Comment 1:

Weird... Is the way it's rewriting the CallExpr.Args field wrong?

Comment 2:

Hold on, the comment triggers this? How? I'm not touching that with a 10-foot pole...

Comment 3:

Is token location information somehow not invalidated or ignored? This is interesting
output too by the way: 
$ cat x.go 
package main
import "crypto/hmac"
func main() {
    go hmac.NewMD5(nil)  // foo
}
$ gofix x.go 
x.go:9:22: expected ')', found newline
x.go:10:2: expected ';', found ','
Note the "go". I'm completely in the dark on this one... I wrote that gofix, but I have
only very basic understanding of AST and it's parser/printer.

Comment 4:

The way to debug this would be to add the test case to
the gofix tests and then put some prints in
src/cmd/gofix/main_test.go's parseFixPrint to
see what the intermediate printing output is.
It will have put the // in the wrong place.
I believe Robert is working on getting //'s into
only right places, so this issue will probably
resolve when that gets fixed.

Comment 6:

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

Comment 7:

go/printer incorrectly reformats the gofix-ed program (the comment has incorrect
position information) and the rewritten program looks like:
package main
import (
    "crypto/hmac"
    "crypto/md5"
)
func main() {
    hmac.New(md5.New // foo
    , nil)
}
which is clearly incorrect:
x.go:9:2: expected 1 expression
x.go:10:7: expected statement, found ')'
This is a duplicate of issue #1505.

Comment 8:

Owner changed to @griesemer.

Status changed to Duplicate.

Merged into issue #1505.