Consider a binary tree node:
type Node struct {
value uint8
left, right *Node
}
If I want to add a function to each Node, should it be:
A: func (n *Node) height() int
or
B: func (n Node) height() int
I want to know which of A or B you would choose and why.
I can model linked lists or recursive structs, but I don't know when the receiver should be a pointer or not-a-pointer.
分享 The main difference is the idea of pass by value vs pass by reference
When you're using example A you're passing be reference, meaning that any changes you apply to n inside of func (n *Node) height will apply to the Node you're using to call Node.height().
By comparison, what you're doing in example B is passing by value where you're really just passing a copy of the Node you're using to call Node.height() so any changes to that Node will not apply outside of the function.
Here is a small playground to demonstrate the differences: http://play.golang.org/p/JodPRPBHDg
Notice in the example that when you call node.incHeight(), it goes to the function:
func (n Node) incHeight() uint8 {
n.value++
return n.value
}
Because this is pass by value the node.value is still the same unless you store the returned value as node.value. However, if you call node.incrementHeight() it goes to the function:
func (n *Node) incrementHeight() {
n.value++
return
}
This will change the value of node.value without needing to return anything because its referencing the original struct, rather than a copy of the struct.
So in answer to which would you choose and why, it really depends on whether you want to be able to make lasting changes to the struct you're using to call the function or if you'd rather just pass a copy that can see and alter the values but only while within the function and not have any lasting effects on the original struct.
分享
