A common dilemma when you define the methods of a structs is how you define your methods. Should you use value receivers or pointer receivers?

So, basically, it boils down to:

func (s *MyStruct) pointerMethod() { } // method on pointer
func (s MyStruct)  valueMethod()   { } // method on value

Luckily, there are some basic rules you need to keep in mind:

  • For a given type, don’t mix value and pointer receivers.
  • If in doubt, use pointer receivers (they are safe and extendable).

You must use pointer receivers

  • if any method needs to mutate the receiver
  • for structs that contain a sync.Mutex or similar synchronizing field (they musn’t be copied)

You probably want to use pointer receivers

  • for large structs or arrays (it can be more efficient)
  • in all other cases

You probably want to use value receivers

  • for map, func and chan types
  • for simple basic types such as int or string
  • for small arrays or structs that are value types, with no mutable fields and no pointers
  • when they need to be concurrency safe (pointer receivers are not concurrency safe)

You may want to use value receivers

  • for slices with methods that do not reslice or reallocate the slice.

You can read more about it in the Golang FAQ.

Related Posts

  • Mutual exclusion lock
  • The beauty of io.Reader in Go
  • Parsing a key pair from a PEM file in Go
  • Assert vs require in testify
  • Handling Unix timestamps in JSON