[WIP] experiment with a new way of effects desugaring #120639

fee1-dead · 2024-02-04T10:12:00Z

cc @rust-lang/project-const-traits. Will write down notes once I have finished.

See if we want T: Tr to desugar into T: Tr, T::Effects: Compat<true>
Fix ICEs on type Assoc: ~const Tr and type Assoc<T: ~const Tr>
add types and traits to minicore test
update rustc-dev-guide

rustbot · 2024-02-04T10:12:07Z

(rustbot has picked a reviewer for you, use r? to override)

oli-obk · 2024-02-04T11:20:30Z

compiler/rustc_hir_analysis/src/collect/item_bounds.rs

@@ -104,6 +104,11 @@ pub(super) fn explicit_item_bounds(
        None => {}
    }

+    // effects desugared associated types have no bounds.


Could feed this query.

Also: don't we need a where Self: Sized bound to avoid breaking all dyn Trait for now?

assoc types don't break object safety (we'd just need to make it default to Runtime)

what breaks object safety would probably be the const host: bool that we now insert onto each method.

bors · 2024-02-06T05:10:03Z

☔ The latest upstream changes (presumably #120392) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2024-02-13T03:33:51Z

☔ The latest upstream changes (presumably #120991) made this pull request unmergeable. Please resolve the merge conflicts.

fee1-dead · 2024-02-26T06:56:53Z

This PR changes T: Tr where Tr is #[const_trait] to also lower into an additional bound of <T as Tr>::{opaque}: Compat<true>. That means potentially more work for the trait solver. Let's see what the perf impact is like, as right now we have Add as a #[const_trait]

@bors try @rust-timer queue

rust-timer · 2024-02-26T06:56:55Z

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

fee1-dead · 2024-02-26T06:57:57Z

@bors try

bors · 2024-02-26T06:58:03Z

⌛ Trying commit f1cd746 with merge c5d6267...

…ring, r=<try> [WIP] experiment with a new way of effects desugaring cc `@rust-lang/project-const-traits.` Will write down notes once I have finished. * [ ] See if we want `T: Tr` to desugar into `T: Tr, T::Effects: Compat<true>` * [ ] Fix ICEs on `type Assoc: ~const Tr` and `type Assoc<T: ~const Tr>` * [ ] add types and traits to minicore test * [ ] update rustc-dev-guide

bors · 2024-02-26T08:17:05Z

💔 Test failed - checks-actions

fee1-dead · 2024-02-27T00:23:15Z

@bors try @rust-timer queue

fee1-dead · 2024-03-03T05:44:06Z

The compile error from perf is interesting. The existence of closures that only implement FnOnce for use in compile time makes us unable to assume that any trait implementation would allow for usage in runtime. (this would extend to future user-defined types if we allow ~const dyn and ~const fn and the such)

This means for any trait that has a const_trait super trait, like trait Curve: Add, we also want this to implicitly mean that trait Curve: Add<Effects: ComptabileWith<Runtime>>. This mostly works fine (?) by adding to the implied bounds, but causes trouble when binders are involved, as seen with those compile failures.

I've opened #121929, and fixing that would probably unblock this. If that cannot be fixed easily, we might need to explore a workaround such as emitting two separate associated types each encoding whether it is compatible with runtime=true and compatible with runtime=false.

bors · 2024-03-07T05:20:27Z

☔ The latest upstream changes (presumably #122113) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2024-04-23T13:35:43Z

☔ The latest upstream changes (presumably #124289) made this pull request unmergeable. Please resolve the merge conflicts.

plus some other fixes.

it causes ICEs since we assume that the only substs are on the trait and the method has no generic parameters

We need to change how we handle it in the trait system before re-enabling it.

This avoids mix up when we create inference params for them and try to equate those to effect inference params.

fee1-dead · 2024-05-13T08:39:17Z

Noting, with initial implementation of Min (still quite wrong in its desugaring with binders and etc), need to investigate the tests/ui/rfcs/rfc-2632-const-trait-impl/super-traits tests.

Consider appending Self::Effects into the clause as well.

rustbot assigned oli-obk Feb 4, 2024

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Feb 4, 2024