strconv: FormatFloat(38.675, 'f', -1, 64) does not give the most precise decimal approximation

[mrwonko]

What version of Go are you using (go version)?

1.7

What operating system and processor architecture are you using (go env)?

Playground

What did you do?

strconv.FormatFloat(38.675, 'f', -1, 64)
https://play.golang.org/p/r2SL1x_Uu6

What did you expect to see?

38.675 doesn't have an exact float64 representation; I expected 38.6749999999999971578290569595992565155029296875.

What did you see instead?

38.675

[ALTree]

The special precision -1 uses the smallest number of digits necessary such that ParseFloat will return f exactly.

The documentation doesn't say that FormatFloat will print the most precise decimal approximation. What it does say is that it'll output just enough digits so that ParseFloat will return f exactly.

[mrwonko]

So "exactly" in this context means a decimal representation whose closest float64 representation is the correct one? I see.

[rsc]

Note that for what you expect, you can use:

strconv.FormatFloat(38.675, 'g', 1000, 64)