There are websites and embedded content that should be calling the Storage Access API but currently are not. We should add the capability to call the API.
<rdar://problem/64549429>
Correction: We should add the capability to call the API *on their behalf*.
Created attachment 402357 [details] Patch
Created attachment 402520 [details] Patch
Comment on attachment 402520 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=402520&action=review This seems ok to me. > Source/WebCore/dom/DocumentStorageAccess.cpp:111 > + // The existence of a frame and page has been checked in requestStorageAccessQuickCheck(). This might be true now, but it might not be true with future use of this code. I think it's a good idea to leave the null checks in place, maybe with assert_not_reached.
(In reply to Alex Christensen from comment #5) > Comment on attachment 402520 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=402520&action=review > > This seems ok to me. Thanks for the review! > > Source/WebCore/dom/DocumentStorageAccess.cpp:111 > > + // The existence of a frame and page has been checked in requestStorageAccessQuickCheck(). > > This might be true now, but it might not be true with future use of this > code. I think it's a good idea to leave the null checks in place, maybe > with assert_not_reached. Right now I RELEASE_ASSERT which should force us to find the problem. Would you prefer an early return with just a debug assert?
Created attachment 402526 [details] Patch
Waiting for EWS to make sure I fixed the Windows build issue.
Committed r263383: <https://trac.webkit.org/changeset/263383> All reviewed patches have been landed. Closing bug and clearing flags on attachment 402526 [details].
Comment on attachment 402526 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=402526&action=review > Source/WebCore/dom/Element.cpp:366 > + if (hasClass() && Quirks::StorageAccessResult::ShouldCancelEvent == document().quirks().triggerOptionalStorageAccessQuirk(eventType, classNames())) > + return false; This is such a weird place to add a quirk.