user0/rust
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Make missing_const_for_fn operate on non-optimized MIR (#14003)

Browse source 
The main reason for this is that we might transform MIR in the
optimization passes in a way that doesn't work with const-eval, but it
is irrelevant since const-eval uses another MIR (`mir_for_ctfe`).

Specifically this came up when adding a new check in debug builds
(https://github.com/rust-lang/rust/pull/134424), which is added as part
of an optimization pass.

changelog: none
This commit is contained in:
Alejandra González 2025-03-28 11:01:32 +00:00 • committed by GitHub
commit d78622eeb8
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
clippy_lints/src/missing_const_for_fn.rs
View file

@ -155,9 +155,9 @@ impl<'tcx> LateLintPass<'tcx> for MissingConstForFn {
return;
}
let mir = cx.tcx.optimized_mir(def_id);
let mir = cx.tcx.mir_drops_elaborated_and_const_checked(def_id);
if let Ok(()) = is_min_const_fn(cx, mir, self.msrv)
if let Ok(()) = is_min_const_fn(cx, &mir.borrow(), self.msrv)
&& let hir::Node::Item(hir::Item { vis_span, .. }) | hir::Node::ImplItem(hir::ImplItem { vis_span, .. }) =
cx.tcx.hir_node_by_def_id(def_id)
{