Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!seismo!mcnc!gatech!hubcap!ncrcae!ncr-sd!matt From: matt@ncr-sd.SanDiego.NCR.COM (Matt Costello) Newsgroups: comp.bugs.misc Subject: Minor bug in Pathalias Message-ID: <1596@ncr-sd.SanDiego.NCR.COM> Date: Wed, 1-Jul-87 17:08:10 EDT Article-I.D.: ncr-sd.1596 Posted: Wed Jul 1 17:08:10 1987 Date-Received: Fri, 3-Jul-87 00:46:56 EDT Reply-To: matt@ncr-sd.SanDiego.NCR.COM (Matt Costello) Distribution: world Organization: NCR Corporation, Rancho Bernardo Lines: 40 Keywords: pathalias There is a minor bug in the pathalias(1) program. The bug causes pathalias to dump core when it runs out of memory. It is in the routine mymalloc() in the file mem.c; its sccsid is mem.c 8.3 (down!honey) 86/06/08 The problem is that the mymalloc() routine does not check the malloc(3) return value before trying to clear the memory returned by malloc. The published maps are getting big enough that a process size of 1 MB is no longer big enough to run pathalias. This bug is present in both the January 1986 and November 1986 distributions. The following is a context diff of the change for the November 1986 distribution: *** mem.c.orig Fri Feb 6 20:54:24 1987 --- mem.c Wed Jul 1 13:24:38 1987 *************** *** 199,205 if (n >= 1024) { /* for hash table */ rval = malloc(n); /* aligned */ ! strclear(rval, n); return(rval); } --- 199,206 ----- if (n >= 1024) { /* for hash table */ rval = malloc(n); /* aligned */ ! if (rval != 0) ! strclear(rval, n); return(rval); } -- Matt Costello +1 619 485 2926 {sdcsvax,cbosgd,pyramid,nosc.ARPA,ihnp4}!ncr-sd!matt