FE_DIVBYZERO, FE_INEXACT, FE_INVALID, FE_OVERFLOW, FE_UNDERFLOW, FE_ALL_EXCEPT (original) (raw)
| Defined in header | | | | | | | ----------------------------------------------------------------------- | -------------- | ------------- | ----------------- | | ------------- | | #define FE_DIVBYZERO /*implementation defined power of 2*/ | | (since C++11) | | | | | #define FE_INEXACT /*implementation defined power of 2*/ | | (since C++11) | | | | | #define FE_INVALID /*implementation defined power of 2*/ | | (since C++11) | | | | | #define FE_OVERFLOW /*implementation defined power of 2*/ | | (since C++11) | | | | | #define FE_UNDERFLOW /*implementation defined power of 2*/ | | (since C++11) | | | | | #define FE_ALL_EXCEPT FE_DIVBYZERO | FE_INEXACT | \ FE_INVALID | FE_OVERFLOW | \ FE_UNDERFLOW | | (since C++11) |
All these macro constants (except FE_ALL_EXCEPT) expand to integer constant expressions that are distinct powers of 2, which uniquely identify all supported floating-point exceptions. Each macro is defined only if it is supported.
The macro constant FE_ALL_EXCEPT, which expands to the bitwise OR of all other FE_*, is always defined and is zero if floating-point exceptions are not supported by the implementation.
| Constant | Explanation |
|---|---|
| FE_DIVBYZERO | pole error occurred in an earlier floating-point operation |
| FE_INEXACT | inexact result: rounding was necessary to store the result of an earlier floating-point operation |
| FE_INVALID | domain error occurred in an earlier floating-point operation |
| FE_OVERFLOW | the result of the earlier floating-point operation was too large to be representable |
| FE_UNDERFLOW | the result of the earlier floating-point operation was subnormal with a loss of precision |
| FE_ALL_EXCEPT | bitwise OR of all supported floating-point exceptions |
The implementation may define additional macro constants in to identify additional floating-point exceptions. All such constants begin with FE_ followed by at least one uppercase letter.
See math_errhandling for further details.
[edit] Example
Possible output:
1.0/0.0 = inf division by zero reported 1.0/10 = 0.1 inexact result reported sqrt(-1) = -nan invalid result reported