Comments describing the test which tracks the scheduler misbehavior

sdks/rust/tests/procedure-concurrency-client/src/test_handlers.rs

@@ -470,6 +470,8 @@ async fn exec_procedure_concurrent_with_scheduled_reducer(db_name: &str) {
     test_counter.wait_for_all().await;
 }
 
+/// Like [`exec_procedure_reducer_same_client_not_interleaved`], but with the scheduler instead of a client.
+/// Tracks a behavior that we'd like to change.
 async fn exec_scheduled_procedure_scheduled_reducer_not_interleaved(db_name: &str) {
     let test_counter = TestCounter::new();
     let sub_applied_nothing_result = test_counter.add_test("on_subscription_applied_nothing");

sdks/rust/tests/test.rs

@@ -549,6 +549,10 @@ mod rust_procedure_concurrency {
         make_test("procedure-concurrent-with-scheduled-reducer").run()
     }
 
+    /// Similar to [`procedure_reducer_same_client_not_interleaved`],
+    /// but with the scheduler subsystem acting as the "client."
+    ///
+    /// As with the linked test, we're not attached to this behavior.
     #[test]
     fn scheduled_procedure_scheduled_reducer_not_interleaved() {
         make_test("scheduled-procedure-scheduled-reducer-not-interleaved").run()