Skip to main content

floating-point-overflow

floating-point-overflow Condition Type

Class Precedence List:

floating-point-overflow, arithmetic-error, error, serious-condition, condition, t

Description:

The type floating-point-overflow consists of error conditions that occur because of floating-point overflow.

Expanded Reference: floating-point-overflow​

Type Hierarchy​

The floating-point-overflow condition is a subtype of arithmetic-error. It is signaled when a floating-point operation produces a result that is too large to be represented.

(subtypep 'floating-point-overflow 'arithmetic-error)
=> T
=> T
(subtypep 'floating-point-overflow 'error)
=> T
=> T

Triggering Overflow​

Overflow occurs when the result of a floating-point computation exceeds the maximum representable value for that float type.

(handler-case (exp 1000.0d0)
  (floating-point-overflow ()
    :overflow))
=> :OVERFLOW

(handler-case (* most-positive-double-float 2.0d0)
  (floating-point-overflow ()
    :overflow))
=> :OVERFLOW

Inspecting the Error​

Use the arithmetic-error-operation and arithmetic-error-operands accessors to determine what caused the overflow.

(handler-case (exp 1000.0d0)
  (floating-point-overflow (c)
    (list :op (arithmetic-error-operation c)
          :args (arithmetic-error-operands c))))
;; => impl-dependent

Safe Computation Pattern​

A common pattern is to catch overflow and substitute a sentinel value such as the maximum representable float.

(defun safe-exp (x)
  (handler-case (exp x)
    (floating-point-overflow ()
      most-positive-double-float)))

(safe-exp 1.0d0)
=> 2.718281828459045d0
(safe-exp 1000.0d0)
=> 1.7976931348623157d308