Library uses float which causes rounding errors

[Talljoe]

Describe the bug

The Go library uses float for monetary values. Float is widely discouraged as it does not contain the necessary precision to handle currency values.

To Reproduce

Fetch a credit invoice of $-79.99. Multiply by 100.0 and convert to an `int:

// invoice.Subtotal = -79.99
x = int(invoice.Subtotal * 100.0)
assert.equal(-7999, x)

Error: Not equal:
       expected: -7999
       actual  : -7998

Expected behavior

Library should use a Decimal or other fixed-point

Your Environment

• Which version of this library are you using? v3.8.0 f4d23f429fa01778f037d3b88c04c380f8fd745d
• Which version of golang are you using? 1.14.9

[douglasmiller]

Thank you for reporting this @Talljoe. We are aware of this limitation and are investigating solutions that we can employ in our upcoming 4.x client release. Updating the 3.x client would introduce a breaking change to existing implementations, unfortunately. We will leave this issue open and will provide updates when they become available.

[speza]

Have there been any updates on this @douglasmiller?

[wellsync-matt-may]

This was 4 years ago and it's still not fixed in the product? I just submitted a support ticket. This issue results in incorrect billing calculations.