HP Forums

Full Version: A Really Obscure (unknown?) Bug
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

For no good reason I thought that I'd try to see what a 17BII does when you enter a lot, a real lot, of digits. I was expecting an overflow error when I got past integers with 500 digits. But this was not the case.

The calculator accepts a seemingly infinite number of digits and when you get past 500 and hit ENTER it goes nuts and flips the sign of the exponent. So entering a string of 501 1's for example, will give you 1.11111111111E-500. (502 1's would give you 1.11111111111E-499, etc.) Aside from the obvious entry bug, this number should be an underflow condition according to the manual, but it's not and it works like it's a real number. You can take LOG or LN or SQR of it, multiply it by 10 etc.. As far as I know it's impossible to enter a number this small through normal means.

In checking the 17B, 19BII, 27S I found that they all have the same bug. Probably the 19B has the same bug and maybe some others. Of course this bug would never appear in real life, which is why HP probably never bothered to check there entry routines for this problem.

Oh yes! The 28S doesn't exhibit this behaviour. 501+ digits give a "syntax error".

I think I'd lose track long before I got to 500 pushes of the "1" key. I wonder if that means I've got a really obscure bug too....

It's not easy being a nerd. You've got work at it.