feclearexcept - Linux

Overview

feclearexcept clears the floating-point exceptions flags: FE_ALL_EXCEPT (or FE_INVALID, FE_DIVBYZERO, FE_OVERFLOW and FE_UNDERFLOW). It is commonly used to reset the flags before performing floating-point calculations to ensure a clean slate.

Syntax

feclearexcept(int excepts)

Options/Flags

The excepts parameter is a bitmask of the exceptions to clear. It can be any combination of the following values:

• FE_INVALID (1): Invalid operation exception
• FE_DIVBYZERO (2): Division by zero exception
• FE_OVERFLOW (4): Overflow exception
• FE_UNDERFLOW (8): Underflow exception
• FE_ALL_EXCEPT (15): All of the above exceptions

Examples

Clear all exceptions:

#include <fenv.h>

int main() {
  feclearexcept(FE_ALL_EXCEPT);
  return 0;
}

Clear only the division by zero exception:

#include <fenv.h>

int main() {
  feclearexcept(FE_DIVBYZERO);
  return 0;
}

Common Issues

Error handling: If feclearexcept fails to clear the exceptions, it sets the FE_INVALID exception and returns -1. Check the return value to ensure successful clearing of exceptions.

Integration

Using feclearexcept with floating-point precision:

#include <fenv.h>
#include <math.h>

int main() {
  float x = 1e20;
  float y = 1e-20;

  feclearexcept(FE_ALL_EXCEPT);
  float result = x * y;

  if (fetestexcept(FE_UNDERFLOW)) {
    printf("Underflow occurred during calculation.\n");
  }
  return 0;
}

Related Commands

• fegetexcept: Get the floating-point exceptions flags.
• fesetexcept: Set the floating-point exceptions flags.
• raise: Raise a floating-point exception.