func (d Decimal) MustFloat64()

[felipeneuwald]

Implementes func (d Decimal) MustFloat64().
Resolves #204.

[mwoss]

Thanks for your contribution @felipeneuwald!
I was thinking for a moment about your PR for a bit and I'm not convinced whether the MustFloat method should be a part of the decimal library. I'm not sure this wrapper Method would be valuable for this library. Maybe it's just me, not sure :<

Note: Unit tests are missing.

What do you think? @njason

[felipeneuwald]

Hey guys,

There are cases that I don't mind if the returned float64 value is exact or not, and I want to assign the float64 value directly into a struct using a single line of code. See:

package main

import (
	"fmt"

	"github.com/shopspring/decimal"
)

type someType struct {
	a float64
	b float64
}

func main() {
	data1 := someType{
		a: float64(1),
		b: decimalToFloat64(decimal.NewFromInt(2)),
	}
	fmt.Printf("%+v\n", data1)

	myDecimal := decimal.NewFromInt(4)
	myFloat, _ := myDecimal.Float64()
	data2 := someType{
		a: float64(3),
		b: myFloat,
	}
	fmt.Printf("%+v\n", data2)

}

// decimalToFloat64 receives a d decimal.Decimal and returns a float64
func decimalToFloat64(d decimal.Decimal) float64 {
	f, _ := d.Float64()
	return f
}

In that case, the only way to use a single line of code to assign the value to data1.b is by using the function decimalToFloat64. A little bit of extra code is needed for data2.b.

There are similar cases in the standard library. See https://golang.org/pkg/html/template/#Must.

Regarding the unit test, I don't see any case that hasn't been covered in TestFloat64() - but if you guys think this is mandatory to move forward, just let me know, and I'll write the test.

Cheers, Felipe.

[mwoss]

Thanks for showing your point of view on that Felipe!
Let's wait for Jason's opinion, as I want to hear what he thinks about it. :D

[twocs]

In general, when I see a function named like MustFloat64, precedence is that it must panic if the underlying function would fail.
e.g. "panics if the error is non-nil" https://golang.org/src/text/template/helper.go

As the result of this PR would be a function that ignores any errors. I'd recommend a name such as:
InexactFloat64

[felipeneuwald]

@twocs nice catch. Please check now. Cheers.

[mwoss]

Looks good, merging :3