Another little 'must fix' for 6.5.2: someone removed HAVE_KILL
authorTom Lane <tgl@sss.pgh.pa.us>
Tue, 14 Sep 1999 22:33:35 +0000 (22:33 +0000)
committerTom Lane <tgl@sss.pgh.pa.us>
Tue, 14 Sep 1999 22:33:35 +0000 (22:33 +0000)
from the configuration symbols, but neglected to remove #ifdef HAVE_KILL
from async.c in the REL6_5 branch.  Result: cross-backend NOTIFY dead in
the water.

src/backend/commands/async.c

index e1bbfd2b3669abc506b25e90da0f710216693eff..8aaef07ee7e0706fe52f57b4ac547a2a963e86c4 100644 (file)
@@ -510,7 +510,6 @@ AtCommit_Notify()
                                 * signal first, because the other guy can't read
                                 * pg_listener until we unlock it.
                                 */
-#ifdef HAVE_KILL
                                if (kill(listenerPID, SIGUSR2) < 0)
                                {
 
@@ -526,7 +525,6 @@ AtCommit_Notify()
                                        heap_delete(lRel, &lTuple->t_self, NULL);
                                }
                                else
-#endif
                                {
                                        d = heap_getattr(lTuple, Anum_pg_listener_notify,
                                                                         tdesc, &isnull);