strconv: FormatFloat(x, 'g', -1, 64) does not give the most precise decimal approximation

[remyoudompheng]

What steps will reproduce the problem?
1. compile and run

package main

import "strconv"
import "fmt"

func main() {
    fmt.Println(strconv.FormatFloat(383260575764816448, 'g', -1, 64))
}


What is the expected output?

3.8326057576481645e+17

What do you see instead?

3.8326057576481644e+17

Which compiler are you using (5g, 6g, 8g, gccgo)?

6g

Which operating system are you using?

linux

Which revision are you using?  (hg identify)

c956aa39a269 (tip)

Please provide any additional information below.

The obtained result has the same length and can be parsed to the same float64 as the
expected result. However, it is not the closest possible to the original float
(5988446496325257p+6).

[adg]

Comment 1:

Labels changed: added priority-later, packagebug, removed priority-triage.

Status changed to HelpWanted.

[rsc]

Comment 2:

Labels changed: added priority-go1, removed priority-later.

[rsc]

Comment 3:

This issue was closed by revision 6f77cd2.

Status changed to Fixed.