[netperf-dev] netperf4 commit notice r59 - trunk/src

raj at netperf.org raj at netperf.org
Fri Feb 17 17:07:51 PST 2006


Author: raj
Date: 2006-02-17 17:07:49 -0800 (Fri, 17 Feb 2006)
New Revision: 59

Modified:
   trunk/src/netperf.h
   trunk/src/netserver.c
Log:
other misc nonsense, related to the run as a child of inetd stuff



Modified: trunk/src/netperf.h
===================================================================
--- trunk/src/netperf.h	2006-02-16 23:27:04 UTC (rev 58)
+++ trunk/src/netperf.h	2006-02-18 01:07:49 UTC (rev 59)
@@ -63,7 +63,7 @@
 #include "netconfidence.h"
 
 #define NETPERF_DEFAULT_SERVICE_NAME     "netperf4"
-#define NETPERF_DTD_FILE (const xmlChar *)"./netperf_docs.dtd"
+#define NETPERF_DTD_FILE (const xmlChar *)"netperf_docs.dtd"
 #define NETPERF_VERSION (const xmlChar *)"4"
 #define NETPERF_UPDATE  (const xmlChar *)"0"
 #define NETPERF_FIX     (const xmlChar *)"999"

Modified: trunk/src/netserver.c
===================================================================
--- trunk/src/netserver.c	2006-02-16 23:27:04 UTC (rev 58)
+++ trunk/src/netserver.c	2006-02-18 01:07:49 UTC (rev 59)
@@ -998,6 +998,15 @@
 	/* we need to put a warning _somewhere_ but where? */
 	exit(-1);
       }
+      else {
+	/* and since we are a child of inetd, we don't want any stray
+	   stderr stuff from other packages going out on the control
+	   connection, so we should point stderr at where. if there
+	   happens to be anything waiting to be flushed, we probably
+	   best not flush it or fclose stderr lest that push garbage
+	   onto the control connection. raj 2006-02-16 */
+	stderr = where;
+      }
     }
   }
 



More information about the netperf-dev mailing list