https://bugs.kde.org/show_bug.cgi?id=422282
--- Comment #32 from Méven Car <meve...@gmail.com> --- The bug was not introduced by https://invent.kde.org/system/dolphin/-/blob/ac234a9c55aed509b385ef03835a6d0f563e6a22/src/views/dolphinview.cpp, proof of that is that ac234a9c was not part of Dolphin 20.04 (https://invent.kde.org/system/dolphin/-/commits/v20.04.0) in then current TumbleWeed version when the bug was originally reported. The bug was preexisting but the commit made it more likely because now slotItemDeleted can be called later because of the I/O added now delays slotItemDeleted. This should be happening more frequently on network filesystem for instance. The bug origin is that KFileItemModel::slotItemDeleted called due to a KCoreDirLister event, can be called before KFileItemModel::setExpanded is finished running, causing slotItemDeleted trying to do a double free. Typical race condition. There is a risk of removing the wrong items while fixing this. Code mentions https://bugs.kde.org/show_bug.cgi?id=332102#c11 that experienced the same kind of issue. This happens when you close an expanded folders whose content is being removed if I understand correctly. I experienced it myself, but it is not trivial to reproduce. -- You are receiving this mail because: You are watching all bug changes.
