Toolchain upgrade to nightly-2024-05-29 failed

[github-actions]

Updating Rust toolchain from nightly-2024-05-28 to nightly-2024-05-29 requires source changes.
The failed automated run can be found here.
Please review the changes at https://github.com/rust-lang/rust from rust-lang/rust@84b40fc up to rust-lang/rust@da159eb. The log for this commit range is:
rust-lang/rust@da159eb331 Auto merge of #125682 - weihanglo:update-cargo, r=weihanglo
rust-lang/rust@037577a95a Update cargo
rust-lang/rust@274499dd0f Auto merge of #125665 - matthiaskrgr:rollup-srkx0v1, r=matthiaskrgr
rust-lang/rust@faabc74625 Rollup merge of #125637 - nnethercote:rustfmt-fixes, r=GuillaumeGomez
rust-lang/rust@de2bf3687b Rollup merge of #125598 - compiler-errors:proof-tree-builder, r=lcnr
rust-lang/rust@86f6fae93d Rollup merge of #125590 - ChrisDenton:mingw-ci-3, r=Kobzol
rust-lang/rust@78b4cafa6e Rollup merge of #125573 - GuillaumeGomez:migrate-allow-warnings-cmdline-stability, r=jieyouxu
rust-lang/rust@713c852a2f Rollup merge of #117671 - kjetilkjeka:nvptx_c_abi_avoid_direct, r=davidtwco
rust-lang/rust@8c4db851a7 Auto merge of #122662 - Mark-Simulacrum:optional-drop, r=bjorn3
rust-lang/rust@f494036530 Make ProofTreeBuilder actually generic over interner
rust-lang/rust@50a5da16b8 EvalCtxt::tcx() -> EvalCtxt::interner()
rust-lang/rust@7717a306b2 Auto merge of #124650 - oli-obk:pattern_types_syntax, r=nnethercote
rust-lang/rust@ac7e836834 Bless codegen test
rust-lang/rust@eae5031ecb Cache whether a body has inline consts
rust-lang/rust@ddc5f9b6c1 Create const block DefIds in typeck instead of ast lowering
rust-lang/rust@e5cba17b84 Use the HIR instead of mir_keys for determining whether something will have a MIR body.
rust-lang/rust@53e3c3271f Make body-visiting logic reusable
rust-lang/rust@be94ca0bcd Remove a CTFE check that was only ever used to ICE
rust-lang/rust@f1b0ca08a4 Don't format tests/run-make/*/rmake.rs.
rust-lang/rust@4702a1c345 Fix comments.
rust-lang/rust@404d47ec20 Migrate run-make/allow-warnings-cmdline-stability to rmake.rs
rust-lang/rust@56733372b8 Add assert_not_contains to run-make-support library
rust-lang/rust@f989d2f625 Auto merge of #125649 - workingjubilee:rollup-zwoum3k, r=workingjubilee
rust-lang/rust@4aaf9f645e Rollup merge of #125647 - tspiteri:track-lazy_cell_consume, r=workingjubilee
rust-lang/rust@01aa2e8511 Rollup merge of #125640 - fmease:plz-no-stringify, r=estebank
rust-lang/rust@941bf8bee1 Rollup merge of #125551 - clarfonthey:ip-bits, r=jhpratt
rust-lang/rust@fb95fda87f Rollup merge of #125343 - lcnr:eagerly-normalize-added-goals, r=compiler-errors
rust-lang/rust@8e89f83cbb Rollup merge of #125089 - Urgau:non_local_def-suggestions, r=estebank
rust-lang/rust@402a649e75 update tracking issue for lazy_cell_consume
rust-lang/rust@27cdc0df4e Don't suggest turning non-char-literal exprs of ty char into string literals
rust-lang/rust@c0d600385b Auto merge of #125636 - workingjubilee:bump-backtrace-0.3.72, r=workingjubilee
rust-lang/rust@98bfd54b0a eagerly normalize when adding goals
rust-lang/rust@13ce229042 refactor analyse visitor to instantiate states in order
rust-lang/rust@4d5a9bcb86 change selection test to run-pass
rust-lang/rust@87599ddd86 add debug_assert to alias-relate
rust-lang/rust@3ec9d8db27 Sync libstd deps with backtrace
rust-lang/rust@00b759530e Bump backtrace to 0.3.72
rust-lang/rust@d86e122941 Auto merge of #125609 - diondokter:opt-size-char-count, r=thomcc
rust-lang/rust@bcfa67d50d Remove out-of-date comment.
rust-lang/rust@71213fd607 Auto merge of #125539 - matthiaskrgr:cräsh, r=jieyouxu
rust-lang/rust@c7d300442f non_local_defs: point the parent item when appropriate
rust-lang/rust@98273ec612 non_local_defs: point to Self and Trait to give more context
rust-lang/rust@b71952904d non_local_defs: suggest removing leading ref/ptr to make the impl local
rust-lang/rust@ab23fd8dea non_local_defs: improve main without a trait note
rust-lang/rust@d3dfe14b53 non_local_defs: be more precise about what needs to be moved
rust-lang/rust@402580bcd5 non_local_defs: improve exception note for impl and macro_rules!
rust-lang/rust@22095fbd8d non_local_defs: use labels to indicate what may need to be moved
rust-lang/rust@26b873d030 non_local_defs: use span of the impl def and not the impl block
rust-lang/rust@de1c122950 non_local_defs: improve some notes around trait, bounds, consts
rust-lang/rust@06c6a2d9d6 non_local_defs: switch to more friendly primary message
rust-lang/rust@5ad4ad7aee non_local_defs: move out from #[derive(LintDiagnostic)] to manual impl
rust-lang/rust@4c002fce9d Omit non-needs_drop drop_in_place in vtables
rust-lang/rust@bae945201f remove fixed crashes, add fixed crashes to tests, add new cashed found in the meantime
rust-lang/rust@e5d100363a crashes: increment the number of tracked ones
rust-lang/rust@05fa647dc7 Always use the general case char count
rust-lang/rust@f931290949 Update description of install-mingw
rust-lang/rust@19cfe8d9e6 Add "Setup Python" action to github-hosted runners
rust-lang/rust@f63931bc1a Cleanup custom mingw in CI
rust-lang/rust@0d42cf7afe Stabilise ip_bits feature
rust-lang/rust@ead02ba0f1 NVPTX: Avoid PassMode::Direct for args in C abi

[qinheping]

The root cause of the failure is

rust-lang/rust@e5cba17b84 Use the HIR instead of mir_keys for determining whether something will have a MIR body.

Struct constructors have body in mir---has_body return true according to mir_keys---, but have no body according to HIR. So this commit will make the body of struct constructor instances None, and hence panic: https://github.com/model-checking/kani/blob/main/kani-compiler/src/kani_middle/transform/mod.rs#L78.

In #3219, I tried to ignore all struct constructors that have no body in the reachability module. But I don't think it is a good fix.

@celinval, can you confirm if the inconsistency of such has_body check is intended? If no, I will create an issue on Rust side.

[celinval]

@oli-obk can you please confirm?

[qinheping]

Here is an example, the constructor of HidesAPointer has body in mir but not hir.

struct HidesAPointer(*mut u32);

fn hidden_pointer(h: HidesAPointer) {}

fn main() {
    let mut a = 0;
    hidden_pointer(HidesAPointer(&mut a))
}

[oli-obk]

Hm yea, I think I screwed up the logic there. Altough I wonder how that's not an issue for rustc at all

[tautschnig]

Hm yea, I think I screwed up the logic there. Altough I wonder how that's not an issue for rustc at all

is_mir_available doesn't seem to be used a lot, maybe the cases of constructors just never matter for those callers?

[oli-obk]

yea, I gave it a look and it seems we already special case those callers everywhere.

Why do you need is_mir_available anyway? doesn't Kani just -Zalways-emit-mir?

[celinval]

Kani invokes stable-mir's Instace::has_body(), which invokes is_mir_available the same way the rustc collector checks which instance has body. Even with -Zalways-emit-mir, the has_body() should return false for cases such as extern functions, intrinsics, shims, and other things.

@oli-obk is there anything special with struct constructors?

[tautschnig]

@oli-obk Further to what @celinval said: we invoke Instance::body() where we know that such a MIR body will be available, but we now get spuriously get None returned after rust-lang/rust@e5cba17b84 for some (?) constructors. Given those constructors don't seem to have a body just yet in HIR, this is unsurprising, but it still seems wrong for a function called is_mir_available to compute its result off of HIR instead of MIR. We need those Body objects further down the line, so we can't special-case constructors. Why was this change deemed necessary (and right) in the first place, and can it please be reverted?

[oli-obk]

Yea, let's revert it (and add a SMIR test). It was a step in an attempt to eliminate mir_keys entirely, as its existance is in direct conflict with query feeding of MIR queries

[celinval]

@oli-obk are you planning to revert it or do you need any help?

[tautschnig]

For the record: @oli-obk took care of this in rust-lang/rust#126077, waiting for it to be merged.