Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts by compiler-errors · Pull Request #145620 · rust-lang/rust

compiler-errors · 2025-08-19T18:56:25Z

When we have an impossible where clause like dyn Trait<u8>: Sized, this may make a dyn-to-dyn cast like dyn Trait<()> -> dyn trait<u8> to successfully type check as if it were a wide-to-thin ptr cast (discarding metadata):

rust/compiler/rustc_hir_typeck/src/cast.rs

Lines 862 to 865 in 16ad385

    
           // Cast to thin pointer is OK 
        
           if dst_kind == PointerKind::Thin { 
        
               return Ok(CastKind::PtrPtrCast); 
        
           }

In borrowck, we are expecting that the only meaningful dyn-to-dyn cast to be a metadata-preserving wide-to-wide ptr cast, which requires that the principals of the dyn pointers are equal. Borrowck additionally assumes that these principals have already been proven equal modulo regions, and we thus ICE since Trait<u8> and Trait<()> do not unify:

rust/compiler/rustc_borrowck/src/type_check/mod.rs

Lines 1481 to 1524 in 16ad385

    
           // This checks (lifetime part of) vtable validity for pointer casts, 
        
           // which is irrelevant when there are aren't principal traits on 
        
           // both sides (aka only auto traits). 
        
           // 
        
           // Note that other checks (such as denying `dyn Send` -> `dyn 
        
           // Debug`) are in `rustc_hir_typeck`. 
        
           if let ty::Dynamic(src_tty, _src_lt, ty::Dyn) = *src_tail.kind() 
        
               && let ty::Dynamic(dst_tty, dst_lt, ty::Dyn) = *dst_tail.kind() 
        
               && src_tty.principal().is_some() 
        
               && dst_tty.principal().is_some() 
        
           { 
        
               // Remove auto traits. 
        
               // Auto trait checks are handled in `rustc_hir_typeck` as FCW. 
        
               let src_obj = Ty::new_dynamic( 
        
                   tcx, 
        
                   tcx.mk_poly_existential_predicates( 
        
                       &src_tty.without_auto_traits().collect::<Vec<_>>(), 
        
                   ), 
        
                   // FIXME: Once we disallow casting `*const dyn Trait + 'short` 
        
                   // to `*const dyn Trait + 'long`, then this can just be `src_lt`. 
        
                   dst_lt, 
        
                   ty::Dyn, 
        
               ); 
        
               let dst_obj = Ty::new_dynamic( 
        
                   tcx, 
        
                   tcx.mk_poly_existential_predicates( 
        
                       &dst_tty.without_auto_traits().collect::<Vec<_>>(), 
        
                   ), 
        
                   dst_lt, 
        
                   ty::Dyn, 
        
               ); 
        
               debug!(?src_tty, ?dst_tty, ?src_obj, ?dst_obj); 
        
               self.sub_types( 
        
                   src_obj, 
        
                   dst_obj, 
        
                   location.to_locations(), 
        
                   ConstraintCategory::Cast { 
        
                       is_implicit_coercion: false, 
        
                       unsize_to: None, 
        
                   }, 
        
               ) 
        
               .unwrap();

This PR fixes this ICE by checking whether the RHS of the cast is considered to be Sized in the environment of the MIR typeck, and if so then skipping over this dyn->dyn principal compatibility check.

r? @lcnr perhaps?

compiler/rustc_borrowck/src/type_check/mod.rs

lcnr

nit, then r=me

View changes since this review

compiler/rustc_borrowck/src/type_check/mod.rs

…yn casts

rustbot · 2025-08-23T17:13:17Z

This PR was rebased onto a different master 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.

compiler-errors · 2025-08-24T20:37:00Z

@bors r=lcnr

bors · 2025-08-24T20:37:03Z

📌 Commit e0fb6eb has been approved by lcnr

It is now in the queue for this repository.

…=lcnr Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts Fixes rust-lang#141806 When we have an impossible where clause like `dyn Trait<u8>: Sized`, this may make a dyn-to-dyn cast like `dyn Trait<()> -> dyn trait<u8>` to successfully type check as if it were a wide-to-thin ptr cast (discarding metadata): https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_hir_typeck/src/cast.rs#L862-L865 In borrowck, we are expecting that the only meaningful dyn-to-dyn cast to be a metadata-preserving wide-to-wide ptr cast, which requires that the principals of the dyn pointers are equal. Borrowck additionally assumes that these principals have already been proven equal *modulo regions*, and we thus ICE since `Trait<u8>` and `Trait<()>` do not unify: https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_borrowck/src/type_check/mod.rs#L1481-L1524 This PR fixes this ICE by checking whether the RHS of the cast is considered to be Sized in the environment of the MIR typeck, and if so then skipping over this dyn->dyn principal compatibility check. r? `@lcnr` perhaps?

…=lcnr Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts Fixes rust-lang#141806 When we have an impossible where clause like `dyn Trait<u8>: Sized`, this may make a dyn-to-dyn cast like `dyn Trait<()> -> dyn trait<u8>` to successfully type check as if it were a wide-to-thin ptr cast (discarding metadata): https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_hir_typeck/src/cast.rs#L862-L865 In borrowck, we are expecting that the only meaningful dyn-to-dyn cast to be a metadata-preserving wide-to-wide ptr cast, which requires that the principals of the dyn pointers are equal. Borrowck additionally assumes that these principals have already been proven equal *modulo regions*, and we thus ICE since `Trait<u8>` and `Trait<()>` do not unify: https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_borrowck/src/type_check/mod.rs#L1481-L1524 This PR fixes this ICE by checking whether the RHS of the cast is considered to be Sized in the environment of the MIR typeck, and if so then skipping over this dyn->dyn principal compatibility check. r? ``@lcnr`` perhaps?

Rollup of 6 pull requests Successful merges: - #135761 (Dial down detail of B-tree description) - #144373 (remove deprecated Error::description in impls) - #145620 (Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts) - #145783 (add span to struct pattern rest (..)) - #145817 (cg_llvm: Replace the `llvm::Bool` typedef with a proper newtype) - #145820 (raw-dylib-elf: set correct `DT_VERDEFNUM`) r? `@ghost` `@rustbot` modify labels: rollup

…=lcnr Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts Fixes rust-lang#141806 When we have an impossible where clause like `dyn Trait<u8>: Sized`, this may make a dyn-to-dyn cast like `dyn Trait<()> -> dyn trait<u8>` to successfully type check as if it were a wide-to-thin ptr cast (discarding metadata): https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_hir_typeck/src/cast.rs#L862-L865 In borrowck, we are expecting that the only meaningful dyn-to-dyn cast to be a metadata-preserving wide-to-wide ptr cast, which requires that the principals of the dyn pointers are equal. Borrowck additionally assumes that these principals have already been proven equal *modulo regions*, and we thus ICE since `Trait<u8>` and `Trait<()>` do not unify: https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_borrowck/src/type_check/mod.rs#L1481-L1524 This PR fixes this ICE by checking whether the RHS of the cast is considered to be Sized in the environment of the MIR typeck, and if so then skipping over this dyn->dyn principal compatibility check. r? ```@lcnr``` perhaps?

Rollup of 5 pull requests Successful merges: - #135761 (Dial down detail of B-tree description) - #144373 (remove deprecated Error::description in impls) - #145620 (Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts) - #145817 (cg_llvm: Replace the `llvm::Bool` typedef with a proper newtype) - #145820 (raw-dylib-elf: set correct `DT_VERDEFNUM`) r? `@ghost` `@rustbot` modify labels: rollup

…=lcnr Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts Fixes rust-lang#141806 When we have an impossible where clause like `dyn Trait<u8>: Sized`, this may make a dyn-to-dyn cast like `dyn Trait<()> -> dyn trait<u8>` to successfully type check as if it were a wide-to-thin ptr cast (discarding metadata): https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_hir_typeck/src/cast.rs#L862-L865 In borrowck, we are expecting that the only meaningful dyn-to-dyn cast to be a metadata-preserving wide-to-wide ptr cast, which requires that the principals of the dyn pointers are equal. Borrowck additionally assumes that these principals have already been proven equal *modulo regions*, and we thus ICE since `Trait<u8>` and `Trait<()>` do not unify: https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_borrowck/src/type_check/mod.rs#L1481-L1524 This PR fixes this ICE by checking whether the RHS of the cast is considered to be Sized in the environment of the MIR typeck, and if so then skipping over this dyn->dyn principal compatibility check. r? ````@lcnr```` perhaps?

Zalathar · 2025-08-25T05:41:55Z

Tracking down a failure in rollup #145833.

@bors try jobs=dist-various-2

rust-bors · 2025-08-25T05:41:58Z

⌛ Trying commit e0fb6eb with merge 933692e…

To cancel the try build, run the command @bors try cancel.

Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts try-job: dist-various-2

compiler-errors · 2025-08-25T05:48:24Z

@bors try cancel

due to that other pr

rust-bors · 2025-08-25T05:48:26Z

Try build cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/17200204036

Rollup of 10 pull requests Successful merges: - #135761 (Dial down detail of B-tree description) - #145620 (Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts) - #145788 (Fix attribute target checking for macro calls) - #145794 (bootstrap.py: Improve CPU detection on NetBSD) - #145817 (cg_llvm: Replace the `llvm::Bool` typedef with a proper newtype) - #145820 (raw-dylib-elf: set correct `DT_VERDEFNUM`) - #145828 (Update `bitflags` to 2.9.3.) - #145830 (Remove the lifetime from `ExpTokenPair`/`SeqSep`.) - #145836 (Remove outdated bug comments) - #145842 (rustc-dev-guide subtree update) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of #145620 - compiler-errors:fake-dyn-to-dyn, r=lcnr Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts Fixes #141806 When we have an impossible where clause like `dyn Trait<u8>: Sized`, this may make a dyn-to-dyn cast like `dyn Trait<()> -> dyn trait<u8>` to successfully type check as if it were a wide-to-thin ptr cast (discarding metadata): https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_hir_typeck/src/cast.rs#L862-L865 In borrowck, we are expecting that the only meaningful dyn-to-dyn cast to be a metadata-preserving wide-to-wide ptr cast, which requires that the principals of the dyn pointers are equal. Borrowck additionally assumes that these principals have already been proven equal *modulo regions*, and we thus ICE since `Trait<u8>` and `Trait<()>` do not unify: https://github.com/rust-lang/rust/blob/16ad385579cebb6f7d53367c552661b6b51a4a02/compiler/rustc_borrowck/src/type_check/mod.rs#L1481-L1524 This PR fixes this ICE by checking whether the RHS of the cast is considered to be Sized in the environment of the MIR typeck, and if so then skipping over this dyn->dyn principal compatibility check. r? `@lcnr` perhaps?

Rollup of 10 pull requests Successful merges: - rust-lang/rust#135761 (Dial down detail of B-tree description) - rust-lang/rust#145620 (Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts) - rust-lang/rust#145788 (Fix attribute target checking for macro calls) - rust-lang/rust#145794 (bootstrap.py: Improve CPU detection on NetBSD) - rust-lang/rust#145817 (cg_llvm: Replace the `llvm::Bool` typedef with a proper newtype) - rust-lang/rust#145820 (raw-dylib-elf: set correct `DT_VERDEFNUM`) - rust-lang/rust#145828 (Update `bitflags` to 2.9.3.) - rust-lang/rust#145830 (Remove the lifetime from `ExpTokenPair`/`SeqSep`.) - rust-lang/rust#145836 (Remove outdated bug comments) - rust-lang/rust#145842 (rustc-dev-guide subtree update) r? `@ghost` `@rustbot` modify labels: rollup

rustbot assigned lcnr Aug 19, 2025

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 Aug 19, 2025

compiler-errors commented Aug 19, 2025

View reviewed changes

compiler/rustc_borrowck/src/type_check/mod.rs Outdated Show resolved Hide resolved

lcnr reviewed Aug 22, 2025

View reviewed changes

compiler/rustc_borrowck/src/type_check/mod.rs Outdated Show resolved Hide resolved

compiler/rustc_borrowck/src/type_check/mod.rs Outdated Show resolved Hide resolved

Account for impossible bounds making seemingly unsatisfyable dyn-to-d…

c957c4e

…yn casts

compiler-errors force-pushed the fake-dyn-to-dyn branch from 3131f84 to 0ee588a Compare August 23, 2025 17:13

Nits

e0fb6eb

compiler-errors force-pushed the fake-dyn-to-dyn branch from 0ee588a to e0fb6eb Compare August 23, 2025 17:13

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-review Status: Awaiting review from the assignee but also interested parties. labels Aug 24, 2025

jhpratt mentioned this pull request Aug 24, 2025

Rollup of 6 pull requests #145831

Closed

jhpratt mentioned this pull request Aug 24, 2025

Rollup of 5 pull requests #145832

Closed

Zalathar mentioned this pull request Aug 25, 2025

Rollup of 5 pull requests #145833

Closed

rust-bors bot added a commit that referenced this pull request Aug 25, 2025

Auto merge of #145620 - compiler-errors:fake-dyn-to-dyn, r=<try>

933692e

Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts try-job: dist-various-2

Zalathar mentioned this pull request Aug 25, 2025

Rollup of 10 pull requests #145844

Merged

bors merged commit 8a2568a into rust-lang:master Aug 25, 2025
10 of 11 checks passed

rustbot added this to the 1.91.0 milestone Aug 25, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts#145620

Account for impossible bounds making seemingly unsatisfyable dyn-to-dyn casts#145620
bors merged 2 commits intorust-lang:masterfrom
compiler-errors:fake-dyn-to-dyn

compiler-errors commented Aug 19, 2025 •

edited

Loading

Uh oh!

Uh oh!

lcnr left a comment •

edited by rustbot

Loading

Uh oh!

Uh oh!

Uh oh!

rustbot commented Aug 23, 2025

Uh oh!

compiler-errors commented Aug 24, 2025

Uh oh!

bors commented Aug 24, 2025

Uh oh!

Zalathar commented Aug 25, 2025

Uh oh!

rust-bors bot commented Aug 25, 2025

Uh oh!

compiler-errors commented Aug 25, 2025

Uh oh!

rust-bors bot commented Aug 25, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

	// Cast to thin pointer is OK
	if dst_kind == PointerKind::Thin {
	return Ok(CastKind::PtrPtrCast);
	}

	// This checks (lifetime part of) vtable validity for pointer casts,
	// which is irrelevant when there are aren't principal traits on
	// both sides (aka only auto traits).
	//
	// Note that other checks (such as denying `dyn Send` -> `dyn
	// Debug`) are in `rustc_hir_typeck`.
	if let ty::Dynamic(src_tty, _src_lt, ty::Dyn) = *src_tail.kind()
	&& let ty::Dynamic(dst_tty, dst_lt, ty::Dyn) = *dst_tail.kind()
	&& src_tty.principal().is_some()
	&& dst_tty.principal().is_some()
	{
	// Remove auto traits.
	// Auto trait checks are handled in `rustc_hir_typeck` as FCW.
	let src_obj = Ty::new_dynamic(
	tcx,
	tcx.mk_poly_existential_predicates(
	&src_tty.without_auto_traits().collect::<Vec<_>>(),
	),
	// FIXME: Once we disallow casting `*const dyn Trait + 'short`
	// to `*const dyn Trait + 'long`, then this can just be `src_lt`.
	dst_lt,
	ty::Dyn,
	);
	let dst_obj = Ty::new_dynamic(
	tcx,
	tcx.mk_poly_existential_predicates(
	&dst_tty.without_auto_traits().collect::<Vec<_>>(),
	),
	dst_lt,
	ty::Dyn,
	);

	debug!(?src_tty, ?dst_tty, ?src_obj, ?dst_obj);

	self.sub_types(
	src_obj,
	dst_obj,
	location.to_locations(),
	ConstraintCategory::Cast {
	is_implicit_coercion: false,
	unsize_to: None,
	},
	)
	.unwrap();

Uh oh!

Conversation

compiler-errors commented Aug 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

Uh oh!

Uh oh!

rustbot commented Aug 23, 2025

Uh oh!

compiler-errors commented Aug 24, 2025

Uh oh!

bors commented Aug 24, 2025

Uh oh!

Zalathar commented Aug 25, 2025

Uh oh!

rust-bors bot commented Aug 25, 2025

Uh oh!

compiler-errors commented Aug 25, 2025

Uh oh!

rust-bors bot commented Aug 25, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

compiler-errors commented Aug 19, 2025 •

edited

Loading

lcnr left a comment •

edited by rustbot

Loading