leaks.spec.ts:164 › waitFor should not leak failed once on this PR, but it flakes the same way on other PRs and on main. Nothing to do with the Accept-Language change.

Details

Overall: the "tests 1" report shows 1 failed + 4 flaky; the rest of the matrix (incl. MCP) is green. The single failure doesn't reproduce reliably and hits code this PR never touches.

Pre-existing flake / infra

• [webkit-page] › page/page-leaks.spec.ts:164 › waitFor should not leak — pre-existing flake, WebKit-Linux only. Across the test-results DB this test is bimodal on webkit-page: failed 7 of 460 runs (1.5%), passed 453, all on webkit-ubuntu-22.04-node20. Six of those seven failures are on runs this PR can't be responsible for — PRs #41884, #41889, #41888, and two push/main runs (ecf2e532, 40372734) — over the same day window. This PR only changes WebKit Accept-Language header handling (wkPage.ts) and a browsercontext-locale test; the leak/GC path this test asserts on isn't exercised by that change.

The 4 flaky entries (video hidpi, emulate-media reload, request COOP body, ui-mode-trace) all passed on retry and are unrelated to the diff — no triage needed.

_{Triaged by the Playwright bot - agent run}