Exception Handling and Math Errors

[Joe Hesse]

What do I have to do to the following program to catch the runtime error?
Many thanks,
Joe Hesse

------------------------------------------------
#include <iostream>
#include <stdexcept>
using namespace std;

void f()
{
int x = 1, y = 0, z;
z = x/y; // divide by zero
}

int main()
{
try
{
f(); // function has a runtime error
}
catch(const runtime_error &re)
{
cerr << re.what() << "\n";
}
catch(...)
{
cerr << "Caught ..." << "\n";
}

return 0;
}

------------------------------------------------

 

[Alf P. Steinbach]

What do I have to do to the following program to catch the runtime error?

AFAIK there's no fully platform-independent way to do that. One "nearly"
platform-independent way is to use a C SIGFPE signal (see 'signal') to set
a flag somewhere, which you can then check after the arithmetic. But in
practice this also involves platform-dependent code, so you might as well
go all the way in that direction -- just try to wrap it nicely up.

Note: you can use std::numeric_limits<double> to check what kind of support
your C++ implementation has for various floating point functionality.