[Webkit-unassigned] [Bug 231914] AX: Any addition of children should funnel through AccessibilityObject::addChild

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=231914

--- Comment #16 from Tyler Wilcock <tyler_w at apple.com> ---
> I’ve seen this assert to be useful in the past when we mess up the parent
> child hierarchy for mock elements or other kinds of custom sub classes
> Is it asserting now in your tests?

It is, yeah — specifically for anything table related (parent or child), and for autofill buttons. If it's been useful in the context of mock objects, maybe we can keep it, but would need the conditionals around the ASSERT I added in https://bugs.webkit.org/attachment.cgi?id=441675&action=diff.

It seems we've experienced issues in the past with table parent-child relationships — see AccessibiltyTableRow::parentTable():

AccessibilityTable* AccessibilityTableRow::parentTable() const
{
    // The parent table might not be the direct ancestor of the row unfortunately. ARIA states that role="grid" should
    // only have "row" elements, but if not, we still should handle it gracefully by finding the right table.
    for (AccessibilityObject* parent = parentObject(); parent; parent = parent->parentObject()) {
    ...truncated...
}

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20211019/890ea387/attachment-0001.htm>