[cixzhang]

Maintainer impact

The release-time accessibility contracts can go red on a correct Tooltip. In CI, TooltipTallContent.a11y.browser.spec.ts › "a wheel over a tooltip that fits still scrolls the page" failed with the tooltip no longer in the page at its final check. A red release check on a healthy commit holds the release.

Intended invariant

The contract is unchanged: a wheel over a tooltip that fits scrolls the page instead of being captured, the wheel does not dismiss the tooltip, and the tooltip has no scroll range of its own (spec:AST-025 FR14). Tooltip FR3 also still holds: once the pointer is over neither the trigger nor the tooltip, it closes after the hover bridge.

Change

The wheel scrolls the page 300 px, which carries the trigger and the tooltip out from under the resting pointer. That is a leave, so the tooltip closes about 100 ms later, as FR3 specifies. The test checked "still open" after it had polled for the scroll, so it only passed when that check ran before the hover bridge expired. On a slow runner it doesn't.

The test now records the tooltip's state at the page's first scroll event: the moment the wheel has been passed on to the page, before the pointer has left. The checks are the same (open, and no scroll range of its own), read at the point where they answer the contract's question. Test only: no product code and no Changeset.

Evidence

All runs were in Chromium against the Storybook built for the failing commit, with Playwright 1.61.1.

• Failure before the change: 6/40 failures with the CPU saturated, with the same error and line as CI; 20/20 passes at normal load. A probe showed that after the wheel the trigger is above the viewport and the tooltip has moved beside it, away from the pointer, and that the tooltip closes about 36 ms after the old check ran.
• Success after the change: 40/40 under CPU saturation and 40/40 at normal load. The whole file passes repeated runs.
• Product/runtime behavior verified unchanged: no source change. Closing after the pointer leaves is the specified FR3 behavior.

Scope

• No intended public API, product behavior, visual direction, or policy change.
• Product changes discovered during the work were removed or split.
• Public text and artifacts contain no internal Meta context.

Testing

• pnpm exec playwright test packages/core/src/Tooltip/__tests__/TooltipTallContent.a11y.browser.spec.ts --project chromium -g "fits still scrolls" --repeat-each=40 --fully-parallel, at normal load and under CPU saturation, before and after.
• prettier --check and eslint on the file are clean.

Credit: @ejhammond wrote this contract in #8254. This change only reads its "still open" check at the moment the wheel reaches the page.

Posted by Cindy's agent on her behalf.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
astryx	Ready	Preview	Oct 30, 2026 5:41am UTC

[astracat-bot]

AI review status for this pull request.

Review status	Updated
🟢 Reviewed (for maintainers only)	Oct 30, 2026, 6:04 AM UTC

[astracat-bot]

Reading the tooltip's state at the first scroll event is a sound fix: it takes the check out of its race with the hover-bridge close that the scroll itself triggers, and the page-scroll and scroll-range assertions are unchanged. The updated test passed in Chromium and WebKit, including when a late check made the old version fail.

One optional follow-up: in Chromium this test alone no longer fails if the wheel closes the tooltip a task later (the tall-tooltip wheel tests still catch that). Wheeling only a few pixels keeps the surface under the pointer, so a check after the bridge stays deterministic and keeps that coverage here.

[Automated review]