Implement trait const stability #133999

fee1-dead · 2024-12-07T03:27:51Z

.. By enforcing it in three locations:

When a bound during lowering is ~const or const
When an impl is const
When a MIR call is on a trait's method

Reopens #90080
Reopens #88955
Closes rust-lang/project-const-traits#5
Closes rust-lang/project-const-traits#16

rustbot · 2024-12-07T03:28:00Z

rustbot has assigned @oli-obk.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

rustbot · 2024-12-07T03:28:02Z

HIR ty lowering was modified

cc @fmease

Some changes occurred to the CTFE machinery

cc @rust-lang/wg-const-eval

fee1-dead · 2024-12-07T03:28:26Z

also cc @rust-lang/project-const-traits

compiler/rustc_middle/src/ty/context.rs

oli-obk · 2024-12-07T09:27:11Z

I'm guessing with the current const trait impl we'd have the same issues with making impls unstable as we'd have for nonconst impls?

RalfJung · 2024-12-07T09:28:17Z

Yeah... we should enforce that if a trait is marked const-stable then all its impls (in staged_api crates) are const-stable, like we do for regular stability.

fee1-dead · 2024-12-07T14:31:01Z

The current code doesn't deal with impls' const stability yet, that work can be incremental on top of this (i will personally probably have more cycles to spend until one week or more, so i'd prefer us landing a reviewed version of this first)

RalfJung · 2024-12-07T14:47:45Z

That's fine as long as this is recorded somewhere as a blocker.

compiler/rustc_const_eval/src/check_consts/check.rs

tests/ui/traits/const-traits/staged-api.rs

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

bors · 2024-12-10T14:10:28Z

☔ The latest upstream changes (presumably #134108) made this pull request unmergeable. Please resolve the merge conflicts.

fee1-dead · 2024-12-26T17:09:05Z

@rustbot ready

Rollup merge of rust-lang#135139 - c410-f3r:8-years-rfc, r=jhpratt [generic_assert] Constify methods used by the formatting system cc rust-lang#44838 Starts the "constification" of all the elements required to allow the execution of the formatting system in constant environments. ```rust const _: () = { panic!("{:?}", 1i32); }; ``` Further stuff is blocked by rust-lang#133999.

compiler/rustc_middle/messages.ftl

compiler/rustc_middle/src/middle/stability.rs

oli-obk · 2025-01-07T10:19:06Z

r=me with diagnostic and readability nits resolved

fee1-dead · 2025-01-07T11:26:08Z

@bors r=oli-obk

bors · 2025-01-07T11:26:11Z

📌 Commit 05ca6ae has been approved by oli-obk

It is now in the queue for this repository.

RalfJung · 2025-01-07T11:33:45Z

@bors r-
I just started taking a look concurrently with Oli as I somehow thought it was assigned to me, sorry. I'll have some comments for compiler/rustc_middle/src/middle/stability.rs as that logic should very closely mirror compiler/rustc_const_eval/src/check_consts/check.rs.

RalfJung · 2025-01-07T10:02:02Z

compiler/rustc_middle/src/middle/stability.rs

+                let unstable_feature_allowed = span.allows_unstable(feature)
+                    || implied_feature.is_some_and(|f| span.allows_unstable(f));


For function calls we only honor allows_unstable if the callee is safe-to-expose on stable. So here I guess the question would be, when is a trait safe to expose on stable? Right now I think the answer is "only when it is stable", we don't support something like #[const_stable_indirect] on traits. That means we should not check span.allows_unstable here.

RalfJung · 2025-01-07T11:35:24Z

compiler/rustc_middle/src/middle/stability.rs

+        span: Span,
+        parent_def: Option<LocalDefId>,
+    ) {
+        let Some(ConstStability {


Suggested change

let Some(ConstStability {

// This should work pretty much exactly like the function stability logic in

// `compiler/rustc_const_eval/src/check_consts/check.rs`.

// FIXME: Find some way to not duplicate that logic.

let Some(ConstStability {

RalfJung · 2025-01-07T11:37:48Z

compiler/rustc_middle/src/middle/stability.rs

+            });
+            self.disabled_nightly_features(&mut diag, None, [(String::new(), feature)]);
+            diag.emit();
+        } else if let Some(parent) = parent_def {


I don't understand this -- what is the "parent" here?

The equivalent code for this entire if-else-if in check.rs is

if !feature_enabled || !callee_safe_to_expose_on_stable { self.check_op(ops::FnCallUnstable { def_id: callee, feature, feature_enabled, safe_to_expose_on_stable: callee_safe_to_expose_on_stable, }); }

Here callee_safe_to_expose_on_stable is always false so we should always behave as-if check_op was called.

Either a const fn that contains the ~const Tr bound or the callee of a const method.

we should always behave as-if check_op was called.

Yes, this is why this code is derived from check_op_spanned.

Either a const fn that contains the ~const Tr bound or the callee of a const method.

Sorry I don't follow. If you had said "always the const fn that contains the thing we are stability-checking", i.e., the caller, that would make more sense to me. Is this a typo?

The existing const checks have nothing like this, as far as I can see. I have no idea what it means here that we are skipping the check if there is no parent. We should always know the function we are in and that defines whether we are in recursive-const-stability mode.

Yes, this is why this code is derived from check_op_spanned.

Can you make the structure parallel, i.e. also have a function like that here with essentially the same inputs (without the NonConstOp trait of course). Or ideally share that code rather than duplicate it... maybe we need a compiler/rustc_middle/src/middle/const_stability.rs file for those functions. (stability.rs is already huge.)

RalfJung · 2025-01-07T11:38:04Z

compiler/rustc_middle/src/middle/stability.rs

+                            // other unstably-const features with `const_stable_indirect`
+                            stab.is_const_stable() || stab.const_stable_indirect
+                        }
+                    };


What is this logic? This has to be a duplicate of something in check.rs, right?

This is duplicating enforce_recursive_const_stability.

That logic is fairly critical and definitely should not be duplicated.

bors · 2025-01-09T19:48:59Z

☔ The latest upstream changes (presumably #135286) made this pull request unmergeable. Please resolve the merge conflicts.

rustbot · 2025-01-10T11:04:55Z

The rustc-dev-guide subtree was changed. If this PR only touches the dev guide consider submitting a PR directly to rust-lang/rustc-dev-guide otherwise thank you for updating the dev guide with your changes.

cc @BoxyUwU, @jieyouxu, @Kobzol

rust-log-analyzer · 2025-01-10T11:13:55Z

The job x86_64-gnu-tools failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

   Compiling compiler_builtins v0.1.140
error: trait has missing const stability attribute
  --> library/core/src/intrinsics/fallback.rs:11:1
   |
11 | / pub trait CarryingMulAdd: Copy + 'static {
12 | |     type Unsigned: Copy + 'static;
14 | |         self,
...  |
18 | |     ) -> (Self::Unsigned, Self);
19 | | }

BoxyUwU · 2025-01-10T14:03:12Z

ive never seen a better diff in my life

RalfJung · 2025-01-10T14:11:16Z

Yeah something went wrong with the rebase, this now deletes the dev guide.^^

bors · 2025-01-10T15:05:14Z

☔ The latest upstream changes (presumably #135327) made this pull request unmergeable. Please resolve the merge conflicts.

fee1-dead · 2025-01-12T08:44:21Z

See https://rust-lang.zulipchat.com/#narrow/channel/419616-t-compiler.2Fproject-const-traits/topic/activity/near/493169956

rustbot assigned oli-obk Dec 7, 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 Dec 7, 2024

compiler-errors mentioned this pull request Dec 7, 2024

Constify PartialEq #133995

Open

compiler-errors self-assigned this Dec 7, 2024

oli-obk reviewed Dec 7, 2024

View reviewed changes

compiler/rustc_middle/src/ty/context.rs Outdated Show resolved Hide resolved

RalfJung reviewed Dec 7, 2024

View reviewed changes

compiler/rustc_const_eval/src/check_consts/check.rs Outdated Show resolved Hide resolved

tests/ui/traits/const-traits/staged-api.rs Outdated Show resolved Hide resolved

compiler-errors reviewed Dec 7, 2024

View reviewed changes

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs Outdated Show resolved Hide resolved

fee1-dead added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Dec 26, 2024

fee1-dead force-pushed the traitconststab branch from fa450eb to ee3f3df Compare December 26, 2024 15:08

This comment has been minimized.

Sign in to view

fee1-dead force-pushed the traitconststab branch from ee3f3df to 9c55da1 Compare December 26, 2024 17:06

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Dec 26, 2024

This comment has been minimized.

Sign in to view

fee1-dead force-pushed the traitconststab branch from 9c55da1 to 080ae25 Compare December 27, 2024 03:01

This comment has been minimized.

Sign in to view

fee1-dead force-pushed the traitconststab branch from 080ae25 to d60bc65 Compare December 27, 2024 03:19

fee1-dead mentioned this pull request Dec 30, 2024

Const traits rust-lang/rust-project-goals#106

Open

2 tasks

oli-obk requested changes Jan 7, 2025

View reviewed changes

compiler/rustc_middle/messages.ftl Outdated Show resolved Hide resolved

compiler/rustc_middle/src/middle/stability.rs Outdated Show resolved Hide resolved

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jan 7, 2025

fee1-dead force-pushed the traitconststab branch from d60bc65 to 05ca6ae Compare January 7, 2025 11:25

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jan 7, 2025

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Jan 7, 2025

This comment has been minimized.

Sign in to view

RalfJung reviewed Jan 7, 2025

View reviewed changes

fee1-dead added 2 commits January 10, 2025 19:04

Implement trait const stability

f9de1b6

fix rustc_const_stable_indirect message

2d0c0e7

fee1-dead force-pushed the traitconststab branch from 05ca6ae to 2d0c0e7 Compare January 10, 2025 11:04

rustbot added the A-rustc-dev-guide Area: rustc-dev-guide label Jan 10, 2025

compiler-errors assigned RalfJung and unassigned oli-obk and compiler-errors Jan 12, 2025

RalfJung mentioned this pull request Jan 13, 2025

Enforce syntactical stability of const traits in HIR #135423

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Implement trait const stability #133999

Implement trait const stability #133999

fee1-dead commented Dec 7, 2024

rustbot commented Dec 7, 2024

rustbot commented Dec 7, 2024

fee1-dead commented Dec 7, 2024

oli-obk commented Dec 7, 2024

RalfJung commented Dec 7, 2024

fee1-dead commented Dec 7, 2024

RalfJung commented Dec 7, 2024

bors commented Dec 10, 2024

This comment has been minimized.

fee1-dead commented Dec 26, 2024

This comment has been minimized.

This comment has been minimized.

oli-obk commented Jan 7, 2025

fee1-dead commented Jan 7, 2025

bors commented Jan 7, 2025

RalfJung commented Jan 7, 2025 •

edited

Loading

This comment has been minimized.

RalfJung Jan 7, 2025 •

edited

Loading

fee1-dead Jan 10, 2025

RalfJung Jan 7, 2025

RalfJung Jan 7, 2025 •

edited

Loading

fee1-dead Jan 10, 2025

RalfJung Jan 10, 2025 •

edited

Loading

RalfJung Jan 10, 2025

RalfJung Jan 7, 2025

fee1-dead Jan 10, 2025

RalfJung Jan 10, 2025

bors commented Jan 9, 2025

rustbot commented Jan 10, 2025

rust-log-analyzer commented Jan 10, 2025

BoxyUwU commented Jan 10, 2025

RalfJung commented Jan 10, 2025 via email

bors commented Jan 10, 2025

fee1-dead commented Jan 12, 2025

		let unstable_feature_allowed = span.allows_unstable(feature)
		\|\| implied_feature.is_some_and(\|f\| span.allows_unstable(f));

-        let Some(ConstStability {
+        // This should work pretty much exactly like the function stability logic in
+        // `compiler/rustc_const_eval/src/check_consts/check.rs`.
+        // FIXME: Find some way to not duplicate that logic.
+        let Some(ConstStability {

Implement trait const stability #133999

Are you sure you want to change the base?

Implement trait const stability #133999

Conversation

fee1-dead commented Dec 7, 2024

rustbot commented Dec 7, 2024

rustbot commented Dec 7, 2024

fee1-dead commented Dec 7, 2024

oli-obk commented Dec 7, 2024

RalfJung commented Dec 7, 2024

fee1-dead commented Dec 7, 2024

RalfJung commented Dec 7, 2024

bors commented Dec 10, 2024

This comment has been minimized.

fee1-dead commented Dec 26, 2024

This comment has been minimized.

This comment has been minimized.

oli-obk commented Jan 7, 2025

fee1-dead commented Jan 7, 2025

bors commented Jan 7, 2025

RalfJung commented Jan 7, 2025 • edited Loading

This comment has been minimized.

RalfJung Jan 7, 2025 • edited Loading

Choose a reason for hiding this comment

fee1-dead Jan 10, 2025

Choose a reason for hiding this comment

RalfJung Jan 7, 2025

Choose a reason for hiding this comment

RalfJung Jan 7, 2025 • edited Loading

Choose a reason for hiding this comment

fee1-dead Jan 10, 2025

Choose a reason for hiding this comment

RalfJung Jan 10, 2025 • edited Loading

Choose a reason for hiding this comment

RalfJung Jan 10, 2025

Choose a reason for hiding this comment

RalfJung Jan 7, 2025

Choose a reason for hiding this comment

fee1-dead Jan 10, 2025

Choose a reason for hiding this comment

RalfJung Jan 10, 2025

Choose a reason for hiding this comment

bors commented Jan 9, 2025

rustbot commented Jan 10, 2025

rust-log-analyzer commented Jan 10, 2025

BoxyUwU commented Jan 10, 2025

RalfJung commented Jan 10, 2025 via email

bors commented Jan 10, 2025

fee1-dead commented Jan 12, 2025

RalfJung commented Jan 7, 2025 •

edited

Loading

RalfJung Jan 7, 2025 •

edited

Loading

RalfJung Jan 7, 2025 •

edited

Loading

RalfJung Jan 10, 2025 •

edited

Loading