Lens over all enum variants at once?

[maackle]

This is a weird one, and I don't even know what the equivalent in Haskell lenses would be...but I often wind up with enums like this:

enum E {
    A(A),
    B(B),
}

struct A {
    x: u8,
    y: bool,
}

struct B {
    x: u8,
    z: String
}

i.e. all variants of E have a field x: u8.

It would be great to write an optic which works like a lens over E which always gets me the x, no matter which variant it's from.

This could work if we could define LensRef and LensMut in terms of a Fn(&mut T) -> &mut Image instead of a field (if T: Clone, then we can implement Lens too). But I wonder if there's also a more direct but weird way, like "traverse all prisms of all enum variants", then apply the x lens to each, the result being a Prism which always has an image (since exactly one Prism will match). I don't know how to express that though.

It's the equivalent of the pattern match E::A(A { x, .. }) | E::B(B { x, .. })

[TOETOE55]

I think this has no equivalent expression in haskell, a reason is that haskell does not allow fields with the same name.

I have no idea but to impl LensX by ourselves.😭

[Aprabhat19]

Hey I was recently working on a POC that is very similar to the use case. I worked on lenses that were supposed to peek inside themselves ,and get the variants of a struct field