Unit math operations on integer Decimal values return integer rather than Decimal

[glorieux]

Hello,

There seems to be an issue when using the math functions with Decimal values.

Here is a quick example to illustrate the problem:

> Cldr.Unit.add(Cldr.Unit.new!(:unit, Decimal.new(42)), Cldr.Unit.new!(:unit, Decimal.new(2))) 
|> Cldr.Unit.value()
44

In that case I would expect the value to be Decimal.new("44").

It seems to work fine if I introduce a float:

> Cldr.Unit.add(Cldr.Unit.new!(:unit, Decimal.new(42)), Cldr.Unit.new!(:unit, Decimal.from_float(2.2))) 
|> Cldr.Unit.value()
Decimal.new("44.2")

I've narrowed the origin of the problem to that call to Cldr.Math.maybe_integer() on every math operations.

I would expect the returns value to be a Decimal if I follow the logic from cldr_utils math functions documentation:

 Adds two numbers together.

 The numbers can be integers, floats or Decimals.
 The type of the return will be Decimal if the
  either of the arguments is a Decimal.

 If both arguments are integers, the result will
 be an integer. If either of the arguments is a float,
 the result will be a float.

Is there any reason those calls to Cldr.Math.maybe_integer() are required?
If not how do you want to proceed to remove them (I could make a pull request)?

In any case thanks for that library it already saved me so much time and headaches 🙏.