In https://bugs.webkit.org/show_bug.cgi?id=235246, we added: else if (attrName == disabledAttr) postNotification(element, AXObjectCache::AXDisabledStateChanged); So that we could update the isolated tree when this happens. However, posting a notification here is unnecessary work, since we can just call updateIsolatedTree directly instead.
<rdar://problem/87947110>
Created attachment 449770 [details] Patch
Comment on attachment 449770 [details] Patch I’ll take your word that this was not required. Is there some discipline where we can verify such things with tests, not just with knowledge of the code?
I'm not certain that a notification is not required. Clients may need to know that something is now disabled or not. Even if some clients are not using it at the moment, we should post the notification for API consistency.