The websiteDataStore.deviceIdHashSaltStorage NULL check in UserMediaPermissionRequestManagerProxy::enumerateMediaDevicesForFrame is in the wrong place.
<rdar://problem/47179650>
Created attachment 358825 [details] Patch
Comment on attachment 358825 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=358825&action=review > Source/WebKit/UIProcess/UserMediaPermissionRequestManagerProxy.cpp:464 > + return; Can we make websiteDataStore return a reference?
Created attachment 358841 [details] Patch
Comment on attachment 358841 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=358841&action=review > Source/WebKit/UIProcess/UserMediaPermissionRequestManagerProxy.cpp:459 > + return; Do we need that check here?
Created attachment 358842 [details] Patch
Comment on attachment 358841 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=358841&action=review >> Source/WebKit/UIProcess/UserMediaPermissionRequestManagerProxy.cpp:459 >> + return; > > Do we need that check here? It is inside of an asynchronous completion handler, so I assume it is possible for the page to have been invalidated.
Comment on attachment 358842 [details] Patch Clearing flags on attachment: 358842 Committed r239850: <https://trac.webkit.org/changeset/239850>
All reviewed patches have been landed. Closing bug.