Created attachment 338543 [details] Backtrace The 2.20.1 update just hit Fedora. Here is a strange crash report; I have never seen anything like before, except perhaps for bug #183348. As usual, a better backtrace is attached. Thread no. 1 (10 frames) #0 WTFCrash at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/WTF/wtf/Assertions.cpp:271 #1 WebKit::WebProcess::ensureWebToStorageProcessConnection at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/WebKit/WebProcess/WebProcess.cpp:1205 #2 WebKit::WebDatabaseProvider::idbConnectionToServerForSession at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/WebKit/WebProcess/Databases/WebDatabaseProvider.cpp:82 #3 WebCore::Page::idbConnection at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/WebCore/page/Page.cpp:2245 #4 WebCore::Document::idbConnectionProxy at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/WebCore/dom/Document.cpp:3155 #5 WebCore::DOMWindowIndexedDatabase::indexedDB at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/WebCore/Modules/indexeddb/DOMWindowIndexedDatabase.cpp:112 #6 WebCore::jsDOMWindowIndexedDB at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/x86_64-redhat-linux-gnu/DerivedSources/WebCore/JSDOMWindow.cpp:7923 #7 JSC::PropertySlot::customGetter at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/JavaScriptCore/runtime/PropertySlot.cpp:50 #8 JSC::PropertySlot::getValue at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/JavaScriptCore/runtime/PropertySlot.h:408 #9 JSC::JSValue::get at /usr/src/debug/webkitgtk4-2.20.1-1.fc27.x86_64/Source/JavaScriptCore/runtime/JSCJSValueInlines.h:812
Yes, we would probably need the same approach in ensureWebToStorageProcessConnection(). The web process might be calling this right after the UI process has closed the connection.
*** This bug has been marked as a duplicate of bug 183348 ***