Go interfaces and nil

This little subtlety in Go caught me today.  I was trying determine if what an interface referenced was nil and I wrote a bit of code like this (this is not the actual code I wrote, just a simple version to demonstrate the issue):

type I interface {

}

func main() {

var p *int

var i I

i = p

if i != nil {

fmt.Println("p + 1 ==", *i.(*int) + 1)

}

}

Sadly, it did not have the intended outcome and the program panicked:

panic: runtime error: invalid memory address or nil pointer dereference

What happened?  I tested to see if the interface was nil but the program still panicked over a nil pointer?  That can't be right.  So I wrote a small test program and realized something pretty simple.  You need to test both that the interface isn't nil and also check to make sure that what it references isn't nil:

type I interface {
}

func main() {
var p *int
var i I
i = p
if i != nil && i.(*int) != nil {
fmt.Println("p + 1 ==", *i.(*int) + 1)
}
}


I've been writing Go for over a year and I'd never really thought about this before.. or been caught by a defect because of it.