Rollup merge of #99861 - lcnr:orphan-check-cg, r=jackh726

orphan check: rationalize our handling of constants

cc `@rust-lang/types` `@rust-lang/project-const-generics` on whether you agree with this reasoning.

r? types

compiler/rustc_trait_selection/src/traits/coherence.rs

@@ -734,7 +734,21 @@ impl<'tcx> TypeVisitor<'tcx> for OrphanChecker<'tcx> {
         result
     }
 
-    // FIXME: Constants should participate in orphan checking.
+    /// All possible values for a constant parameter already exist
+    /// in the crate defining the trait, so they are always non-local[^1].
+    ///
+    /// Because there's no way to have an impl where the first local
+    /// generic argument is a constant, we also don't have to fail
+    /// the orphan check when encountering a parameter or a generic constant.
+    ///
+    /// This means that we can completely ignore constants during the orphan check.
+    ///
+    /// See `src/test/ui/coherence/const-generics-orphan-check-ok.rs` for examples.
+    ///
+    /// [^1]: This might not hold for function pointers or trait objects in the future.
+    /// As these should be quite rare as const arguments and especially rare as impl
+    /// parameters, allowing uncovered const parameters in impls seems more useful
+    /// than allowing `impl<T> Trait<local_fn_ptr, T> for i32` to compile.
     fn visit_const(&mut self, _c: ty::Const<'tcx>) -> ControlFlow<Self::BreakTy> {
         ControlFlow::CONTINUE
     }