Unrolled build for rust-lang#136522

Rollup merge of rust-lang#136522 - compiler-errors:dyn_compatible_for_dispatch, r=oli-obk

Remove `feature(dyn_compatible_for_dispatch)` from the compiler

This PR proposes the removal of `feature(dyn_compatible_for_dispatch)` from the compiler.

* As far as I can tell from the tracking issue, there's very little demand for this feature. I think that if this feature becomes useful in the future, then a fresh implementation from a fresh set of eyes, with renewed understanding of how this feature fits into the picture of Rust as it exists **today** would be great to have; however, in the absence of this demand, I don't see a particularly good reason to keep this implementation around.

* The RFC didn't receive very much discussion outside of the lang team, and while the discussion it received seemed to suggest that this feature was aiming to simplify the language and improve expressibility, I don't think this feature has really demonstrated either of those goals in practice. Furthermore, nobody seems to have owned this feature for quite some time or express desire to push for its stabilization.

* Relatedly, I find some of the RFC discussion like "when we make things impossible it's often presumptuous"[^1] and "I tend to want to take a 'we are all adults here' attitude toward unsafe code"[^2] to be particularly uncompelling. Of course this is no criticism to the authors of those comments since they're pretty old comments now, but type soundness is (IMO) the primary goal of the types team. This feature doesn't really do much other than further complicating the story of where we must validate object safety for soundness, along making dyn-incompatible trait object types *almost* seem useful, but very much remain UB to create and misleading to users who don't know better.

* Dyn compatibility's story has gotten more complicated since the feature was proposed in 2017, and now it needs to interact with things like associated consts, GATs, RPITITs, trait upcasting, `dyn*`, etc. While some of this is exercised in the codebase today, I'm not confident all of the corners of this feature have been hammered out. Reducing the "surface area" for what can go wrong in the compiler, especially around a side of the language (`dyn Trait`) that has been known to be particularly unsound in the past, seems good enough motivation to get rid of this for now.

[^1]: rust-lang/rfcs#2027 (comment)
[^2]: rust-lang/rfcs#2027 (comment)

cc `@rust-lang/types` `@rust-lang/lang`

Tracking:

- rust-lang#43561

r? types

Author: rust-timer

compiler/rustc_feature/src/removed.rs

@@ -100,6 +100,15 @@ declare_features! (
      Some("renamed to `doc_notable_trait`")),
     /// Allows using `#[unsafe_destructor_blind_to_params]` (RFC 1238).
     (removed, dropck_parametricity, "1.38.0", Some(28498), None),
+    /// Allows making `dyn Trait` well-formed even if `Trait` is not dyn compatible[^1].
+    /// In that case, `dyn Trait: Trait` does not hold. Moreover, coercions and
+    /// casts in safe Rust to `dyn Trait` for such a `Trait` is also forbidden.
+    ///
+    /// Renamed from `object_safe_for_dispatch`.
+    ///
+    /// [^1]: Formerly known as "object safe".
+    (removed, dyn_compatible_for_dispatch, "1.83.0", Some(43561),
+     Some("removed, not used heavily and represented additional complexity in dyn compatibility")),
     /// Uses generic effect parameters for ~const bounds
     (removed, effects, "1.84.0", Some(102090),
      Some("removed, redundant with `#![feature(const_trait_impl)]`")),

compiler/rustc_feature/src/unstable.rs

@@ -270,14 +270,6 @@ declare_features! (
     (unstable, doc_notable_trait, "1.52.0", Some(45040)),
     /// Allows using the `may_dangle` attribute (RFC 1327).
     (unstable, dropck_eyepatch, "1.10.0", Some(34761)),
-    /// Allows making `dyn Trait` well-formed even if `Trait` is not dyn compatible[^1].
-    /// In that case, `dyn Trait: Trait` does not hold. Moreover, coercions and
-    /// casts in safe Rust to `dyn Trait` for such a `Trait` is also forbidden.
-    ///
-    /// Renamed from `object_safe_for_dispatch`.
-    ///
-    /// [^1]: Formerly known as "object safe".
-    (unstable, dyn_compatible_for_dispatch, "1.83.0", Some(43561)),
     /// Allows using the `#[fundamental]` attribute.
     (unstable, fundamental, "1.0.0", Some(29635)),
     /// Allows using `#[link_name="llvm.*"]`.

compiler/rustc_hir_analysis/src/coherence/mod.rs

@@ -199,11 +199,7 @@ fn check_object_overlap<'tcx>(
 
         for component_def_id in component_def_ids {
             if !tcx.is_dyn_compatible(component_def_id) {
-                // Without the 'dyn_compatible_for_dispatch' feature this is an error
-                // which will be reported by wfcheck. Ignore it here.
-                // This is tested by `coherence-impl-trait-for-trait-dyn-compatible.rs`.
-                // With the feature enabled, the trait is not implemented automatically,
-                // so this is valid.
+                // This is a WF error tested by `coherence-impl-trait-for-trait-dyn-compatible.rs`.
             } else {
                 let mut supertrait_def_ids = elaborate::supertrait_def_ids(tcx, component_def_id);
                 if supertrait_def_ids

compiler/rustc_trait_selection/src/traits/dyn_compatibility.rs

@@ -538,10 +538,10 @@ fn receiver_for_self_ty<'tcx>(
 /// a pointer.
 ///
 /// In practice, we cannot use `dyn Trait` explicitly in the obligation because it would result in
-/// a new check that `Trait` is dyn-compatible, creating a cycle (until dyn_compatible_for_dispatch
-/// is stabilized, see tracking issue <https://github.com/rust-lang/rust/issues/43561>).
-/// Instead, we fudge a little by introducing a new type parameter `U` such that
+/// a new check that `Trait` is dyn-compatible, creating a cycle.
+/// Instead, we emulate a placeholder by introducing a new type parameter `U` such that
 /// `Self: Unsize<U>` and `U: Trait + ?Sized`, and use `U` in place of `dyn Trait`.
+///
 /// Written as a chalk-style query:
 /// ```ignore (not-rust)
 /// forall (U: Trait + ?Sized) {
@@ -572,8 +572,6 @@ fn receiver_is_dispatchable<'tcx>(
 
     // the type `U` in the query
     // use a bogus type parameter to mimic a forall(U) query using u32::MAX for now.
-    // FIXME(mikeyhew) this is a total hack. Once dyn_compatible_for_dispatch is stabilized, we can
-    // replace this with `dyn Trait`
     let unsized_self_ty: Ty<'tcx> =
         Ty::new_param(tcx, u32::MAX, rustc_span::sym::RustaceansAreAwesome);
 

compiler/rustc_trait_selection/src/traits/select/candidate_assembly.rs

@@ -859,13 +859,7 @@ impl<'cx, 'tcx> SelectionContext<'cx, 'tcx> {
                         }
 
                         if let Some(principal) = data.principal() {
-                            if !self.infcx.tcx.features().dyn_compatible_for_dispatch() {
-                                principal.with_self_ty(self.tcx(), self_ty)
-                            } else if self.tcx().is_dyn_compatible(principal.def_id()) {
-                                principal.with_self_ty(self.tcx(), self_ty)
-                            } else {
-                                return;
-                            }
+                            principal.with_self_ty(self.tcx(), self_ty)
                         } else {
                             // Only auto trait bounds exist.
                             return;

compiler/rustc_trait_selection/src/traits/wf.rs

@@ -904,19 +904,14 @@ impl<'a, 'tcx> TypeVisitor<TyCtxt<'tcx>> for WfPredicates<'a, 'tcx> {
                 // FIXME(#27579) RFC also considers adding trait
                 // obligations that don't refer to Self and
                 // checking those
-
-                let defer_to_coercion = tcx.features().dyn_compatible_for_dispatch();
-
-                if !defer_to_coercion {
-                    if let Some(principal) = data.principal_def_id() {
-                        self.out.push(traits::Obligation::with_depth(
-                            tcx,
-                            self.cause(ObligationCauseCode::WellFormed(None)),
-                            self.recursion_depth,
-                            self.param_env,
-                            ty::Binder::dummy(ty::PredicateKind::DynCompatible(principal)),
-                        ));
-                    }
+                if let Some(principal) = data.principal_def_id() {
+                    self.out.push(traits::Obligation::with_depth(
+                        tcx,
+                        self.cause(ObligationCauseCode::WellFormed(None)),
+                        self.recursion_depth,
+                        self.param_env,
+                        ty::Binder::dummy(ty::PredicateKind::DynCompatible(principal)),
+                    ));
                 }
             }
 

src/tools/tidy/src/issues.txt

@@ -3827,7 +3827,6 @@ ui/suggestions/issue-103646.rs
 ui/suggestions/issue-104086-suggest-let.rs
 ui/suggestions/issue-104287.rs
 ui/suggestions/issue-104327.rs
-ui/suggestions/issue-104328.rs
 ui/suggestions/issue-104961.rs
 ui/suggestions/issue-105226.rs
 ui/suggestions/issue-105494.rs

tests/ui/coherence/coherence-unsafe-trait-object-impl.rs

@@ -1,10 +1,7 @@
-//@ check-pass
-
 // Regression test for #128176. Previously we would call `type_of` on the `1` anon const
 // before the anon const had been lowered and had the `type_of` fed with a result.
 
 #![feature(generic_const_exprs)]
-#![feature(dyn_compatible_for_dispatch)]
 #![allow(incomplete_features)]
 
 trait X {
@@ -13,6 +10,7 @@ trait X {
 
 const _: () = {
     fn f2<'a>(arg: Box<dyn X<Y<1> = &'a ()>>) {}
+    //~^ ERROR the trait `X` is not dyn compatible
 };
 
 fn main() {}