The value used during the assertion.
The target value.
The maximum space between two approximately equal floating point numbers measured in units of least precision. A higher number means more approximation.
The error message to display.
The file name where the error occurred. The value is added automatically at the call site.
The line where the error occurred. The value is added automatically at the call site.
DUnitAssertError if the assertation fails.
float smallestFloatSubnormal = float.min_normal * float.epsilon; smallestFloatSubnormal.assertApprox(-smallestFloatSubnormal); double smallestDoubleSubnormal = double.min_normal * double.epsilon; smallestDoubleSubnormal.assertApprox(-smallestDoubleSubnormal); 0.0f.assertApprox(-0.0f); (-0.0f).assertApprox(0.0f); 0.0.assertApprox(-0.0); (-0.0).assertApprox(0.0); 2.0f.assertApprox(1.999999f); 1.999999f.assertApprox(2.0f); 2.0.assertApprox(1.999999999999999); 1.999999999999999.assertApprox(2.0); // The following tests pass but are open for debate whether or not they should. float.max.assertApprox(float.infinity); float.infinity.assertApprox(float.max); double.infinity.assertApprox(double.max); double.max.assertApprox(double.infinity); float.nan.assertApprox(float.nan); double.nan.assertApprox(double.nan); // Assert a DUnitAssertError is thrown if assertApprox fails. 10f.assertApprox(0f).assertThrow!(DUnitAssertError)("Failed asserting approximately equal");
Assert that two floating point values are approximately equal.