The feclearexcept routine clears the floating-point exception flags specified by excepts, whereas feraiseexcept raises the specified exceptions. Raising an exception causes the corresponding flag to be set, and a SIGFPE is delivered to the process if the exception is unmasked.
The fetestexcept function determines which flags are currently set, of those specified by excepts.
The fegetexceptflag function stores the state of the exception flags specified in excepts in the opaque object pointed to by flagp. Similarly, fesetexceptflag changes the specified exception flags to reflect the state stored in the object pointed to by flagp. Note that the flags restored with fesetexceptflag must be a (not necessarily proper) subset of the flags recorded by a prior call to fegetexceptflag.
For all of these functions, the possible types of exceptions include those described in fenv(3). Some architectures may define other types of floating-point exceptions.
On some architectures, raising an overflow or underflow exception also causes an inexact exception to be raised. In these cases, the overflow or underflow will be raised first.
The fegetexceptflag and fesetexceptflag routines are preferred to fetestexcept and feraiseexcept, respectively, for saving and restoring exception flags. The latter do not re-raise exceptions and may preserve architecture-specific information such as addresses where exceptions occurred.
The feclearexcept, fegetexceptflag, feraiseexcept, and fesetexceptflag functions return 0 upon success, and non-zero otherwise. The fetestexcept function returns the bitwise OR of the values of the current exception flags that were requested.
sigaction(2), feholdexcept(3), fenv(3), feupdateenv(3), fpgetsticky(3), fpresetsticky(3)