Reimplement nullification of walsender timestamp

Make the value null only at pg_stat_activity-output time, as suggested
by Tom Lane, instead of messing with the internal state.  This should
appease buildfarm members with force_parallel_mode=regress, which are
running parallel queries on logical replication walsenders.

The fact that walsenders can run parallel queries should perhaps be
studied more carefully, but for the moment let's get rid of the red
blots in buildfarm.

Backpatch to pg10, like the previous commit.

Discussion: https://postgr.es/m/30804.1578438763@sss.pgh.pa.us

src/backend/access/transam/xact.c

@@ -816,13 +816,6 @@ GetCurrentTransactionStopTimestamp(void)
 void
 SetCurrentStatementStartTimestamp(void)
 {
-	/*
-	 * Skip if on a walsender; this is not needed, and it confuses monitoring
-	 * if we publish non-NULL values.
-	 */
-	if (am_walsender)
-		return;
-
 	if (!IsParallelWorker())
 		stmtStartTimestamp = GetCurrentTimestamp();
 	else

src/backend/utils/adt/pgstatfuncs.c

@@ -724,7 +724,13 @@ pg_stat_get_activity(PG_FUNCTION_ARGS)
 			else
 				nulls[7] = true;
 
-			if (beentry->st_xact_start_timestamp != 0)
+			/*
+			 * Don't expose transaction time for walsenders; it confuses
+			 * monitoring, particularly because we don't keep the time up-to-
+			 * date.
+			 */
+			if (beentry->st_xact_start_timestamp != 0 &&
+				beentry->st_backendType != B_WAL_SENDER)
 				values[8] = TimestampTzGetDatum(beentry->st_xact_start_timestamp);
 			else
 				nulls[8] = true;