docs(locator): drop type markup from ariaSnapshotJSON details

[Simon Knott (Skn0tt)]

The freeform Details list on locator.ariaSnapshotJSON used <[string]> / <[Object]> type markup. That syntax only gets escaped when it goes through the structured property renderer - in plain prose it lands in the MDX as-is and Docusaurus chokes on it (Unexpected character '[').

That's what's breaking the docs roll: microsoft/playwright.dev#2115

Dropped the type annotations and left the field names + descriptions.

[github-actions]

Test results for "MCP"

3 failed
❌ [chromium] › mcp/cli-core.spec.ts:37 › preserves URL with & query params @mcp-macos-latest-chromium
❌ [chromium] › mcp/http.spec.ts:105 › http transport browser lifecycle (isolated) @mcp-ubuntu-latest-chromium
❌ [firefox] › mcp/cli-keyboard.spec.ts:19 › press @mcp-windows-latest-firefox

8060 passed, 1284 skipped

---

Merge workflow run.

[github-actions]

Test results for "tests 1"

1 failed
❌ [playwright-test] › reporter.spec.ts:251 › merged › should not have internal error when steps are finished after timeout @windows-latest-node22

3 flaky

⚠️ [chromium-library] › library/popup.spec.ts:260 › should not throw when click closes popup `@chromium-ubuntu-22.04-node22`
⚠️ [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`

51235 passed, 1190 skipped

---

Merge workflow run.

[github-actions]

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

🟢 Clear — all three failures are pre-existing flakes

This PR only edits docs prose and the generated types.d.ts for locator.ariaSnapshotJSON — it can't reach the MCP CLI/HTTP tests that failed. All three have failed on unrelated SHAs in the results DB.

Details

The diff touches docs/src/api/class-locator.md and the two types.d.ts copies — comment text only, no product code path. None of the failing MCP tests exercise anything this PR changes.

Pre-existing flake / infra

• [chromium] › mcp/cli-core.spec.ts:37 › preserves URL with & query params (mcp-macos-latest-chromium) — failed 1 of 677 runs on this macOS/chromium bot, on a SHA unrelated to this PR; passed every other time.
• [chromium] › mcp/http.spec.ts:105 › http transport browser lifecycle (isolated) (mcp-ubuntu-latest-chromium) — a known bimodal flake: 16 of 682 runs failed on this bot across unrelated SHAs, and it also flips on firefox, webkit and msedge. Textbook lifecycle-timing flake.
• [firefox] › mcp/cli-keyboard.spec.ts:19 › press (mcp-windows-latest-firefox) — failed 1 of 428 runs on this windows/firefox bot on an unrelated SHA; passed every other time.

_{Triaged by the Playwright bot - agent run}

[github-actions]

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

🟢 CI is clear — all four failures are pre-existing flakes

This PR is docs-only (it drops <[string]>/<[Object]> type markup from the ariaSnapshotJSON prose and the generated types.d.ts). It touches no product code path, so none of the failing MCP or reporter tests can be reached by the diff. Every failure has a flake record on unrelated SHAs in the aggregated CI results.

Details

Overall: the diff is confined to docs/src/api/class-locator.md and the two generated types.d.ts doc comments. The failing tests exercise the MCP CLI/HTTP transport and the test-runner reporter — nothing the change reaches. Each is a known flake, so I read all four as pre-existing.

Pre-existing flake / infra

• [chromium] › mcp/cli-core.spec.ts:37 › preserves URL with & query params — flaky. Across the results DB this test has a within-run retry rescue on chromium (0 final failures across ~683 runs, 1 rescued) on SHAs unrelated to this PR.
• [chromium] › mcp/http.spec.ts:105 › http transport browser lifecycle (isolated) — heavily flaky. Its verdict flips across runs on every engine: chromium 6 failed / 677 passed (plus 10 retry rescues), webkit 3 failed, firefox 2 failed, all on unrelated SHAs.
• [firefox] › mcp/cli-keyboard.spec.ts:19 › press — flaky. Final verdict flips across runs — 1 failed / 675 passed on firefox, with matching one-off failures on chrome and webkit, none tied to this PR.
• [playwright-test] › reporter.spec.ts:251 › merged › should not have internal error when steps are finished after timeout — flaky. 0 final failures across 955 runs but 3 within-run retry rescues, i.e. it fails-then-passes intermittently regardless of the change under test.

The three tests the report already flagged as flaky (popup.spec.ts:260, page-emulate-media.spec.ts:144, page-event-request.spec.ts:181) passed on retry and need no triage.

I'm a first pass, not the final word — but with a docs-only diff and a flake history for each failure, I don't see anything caused by this PR.

_{Triaged by the Playwright bot - agent run}