traits: Allow escaping self types in ExistentialTraitRef::with_self_ty by Dnreikronos · Pull Request #157280 · rust-lang/rust

Dnreikronos · 2026-06-02T02:25:27Z

WF-checking recurses through higher-ranked binders without instantiating them, so a dyn Trait nested inside a for<'a> binder reaches the ty::Dynamic arm of WfPredicates::visit_ty while still carrying escaping bound vars. Building the principal trait ref there via ExistentialTraitRef::with_self_ty hands that escaping self type to a debug_assert!(!self_ty.has_escaping_bound_vars()), which ICEs once the assertion is enabled. The assert was accidentally disabled in 2018 refactor #53816, and the // FIXME(#157122) left in its place asks whether to remove it or fix the fallout.

An escaping bound var is expected for trait objects, and we already catch trait refs with escaping bound vars at the places that actually use them — so creating one here is fine. Rather than work around the assert in WF, this removes the assert and its stale FIXME from with_self_ty.

Two regression tests pin the behavior:

wf-dyn-in-hrtb-bound-issue-157122.rs reproduces the original ICE (dyn nested in an HRTB bound).
wf-dyn-in-hrtb-bound-const-mismatch.rs checks that the ConstArgHasType obligation this arm reads off still fires, so an ill-typed const argument on such a dyn keeps erroring instead of compiling clean.

rustbot · 2026-06-02T02:25:30Z

changes to the core type system

cc @lcnr

rustbot · 2026-06-02T02:25:33Z

r? @mejrs

rustbot has assigned @mejrs.
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

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 73 candidates
Random selection from 17 candidates

WF-checking walks through higher-ranked binders without instantiating them, so a `dyn Trait` nested inside a `for<'a>` bound reaches the `ty::Dynamic` arm of `WfPredicates::visit_ty` while still carrying escaping bound vars. Feeding that type into `ExistentialTraitRef::with_self_ty` violated its no-escaping-self precondition and tripped the assertion that guards it. The trait ref built there is only used to read off `ConstArgHasType` clauses, which constrain the trait's own const arguments and never mention the `Self` type. Substitute a fresh placeholder self type when the real one escapes: the assertion holds and the const-argument check is still performed. Re-enable the `with_self_ty` debug assertion now that its precondition is upheld.

One test covers the original ICE (a `dyn Trait` nested in a `for<'a>` bound, distilled from itertools). The other locks in that a const argument on such a nested `dyn` is still type-checked, so the placeholder-self-type fix cannot silently regress into dropping the `ConstArgHasType` check.

rustbot · 2026-06-02T02:49:28Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

mejrs · 2026-06-02T09:27:35Z

I'm not confident reviewing this.

r? @fmease perhaps?

Dnreikronos · 2026-06-02T12:38:32Z

I'm not confident reviewing this.

r? @fmease perhaps?

Hi, @mejrs!
I'm not 100% confident that this implementation that I did it's the best approach as possible to solve the problem. So I'm totally open/available to discuss about it

mejrs · 2026-06-02T13:30:34Z

Oh it's not about what and how you implemented this, I'm very unfamiliar with this part of the compiler so I'm ill-suited to review PRs touching it.

lcnr

sorry for the late review, RustWeek was a thing

View changes since this review

WF-checking can reach ExistentialTraitRef::with_self_ty with a dyn principal that still carries escaping bound vars, e.g. a dyn nested in a for<'a> bound, which tripped the debug assertion forbidding escaping self types and ICEd when compiling itertools. Building such a trait ref is fine: escaping bound vars are caught where it is actually used. Remove the assertion rather than working around it, which lets the WF visitor pass the type through unchanged and drops the placeholder-self-type detour.

lcnr · 2026-06-10T12:15:48Z

@bors delegate+

r=me after updating the PR name

rust-bors · 2026-06-10T12:15:51Z

✌️ @Dnreikronos, you can now approve this pull request!

If @lcnr told you to "r=me" after making some further change, then please make that change and post @bors r=lcnr.

View changes since this delegation.

Dnreikronos · 2026-06-10T12:55:08Z

just reenamed the PR.
the old name still described the placeholder self type approach we dropped.

@bors r=lcnr

rust-bors · 2026-06-10T12:55:16Z

📌 Commit 3669c24 has been approved by lcnr

It is now in the queue for this repository.

🌲 The tree is currently closed for pull requests below priority 5. This pull request will be tested once the tree is reopened.

…escaping_bound_vars_assert, r=lcnr traits: Allow escaping self types in ExistentialTraitRef::with_self_ty Fixes rust-lang#157122 WF-checking recurses through higher-ranked binders without instantiating them, so a `dyn Trait` nested inside a `for<'a>` binder reaches the `ty::Dynamic` arm of `WfPredicates::visit_ty` while still carrying escaping bound vars. Building the principal trait ref there via `ExistentialTraitRef::with_self_ty` hands that escaping self type to a `debug_assert!(!self_ty.has_escaping_bound_vars())`, which ICEs once the assertion is enabled. The assert was accidentally disabled in 2018 refactor rust-lang#53816, and the `// FIXME(rust-lang#157122)` left in its place asks whether to remove it or fix the fallout. An escaping bound var is expected for trait objects, and we already catch trait refs with escaping bound vars at the places that actually use them — so creating one here is fine. Rather than work around the assert in WF, this removes the assert and its stale FIXME from `with_self_ty`. Two regression tests pin the behavior: - `wf-dyn-in-hrtb-bound-issue-157122.rs` reproduces the original ICE (`dyn` nested in an HRTB bound). - `wf-dyn-in-hrtb-bound-const-mismatch.rs` checks that the `ConstArgHasType` obligation this arm reads off still fires, so an ill-typed const argument on such a `dyn` keeps erroring instead of compiling clean.

…uwer Rollup of 14 pull requests Successful merges: - #157280 (traits: Allow escaping self types in ExistentialTraitRef::with_self_ty) - #157282 (Fix post-monomorphization error note race in the parallel frontend) - #157352 (Make the retained dep graph deterministic under the parallel frontend) - #157601 (Emit error for unused target expression in glob and list delegations) - #157626 (Autogenerate unstable compiler flag stubs for unstable-book) - #157647 (Start using comptime for reflection intrinsics and their wrapper functions) - #157013 (Ensure inferred let pattern types are well-formed) - #157288 (platform support: add SNaN erratum to MIPS targets) - #157355 (Add `or_try_*` variants for `HashMap` and `BTreeMap` Entry APIs) - #157577 (Fix parser error recovery treating 'dyn' as a strict keyword in Rust …) - #157670 (Rename `errors.rs` file to `diagnostics.rs` (4/N)) - #157691 (Move symbol hiding code to a new file) - #157700 (Rename `errors.rs` file to `diagnostics.rs` (5/N)) - #157703 (Fix doc link to Instant sub in saturating caveat) Failed merges: - #157699 (Arg splat experiment - hir FnDecl impl)

…escaping_bound_vars_assert, r=lcnr traits: Allow escaping self types in ExistentialTraitRef::with_self_ty Fixes rust-lang#157122 WF-checking recurses through higher-ranked binders without instantiating them, so a `dyn Trait` nested inside a `for<'a>` binder reaches the `ty::Dynamic` arm of `WfPredicates::visit_ty` while still carrying escaping bound vars. Building the principal trait ref there via `ExistentialTraitRef::with_self_ty` hands that escaping self type to a `debug_assert!(!self_ty.has_escaping_bound_vars())`, which ICEs once the assertion is enabled. The assert was accidentally disabled in 2018 refactor rust-lang#53816, and the `// FIXME(rust-lang#157122)` left in its place asks whether to remove it or fix the fallout. An escaping bound var is expected for trait objects, and we already catch trait refs with escaping bound vars at the places that actually use them — so creating one here is fine. Rather than work around the assert in WF, this removes the assert and its stale FIXME from `with_self_ty`. Two regression tests pin the behavior: - `wf-dyn-in-hrtb-bound-issue-157122.rs` reproduces the original ICE (`dyn` nested in an HRTB bound). - `wf-dyn-in-hrtb-bound-const-mismatch.rs` checks that the `ConstArgHasType` obligation this arm reads off still fires, so an ill-typed const argument on such a `dyn` keeps erroring instead of compiling clean.

Rollup of 23 pull requests Successful merges: - #157280 (traits: Allow escaping self types in ExistentialTraitRef::with_self_ty) - #157282 (Fix post-monomorphization error note race in the parallel frontend) - #157352 (Make the retained dep graph deterministic under the parallel frontend) - #157601 (Emit error for unused target expression in glob and list delegations) - #157611 (Update `browser-ui-test` version to `0.24.0`) - #157620 (Add a strategy FnMut to inject behavior into the flush cycle) - #157645 (Windows TLS - Only register the `atexit` hook when `cleanup` can be unloaded) - #157647 (Start using comptime for reflection intrinsics and their wrapper functions) - #157719 (resolve: Partially revert "Remove a special case for dummy imports") - #156497 (fix-155516: Don't suggest wrong unwrap expect) - #156583 (Support defaults for static EIIs) - #157013 (Ensure inferred let pattern types are well-formed) - #157230 (borrowck: avoid ICE describing fields on generic params) - #157288 (platform support: add SNaN erratum to MIPS targets) - #157350 (compiletest: ignore SVG `y` offset in by-lines comparison) - #157355 (Add `or_try_*` variants for `HashMap` and `BTreeMap` Entry APIs) - #157384 (Add `#[rustc_dump_generics]` attribute) - #157577 (Fix parser error recovery treating 'dyn' as a strict keyword in Rust 2015 when used in `dyn + dyn`) - #157670 (Rename `errors.rs` file to `diagnostics.rs` (4/N)) - #157691 (Move symbol hiding code to a new file) - #157697 (Add more tests that exercise the well-formedness checking of lazy type aliases) - #157700 (Rename `errors.rs` file to `diagnostics.rs` (5/N)) - #157703 (Fix doc link to Instant sub in saturating caveat) Failed merges: - #157699 (Arg splat experiment - hir FnDecl impl)

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. labels Jun 2, 2026

rustbot assigned mejrs Jun 2, 2026

Dnreikronos added 2 commits June 1, 2026 23:48

Dnreikronos force-pushed the existential_trait_ref_escaping_bound_vars_assert branch from 87f6783 to 0676d4a Compare June 2, 2026 02:49

rustbot assigned fmease and unassigned mejrs Jun 2, 2026

oli-obk added T-types Relevant to the types team, which will review and decide on the PR/issue. and removed T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jun 2, 2026

lcnr reviewed Jun 9, 2026

View reviewed changes

Comment thread compiler/rustc_type_ir/src/predicate.rs Outdated

lcnr reviewed Jun 9, 2026

View reviewed changes

This comment was marked as outdated.

Sign in to view

fmease assigned lcnr and unassigned fmease Jun 9, 2026

Dnreikronos changed the title ~~traits: Use a placeholder self type for escaping dyn principals in WF~~ traits: Allow escaping self types in ExistentialTraitRef::with_self_ty Jun 10, 2026

rust-bors Bot 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-review Status: Awaiting review from the assignee but also interested parties. labels Jun 10, 2026

JonathanBrouwer mentioned this pull request Jun 10, 2026

Rollup of 14 pull requests #157708

Closed

JonathanBrouwer mentioned this pull request Jun 10, 2026

Rollup of 17 pull requests #157720

Closed

jhpratt mentioned this pull request Jun 10, 2026

Rollup of 23 pull requests #157726

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

traits: Allow escaping self types in ExistentialTraitRef::with_self_ty#157280

traits: Allow escaping self types in ExistentialTraitRef::with_self_ty#157280
Dnreikronos wants to merge 3 commits into
rust-lang:mainfrom
Dnreikronos:existential_trait_ref_escaping_bound_vars_assert

Dnreikronos commented Jun 2, 2026 •

edited by rustbot

Loading

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

mejrs commented Jun 2, 2026

Uh oh!

Dnreikronos commented Jun 2, 2026

Uh oh!

mejrs commented Jun 2, 2026

Uh oh!

Uh oh!

lcnr left a comment •

edited by rustbot

Loading

Uh oh!

This comment was marked as outdated.

lcnr commented Jun 10, 2026

Uh oh!

rust-bors Bot commented Jun 10, 2026

Uh oh!

Dnreikronos commented Jun 10, 2026 •

edited

Loading

Uh oh!

rust-bors Bot commented Jun 10, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

Uh oh!

Conversation

Dnreikronos commented Jun 2, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

rustbot commented Jun 2, 2026

Uh oh!

mejrs commented Jun 2, 2026

Uh oh!

Dnreikronos commented Jun 2, 2026

Uh oh!

mejrs commented Jun 2, 2026

Uh oh!

Uh oh!

lcnr 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!

This comment was marked as outdated.

lcnr commented Jun 10, 2026

Uh oh!

rust-bors Bot commented Jun 10, 2026

Uh oh!

Dnreikronos commented Jun 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rust-bors Bot commented Jun 10, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

Dnreikronos commented Jun 2, 2026 •

edited by rustbot

Loading

lcnr left a comment •

edited by rustbot

Loading

Dnreikronos commented Jun 10, 2026 •

edited

Loading