[Webkit-unassigned] [Bug 261680] AX: columnHeaders() and rowHeaders() are performed on the main thread in ITM

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=261680

--- Comment #8 from Tyler Wilcock <tyler_w at apple.com> ---
Comment on attachment 467748
  --> https://bugs.webkit.org/attachment.cgi?id=467748
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=467748&action=review

> Source/WebCore/accessibility/AXObjectCache.cpp:2444
> +        postNotification(element, AXObjectCache::AXScopeChanged);

Because we are inside AXObjectCache, I don't think we need the AXObjectCache:: prefix here.

> Source/WebCore/accessibility/AccessibilityObjectInterface.h:1789
> +inline bool AXCoreObject::isTableCellInSameRowGroup(AXCoreObject* otherTableCell)
> +{
> +    return rowGroupAncestorID() ? rowGroupAncestorID() == otherTableCell->rowGroupAncestorID() : false;
> +}

Seems like we unconditionally dereference otherTableCell. I think we either need a null-check or to pass a reference instead of a pointer.

Also, since rowGroupAncestorID() does a lot of iteration, let's compute it only once. e.g.:

AXID ancestorID = rowGroupAncestorID();
return ancestorID ? ancestorID == otherTableCell->rowGroupAncestorID() : false;

> Source/WebCore/accessibility/AccessibilityObjectInterface.h:1798
> +    auto colRange = columnIndexRange();
> +    auto otherColRange = tableCell->columnIndexRange();
> +
> +    if (colRange.first <= (otherColRange.first + otherColRange.second))
> +        return true;
> +    return false;

I know you didn't originally write this code, but let's bring it up to our style standards by fully spelling out columnRange vs. colRange, same for otherColRange.

-- 
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/20230919/295007b0/attachment.htm>