asked on

Borland 4.02 Floating Point Overflow Error

I'm new to C++ programming and recently purchased the C++ Starter Kitfrom Sam's publishing for a C++ class I'm taking. The "kit" comes with the Borland C++ compiler v. 4.02, a book and a bunch of info on CD-ROM. The problem is I'm having trouble entering real numbers into float variables. For example:

#include <iostream.h>

void main ()
{
float z;
cout << "Enter a real number: ";
cin >> z;
cout << "The nuber is: " << z << endl;
}

The above program will compile and even run as long as an integer is entered. But, if a real number is entered (ie. 1.2) a run-time error "Floating Point: Overflow" occurs. I can assign real numbers in the program, but not enter them from the terminal. Even more strange is that I installed the compiler on my laptop and the program runs fine.
The technical support for SAMs publishing have no idea what's going on and I think Borland charges for technical support. I suspect that there is a problem with the compiler setup. Can anyone out there helpme. My desktop has a Cyrix 6x86-100 (P120+) and the laptop an Intel
P120.

Thanks,
Mark

AVaulin

Try use double z;

ASKER CERTIFIED SOLUTION

8051

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

8051

OK, go menu->options->compiler->advanced code generation.
1. Floating Point (*) 287/387
2. Instruction Set (*) 386
3. Options [v] Fast floating point

It works perfect.

Regards

mshuman

ASKER

My compiler (Borland 4.02) doesn't have the same options that yours does. But, I've tried every combination of CPU and floating point options there are. Also, I've been doing some research... The error only occurs if I compile the program as an EasyWin executable, if I compile it as a standard DOS executable the program executes without the error. Another strange thing is that I have compiled the program as an EasyWin executable and have run the program on a number of other computers running either Win 3.1 or Win 95 and all have run the program without a problem -- though I cannot run it on the computer on which it was compiled. Again the error only occurs if a real number is entered. If I enter an integer, the program works.

Mark

mshuman

ASKER

Oh well, I figured out the problem myself. Here's how, because I don't think anyone could have done this without being at my computer.
1. Started Win 95 in Safe Mode - program ran without error
2. Started Win 95 by with step-by-step confirmation
a. No registry load - program had error
b. No config.sys - program had error
c. No windows drivers - no error
3. Start Windows 95 in normal mode and ran program after removing programs from memory (cntrl-alt-del to see programs)...
Program ran without error after removing em_exec.exe (Logitech trackball driver), started em_exec.exe again -- program had error.
4. Therefore program somehow conflicts with Logitech Trackball Driver -- I'm going to contact Logitech now.

Thanks for your help,

Mark

dclane

In Floating point ( and all variables ) you must 'scope' your variables if that does not work. Try Math control functions.

DCLANE