https://bugs.kde.org/show_bug.cgi?id=419749
--- Comment #8 from David Edmundson <k...@davidedmundson.co.uk> --- There's a chance this is surfaced by my 80d3f148e0a77f1e38dd2b79c7747e37ac67a5ed *if* kwayland called destroy() then setGeometry() before create() the receiving Qt client would delete the QWaylandScreen object, release their resource, but then process an event dispatched to their now deleted proxy object, that sequence of events would match the backtrace. If you can reproduce reliably we could put debug in KWaylandServer to make sure our global is valid when sending any output events? If that's the case it can be guarded on the client by but really we should just fix kwin sending such events. -- You are receiving this mail because: You are watching all bug changes.