I agree. The problem actually appears to be that, while processing of an event is held off when the associated file is still changing, these changes still generate new events that are also queued. So when a file finally stops changing suddenly a whole bunch of queued events for that file become ready for processing.
I agree. The problem actually appears to be that, while processing of an event is held off when the associated file is still changing, these changes still generate new events that are also queued. So when a file finally stops changing suddenly a whole bunch of queued events for that file become ready for processing.