MGCA: Support struct expressions without intermediary anon consts #149114

BoxyUwU · 2025-11-19T21:36:30Z

r? oli-obk

tracking issue: #132980
based on: #150025

High level goal

Under feature(min_generic_const_args) this PR adds another kind of const argument. A struct/variant construction const arg kind. We represent the values of the fields as themselves being const arguments which allows for uses of generic parameters subject to the existing restrictions present in min_generic_const_args:

fn foo<const N: Option<u32>>() {}

trait Trait {
    #[type_const]
    const ASSOC: usize;
}

fn bar<T: Trait, const N: u32>() {
    // the initializer of `_0` is a `N` which is a legal const argument
    // so this is ok.
    foo::<{ Some::<u32> { 0: N } }>();

    // this is allowed as mgca supports uses of assoc consts in the
    // type system. ie `<T as Trait>::ASSOC` is a legal const argument
    foo::<{ Some::<u32> { 0: <T as Trait>::ASSOC } }>();

    // this on the other hand is not allowed as `N + 1` is not a legal
    // const argument
    foo::<{ Some::<u32> { 0: N + 1 } }>();
}

This PR does not support uses of const ctors, e.g. None. And also does not support tuple constructors, e.g. Some(N). I believe that it would not be difficult to add support for such functionality after this PR lands so have left it out deliberately.

We currently require that all generic parameters on the type being constructed be explicitly specified. I haven't really looked into why that is but it doesn't seem desirable to me as it should be legal to write Some { ... } in a const argument inside of a body and have that desugar to Some::<_> { ... }. Regardless this can definitely be a follow-up PR and I assume this is some underlying consistency with the way that elided args are handled with type paths elsewhere.

This PRs implementation of supporting struct expressions is somewhat incomplete. We don't handle Foo { ..expr } at all and aren't handling privacy/stability. The printing of ConstArgKind::Struct HIR nodes doesn't really exist either :')

I've tried to keep the implementation here somewhat deliberately incomplete as I think a number of these issues are actually quite small and self contained after this PR lands and I'm hoping it could be a good set of issues to mentor newer contributors on 🤔 I just wanted the "bare minimum" required to actually demonstrate that the previous changes are "necessary".

`ValTree` now recurse through `ty::Const`

In order to actually represent struct/variant construction in ty::Const without going through an anon const we would need to introduce some new ConstKind variant. Let's say some hypothetical ConstKind::ADT(Ty<'tcx>, List<Const<'tcx>>).

This variant would represent things the same way that ValTree does with the first element representing the VariantIdx of the enum (if its an enum), and then followed by a list of field values in definition order.

This could work but there are a few reasons why it's suboptimal.

First it would mean we have a second kind of Const that can be normalized. Right now we only have ConstKind::Unevaluated which possibly needs normalization. Similarly with TyKind we only have TyKind::Alias. If we introduced ConstKind::ADT it would need to be normalized to a ConstKind::Value eventually. This feels to me like it has the potential to cause bugs in the long run where only ConstKind::Unevaluated is handled by some code paths.

Secondly it would make type equality/inference be kind of... weird... It's desirable for Some { 0: ?x } eq Some { 0: 1_u32 } to result in ?x=1_u32. I can't see a way for this to work with this ConstKind::ADT design under the current architecture for how we represent types/consts and generally do equality operations.

We would need to wholly special case these two variants in type equality and have a custom recursive walker separate from the existing architecture for doing type equality. It would also be somewhat unique in that it's a non-rigid ty::Const (it can be normalized more later on in type inference) while also having somewhat "structural" equality behaviour.

Lastly, it's worth noting that its not actually ConstKind::ADT that we want. It's desirable to extend this setup to also support tuples and arrays, or even references if we wind up supporting those in const generics. Therefore this isn't really ConstKind::ADT but a more general ConstKind::ShallowValue or something to that effect. It represents at least one "layer" of a types value :')

Instead of doing this implementation choice we instead change ValTree::Branch:

enum ValTree<'tcx> {
    Leaf(ScalarInt),
    // Before this PR:
    Branch(Box<[ValTree<'tcx>]>),
    // After this PR
    Branch(Box<[Const<'tcx>]>),
}

The representation for so called "shallow values" is now the same as the representation for the entire full value. The desired inference/type equality behaviour just falls right out of this. We also don't wind up with these shallow values actually being non-rigid. And ValTree already supports references/tuples/arrays so we can handle those just fine.

I think in the future it might be worth considering inlining ValTree into ty::ConstKind. E.g:

enum ConstKind {
    Scalar(Ty<'tcx>, ScalarInt),
    ShallowValue(Ty<'tcx>, List<Const<'tcx>>),
    Unevaluated(UnevaluatedConst<'tcx>),
    ...
}

This would imply that the usage of ValTrees in patterns would now be using ty::Const but they already kind of are anyway and I think that's probably okay in the long run. It also would mean that the set of things we could represent in const patterns is greater which may be desirable in the long run for supporting things such as const patterns of const generic parameters.

Regardless, this PR doesn't actually inline ValTree into ty::ConstKind, it only changes Branch to recurse through Const. This change could be split out of this PR if desired.

I'm not sure if there'll be a perf impact from this change. It's somewhat plausible as now all const pattern values that have nesting will be interning a lot more Tys. We shall see :>

Forbidding generic parameters under mgca

Under mgca we now allow all const arguments to resolve paths to generic parameters. We then later actually validate that the const arg should be allowed to access generic parameters if it did wind up resolving to any.

This winds up just being a lot simpler to implement than trying to make name resolution "keep track" of whether we're inside of a non-anon-const const arg and then encounter a const { ... } indicating we should now stop allowing resolving to generic parameters.

It's also somewhat in line with what we'll need for a feature(generic_const_args) where we'll want to decide whether an anon const should have any generic parameters based off syntactically whether any generic parameters were used. Though that design is entirely hypothetical at this point :)

rustbot · 2025-11-19T21:36:34Z

oli-obk is not on the review rotation at the moment.
They may take a while to respond.

BoxyUwU · 2025-11-20T01:16:22Z

@bors try @rust-timer queue

MGCA: Support struct expressions without intermediary anon consts

compiler/rustc_middle/src/arena.rs

compiler/rustc_ast/src/ast.rs

BoxyUwU · 2025-11-20T01:48:27Z

compiler/rustc_ty_utils/src/consts.rs

            let (variant_idx, branches) = if def.is_enum() {
                let (head, rest) = branches.split_first().unwrap();
-                (VariantIdx::from_u32(head.unwrap_leaf().to_u32()), rest)
+                (VariantIdx::from_u32(head.to_value().valtree.unwrap_leaf().to_u32()), rest)


the addition of all the .to_value().valtree everywhere is kind of unfortunate

BoxyUwU · 2025-11-20T01:49:05Z

compiler/rustc_mir_build/src/thir/constant.rs

            // We need a branch for each "level" of the data structure.
            let bytes = ty::ValTree::from_raw_bytes(tcx, byte_sym.as_byte_str());
-            ty::ValTree::from_branches(tcx, [bytes])
+            ty::ValTree::from_branches(tcx, [ty::Const::new_value(tcx, bytes, *inner_ty)])


same with all the extra calls to ty::Const::new_x

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

rust-bors · 2025-11-20T03:33:56Z

☀️ Try build successful (CI)
Build commit: eeb4682 (eeb4682d365158169a665813e65ad6eabb62448b, parent: d2f887349fe3ea079a4f89b020ce6df1993e1e06)

rust-timer · 2025-11-20T04:53:15Z

Finished benchmarking commit (eeb4682): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

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.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.3%	[0.1%, 0.7%]	18
Regressions ❌ (secondary)	0.4%	[0.1%, 0.7%]	8
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.5%, -0.1%]	9
All ❌✅ (primary)	0.3%	[0.1%, 0.7%]	18

Max RSS (memory usage)

Results (primary 1.2%, secondary 1.0%)

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

	mean	range	count
Regressions ❌ (primary)	1.2%	[1.0%, 1.4%]	2
Regressions ❌ (secondary)	1.0%	[1.0%, 1.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.2%	[1.0%, 1.4%]	2

Cycles

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)	3.0%	[3.0%, 3.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.2%	[-3.2%, -3.2%]	1
All ❌✅ (primary)	-	-	0

Binary size

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)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Bootstrap: 474.035s -> 472.627s (-0.30%)
Artifact size: 388.78 MiB -> 388.92 MiB (0.04%)

BoxyUwU · 2025-11-20T06:30:55Z

@bors try @rust-timer queue

lcnr · 2025-12-16T13:31:20Z

compiler/rustc_hir_typeck/src/lib.rs


+    if let DefKind::AnonConst = tcx.def_kind(def_id)
+        && let ty::AnonConstKind::MCG = tcx.anon_const_kind(def_id)
+        && let Err(e) = tcx.hir_const_arg_anon_check_invalid_param_uses(def_id)


check_const_arg_invalid_param_uses?

lcnr · 2025-12-16T13:59:53Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

+            && let Err(e) = tcx.hir_const_arg_anon_check_invalid_param_uses(anon.def_id)
+        {
+            return ty::Const::new_error(tcx, e);
+        }


is this necessary?

lcnr · 2025-12-16T14:06:06Z

compiler/rustc_hir_typeck/src/lib.rs

+        && let ty::AnonConstKind::MCG = tcx.anon_const_kind(def_id)
+        && let Err(e) = tcx.hir_const_arg_anon_check_invalid_param_uses(def_id)
+    {
+        e.raise_fatal();


kinda meh:

three options

wrap with Result<TypeckResults, ErrorGuaranteed>

meh, not used by other code

make a dummy TypeckResults

fragile, we don't have good test coverage for how e.g. lints access the typeck results, likely to cause a bunch of rare ICE

change HirTyLowering to disable lowering anything to ty::Param/ConstKind

similar to "forbid lower to infer"

seems best, add fixme for this pls :3

lcnr · 2025-12-16T14:08:08Z

compiler/rustc_middle/src/hir/map.rs

        None
    }
+
+    pub fn hir_const_arg_anon_check_invalid_param_uses(


pretty iffy, idk, add fixme that we should remove it medium-term

lcnr · 2025-12-16T14:22:00Z

compiler/rustc_ast_lowering/src/lib.rs

+                    expr.id,
+                    &se.qself,
+                    &se.path,
+                    ParamMode::Explicit,


add FIXME that we should use Optional here and make sure the ItemCtxt still errors if u do it in a signature

lcnr · 2025-12-16T14:23:43Z

compiler/rustc_ast_lowering/src/lib.rs

+
+                let fields = self.arena.alloc_from_iter(se.fields.iter().map(|f| {
+                    let hir_id = self.lower_node_id(f.id);
+                    self.lower_attrs(hir_id, &f.attrs, f.span, Target::ExprField);


add fixme to look into this, I expect this to be broken as it causes us to silently allow attrs

lcnr · 2025-12-16T14:27:47Z

compiler/rustc_ast_lowering/src/lib.rs

+                        self.lower_anon_const_to_const_arg_direct(anon_const)
+                    } else {
+                        self.lower_expr_to_const_arg_direct(&f.expr)


this naming sucks :3 we should rename the lower_ty and lower_ty_direct to lower_ty_and_intern and lower_ty.

Same with const stuff.

lcnr · 2025-12-16T14:32:42Z

compiler/rustc_hir/src/intravisit.rs

 ) -> V::Result {
    let ConstArg { hir_id, kind } = const_arg;
    try_visit!(visitor.visit_id(*hir_id));
+


why whitespace

lcnr · 2025-12-16T14:37:31Z

compiler/rustc_middle/src/hir/mod.rs

            Node::Block(_)
            | Node::Arm(_)
            | Node::ExprField(_)
+            | Node::ConstArgExprField(_)


try to move into an unreachable branch and then we get a test if it ICEs after merge

lcnr · 2025-12-16T14:38:24Z

compiler/rustc_resolve/src/def_collector.rs

        // `MgcaDisambiguation::Direct` is set even when MGCA is disabled, so
        // to avoid affecting stable we have to feature gate the not creating
        // anon consts
        if let MgcaDisambiguation::Direct = constant.mgca_disambiguation


make this a match now that mgca impacts both branches

lcnr · 2025-12-16T14:40:51Z

compiler/rustc_resolve/src/def_collector.rs

+                handle_const_block(self, constant, DefKind::InlineConst);
+                return;


return handle_const_block(self, constant, DefKind::InlineConst);

lcnr · 2025-12-16T14:41:21Z

compiler/rustc_resolve/src/def_collector.rs

+
+                match rest {
+                    StructRest::Base(expr) => self.visit_expr(expr),
+                    _ => (),


??? why no exhaustive

lcnr · 2025-12-16T14:46:12Z

compiler/rustc_resolve/src/def_collector.rs

            _ => self.invocation_parent.parent_def,
        };

        self.with_parent(parent_def, |this| visit::walk_expr(this, expr))


can you rewrite this to only use with_parent if we actually create a new def_id and just walk_expr in the other branches?

lcnr · 2025-12-16T14:47:56Z

compiler/rustc_resolve/src/def_collector.rs

+
+                for init_expr in fields {
+                    if let ExprKind::ConstBlock(ref constant) = init_expr.expr.kind {
+                        handle_const_block(self, constant, DefKind::AnonConst);


please add a comment that we only do this to give const blocks a different DefKind, the rest is just a walk

lcnr · 2025-12-16T14:52:41Z

compiler/rustc_type_ir/src/const_kind.rs

+    /// the fields of the variant.
+    ///
+    /// ZST types are represented as an empty slice.
+    Branch(Box<[I::Const]>),


pls add FIXME to use a List

also add FIXME to struct that we shouldn't intern valtree anymore, we already intern Const and valtree is now always directly inside of Const

lcnr · 2025-12-16T14:54:04Z

compiler/rustc_type_ir/src/inherent.rs

 }

+pub trait ValTree<I: Interner<ValTree = Self>>: Copy + Debug + Hash + Eq {
+    // This isnt' `IntoKind` because then we can't return a reference


add a FIXME that we don't need this once we properly intern instead of using Box<[_]>

lcnr · 2025-12-16T14:58:34Z

compiler/rustc_middle/src/ty/structural_impls.rs

+
+    fn fold_with<F: TypeFolder<TyCtxt<'tcx>>>(self, folder: &mut F) -> Self {
+        let inner: &ty::ValTreeKind<TyCtxt<'tcx>> = &*self;
+        folder.cx().intern_valtree(inner.clone().fold_with(folder))


before reintern, check for eq

lcnr · 2025-12-16T14:58:46Z

compiler/rustc_middle/src/ty/structural_impls.rs

+        folder: &mut F,
+    ) -> Result<Self, F::Error> {
+        let inner: &ty::ValTreeKind<TyCtxt<'tcx>> = &*self;
+        let valtree = folder.cx().intern_valtree(inner.clone().try_fold_with(folder)?);


also check for eq here

lcnr · 2025-12-16T14:59:37Z

compiler/rustc_middle/src/ty/consts/valtree.rs

+                self.valtree
+                    .unwrap_branch()
+                    .into_iter()
+                    .map(|ct| ct.to_value().valtree.unwrap_leaf().to_u8()),


pls add a method to ty::Const to get the leaf and use that

lcnr · 2025-12-16T15:00:22Z

compiler/rustc_middle/src/ty/consts/valtree.rs

+impl<'tcx> ty::ValTreeKind<TyCtxt<'tcx>> {
    #[inline]
-    pub fn unwrap_leaf(&self) -> ScalarInt {
+    fn unwrap_leaf(&self) -> ScalarInt {


pls move as many as possible to rustc_type_ir instead of adding an extension trait

lcnr · 2025-12-16T15:03:03Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

        }
    }

+    fn lower_struct_expr_const_arg(


fn lower_const_arg_struct or lower_const_arg_expr_struct idk, sth that's consistent and not ass :>

lcnr · 2025-12-16T15:04:28Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

+    ) -> Const<'tcx> {
+        let tcx = self.tcx();
+
+        let non_adt_or_variant_res = || {


add FIXME that most of this function should be reused with normal struct expr hir lowering

lcnr · 2025-12-16T15:08:18Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

+            .fields
+            .iter()
+            .map(|field_def| {
+                // FIXME(mgca): we aren't really handling privacy or stability at all but we should


and macro visibility

lcnr · 2025-12-16T15:08:55Z

tests/ui/const-generics/mgca/adt_expr_erroneuous_inits.rs

+#![expect(incomplete_features)]
+#![crate_type = "lib"]
+
+#[derive(Eq, PartialEq, std::marker::ConstParamTy)]


lcnr · 2025-12-16T15:09:24Z

tests/ui/const-generics/mgca/adt_expr_fields_type_check.rs

@@ -0,0 +1,17 @@
+//@ check-pass
+// This should error


change to FIXME

rust-bors · 2025-12-16T15:45:43Z

☀️ Try build successful (CI)
Build commit: b925bed (b925bed27a63ed0045d0e799ff62703dcd9ffb84, parent: 61cc47e367d7be91e13bcd01e4e96e0e233d4f6d)

rust-timer · 2025-12-16T16:26:50Z

Finished benchmarking commit (b925bed): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

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.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.3%	[0.1%, 0.5%]	11
Regressions ❌ (secondary)	0.1%	[0.1%, 0.2%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.6%, -0.1%]	16
All ❌✅ (primary)	0.3%	[0.1%, 0.5%]	11

Max RSS (memory usage)

Results (primary -1.5%, secondary 2.3%)

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)	3.7%	[3.6%, 3.9%]	2
Improvements ✅ (primary)	-1.5%	[-1.5%, -1.5%]	1
Improvements ✅ (secondary)	-0.7%	[-0.7%, -0.7%]	1
All ❌✅ (primary)	-1.5%	[-1.5%, -1.5%]	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)	2.2%	[2.2%, 2.2%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.3%	[-2.3%, -2.3%]	1
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 478.546s -> 479.261s (0.15%)
Artifact size: 390.29 MiB -> 390.43 MiB (0.04%)

rustbot assigned oli-obk Nov 19, 2025

BoxyUwU force-pushed the mgca_adt_exprs branch from 0dbf003 to a019ac7 Compare November 19, 2025 21:45