Path: utzoo!utgpu!watserv1!watmath!att!linac!pacific.mps.ohio-state.edu!zaphod.mps.ohio-state.edu!rpi!crdgw1!uunet!icom!xwkg.Icom.Com!andy
From: andy@xwkg.Icom.Com (Andrew H. Marrinson)
Newsgroups: comp.sources.d
Subject: Re: tgif core dumps all the time -- just mine?
Keywords: tgif bugs
Message-ID: <andy.671749133@xwkg>
Date: 15 Apr 91 20:58:53 GMT
References: <150@mx-1>
Sender: news@icom.icom.com (News Feed)
Organization: Icom Systems, Inc.
Lines: 39

No, it isn't just you.  I had to make the following change to msg.c
(in addition to a handful of other stuff for System V that you
presumably found already):

*** msg.c~      Thu Mar 21 15:46:00 1991
--- msg.c       Thu Mar 21 16:57:32 1991
***************
*** 38,44 ****

     if (*Msg == '\0') { topMsgNumber = msgCount; return; }

!    s = (char *) calloc (strlen (Msg), sizeof(char));
     msg_ptr = (struct MsgRec *) calloc (1, sizeof(struct MsgRec));

     strcpy (s, Msg);
--- 38,44 ----

     if (*Msg == '\0') { topMsgNumber = msgCount; return; }

!    s = (char *) calloc (strlen (Msg) + 1, sizeof(char));
     msg_ptr = (struct MsgRec *) calloc (1, sizeof(struct MsgRec));

     strcpy (s, Msg);

-------------- END OF PATCH -------------

It seems as though some mallocs always allocate a little bit more
memory than you ask for.  I found a virtually identical bug in the CMU
SNMP client.  The System V malloc is not so forgiving!

I haven't sent this bug fix to the author of tgif yet.  I suppose I
ought to...

Hope this helps...
--
		Andrew H. Marrinson
		Icom Systems, Inc.
		Wheeling, IL, USA
		(andy@icom.icom.com)