chore(deps): update @atlaskit/tooltip to v22.2.0

[renovate]

This PR contains the following updates:

Package	Change	Age	Confidence
@atlaskit/tooltip (source)	22.1.0 → 22.2.0

---

Release Notes

atlassian/atlassian-frontend-mirror (@​atlaskit/tooltip)

v22.2.0

Minor Changes

• 7250582895c0b -
  Top-layer adoption work behind the platform-dst-top-layer feature flag. Public adopter APIs are
  intentionally kept narrow while the top-layer API surface settles, with one exception called out
  below.

  Highlights:

  • Pass the full [along, away] legacy popper offset through to the new top-layer
    placement.offset API (via fromLegacyPlacement). Previously only the away axis was
    forwarded, which dropped the along offset for consumers of Popup, PopupSelect,
    Spotlight, and Tooltip when platform-dst-top-layer is enabled.
  • Fix broken import of dialogHeight and dialogWidth from the removed utils module in
    @atlaskit/modal-dialog.

  Public API:

  • @atlaskit/tooltip (minor): add an optional testId?: string field to TriggerProps.
    This is additive (no existing prop changes shape). Required because @atlaskit/button/new (and
    other Pressable-backed primitives) overwrite data-testid from spread, so the legacy
    (triggerProps as any)['data-testid'] workaround is silently absorbed by those consumers. A
    typed testId field flows through their own testId destructure instead, restoring
    data-testid propagation onto the rendered trigger element.
  • @atlaskit/popup, @atlaskit/dropdown-menu (patch): no public type changes. Wider
    aria-haspopup unions that the FF-on path produces are bridged at the package boundary into
    @atlaskit/top-layer with localised FUDGE(top-layer-api) casts, documented in
    packages/design-system/top-layer/notes/decisions/migration-roadmap.md ("Open API decisions
    deferred to a follow-up PR"). They will be widened in a follow-up minor PR once the top-layer
    API is committed.
  • @atlaskit/modal-dialog, @atlaskit/select, @atlaskit/spotlight
    (patch/minor): no public type changes; bug fixes only.

  Merge-readiness fixes (FF-on test wiring + adopter behavior):

  • @atlaskit/popup (minor): wire the compositional PopupContent to delegate to
    PopupContentTopLayer when platform-dst-top-layer is enabled. Previously only the legacy
    Popup component had the FF branch, leaving consumers of the compositional API on the legacy
    popper path.
  • @atlaskit/select (minor): add an onClick handler to the PopupSelect top-layer
    trigger so clicks open/close the menu (mirrors the legacy global click handler in
    popup-select.tsx). Add explicit Escape handling on the menu's onKeyDown so the menu closes
    and focus returns to the trigger.
  • @atlaskit/top-layer (patch): the <dialog> rendered by the Dialog primitive now sets
    aria-modal="true" explicitly. Modern browsers infer modal semantics from .showModal() but
    some assistive tech still keys off the explicit attribute.
  • @atlaskit/top-layer (patch): guard use-anchor-positioning against environments where
    ResizeObserver is not defined (e.g. jest's node environment, used by the post-office test
    suite). The observer is used to wait for the popover's first valid layout before measuring;
    consumers in non-DOM jest environments now get a no-op observer and the scroll/resize listeners
    still apply if the host environment polyfills showPopover. Real browsers always have
    ResizeObserver.
  • @atlaskit/modal-dialog (patch): on the FF-on path, drop the tabIndex={-1} (and unused
    :focus-visible outline) from the modal content wrapper. The native <dialog>.showModal()
    focus-delegate algorithm picks the first focusable descendant (including tabindex=-1), and the
    wrapper was hijacking initial focus from the close button. Also honor shouldReturnFocus={ref}
    on the FF-on path (an unmount-cleanup focuses the ref after dialog.close() so it overrides the
    browser's automatic return-to-trigger). Boolean shouldReturnFocus={false} is not yet honored
    on the FF-on path — see top-layer/notes/merge-blockers.md.
  • @atlaskit/datetime-picker (patch): on the FF-on path, set mode="manual" on the
    Popup.Content rendered by both internal/menu-top-layer.tsx (date-picker calendar) and
    internal/fixed-layer-menu-top-layer.tsx (time-picker menu). With the default mode="auto",
    the same click event that opens the menu (which targets the react-select combobox input —
    outside the popover element) bubbles to the browser's native popover light-dismiss handler and
    immediately closes the menu. react-select / DateTimePicker already own outside-click and Esc
    dismissal via their own state, so opting out of the native auto-dismiss is the correct
    integration. Also extend the existing Esc → trigger-focus restoration in
    components/date-picker.tsx to the FF-on path (manual mode disables the browser's built-in
    focus return, and the legacy code path was already handling this for itself behind an FF
    negation).
  • @atlaskit/popup (no public API change): no source changes — only FF-on Playwright
    spec/example fixes drove the suite from 21/3/2 to 27/0/0. Notable: the two test.fixme'd
    nested-popover cases were not browser limitations; popover="auto" chains correctly via DOM
    ancestry (the original fixmes had the wrong testId selector). Added testId props to two
    examples (16-popup-with-a11y-props, 18-should-fit-container) so default-shape tests can
    reach the trigger.
  • Test alignment for FF-on Playwright suites across popup, select, datetime-picker,
    inline-dialog, inline-message, and modal-dialog: selector updates to match the new
    top-layer testId convention (${testId}--content, [role="dialog"][aria-label="calendar"]),
    per-spec skipAxeCheck() for example-level color-contrast violations unrelated to the
    migration, and focus assertions adjusted to match native <dialog> / Popup.Content auto-focus
    semantics (focus lands on the first focusable child, not the dialog container itself).
  • @atlassian/capacity-planning-capacity-graph, @atlaskit/color-picker,
    @atlassian/timeline-table, @atlassian/global-side-navigation (patch): scope fg
    mocks in unit tests so platform-dst-top-layer returns false. JSDOM does not implement the
    native Popover API (showPopover/hidePopover/toggle events), so leaving the gate ON in unit
    tests caused popover content to remain in the DOM after close and broke close-behaviour
    assertions. Browser coverage for the FF-on path is provided by the Playwright suites listed
    above.
  • @atlaskit/dropdown-menu (no public API change): test/example-only fixes for the FF-on
    Playwright suite. Added role="menuitem" to the nested-trigger ButtonItem in
    examples/93-testing-nested-keyboard-navigation-top-layer.tsx to satisfy axe's
    aria-required-children rule on the parent menu. Added a test.beforeEach(skipAxeCheck) to
    dropdown-menu.spec.tsx (FF-on suite) for example-level color-contrast violations on the
    pre-existing color.text.selected/color.background.selected token pair (3.91:1). Replaced a
    deadlocking await expect(moveItem).not.toBeFocused() pre-open assertion (Playwright's
    auto-wait blocks 5s on the absent element) with await expect(moveItem).not.toBeVisible().
    Suite result: 22/22 passing.

Patch Changes

• Updated dependencies

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • Between 12:00 AM and 03:59 AM (* 0-3 * * *)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[netlify]

✅ Deploy Preview for atlassify canceled.

Name	Link
🔨 Latest commit	8d8a3ce
🔍 Latest deploy log	https://app.netlify.com/projects/atlassify/deploys/6a090be5bc9de5000894851b

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud