Go functions and methods can return multiple values.
func learnMultiple(x, y int) (sum, prod int) {
return x + y, x * y // return two values
}
sum, prod := learnMultiple(10, 50)
Is it similar to returning a tuple?
I come from ruby land where I can return an array and a
sum, prod = ["60","500"]
We can easily look at some compiled code to confirm what's going on behind the scenes.
Consider this snippet:
func f() (a, b byte) {
return 'x', 'y'
}
func main() {
a, b := f()
println(a, b)
}
If we disassemble the created ELF binary, you'll see something like this (inlining was disabled so we could see the call happening):
0000000000400c00 <main.f>:
400c00: c6 44 24 08 78 movb $0x78,0x8(%rsp)
400c05: c6 44 24 09 79 movb $0x79,0x9(%rsp)
400c0a: c3 retq
0000000000400c10 <main.main>:
(...)
400c25: 48 83 ec 10 sub $0x10,%rsp
400c29: e8 d2 ff ff ff callq 400c00 <main.f>
400c2e: 48 0f b6 1c 24 movzbq (%rsp),%rbx
400c33: 48 89 d8 mov %rbx,%rax
400c36: 48 0f b6 5c 24 01 movzbq 0x1(%rsp),%rbx
(...)
So f just puts the result bytes in the stack, and main picks them back and puts them in working registers. A different compiler might also choose to pass these values around between the two scopes straight into registers.
This is all similar to what a compiler for the C language would do, except its specification defines a single return value only.
It is similar to returning a tuple, in the sense that some languages (like python) use tuples to implement multiple return values.
But in Go there is no such thing as a tuple. A function that returns one value allocates one slot on the stack to hold it. A function that returns two values allocates two slots on the stack to hold them. And so on…