Defined in header <math.h>  

#define MATH_ERRNO 1  (since C99)  
#define MATH_ERREXCEPT 2  (since C99)  
#define math_errhandling /*implementation defined*/  (since C99) 
The macro constant math_errhandling
expands to an expression of type int
that is either equal to MATH_ERRNO
, or equal to MATH_ERREXCEPT
, or equal to their bitwise OR (MATH_ERRNO  MATH_ERREXCEPT
).
The value of math_errhandling
indicates the type of error handling that is performed by the floatingpoint operators and functions:
Constant  Explanation 

MATH_ERREXCEPT  indicates that floatingpoint exceptions are used: at least FE_DIVBYZERO , FE_INVALID , and FE_OVERFLOW are defined in <fenv.h> . 
MATH_ERRNO  indicates that floatingpoint operations use the variable errno to report errors. 
If the implementation supports IEEE floatingpoint arithmetic (IEC 60559), math_errhandling & MATH_ERREXCEPT
is required to be nonzero.
The following floatingpoint error conditions are recognized:
Condition  Explanation  errno  floatingpoint exception  Example 

Domain error  the argument is outside the range in which the operation is mathematically defined (the description of each function lists the required domain errors) 
EDOM 
FE_INVALID 
acos(2) 
Pole error  the mathematical result of the function is exactly infinite or undefined 
ERANGE 
FE_DIVBYZERO 
log(0.0) , 1.0/0.0 
Range error due to overflow  the mathematical result is finite, but becomes infinite after rounding, or becomes the largest representable finite value after rounding down 
ERANGE 
FE_OVERFLOW 
pow(DBL_MAX,2) 
Range error due to underflow  the result is nonzero, but becomes zero after rounding, or becomes subnormal with a loss of precision 
ERANGE or unchanged (implementationdefined) 
FE_UNDERFLOW or nothing (implementationdefined) 
DBL_MIN/2 
Inexact result  the result has to be rounded to fit in the destination type  unchanged 
FE_INEXACT or nothing (unspecified) 
sqrt(2) , 1.0/10.0 
Whether FE_INEXACT
is raised by the mathematical library functions is unspecified in general, but may be explicitly specified in the description of the function (e.g. rint
vs nearbyint
).
Before C99, floatingpoint exceptions were not specified, EDOM
was required for any domain error, ERANGE
was required for overflows and implementationdefined for underflows.
#include <stdio.h> #include <fenv.h> #include <math.h> #include <errno.h> #pragma STDC FENV_ACCESS ON int main(void) { printf("MATH_ERRNO is %s\n", math_errhandling & MATH_ERRNO ? "set" : "not set"); printf("MATH_ERREXCEPT is %s\n", math_errhandling & MATH_ERREXCEPT ? "set" : "not set"); feclearexcept(FE_ALL_EXCEPT); errno = 0; printf("log(0) = %f\n", log(0)); if(errno == ERANGE) perror("errno == ERANGE"); if(fetestexcept(FE_DIVBYZERO)) puts("FE_DIVBYZERO (pole error) reported"); }
Possible output:
MATH_ERRNO is set MATH_ERREXCEPT is set log(0) = inf errno = ERANGE: Numerical result out of range FE_DIVBYZERO (pole error) reported
(C99)  floatingpoint exceptions (macro constant) 
macro which expands to POSIXcompatible threadlocal error number variable (macro variable) 

C++ documentation for math_errhandling 
© cppreference.com
Licensed under the Creative Commons AttributionShareAlike Unported License v3.0.
http://en.cppreference.com/w/c/numeric/math/math_errhandling