Fix memory leak in pgoutput for the WAL sender

RelationSyncCache, the hash table in charge of tracking the relation
schemas sent through pgoutput, was forgetting to free the TupleDesc
associated to the two slots used to store the new and old tuples,
causing some memory to be leaked each time a relation is invalidated
when the slots of an existing relation entry are cleaned up.

This is rather hard to notice as the bloat is pretty minimal, but a
long-running WAL sender would be in trouble over time depending on the
workload.  sysbench has proved to be pretty good at showing the problem,
coupled with some memory monitoring of the WAL sender.

Issue introduced in 52e4f0c, that has added row filters for tables
logically replicated.

Author: Boyu Yang
Reviewed-by: Michael Paquier, Hou Zhijie
Discussion: https://postgr.es/m/DM3PR84MB3442E14B340E553313B5C816E3252@DM3PR84MB3442.NAMPRD84.PROD.OUTLOOK.COM
Backpatch-through: 15

Author: michaelpq

src/backend/replication/pgoutput/pgoutput.c

@@ -2095,10 +2095,34 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
 		 * Tuple slots cleanups. (Will be rebuilt later if needed).
 		 */
 		if (entry->old_slot)
+		{
+			TupleDesc	desc = entry->old_slot->tts_tupleDescriptor;
+
+			Assert(desc->tdrefcount == -1);
+
 			ExecDropSingleTupleTableSlot(entry->old_slot);
+
+			/*
+			 * ExecDropSingleTupleTableSlot() would not free the TupleDesc, so
+			 * do it now to avoid any leaks.
+			 */
+			FreeTupleDesc(desc);
+		}
 		if (entry->new_slot)
+		{
+			TupleDesc	desc = entry->new_slot->tts_tupleDescriptor;
+
+			Assert(desc->tdrefcount == -1);
+
 			ExecDropSingleTupleTableSlot(entry->new_slot);
 
+			/*
+			 * ExecDropSingleTupleTableSlot() would not free the TupleDesc, so
+			 * do it now to avoid any leaks.
+			 */
+			FreeTupleDesc(desc);
+		}
+
 		entry->old_slot = NULL;
 		entry->new_slot = NULL;