fix ICE when freshTy(0) leaks into sizedness and fudge paths #151732

enthropy7 · 2026-01-27T10:54:16Z

using -Znext-solver=globally with dyn A<Ty = i32, CT = 0> was able to trigger a couple of different ICEs: first when trait_object_dummy_self (FreshTy(0)) leaked into sizedness fast paths and object candidates, and then when fudge_inference_if_ok ran into Fresh* vars and hit an unreachable! (i saw second one while fixing the original issue). this patch makes those code paths either bail out early or just treat Fresh* as regular types (including forwarding them through InferenceFudger)

also, of course i added a test

about the issue: when we lower dyn A<Ty = i32, CT = 0> with the new trait solver, we use an internal dummy Self called trait_object_dummy_self, represented as FreshTy(0) (and FreshIntTy / FreshFloatTy for the other cases). that means the solver ends up with goals like FreshTy(0): MetaSized or FreshTy(0): A<…>, even though FreshTy(0) is just the Self of this dyn object, not a real user type. several fast paths and bits of diagnostics assumed they would never see Fresh* there (has_trivial_sizedness, sizedness assembly, for_each_relevant_impl, fudge_inference_if_ok) and buged or hit unreachable!, so once trait_object_dummy_self started going through those paths, we got ICEs instead of normal errors.

fixes #151709

FreshTy is used as trait_object_dummy_self for dyn types and should not reach has_trivial_sizedness or assemble_object_bound_candidates. add checks to skip the fast path and return early for FreshTy types to prevent ICEs in the new trait solver.

rustbot · 2026-01-27T10:54:21Z

r? @BoxyUwU

rustbot has assigned @BoxyUwU.
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

fmease

Thanks for trying to fix this but this is not the right fix. trait_object_dummy_self should never escape HIR ty lowering, it's lower_trait_object_ty's fault that it currently does. This issue has been known for around 2 years and there are other similar GH issues which are still open.

See e.g., this source code comment I've recently added:

rust/compiler/rustc_hir_analysis/src/hir_ty_lowering/dyn_trait.rs

Lines 61 to 65 in 78df2f9

    
           // FIXME(mgca): We actually leak `trait_object_dummy_self` if the type of any assoc 
        
           //              const mentions `Self` (in "Self projections" which we intentionally 
        
           //              allow). That's because we query feed the instantiated type to `type_of`. 
        
           //              That's really bad, dummy self should never escape lowering! It leads us 
        
           //              to accept less code we'd like to support and can lead to ICEs later.

Or this one:

rust/tests/ui/const-generics/associated-const-bindings/dyn-compat-self-const-projections-in-assoc-const-ty.rs

Lines 1 to 5 in 78df2f9

    
           // FIXME(mgca): Ideally this would compile -- at least if the user annotated the instantiated type 
        
           //              of the assoc const (but we don't have the syntax for this (yet)). In any case, we 
        
           //              should not leak `trait_object_dummy_self` (defined as `FreshTy(0)` under the hood) 
        
           //              to the rest of the compiler and by extension the user via diagnostics. 
        
           //@ known-bug: unknown

The proper fix will be different, I don't yet have a clear picture in my head though.

View changes since this review

rustbot · 2026-01-27T11:22:13Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

fmease · 2026-01-27T11:29:46Z

I don't think it makes sense to keep this PR open since the final approach will look completely different. Therefore I'm going to close this. Boxy, please correct me if I'm wrong.

CC https://github.com/rust-lang/rust/blob/main/tests/ui/const-generics/associated-const-bindings/dyn-compat-self-const-projections-in-assoc-const-ty.rs and maybe also #123140.

These kinds of issues are intentionally kept unsolved right now (see e.g., PR description of #150843).

rust-log-analyzer · 2026-01-27T11:33:05Z

The job aarch64-gnu-llvm-20-1 failed! Check out the build log: (web) (plain enhanced) (plain)

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

test [ui] tests/ui/asm/aarch64/type-check-2-2.rs ... ok
test [ui] tests/ui/asm/aarch64/type-check-3.rs ... ok
test [ui] tests/ui/asm/aarch64/ttbr0_el2.rs ... ok
test [ui] tests/ui/asm/aarch64/type-check-2.rs ... ok
test [ui] tests/ui/asm/aarch64v8r.rs#hf ... ok
test [ui] tests/ui/asm/aarch64/type-f16.rs ... ok
test [ui] tests/ui/asm/aarch64v8r.rs#sf ... ok
test [ui] tests/ui/asm/arm-low-dreg.rs ... ok
test [ui] tests/ui/asm/asm-with-nested-closure.rs ... ok
test [ui] tests/ui/asm/binary_asm_labels.rs ... ignored, only executed when the architecture is x86_64
test [ui] tests/ui/asm/bad-template.rs#x86_64 ... ok
test [ui] tests/ui/asm/cfg-parse-error.rs ... ok
---
test [ui] tests/ui/extern/issue-64655-extern-rust-must-allow-unwind.rs#fat2 ... ok
test [ui] tests/ui/extern/issue-64655-extern-rust-must-allow-unwind.rs#fat3 ... ok
test [ui] tests/ui/extern/issue-80074.rs ... ok
test [ui] tests/ui/extern/issue-95829.rs ... ok
test [ui] tests/ui/extern/lgamma-linkage.rs ... ok
test [ui] tests/ui/extern/no-mangle-associated-fn.rs ... ok
test [ui] tests/ui/extern/not-in-block.rs ... ok
test [ui] tests/ui/extern/unsized-extern-derefmove.rs ... ok
test [ui] tests/ui/extern/windows-tcb-trash-13259.rs ... ok
test [ui] tests/ui/feature-gates/allow-features-empty.rs ... ok
---
test [ui] tests/ui/imports/ambiguous-2.rs ... ok
test [ui] tests/ui/imports/ambiguous-4.rs ... ok
test [ui] tests/ui/imports/ambiguous-7.rs ... ok
test [ui] tests/ui/imports/ambiguous-9.rs ... ok
test [ui] tests/ui/imports/ambiguous-panic-glob-vs-multiouter.rs ... ok
test [ui] tests/ui/imports/ambiguous-panic-globvsglob.rs ... ok
test [ui] tests/ui/imports/ambiguous-8.rs ... ok
test [ui] tests/ui/imports/ambiguous-panic-no-implicit-prelude.rs ... ok
test [ui] tests/ui/imports/ambiguous-glob-vs-expanded-extern.rs ... ok
test [ui] tests/ui/imports/ambiguous-panic-non-prelude-core-glob.rs ... ok
test [ui] tests/ui/imports/ambiguous-panic-non-prelude-std-glob.rs ... ok
---
test [ui] tests/ui/zero-sized/zero-sized-btreemap-insert.rs ... ok

failures:

---- [ui] tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs stdout ----
Saved the actual stderr to `/checkout/obj/build/aarch64-unknown-linux-gnu/test/ui/traits/next-solver/fresh-ty-has-trivial-sizedness/fresh-ty-has-trivial-sizedness.stderr`
diff of stderr:

1 error[E0658]: associated const equality is incomplete
-   --> $DIR/fresh-ty-has-trivial-sizedness.rs:12:29
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:29
3    |
4 LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
5    |                             ^^^^^^

9    = note: this compiler was built on YYYY-MM-DD; consider upgrading it if it is out of date
10 
11 error[E0038]: the trait `A` is not dyn compatible
-   --> $DIR/fresh-ty-has-trivial-sizedness.rs:12:17
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:17
13    |
14 LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
15    |                 ^^^^^^^^^^^^^^^^^^^ `A` is not dyn compatible

16    |
17 note: for a trait to be dyn compatible it needs to allow building a vtable
18       for more information, visit <https://doc.rust-lang.org/reference/items/traits.html#dyn-compatibility>
-   --> $DIR/fresh-ty-has-trivial-sizedness.rs:8:11
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:6:11
20    |
21 LL | trait A {
22    |       - this trait is not dyn compatible...

23 LL |     type Ty;
24 LL |     const CT: Self::Ty;
-    |           ^^ ...because it contains this associated `const`
+    |           ^^ ...because it contains associated const `CT`
26    = help: consider moving `CT` to another trait
27 
28 error[E0277]: the size for values of type `FreshTy(0)` cannot be known

-   --> $DIR/fresh-ty-has-trivial-sizedness.rs:12:34
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:34
30    |
31 LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
32    |                                  ^ doesn't have a known size

33    |
34    = help: the nightly-only, unstable trait `MetaSized` is not implemented for `FreshTy(0)`
35 note: required by a bound in `A`
---
39 LL | |     type Ty;

42    | |_^ required by this bound in `A`
43 
44 error[E0277]: the trait bound `FreshTy(0): A` is not satisfied
-   --> $DIR/fresh-ty-has-trivial-sizedness.rs:12:34
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:34
46    |
47 LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
48    |                                  ^ the trait `A` is not implemented for `FreshTy(0)`

49    |
50 help: this trait has no implementations, consider adding one
-   --> $DIR/fresh-ty-has-trivial-sizedness.rs:6:1
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:4:1
52    |
53 LL | trait A {
54    | ^^^^^^^

Note: some mismatched output was normalized before being compared
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:29
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:17
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:6:11
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:34
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:4:1
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:34
-   --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:4:1
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:29
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:17
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:6:11
+    |           ^^ ...because it contains associated const `CT`
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:34
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:4:1
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:10:34
+   --> $DIR/fresh-ty-has-trivial-sizedness.rs:4:1


The actual stderr differed from the expected stderr
To update references, rerun the tests and pass the `--bless` flag
To only update this specific test, also pass `--test-args traits/next-solver/fresh-ty-has-trivial-sizedness.rs`

error: 1 errors occurred comparing output.
status: exit status: 1
command: env -u RUSTC_LOG_COLOR RUSTC_ICE="0" RUST_BACKTRACE="short" "/checkout/obj/build/aarch64-unknown-linux-gnu/stage2/bin/rustc" "/checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs" "-Zthreads=1" "-Zsimulate-remapped-rust-src-base=/rustc/FAKE_PREFIX" "-Ztranslate-remapped-path-to-local-path=no" "-Z" "ignore-directory-in-diagnostics-source-blocks=/cargo" "-Z" "ignore-directory-in-diagnostics-source-blocks=/checkout/vendor" "--sysroot" "/checkout/obj/build/aarch64-unknown-linux-gnu/stage2" "--target=aarch64-unknown-linux-gnu" "--check-cfg" "cfg(test,FALSE)" "--error-format" "json" "--json" "future-incompat" "-Ccodegen-units=1" "-Zui-testing" "-Zdeduplicate-diagnostics=no" "-Zwrite-long-types-to-disk=no" "-Cstrip=debuginfo" "--emit" "metadata" "-C" "prefer-dynamic" "--out-dir" "/checkout/obj/build/aarch64-unknown-linux-gnu/test/ui/traits/next-solver/fresh-ty-has-trivial-sizedness" "-A" "unused" "-W" "unused_attributes" "-A" "internal_features" "-A" "unused_parens" "-A" "unused_braces" "-Crpath" "-Cdebuginfo=0" "-Lnative=/checkout/obj/build/aarch64-unknown-linux-gnu/native/rust-test-helpers" "-Znext-solver=globally"
stdout: none
--- stderr -------------------------------
error[E0658]: associated const equality is incomplete
##[error]  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:29
   |
LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
   |                             ^^^^^^
   |
   = note: see issue #132980 <https://github.com/rust-lang/rust/issues/132980> for more information
   = help: add `#![feature(min_generic_const_args)]` to the crate attributes to enable
   = note: this compiler was built on YYYY-MM-DD; consider upgrading it if it is out of date

error[E0038]: the trait `A` is not dyn compatible
##[error]  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:17
   |
LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
   |                 ^^^^^^^^^^^^^^^^^^^ `A` is not dyn compatible
   |
note: for a trait to be dyn compatible it needs to allow building a vtable
      for more information, visit <https://doc.rust-lang.org/reference/items/traits.html#dyn-compatibility>
  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:6:11
   |
LL | trait A {
   |       - this trait is not dyn compatible...
LL |     type Ty;
LL |     const CT: Self::Ty;
   |           ^^ ...because it contains associated const `CT`
   = help: consider moving `CT` to another trait

error[E0277]: the size for values of type `FreshTy(0)` cannot be known
##[error]  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:34
   |
LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
   |                                  ^ doesn't have a known size
   |
   = help: the nightly-only, unstable trait `MetaSized` is not implemented for `FreshTy(0)`
note: required by a bound in `A`
  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:4:1
   |
LL | / trait A {
LL | |     type Ty;
LL | |     const CT: Self::Ty;
LL | | }
   | |_^ required by this bound in `A`

error[E0277]: the trait bound `FreshTy(0): A` is not satisfied
##[error]  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:10:34
   |
LL |     let _: &dyn A<Ty = i32, CT = 0> = &();
   |                                  ^ the trait `A` is not implemented for `FreshTy(0)`
   |
help: this trait has no implementations, consider adding one
  --> /checkout/tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs:4:1
   |
LL | trait A {
---
Some errors have detailed explanations: E0038, E0277, E0658.
For more information about an error, try `rustc --explain E0038`.
------------------------------------------

---- [ui] tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs stdout end ----

failures:
    [ui] tests/ui/traits/next-solver/fresh-ty-has-trivial-sizedness.rs

test result: FAILED. 20131 passed; 1 failed; 332 ignored; 0 measured; 21 filtered out; finished in 553.37s

rustbot assigned BoxyUwU Jan 27, 2026

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. WG-trait-system-refactor The Rustc Trait System Refactor Initiative (-Znext-solver) labels Jan 27, 2026

lcnr self-assigned this Jan 27, 2026

fmease requested changes Jan 27, 2026

View reviewed changes

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 27, 2026

fmease closed this Jan 27, 2026

rustbot removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Jan 27, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix ICE when freshTy(0) leaks into sizedness and fudge paths #151732

fix ICE when freshTy(0) leaks into sizedness and fudge paths #151732

enthropy7 commented Jan 27, 2026 •

edited

Loading

Uh oh!

rustbot commented Jan 27, 2026

Uh oh!

fmease left a comment •

edited

Loading

Uh oh!

rustbot commented Jan 27, 2026

Uh oh!

fmease commented Jan 27, 2026

Uh oh!

rust-log-analyzer commented Jan 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

	// FIXME(mgca): We actually leak `trait_object_dummy_self` if the type of any assoc
	// const mentions `Self` (in "Self projections" which we intentionally
	// allow). That's because we query feed the instantiated type to `type_of`.
	// That's really bad, dummy self should never escape lowering! It leads us
	// to accept less code we'd like to support and can lead to ICEs later.

	// FIXME(mgca): Ideally this would compile -- at least if the user annotated the instantiated type
	// of the assoc const (but we don't have the syntax for this (yet)). In any case, we
	// should not leak `trait_object_dummy_self` (defined as `FreshTy(0)` under the hood)
	// to the rest of the compiler and by extension the user via diagnostics.
	//@ known-bug: unknown

Uh oh!

fix ICE when freshTy(0) leaks into sizedness and fudge paths #151732

fix ICE when freshTy(0) leaks into sizedness and fudge paths #151732

Conversation

enthropy7 commented Jan 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jan 27, 2026

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jan 27, 2026

Uh oh!

fmease commented Jan 27, 2026

Uh oh!

rust-log-analyzer commented Jan 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

enthropy7 commented Jan 27, 2026 •

edited

Loading

fmease left a comment •

edited

Loading