So really, the above is what triggers the bug, but I think the real bug is that when udev-acl later gets called to update the permissions, it simply keeps the existing group acl instead of updating it for the new MODE=0660 in the new udev rule.
So really, the above is what triggers the bug, but I think the real bug is that when udev-acl later gets called to update the permissions, it simply keeps the existing group acl instead of updating it for the new MODE=0660 in the new udev rule.