Crash in ServicesOverlayController::~ServicesOverlayController No need to uninstall the PageOverlay, because in WebPage teardown the PageOverlay has already been destroyed. <rdar://problem/17622172>
Created attachment 235090 [details] Patch v1
Comment on attachment 235090 [details] Patch v1 View in context: https://bugs.webkit.org/attachment.cgi?id=235090&action=review > Source/WebKit2/WebProcess/WebPage/mac/ServicesOverlayController.mm:-86 > - m_webPage->uninstallPageOverlay(m_servicesOverlay, PageOverlay::FadeMode::DoNotFade); you should assert that your back pointer was nulled out (and let it be nulled out! fix the early return in willMoveToWebPage).
(In reply to comment #2) > (From update of attachment 235090 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=235090&action=review > > > Source/WebKit2/WebProcess/WebPage/mac/ServicesOverlayController.mm:-86 > > - m_webPage->uninstallPageOverlay(m_servicesOverlay, PageOverlay::FadeMode::DoNotFade); > > you should assert that your back pointer was nulled out (and let it be nulled out! fix the early return in willMoveToWebPage). Nevermind!
(In reply to comment #2) > (From update of attachment 235090 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=235090&action=review > > > Source/WebKit2/WebProcess/WebPage/mac/ServicesOverlayController.mm:-86 > > - m_webPage->uninstallPageOverlay(m_servicesOverlay, PageOverlay::FadeMode::DoNotFade); > > you should assert that your back pointer was nulled out (and let it be nulled out! fix the early return in willMoveToWebPage). Clarified on IRC - Tim misread the early return - It's correct. But he also was assuming that willMoveToWebPage was called with a null WebPage on WebPage teardown. It's not. Filed https://bugs.webkit.org/show_bug.cgi?id=135024 for that
http://trac.webkit.org/changeset/171194