Update comment about where Ty:

We no longer move trait predicates where the self ty is a ty param to
"the bounds of a ty param".

Author: fmease

compiler/rustc_hir_analysis/src/collect/predicates_of.rs

@@ -266,20 +266,16 @@ fn gather_explicit_predicates_of(tcx: TyCtxt<'_>, def_id: LocalDefId) -> ty::Gen
         match predicate.kind {
             hir::WherePredicateKind::BoundPredicate(bound_pred) => {
                 let ty = icx.lowerer().lower_ty_maybe_return_type_notation(bound_pred.bounded_ty);
-
                 let bound_vars = tcx.late_bound_vars(predicate.hir_id);
-                // Keep the type around in a dummy predicate, in case of no bounds.
-                // That way, `where Ty:` is not a complete noop (see #53696) and `Ty`
-                // is still checked for WF.
+
+                // This is a `where Ty:` (sic!).
                 if bound_pred.bounds.is_empty() {
                     if let ty::Param(_) = ty.kind() {
-                        // This is a `where T:`, which can be in the HIR from the
-                        // transformation that moves `?Sized` to `T`'s declaration.
-                        // We can skip the predicate because type parameters are
-                        // trivially WF, but also we *should*, to avoid exposing
-                        // users who never wrote `where Type:,` themselves, to
-                        // compiler/tooling bugs from not handling WF predicates.
+                        // We can skip the predicate because type parameters are trivially WF.
                     } else {
+                        // Keep the type around in a dummy predicate. That way, it's not a complete
+                        // noop (see #53696) and `Ty` is still checked for WF.
+
                         let span = bound_pred.bounded_ty.span;
                         let predicate = ty::Binder::bind_with_vars(
                             ty::ClauseKind::WellFormed(ty.into()),