Quietens Spaceman when debugging UNIT_TESTS locally (#75570)

The contract is removed by virtue of testing.

code/controllers/subsystem/air.dm

@@ -615,7 +615,7 @@ SUBSYSTEM_DEF(air)
 // In an ideal world, we would have absolutely zero active turfs 99.99% of the time, but that's not the case. `log_mapping()` during world initialize triggers a CI fail.
 #ifdef UNIT_TESTS
 	return
-#endif
+#else
 	// Associated lists, left-hand-side is the z-level or z-trait, right-hand-side is the number of active turfs associated with that.
 	var/list/tally_by_level = list()
 	// Discriminate for certain z-traits, stuff like "Linkage" is not helpful.
@@ -672,6 +672,7 @@ SUBSYSTEM_DEF(air)
 
 	message_to_log += "End of active turf list."
 	log_mapping(message_to_log.Join("\n"))
+#endif
 
 /turf/open/proc/resolve_active_graph()
 	. = list()

code/controllers/subsystem/garbage.dm

@@ -404,7 +404,7 @@ SUBSYSTEM_DEF(garbage)
 			#ifdef REFERENCE_TRACKING
 			if (QDEL_HINT_FINDREFERENCE) //qdel will, if REFERENCE_TRACKING is enabled, display all references to this object, then queue the object for deletion.
 				SSgarbage.Queue(D)
-				D.find_references() //This breaks ci. Consider it insurance against somehow pring reftracking on accident
+				INVOKE_ASYNC(D, TYPE_PROC_REF(/datum, find_references))
 			if (QDEL_HINT_IFFAIL_FINDREFERENCE) //qdel will, if REFERENCE_TRACKING is enabled and the object fails to collect, display all references to this object.
 				SSgarbage.Queue(D)
 				SSgarbage.reference_find_on_fail[text_ref(D)] = TRUE

code/controllers/subsystem/minor_mapping.dm

@@ -9,10 +9,11 @@ SUBSYSTEM_DEF(minor_mapping)
 /datum/controller/subsystem/minor_mapping/Initialize()
 	#ifdef UNIT_TESTS // This whole subsystem just introduces a lot of odd confounding variables into unit test situations, so let's just not bother with doing an initialize here.
 	return SS_INIT_NO_NEED
-	#endif // the mice are easily the bigger problem, but let's just avoid anything that could cause some bullshit.
+	#else
 	trigger_migration(CONFIG_GET(number/mice_roundstart))
 	place_satchels()
 	return SS_INIT_SUCCESS
+	#endif // the mice are easily the bigger problem, but let's just avoid anything that could cause some bullshit.
 
 /// Spawns some critters on exposed wires, usually but not always mice
 /datum/controller/subsystem/minor_mapping/proc/trigger_migration(to_spawn=10)