> Seriously though, it would be great to be able to reproduce it. In
> this case it was obvious that the size was wrong, but it may be more
> usual that it's just lower than it should be, so no one would ever
> notice.
>
> I wonder if saving the session would allow Kim to analyse the file and
> debug it if it happened again.
I've encountered half a dozen or so instances of cumulative file sizes < 0 in the Statistics Panel over the last year or so; the last one a few weeks ago. I can't reproduce them nor predict them, so I haven't bothered to report them. But, they do happen.
Even rarer are ZTreeWin product crashes. I can't reproduce nor predict those either. But unfortunately they do happen, albeit very rarely.