Hi Everybody,
A couple of times recently, I had reports of error #9 in wret.exe (often seen while running autolearn.exe). The errors had started recently, and wouldn’t go away. It turned out (for the two users I know of) to have been a very simple problem with a very simple solution.
When I wrote wret.exe, I dimensioned several variables, representing the “number” of the forecast, up to 9999. It didn’t occur to me that anyone would have 10,000 forecasts archived, but that has happened with some users now! If yu make 20 forecasts a day (I think that’s a lot!), it happens in less than a year and a half!
What you can do is just move the older forecasts (ending in .wxf) into another folder. Autolearn should not be using anything older than 4 months or so anyway (it can optionally use data from a year ago, but that’s already analyzed results stored in correc.txt). I plan to come up with better messaging, at least. I think you’ll find autolearn/wret to run faster when it doesn’t have to look at all those really old forecasts.
I do NOT recommend deleting them, as they could be useful in some ways later on. Just store them somewhere separate.
Tom