Fix PING-PONG handling when processing backlog in read buffers

[ngircd-alex.git] / src / ngircd / conn-func.c

diff --git a/src/ngircd/conn-func.c b/src/ngircd/conn-func.c
index 72d38b8621ff559c9c93903989e7c091b018be79..3cf8a3a62b27d43ce32fcc1101f765dac3775fb7 100644 (file)
--- a/src/ngircd/conn-func.c
+++ b/src/ngircd/conn-func.c
@@ -45,13 +45,17 @@ Conn_UpdateIdle(CONN_ID Idx)
 /**
  * Update "ping timestamp", the time of the last outgoing PING request.
  *
+ * the value 0 signals a newly connected client including servers during the
+ * initial "server burst"; and 1 means that no PONG is pending for a PING.
+ *
  * @param Idx Connection index.
+ * @param TimeStamp 0, 1, or time stamp.
  */
 GLOBAL void
-Conn_UpdatePing(CONN_ID Idx)
+Conn_UpdatePing(CONN_ID Idx, time_t TimeStamp)
 {
        assert(Idx > NONE);
-       My_Connections[Idx].lastping = time(NULL);
+       My_Connections[Idx].lastping = TimeStamp;
 }
 
 /*