Bug: Wrong conversion ToUint64

[tdrozdovsky]

...
inValue = "18446744073709551615"
log.Printf("inValue = %v", inValue )
outValue := cast.ToUint64(inValue )
log.Printf("outValue = %v", outValue )
...

Result for v1.4.1:

INFO[2022-06-06T14:01:00+03:00] inValue = 18446744073709551615               
INFO[2022-06-06T14:01:00+03:00] outValue = 18446744073709551615

Result for v1.5.0:

INFO[2022-06-06T14:01:00+03:00] inValue = 18446744073709551615               
INFO[2022-06-06T14:01:00+03:00] outValue = 0

[abukosek]

The problem appears to be in the ToUint* functions and was introduced in commit 2b0eb0f724e320b655240e331aef36d1175986c2.

Previously, the code was:

	case string:
		v, err := strconv.ParseUint(s, 0, 0)
		if err == nil {
			return uint(v), nil
		}
		return 0, fmt.Errorf("unable to cast %#v to uint: %s", i, err)

But in that commit it was replaced with:

	case string:
		v, err := strconv.ParseInt(trimZeroDecimal(s), 0, 0)
		if err == nil {
			if v < 0 {
				return 0, errNegativeNotAllowed
			}
			return uint(v), nil
		}
		return 0, fmt.Errorf("unable to cast %#v of type %T to uint", i, i)

All ToUint* functions have this bug (bad copy&paste?).

The previous code should be restored -- the only new addition is trimZeroDecimal(s), which can be added to the old code as well. The check for negative values is also not needed if using ParseUint as before.