[krbdev.mit.edu #6698] Incremental propagation log time stamp issue

[Jason Rogers via RT]

Greetings,

There is an issue in 1.7.1 and 1.8 (at least) such that the incremental propagation log time stamp returned, through kproplog, on  a 64bit Centos/RHEL platform is nonsensical, while it is correct on a 32bit platform.

The following change in casting corrects this:

@@ -386,6 +386,7 @@
     char                *dbprinc;
     kdb_ent_header_t    *indx_log;
     kdb_incr_update_t   upd;
+    time_t              tstamp;

     if (entry && (entry < ulog->kdb_num))
         start_sno = ulog->kdb_last_sno - entry;
@@ -445,9 +446,11 @@

         if (indx_log->kdb_time.seconds == 0L)
             (void) printf(_("\tUpdate time stamp : None\n"));
-        else
+        else  {
+            tstamp = (time_t) indx_log->kdb_time.seconds;
             (void) printf(_("\tUpdate time stamp : %s"),
-                          ctime((time_t *)&(indx_log->kdb_time.seconds)));
+                         ctime(&tstamp));
+        }

         (void) printf(_("\tAttributes changed : %d\n"),
                       upd.kdb_update.kdbe_t_len);


Thank you,

Jason Rogers
Sine Nomine Associates