StringFixed fixed rounding does not work as expected for fixed numbers

[matejsp]

StringFixed should return with fixed precision but it is not working for fixed numbers as expected.

t.Run("StringFixed", func(t *testing.T) {
    a1 := udecimal.MustParse("10")
    a2 := udecimal.MustParse("1")
    a3 := udecimal.MustParse("1.1")
    a4 := udecimal.MustParse("1.11")

    assert.Equal(t, a1.StringFixed(3), "10.000") // returns 10 ...
    assert.Equal(t, a2.StringFixed(3), "1.000") // return 1
    assert.Equal(t, a3.StringFixed(3), "1.100") // return 1.100
    assert.Equal(t, a4.StringFixed(3), "1.110") // return 1.110
})

[quagmt]

Thanks for reporting this. Looks like I missed this case. However, please notice that this StringFixed version behaves a little bit different than shopspring/decimal. In shopspring/decimal, StringFixed will round the result before printing, while this library just keep the result as it is (any rouding should be performed before calling this method)

   // shopspring/decimal
   a := decimal.RequireFromString("1.555")
   fmt.Println(a.StringFixed(2)) // 1.56

   // udecimal
   b := udecimal.MustParse("1.555")
   fmt.Println(b.StringFixed(2)) // 1.555

[matejsp]

Yes I noticed that I need to round before and that's fine. Another difference I noticed that shopspring supports negative precision (so you can round to -1 => 10s or -2 => 100s). But your round supports only rounding on decimal part.

// StringFixed returns a rounded fixed-point string with places digits after
// the decimal point.
//
// Example:
//
//  NewFromFloat(0).StringFixed(2) // output: "0.00"
//  NewFromFloat(0).StringFixed(0) // output: "0"
//  NewFromFloat(5.45).StringFixed(0) // output: "5"
//  NewFromFloat(5.45).StringFixed(1) // output: "5.5"
//  NewFromFloat(5.45).StringFixed(2) // output: "5.45"
//  NewFromFloat(5.45).StringFixed(3) // output: "5.450"
//  NewFromFloat(545).StringFixed(-1) // output: "550"
func (d Decimal) StringFixed(places int32) string {
    rounded := d.Round(places)
    return rounded.string(false)
}

[quagmt]

All rounding methods don't support negative value. I'm not sure if rounding to power of 10 is necessary or not. Might be an improvement for v2 if it's actually needed.