Comment 15 for bug 538524

Ok, this was a red herring because the code in the plymouth pre-stop job *never* matches: $UPSTART_EVENTS is not set when the job is being stopped, instead the variable set is $UPSTART_STOP_EVENTS. So the plymouth job itself is buggy, but that doesn't bring us any closer to understanding this failure (and in fact, when I query $UPSTART_STOP_EVENTS correctly, and call 'plymouth deactivate' before calling 'plymouth quit --retain-splash', the console hangs and I have to issue an Alt+SysRq+K to get it back - after which kdm *does* start up).