More for ITS#3671, check for Abandon before replaying queued events

diff --git a/servers/slapd/overlays/syncprov.c b/servers/slapd/overlays/syncprov.c
index f184dd2815c7e57e2159013a32ba751573fb7380..6d7ca76f437f68e4ac18fad8651735732765b278 100644 (file)
--- a/servers/slapd/overlays/syncprov.c
+++ b/servers/slapd/overlays/syncprov.c
@@ -842,9 +842,11 @@ syncprov_sendresp( Operation *op, opcookie *opc, syncops *so, Entry **e, int mod
        op->o_private = sop.o_private;
        rs.sr_ctrls = NULL;
        /* Check queue again here; if we were hanging in a send and eventually
-        * recovered, there may be more to send now.
+        * recovered, there may be more to send now. But don't check if the
+        * original psearch has been abandoned.
         */
-       if ( rs.sr_err == LDAP_SUCCESS && queue && so->s_res ) {
+       if ( !so->s_op->o_abandon && rs.sr_err == LDAP_SUCCESS && queue
+               && so->s_res ) {
                ldap_pvt_thread_mutex_lock( &so->s_mutex );
                rs.sr_err = syncprov_qplay( &sop, on, so );
                ldap_pvt_thread_mutex_unlock( &so->s_mutex );