Index: ChangeLog
===================================================================
RCS file: /surf1/CVS/squid/ChangeLog,v
retrieving revision 1.444.2.11
retrieving revision 1.444.2.12
diff -u -r1.444.2.11 -r1.444.2.12
--- ChangeLog	1998/12/09 03:44:10	1.444.2.11
+++ ChangeLog	1998/12/14 04:05:25	1.444.2.12
@@ -105,6 +105,9 @@
 	  store_digest's stuck in memory.
 	- Fixed clientProcessOnlyIfCachedMiss so it unlocks and
 	  unregisters from "cache hit" entries.
+	* Changes below have been made to SQUID_2_1_PATCH3
+	- Fixed memory leak in clientHandleIMSReply for
+	  storeClientCopy failures.
 
 Changes to Squid-2.0 (October 2, 1998):
 
Index: src/client_side.c
===================================================================
RCS file: /surf1/CVS/squid/src/client_side.c,v
retrieving revision 1.420.2.14
retrieving revision 1.420.2.15
diff -u -r1.420.2.14 -r1.420.2.15
--- client_side.c	1998/12/09 20:50:41	1.420.2.14
+++ client_side.c	1998/12/11 05:45:02	1.420.2.15
@@ -1,6 +1,6 @@
 
 /*
- * $Id: client_side.c,v 1.420.2.14 1998/12/09 20:50:41 wessels Exp $
+ * $Id: client_side.c,v 1.420.2.15 1998/12/11 05:45:02 wessels Exp $
  *
  * DEBUG: section 33    Client-side Routines
  * AUTHOR: Duane Wessels
@@ -372,38 +372,10 @@
     int recopy = 1;
     const http_status status = mem->reply->sline.status;
     debug(33, 3) ("clientHandleIMSReply: %s, %d bytes\n", url, (int) size);
-    if (size < 0 && entry->store_status != STORE_ABORTED)
-#if OLD_CODE
-	/*
-	 * This storeAbort() call causes wrong behaviour in some situations.
-	 * If there are multiple clients hanging on this entry, the second
-	 * client will not want the entry to be aborted.  This shows up in
-	 * the stack trace below:
-	 *
-	 * #0  0xef5f452c in kill ()
-	 * #1  0xef5ba620 in abort ()
-	 * #2  0x7d680 in __eprintf (string=0x82a70 "%s:%u: failed assertion `%s'\n", expression=0x82a90 "client_side.c", line=1111, filename=0x83468 "size > 0")
-	 * #3  0x2ad7c in clientCacheHit (data=0x31ae318, buf=0x17f1450 "", size=0) at client_side.c:1111
-	 * #4  0x65cd4 in storeClientCopy2 (e=0x4c48c0, sc=0x2eedd70) at store_client.c:264
-	 * #5  0x665b0 in InvokeHandlers (e=0x4c48c0) at store_client.c:510
-	 * #6  0x63c90 in storeAbort (e=0x4c48c0, cbflag=1) at store.c:626
-	 * #7  0x29190 in clientHandleIMSReply (data=0x31e9518, buf=0x2268810 "", size=-1) at client_side.c:343
-	 * #8  0x66450 in storeUnregister (e=0x2910c, data=0x31e9518) at store_client.c:467
-	 * #9  0x29a54 in httpRequestFree (data=0x31e9518) at client_side.c:579
-	 * #10 0x29eb4 in connStateFree (fd=534528, data=0x314d0d0) at client_side.c:667
-	 * #11 0x2f27c in commCallCloseHandlers (fd=42) at comm.c:501
-	 * #12 0x2f410 in comm_close (fd=42) at comm.c:565
-	 * #13 0x2d054 in clientReadRequest (fd=42, data=0x314d0d0) at client_side.c:2038
-	 * #14 0x309c8 in comm_poll (msec=434) at comm_select.c:354
-	 * #15 0x4da64 in main (argc=5, argv=0xeffffd4c) at main.c:587
-	 */
-	storeAbort(entry, 1);
-#endif
-    /*
-     * Lets just try to return here and see what kind of problems that
-     * causes
-     */
-    return;
+    if (size < 0 && entry->store_status != STORE_ABORTED) {
+	memFree(MEM_CLIENT_SOCK_BUF, buf);
+	return;
+    }
     if (entry->store_status == STORE_ABORTED) {
 	debug(33, 3) ("clientHandleIMSReply: ABORTED '%s'\n", url);
 	/* We have an existing entry, but failed to validate it */