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

[compiler-errors]

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):

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?

[lcnr]

nit, then r=me

View changes since this review

[rustbot]

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]

@bors r=lcnr

[bors]

📌 Commit e0fb6eb has been approved by lcnr

It is now in the queue for this repository.

[Zalathar]

Tracking down a failure in rollup #145833.

@bors try jobs=dist-various-2

[rust-bors]

⌛ Trying commit e0fb6eb with merge 933692e…

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

[compiler-errors]

@bors try cancel

due to that other pr

[rust-bors]

Try build cancelled. Cancelled workflows:

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