mGCA: Make trait object types with type-level associated consts dyn compatible if all of the latter are specified via bindings #150843

fmease · 2026-01-08T22:07:15Z

Under feature min_generic_const_args (mGCA) (#132980), render traits with non-parametrized type-level associated constants (i.e., #[type_const] ones) dyn compatible but force the user to specify all type-level associated consts in the trait object type via bindings (either directly, via supertrait bounds and/or behind trait aliases) just like associated types, their sibling.

Fixes #130300 (feature request).
Fixes #136063 (bug).
Fixes #137260 (bug).
Fixes #137514 (bug).

While I'm accounting for most illegal Self references via const projections & params, I'm intentionally ignoring RUST-123140 in this PR which is to be tackled some other time.

Blockers

~~This PR basically sits atop PR RUST-150799 (but squashed)~~ (rebased)
~~Do a perf run~~ (done)
~~Merge RUST-150844 (split out of this PR)~~ (rebased)
Crate rustc-demangle needs to be updated first, mainly for v0 mangling
- Merge Support v0 demangling of trait object types with assoc const bindings rustc-demangle#87 once sb gives me the OK
- Release a new version
- Update all r-l/r packages that depend on it I'm pretty sure
Apart from v0 mangling, legacy mangling (no longer the default) ICEs on such dyn traits
- we might want to address that
There are some minor diagnostic tweaks I'd still like to perform (see FIXMEs)

rustbot · 2026-01-08T22:07:18Z

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

This PR modifies tests/ui/issues/. If this PR is adding new tests to tests/ui/issues/,
please refrain from doing so, and instead add it to more descriptive subdirectories.

mGCA: Make trait object types with type associated consts dyn compatible if the latter are specified via bindings

rustbot · 2026-01-08T22:36:09Z

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

rust-timer · 2026-01-10T06:27:04Z

Finished benchmarking commit (e81452f): comparison URL.

Overall result: ❌ regressions - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.0%	[0.0%, 0.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 1.0%, secondary 0.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	1.0%	[1.0%, 1.0%]	1
Regressions ❌ (secondary)	1.7%	[1.7%, 1.7%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.6%	[-0.6%, -0.6%]	1
All ❌✅ (primary)	1.0%	[1.0%, 1.0%]	1

Cycles

Results (secondary 0.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.7%	[2.7%, 2.7%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.7%	[-2.7%, -2.7%]	1
All ❌✅ (primary)	-	-	0

Binary size

Results (secondary 0.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.1%	[0.0%, 0.1%]	9
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Bootstrap: 472.536s -> 473.075s (0.11%)
Artifact size: 391.23 MiB -> 391.29 MiB (0.01%)

@BoxyUwU

mGCA: Move tests for assoc const bindings (formerly ACE) into dedicated directory & replace more mentions of ACE Split out of PR #150843. As discussed. Somewhat obvious underlying principle: If the test checks basic or core parts of assoc const bindings and nothing else, move it, otherwise leave it even if it contains ACEs. Motivation: It makes a lot easier for me to continue working on ACE efficiently. r? @BoxyUwU

@BoxyUwU

mGCA: Move tests for assoc const bindings (formerly ACE) into dedicated directory & replace more mentions of ACE Split out of PR rust-lang/rust#150843. As discussed. Somewhat obvious underlying principle: If the test checks basic or core parts of assoc const bindings and nothing else, move it, otherwise leave it even if it contains ACEs. Motivation: It makes a lot easier for me to continue working on ACE efficiently. r? @BoxyUwU

@BoxyUwU

mGCA: Move tests for assoc const bindings (formerly ACE) into dedicated directory & replace more mentions of ACE Split out of PR rust-lang/rust#150843. As discussed. Somewhat obvious underlying principle: If the test checks basic or core parts of assoc const bindings and nothing else, move it, otherwise leave it even if it contains ACEs. Motivation: It makes a lot easier for me to continue working on ACE efficiently. r? @BoxyUwU

…ll assoc consts to be specified via bindings

…soc consts

We used to lower such bad defaulted const args in trait object types to `{type error}`; now correctly lower them to `{const error}`. The added tests used to ICE prior to this change.

…wise ref `Self` The added test used to ICE prior to this change.

This fully rewords the diagnostic that was previously only emitted for assoc ty bindings. That's because it incorrectly called trait aliases *type aliases* and didn't really make it clear what the root cause is. The added test used to ICE prior to this change. I've double-checked that the preexisting test I've modified still ICEs in nightly-2025-03-29.

BoxyUwU · 2026-01-15T15:37:38Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/dyn_trait.rs

+        // Collect all associated items that weren't specified and compute the list of projection
+        // bounds which we'll later turn into existential ones.
+        // For why we keep projections whose associated item has a `Self: Sized` bound,
+        // see <https://github.com/rust-lang/rust/pull/140684>.


can we say sth like "we keep them around so that we the resulting ty after HIR ty lowering still contains the user written types on the rhs of the assoc item binding, see https://github.com/...". clicking the link is annoying 😅

BoxyUwU · 2026-01-15T16:00:41Z

tests/ui/const-generics/associated-const-bindings/dyn-compat-non-type-assoc-const.rs

+}
+
+fn main() {
+    let _: dyn Trait; //~ ERROR the trait `Trait` is not dyn compatible


can you add another case with dyn Trait<K = 0> and check that errors

BoxyUwU · 2026-01-15T16:01:57Z

tests/ui/const-generics/associated-const-bindings/dyn-compat-symbol-mangling.rs

@@ -0,0 +1,27 @@
+// Ensure that we can successfully mangle & demangle trait object types w/ assoc const bindings.
+
+// FIXME(fmease): Legacy mangling still crashes.


FIXME(mgca)

i wanna know this when looking at mgca issues not jsut when looking at your FIXMEs :3

BoxyUwU · 2026-01-15T16:15:13Z

tests/ui/const-generics/associated-const-bindings/dyn-compat-self-const-projections.rs

+
+// This is a run-pass test to ensure that codegen can actually deal with such method instances
+// (e.g., const projections normalize flawlessly to something concrete, symbols get mangled
+// properly properly) and to ensure that the generated code "received" the correct values from


properly properly

BoxyUwU · 2026-01-15T16:30:17Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/errors.rs

+            // FIXME(fmease): Create `ty::AssocTag::descr`.
+            Descr::Tag(ty::AssocTag::Type) => "associated type",
+            Descr::Tag(ty::AssocTag::Const) => "associated constant",
+            _ => "associated item",


exhaustive match + unreachable arm for ty::AssocTag::Fn?

BoxyUwU · 2026-01-15T16:36:00Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/errors.rs

            where_constraints.push(principal_span);
        }

+        // FIXME(fmease): Get rid of this note outright, it is utter nonsense! Maybe mentioning


BoxyUwU

I've looked over all the commits, got some nits and you've still got the other thing to fix with type consts' types having Self in them until approval though

View changes since this review

fmease assigned BoxyUwU Jan 8, 2026

fmease added this to Associated Const Equality (ACE) Jan 8, 2026

fmease added the F-min_generic_const_args `#![feature(min_generic_const_args)]` label Jan 8, 2026

fmease moved this to In Progress in Associated Const Equality (ACE) Jan 8, 2026

This comment was marked as outdated.

Sign in to view

This comment has been minimized.

Sign in to view

rust-bors bot added a commit that referenced this pull request Jan 8, 2026

Auto merge of #150843 - fmease:dyn-ace, r=<try>

ed12841

mGCA: Make trait object types with type associated consts dyn compatible if the latter are specified via bindings

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jan 8, 2026

This comment has been minimized.

Sign in to view

This comment was marked as outdated.

Sign in to view

fmease force-pushed the dyn-ace branch from d8c1ca1 to c11d680 Compare January 8, 2026 22:34

This comment was marked as outdated.

Sign in to view

This comment has been minimized.

Sign in to view

rust-bors bot added a commit that referenced this pull request Jan 8, 2026

Auto merge of #150843 - fmease:dyn-ace, r=<try>

c1345e1

mGCA: Make trait object types with type associated consts dyn compatible if the latter are specified via bindings

fmease force-pushed the dyn-ace branch from c11d680 to 5242b36 Compare January 8, 2026 22:36

rustbot added the T-clippy Relevant to the Clippy team. label Jan 8, 2026

fmease changed the title ~~mGCA: Make trait object types with type associated consts dyn compatible if the latter are specified via bindings~~ mGCA: Make trait object types with type-level associated consts dyn compatible if the latter are specified via bindings Jan 8, 2026

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jan 10, 2026

fmease mentioned this pull request Jan 10, 2026

On conflicting assoc item bindings in trait object types, two similar errors are emitted #150936

Open

This comment has been minimized.

Sign in to view

fmease force-pushed the dyn-ace branch from 2c1bb00 to a292483 Compare January 11, 2026 17:28

fmease mentioned this pull request Jan 11, 2026

ui: add test for normalizing const projections with assoc const equality #150975

Merged

fmease force-pushed the dyn-ace branch 3 times, most recently from 19c584d to fb5b179 Compare January 11, 2026 18:44

fmease force-pushed the dyn-ace branch from fb5b179 to 964d80f Compare January 13, 2026 09:11

fmease added 9 commits January 15, 2026 16:59

Under mGCA make traits with assoc consts dyn compatible but require a…

2c86982

…ll assoc consts to be specified via bindings

Fix v0 symbol mangling for assoc const bindings

984fe74

Under mGCA permit certain const projections in dyn-compatible traits

f5b3203

Generalize diag for missing assoc types to account for assoc consts

c095a55

Generalize diag for conflicting assoc type bindings to account for as…

7228378

…soc consts

Introduce AssocTag::descr & refactor in the vicinity

655ca9c

Fix handling of const params defaults that ref Self & generalize diag

da96ef8

We used to lower such bad defaulted const args in trait object types to `{type error}`; now correctly lower them to `{const error}`. The added tests used to ICE prior to this change.

Force users to specify assoc consts from supertraits that would other…

b55ce90

…wise ref `Self` The added test used to ICE prior to this change.

fmease force-pushed the dyn-ace branch from 964d80f to 7e656f3 Compare January 15, 2026 15:59

BoxyUwU reviewed Jan 15, 2026

View reviewed changes

BoxyUwU approved these changes Jan 15, 2026

View reviewed changes

		@@ -0,0 +1,27 @@
		// Ensure that we can successfully mangle & demangle trait object types w/ assoc const bindings.

		// FIXME(fmease): Legacy mangling still crashes.

Uh oh!

mGCA: Make trait object types with type-level associated consts dyn compatible if all of the latter are specified via bindings #150843

Are you sure you want to change the base?

mGCA: Make trait object types with type-level associated consts dyn compatible if all of the latter are specified via bindings #150843

Uh oh!

Conversation

fmease commented Jan 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Blockers

Uh oh!

rustbot commented Jan 8, 2026

Uh oh!

This comment was marked as outdated.

This comment was marked as outdated.

This comment has been minimized.

This comment has been minimized.

This comment was marked as outdated.

This comment was marked as outdated.

This comment was marked as outdated.

This comment was marked as outdated.

This comment has been minimized.

rustbot commented Jan 8, 2026

Uh oh!

rust-timer commented Jan 10, 2026

Overall result: ❌ regressions - no action needed

Uh oh!

This comment has been minimized.

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

BoxyUwU left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

fmease commented Jan 8, 2026 •

edited

Loading

BoxyUwU left a comment •

edited by rustbot

Loading