const-generics about lang-team

[lcnr]

2021-04-07:

news since the last update:

• feature(min_const_generics) has been stabilized
• we now have regular weekly meetings every tuesday
• feature(const_generics_defaults) is now on nightly, allowing for

#![feature(const_generics_defaults)]
struct WithDefault<const N: usize = 3> {
    inner: [u8; N],
}

current topics:

• complex generic const arguments, i.e. fn split_first<T, const N: usize>(arr: [T; N]) -> (T, [T; N - 1])
  • available on nightly using feature(const_evaluatable_checked) which is still quite broken
  • there are still some big design and implementation challenges
• working towards allowing more const param types, such as fn foo<const CONFIG: Config>
• working towards stabilization of some smaller improvements, see our meeting notes for more details

[varkor]

In the past, there have been various "procedural RFCs" that specify functionality that has already been deemed desirable, and therefore not required the full comment period (e.g. rust-lang/rfcs#2300, rust-lang/rfcs#2591). Perhaps this is another such situation.

[nikomatsakis]

• May be in code complete status, planning a meeting to go over and create a comprehensive test suite

[nikomatsakis]

2020-10-26:

• lcnr and I prepared a list of test cases last week.
• What is the next procedural step?
  • Stabilization report seems appropriate but is not a great place for lasting documentation
  • RFC would be more visible but we don't really need the "comment requirement"

[lcnr]

2021-05-18

• working on a vision doc for const generics
• experimenting with a hack to implement Default for generic arrays
• const parameter defaults is getting close to being ready for stabilization
• some (small) further work on complex generic const arguments and val trees(rust-lang/rust#83234) as well as the smaller improvements mentioned last time

[BoxyUwU]

2021-07-06

• started a WIP document to help figure out structural equality (structural equality is a blocker for allowing more than just integers, char and bool as const param types)
• various small fixes for const_generics_defaults made, would be a good idea to figure out if we can stabilise this now