From 8d7ec96c00994e3e4da652201792ae2040c979c1 Mon Sep 17 00:00:00 2001
From: Nathaniel McCallum <nathaniel@mccallum.life>
Date: Fri, 12 Sep 2025 18:35:21 -0400
Subject: [PATCH] add `[const] PartialEq` bound to `PartialOrd`

This change is included for discussion purposes.

The PartialOrd bound on PartialEq is not strictly necessary. It
is, rather, logical: anything which is orderable should by
definition have equality. Is the same true for constness? Should
every type which is const orderable also have const equality?
---
 library/core/src/cmp.rs | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/library/core/src/cmp.rs b/library/core/src/cmp.rs
index 95896ab144187..7f369d19c3d12 100644
--- a/library/core/src/cmp.rs
+++ b/library/core/src/cmp.rs
@@ -1351,7 +1351,9 @@ pub macro Ord($item:item) {
 #[rustc_diagnostic_item = "PartialOrd"]
 #[allow(multiple_supertrait_upcastable)] // FIXME(sized_hierarchy): remove this
 #[rustc_const_unstable(feature = "const_cmp", issue = "143800")]
-pub const trait PartialOrd<Rhs: PointeeSized = Self>: PartialEq<Rhs> + PointeeSized {
+pub const trait PartialOrd<Rhs: PointeeSized = Self>:
+    [const] PartialEq<Rhs> + PointeeSized
+{
     /// This method returns an ordering between `self` and `other` values if one exists.
     ///
     /// # Examples