I wrote:
> Hence, I present the attached, which also tweaks things to avoid an
> extra pq_flush in the end-of-command code path, and improves the
> comments to discuss the issue of NOTIFYs sent by procedures.
Hearing no comments, I pushed that.
> I'm inclined to think we should flat-out reject LISTEN in any process
> that is not attached to a frontend, at least until somebody takes the
> trouble to add infrastructure that would let it be useful. I've not
> done that here though; I'm not quite sure what we should test for.
After a bit of looking around, it seems that MyBackendType addresses
that problem nicely, so I propose the attached.
regards, tom lane
diff --git a/src/backend/tcop/utility.c b/src/backend/tcop/utility.c
index 27fbf1f3aa..bf085aa93b 100644
--- a/src/backend/tcop/utility.c
+++ b/src/backend/tcop/utility.c
@@ -803,6 +803,23 @@ standard_ProcessUtility(PlannedStmt *pstmt,
ListenStmt *stmt = (ListenStmt *) parsetree;
CheckRestrictedOperation("LISTEN");
+
+ /*
+ * We don't allow LISTEN in background processes, as there is
+ * no mechanism for them to collect NOTIFY messages, so they'd
+ * just block cleanout of the async SLRU indefinitely.
+ * (Authors of custom background workers could bypass this
+ * restriction by calling Async_Listen directly, but then it's
+ * on them to provide some mechanism to process the message
+ * queue.) Note there seems no reason to forbid UNLISTEN.
+ */
+ if (MyBackendType != B_BACKEND)
+ ereport(ERROR,
+ (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+ /* translator: %s is name of a SQL command, eg LISTEN */
+ errmsg("cannot execute %s within a background process",
+ "LISTEN")));
+
Async_Listen(stmt->conditionname);
}
break;