The specification for java.lang.Math talks about the accuracy of the methods there in terms of ulps (units in the last place). Ulps are a common way of describing relative error in floating-point functions. However, none of the Math, StrictMath, Float, and Double classes include methods in calculate the ulp of a floating-point value.
- relates to
-
JDK-4455590 JavaDoc: comments for Math, StrictMath don't provide enough detail
-
- Closed
-
-
JDK-4893344 Regression tests for java.lang.Math would benefit from minor refactoring
-
- Resolved
-
-
-
- Closed
-