Bug 241022 – AX: inert attribute doesn't cause display:contents element to be ignored

• Component: Accessibility
• Status: RESOLVED FIXED

[Tyler Wilcock]

AX: inert attribute doesn't cause display:contents element to be ignored

[Radar WebKit Bug Importer]

<rdar://problem/94034984>

[Tyler Wilcock]

Created attachment 459814 [details]
Patch

[Tyler Wilcock]

Created attachment 459914 [details]
Patch

[Andres Gonzalez]

(In reply to Tyler Wilcock from comment #3)
> Created attachment 459914 [details]
> Patch

--- a/Source/WebCore/accessibility/AccessibilityObject.cpp
+++ a/Source/WebCore/accessibility/AccessibilityObject.cpp

+const RenderStyle* AccessibilityObject::style() const

Doesn't this belong to the AXNodeObject since it requires an underlying element? I.e., AccessibilityObject::style should always return nullptr, and AccessibilityNodeObject::style should be this implementation?

--- a/LayoutTests/accessibility/node-only-inert-object.html
+++ a/LayoutTests/accessibility/node-only-inert-object.html

+            await waitFor(() => accessibilityController.accessibleElementById("button") );

Is the space before the last ) needed?

[Tyler Wilcock]

Created attachment 459915 [details]
Patch

[Tyler Wilcock]

(In reply to Andres Gonzalez from comment #4)
> (In reply to Tyler Wilcock from comment #3)
> > Created attachment 459914 [details]
> > Patch
> 
> --- a/Source/WebCore/accessibility/AccessibilityObject.cpp
> +++ a/Source/WebCore/accessibility/AccessibilityObject.cpp
> 
> +const RenderStyle* AccessibilityObject::style() const
> 
> Doesn't this belong to the AXNodeObject since it requires an underlying
> element? I.e., AccessibilityObject::style should always return nullptr, and
> AccessibilityNodeObject::style should be this implementation?
I don't think so, since it uses AccessibilityObject::element() to get the style. Non AXNodeObjects can and do provide a node() / element() / style(). For example, AccessibilityListBoxOption is not an AXNodeObject but provides a node().

Another theoretical example: AccessibilityScrollView isn't a node object, but could provide a node() / element() through AccessibilityScrollView::m_frameOwnerElement.

So I think letting different AccessibilityObject subclasses provide node() / element() / style() however they can is OK.

> --- a/LayoutTests/accessibility/node-only-inert-object.html
> +++ a/LayoutTests/accessibility/node-only-inert-object.html
> 
> +            await waitFor(() =>
> accessibilityController.accessibleElementById("button") );
> 
> Is the space before the last ) needed?
It isn't — fixed.

[Tyler Wilcock]

Created attachment 459916 [details]
Patch

[EWS]

Committed r295090 (251185@main): <https://commits.webkit.org/251185@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 459916 [details].