Last rune of golang unicode/norm iterator not being read - unicode

Question

I'm using the golang.org/x/text/unicode/norm package to iterate over runes in a []byte. I've chosen this approach as I need to inspect each rune and maintain information about the sequence of runes. The last call to iter.Next() does not read the last rune. It gives 0 bytes read on the last rune.
Here is the code:
package main
import (
"fmt"
"unicode/utf8"
"golang.org/x/text/unicode/norm"
)
func main() {
var (
n int
r rune
it norm.Iter
out []byte
)
in := []byte(`test`)
fmt.Printf("%s\n", in)
fmt.Println(in)
it.Init(norm.NFD, in)
for !it.Done() {
ruf := it.Next()
r, n = utf8.DecodeRune(ruf)
fmt.Printf("bytes read: %d. val: %q\n", n, r)
buf := make([]byte, utf8.RuneLen(r))
utf8.EncodeRune(buf, r)
out = norm.NFC.Append(out, buf...)
}
fmt.Printf("%s\n", out)
fmt.Println(out)
}
This produces the following output:
test
[116 101 115 116]
bytes read: 1. val: 't'
bytes read: 1. val: 'e'
bytes read: 1. val: 's'
bytes read: 0. val: '�'
tes�
[116 101 115 239 191 189]

Answer 1

It is possible this is a bug in golang.org/x/text/unicode/norm and its Init() function.
In the package's test and example that I see all use InitString. So as a workaround, if you change:
it.Init(norm.NFD, in)
to:
it.InitString(norm.NFD, `test`)
things will work as expected.
I would suggest opening up a bug report, but beware that since this is in the "/x" directory that package is considered experimental by go developers.
(BTW, I used my the go debugger to help me track down what's going on, but I should say its use was far the kind of debugger I'd like to see.)