Rfcs: Support type transformations in FRU (functional record update) syntax

[abonander]

I remember discussing this a while ago. I thought I had an issue open somewhere for this but I guess not.

It's currently not allowed to use FRU (functional record update) syntax to return a transformed type, i.e.:

struct Foo<T> {
    val: T,
    tag: &'static str,
}

impl<T> Foo<T> {
    fn transform<U>(self, val: U) -> Foo<U> {
        Foo { val: val, .. self }
    }
}

will throw a "mismatched type" error with the span of self in the FRU literal because it is of type Foo<T> and not Foo<U>, even though the remaining fields of self are compatible.

According to RFC 736, FRU is supposed to be just syntactic sugar for a full struct literal, but a full struct literal works in this case because it doesn't place any restriction on the type of self.

Possible solutions:

• Move fields nominally (foo: <tail expr>.foo, bar: <tail expr>.bar, etc.) and typecheck them individually, not typechecking the tail expression at all. This could allow FRU to work even between otherwise incompatible types which have compatible fields. I like this one more but it may be too implicit for some, and I gather that we don't want to add too much meaning to field names.

• Only check equality of the outermost type constructor between the FRU literal and the tail expression, i.e. check that Foo<T> ~= Foo<U>; then, typecheck the moved fields individually. This would allow the above use-case but disallow using an entirely different type as the source or destination.

As far as I understand it, this should be backwards-compatible since it would strictly allow more code to compile.

Edit: clarify generic params

[mark-i-m]

Sorry if this is a silly question, but what are the precise difference between Foo<T>, Foo<T_> and Foo<_>?

[KodrAus]

@mark-i-m, in this case T_ is just another type parameter. It's not the same as T, it could also be called K or NotT or anything else. The _ on its own means the type parameter is inferred, so we don't give it an explicit name in our code, but it's still there. I usually read Foo<_> as _Foo of Whatever_.

[mark-i-m]

@KodrAus Got it :smile: I wasn't sure if T_ was some sort semi-inferred generic or something... Thanks!

[eddyb]

Does anyone plan to write an RFC on this?

[abonander]

I can, but it'd be basically just the OP here.

[abonander]

Also, not sure on which of the hypothesized solutions is better, that would have to be fleshed out before an RFC draft wouldn't it?

[eddyb]

@abonander I'd go with the more conservative route, i.e. matching fields not by name but rather by the fact that they're defined in the same struct.

[burdges]

We might want structural records back eventually, so avoid messing them up if possible.

[abonander]

Based on the last comment on that thread and the fact that the compiler now does rearrange fields, structural records probably won't be making a return. If they do, it seems to be a completely orthogonal feature.

[burdges]

I doubt the compiler rearranging fields poses any serious concern, but sure the syntax do not obviously conflict since Foo { .. } is quite different from { .. } or whatever.

[jturner314]

I put together a draft of an RFC for this. The trickiest part was thinking through the implications for type inference.

Does anyone have thoughts before I submit a PR? I haven't written an RFC before. In particular, I'm unsure if some of my terminology is correct/understandable, if the type inference sections make sense, and if I've provided sufficient detail. Also, am I missing any obvious edge cases?

[abonander]

@jturner314 At a glance it looks okay to me, though I don't care so much about the details as long as something gets implemented. 😄

You can try posting it to internals.rust-lang.org for early comments but the RFC process itself is supposed to be the actual avenue for developing a proposal. Any issues should get shaken out at that point.

[Centril]

@jturner314 My initial reading is that this is a solidly written and well thought out RFC; Nice work, especially as your first RFC. I left some notes that are nits.

[eddyb]

@jturner314 Looks great, please do open a PR!

cc @nikomatsakis

[jturner314]

I've submitted PR #2528. I made some changes from my initial version (primarily wording and more examples). Thanks for everyone's comments!

[Centril]

Closing in favor of https://github.com/rust-lang/rfcs/pull/2528.