Rollup merge of #148719 - Nadrieril:poly-drop-glue, r=saethlin

Allow unnormalized types in drop elaboration

I work on a [rustc driver](https://github.com/AeneasVerif/charon) that aims to extract the full polymorphic MIR of a crate. Currently the one thing I can't get is drop glue because it fails on unnormalizable types like the fields of `Cow` or `NonZero`.

This PR removes the one check for unnormalized types in drop elaboration. It's a `span_delay_bug` so only there to help catch mistakes. I think that's fine to remove? If something downstream relies on types being normalized the right approach is for MIR validation to check for normalized types, not that one random check.

Author: GuillaumeGomez

compiler/rustc_mir_transform/src/elaborate_drop.rs

@@ -511,19 +511,12 @@ where
                 let tcx = self.tcx();
 
                 assert_eq!(self.elaborator.typing_env().typing_mode, ty::TypingMode::PostAnalysis);
-                let field_ty = match tcx.try_normalize_erasing_regions(
-                    self.elaborator.typing_env(),
-                    field.ty(tcx, args),
-                ) {
-                    Ok(t) => t,
-                    Err(_) => Ty::new_error(
-                        self.tcx(),
-                        self.tcx().dcx().span_delayed_bug(
-                            self.elaborator.body().span,
-                            "Error normalizing in drop elaboration.",
-                        ),
-                    ),
-                };
+                let field_ty = field.ty(tcx, args);
+                // We silently leave an unnormalized type here to support polymorphic drop
+                // elaboration for users of rustc internal APIs
+                let field_ty = tcx
+                    .try_normalize_erasing_regions(self.elaborator.typing_env(), field_ty)
+                    .unwrap_or(field_ty);
 
                 (tcx.mk_place_field(base_place, field_idx, field_ty), subpath)
             })