relates to:

• SR-14974 Decimal.ulp produces inaccurate results

Issue Description:

The documentation for Decimal's .nextUp says: "The least representable value that is greater than this decimal." Likewise, Decimal's .nextDown is described as: "The greatest representable value that is less than this decimal."

Perhaps I'm misunderstanding what this means (quite possible!), but that's not at all what these properties seem to do:

> Decimal(2).nextUp
$R1: Decimal = 3.000000

It's not just my computer being weird, either. There's unit tests which check this: https://github.com/apple/swift/blob/a73a8087968f9111149073107c5242d83635107a/test/stdlib/TestDecimal.swift#L302-L303

Are there not many other representable Decimal values greater than Decimal(2) and less than Decimal(3)? It's just adding 1.0.

The other FloatingPoint types have the same documentation string, and seem to work as advertised:

> Double(2).nextUp
$R2: Double = 2.0000000000000004