12

I'm using an enum for polymorphism, similar to the following:

enum MyType {
    Variant1 { a: i32, b: i32 },
    Variant2 { a: bool, b: bool },
}

Is there clean way to use existing structs for Variant1 and Variant2? I've done the following:

struct Variant1 {
    a: i32,
    b: i32,
}

struct Variant2 {
    a: bool,
    b: bool,
}

enum MyType {
    Variant1(Variant1),
    Variant2(Variant2),
}

but it feels pretty clunky. I was wondering if there's a better way to accomplish a similar thing.

Shepmaster
  • 388,571
  • 95
  • 1,107
  • 1,366
anderspitman
  • 9,230
  • 10
  • 40
  • 61

2 Answers2

15

There have been multiple RFCs about making enum variants their own type:

Unfortunately, they have been postponed or not yet decided on as this is not considered a language priority.

This means that right now, the way you have proposed is the only way to have types for enum variants.

Shepmaster
  • 388,571
  • 95
  • 1,107
  • 1,366
mcarton
  • 27,633
  • 5
  • 85
  • 95
  • 1
    Can the boilerplate be reduced / sugared a little with macros then? Or is it too complicated because they would span accross `enum` *and* `struct` *and* maybe [traits](https://stackoverflow.com/q/51567350/3719101) definitions ? – iago-lito May 24 '19 at 10:51
2

No, that's exactly how that works.

Sebastian Redl
  • 69,373
  • 8
  • 123
  • 157