Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 alpha 4/15/85; site enea.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!gatech!ut-sally!seismo!mcvax!enea!peno
From: peno@enea.UUCP (Pekka Nousiainen)
Newsgroups: net.unix-wizards
Subject: Bourne again (more on memory allocation)
Message-ID: <1107@enea.UUCP>
Date: Sun, 1-Dec-85 23:06:56 EST
Article-I.D.: enea.1107
Posted: Sun Dec  1 23:06:56 1985
Date-Received: Thu, 5-Dec-85 04:37:26 EST
Reply-To: peno@enea.UUCP (Pekka Nousiainen)
Distribution: net
Organization: Enea Data, Sweden
Lines: 13
Keywords: sh,alloc

The recent discussion on the "wacky" memory allocation of sh
got me into looking again at our Genix 4.1 sh which always hangs
in long scripts.  Sure enough, where it loops is in the statement

	WHILE !busy(q = p->word) DO p->word = q->word OD

at the beginning of alloc (really malloc, in blok.c).  Increasing
the constants in brkincr.h seemed to help a bit but I'd like to
have a real fix.  Our machine is pretty flakey in all respects so
this may not occur as often on other Genix 4.1 implementations.

---
Address:  ...!mcvax!enea!peno.