Wishlist: marker for open-ended enums

[aturon]

Previous RFC:

A mechanism for declaring an enum to be extensible, which prevents exhausting matching in downstream crates.

[clarfonthey]

Hey, I think that putting out a new RFC for something like this might like up with 2017's goals, considering how it'll make a much-desired feature for enums simple instead of cryptic (via #[doc(hidden)]).

What do people think?

[Mark-Simulacrum]

There's quite a bit of interesting discussion in rust-lang/rust/issues/32770; worth taking a look at if anyone is planning to write an RFC.

[clarfonthey]

I'd be more than willing to modify the existing RFC that was postponed to change the dedicated syntax to a special attribute. Mostly I'm just curious if it'd make sense to make one now or if the core team would rather make the change later.

[Mark-Simulacrum]

cc @rust-lang/lang, @clarcharr seems willing to drive an RFC on this topic. Can we get a preliminary go-ahead or otherwise?

[clarfonthey]

Clarifying a bit on the idea I have:

This would add an #[non_exhaustive] attribute that can be added to enums to indicate that matches outside the current crate must include a wildcard pattern. We'd remove this restriction for inside the crate because if the enum is changed, it's reasonable for the crate author to update code elsewhere inside the same crate.

[clarfonthey]

So I decided to write this up anyway and we'll see what people think. It's at #2008.

[leoyvens]

Given we have an accepted RFC, I believe this issue can be closed.