---------------------
PatchSet 3108 
Date: 2001/10/08 14:00:59
Author: akroonmaa
Branch: chunked_mempools
Tag: (none) 
Log:
due to our perf optimisation some pools don't reach flush threshold and can
stay at 0 hwater mark at a time of mem stats request. We better rely on count
of allocs ever made here.

Members: 
	src/mem.c:1.9.8.12->1.9.8.13 

Index: squid/src/mem.c
===================================================================
RCS file: /cvsroot/squid-sf//squid/src/mem.c,v
retrieving revision 1.9.8.12
retrieving revision 1.9.8.13
diff -u -r1.9.8.12 -r1.9.8.13
--- squid/src/mem.c	1 Oct 2001 11:06:43 -0000	1.9.8.12
+++ squid/src/mem.c	8 Oct 2001 14:00:59 -0000	1.9.8.13
@@ -1,6 +1,6 @@
 
 /*
- * $Id: mem.c,v 1.9.8.12 2001/10/01 11:06:43 akroonmaa Exp $
+ * $Id: mem.c,v 1.9.8.13 2001/10/08 14:00:59 akroonmaa Exp $
  *
  * DEBUG: section 13    High Level Memory Pool Management
  * AUTHOR: Harvest Derived
@@ -503,7 +503,7 @@
 	memPoolGetStats(&mp_stats, pool);
 	if (!mp_stats.pool)	/* pool destroyed */
 	    continue;
-	if (mp_stats.pool->meter.inuse.hwater_level > 0)	/* this pool has been used */
+	if (mp_stats.pool->meter.gb_saved.count > 0)	/* this pool has been used */
 	    memPoolReport(&mp_stats, mp_total.TheMeter, e);
 	else
 	    not_used++;