Take a look at #7079
ParseAddress parses according to RFC 5322, which only allows US-ASCII characters.
A possible workaround for you: Use the mail.AddressParser struct, to encode your umlauts, guess there should be some examples in the web.

Yes, only ASCII is allowed there. Perhaps we should return a more clear error message if we see any octet greater than 127 in the input.

Ok, took a stab on better error message at https://go-review.googlesource.com/14312

Any feedback on the text would be great!

CL https://golang.org/cl/14312 mentions this issue.

@ulrichSchreiner hi! Did you solved this issue? =)

@gobwas take a look at my answer

@stemar94 i saw your answer, thanks. There are no examples in the web. How could mail.AddressParser encode non ASCII chars? I think, by the spec, it should decode, not encode.

m := "Иван Иванов <ivanov@example.com>"
parser := mail.AddressParser{WordDecoder: &mime.WordDecoder{}}
encodedM := HowCouldIEncodeThisToRFC5322(m) // ?
parser.Parse(m)

try this:

    s := "aä <aa.bb@cc.com>"
    se := mime.QEncoding.Encode("utf-8", s)

    m, e := mail.ParseAddress(se)
    if e != nil {
        panic(e)
    }
    fmt.Printf("%v\n",m.Address)
    fmt.Printf("%v\n",m.Name)
    dec := new(mime.WordDecoder)
    decname, err := dec.Decode(m.Name+"?=")
    if err != nil {
        panic(err)
    }
    fmt.Printf("%v\n",decname)

but only for names with illegal characters :-)

@ulrichSchreiner your example just not panics, but m.Name is still encoded like:

fmt.Printf("%s %s", m.Name, m.Address) // => ivanov@example.com =?utf-8?q?=D0=98=D0=B2=D0=B0=D0=BD_=D0=98=D0=B2=D0=B0=D0=BD=D0=BE=D0=B2?=

I mean, is it possible to parse without second part of your code? )

if you want the name, you have to decode the name with the WordDecoder: http://play.golang.org/p/OzcpaQh9jj

@ulrichSchreiner okay. I've got this. Thank you!