Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!decwrl!glacier!hplabs!sdcrdcf!sdcsvax!little
From: little@sdcsvax.UUCP
Newsgroups: net.lang.c++
Subject: c++ on a 3B2
Message-ID: <1704@sdcsvax.UUCP>
Date: Wed, 23-Apr-86 13:49:35 EST
Article-I.D.: sdcsvax.1704
Posted: Wed Apr 23 13:49:35 1986
Date-Received: Sat, 26-Apr-86 04:09:20 EST
Organization: EECS Dept. U.C. San Diego
Lines: 28
Keywords: free store, 3B2, memory


	Has anyone brought up c++ on a 3B2?  We are trying to install
	version 1.0 on a 3B2 with 2 Meg internal memory.  "make scratch" 
	works okay, generating a new cfront.  The problem is when 
	we try to use this cfront to generate "..c" files when doing a 
	"make cfront", for instance.  It starts compiling dcl.c, 
	then dies with a "free store exhausted" message.  The same 
	thing happened when we tried to make a main..c by hand using 
	our cfront.

	We have incorporated the set of three fixes posted by Jeff 
	Warwick in the files dcl..c, dcl2..c, and expr..c (in the
	scratch directory, so our scratch cfront should have them
	in it.)  
	
	Could our problem be with the alloc.c routines? If so, 
	does anybody have any experience fixing them?  I started
	looking into replacing the malloc() and free() calls with the
	standard ones, but it looks as though that may not be as 
	straightforward as it at first seems.  There are some variables
	such as NFn, NFe, NFs that are fiddled with in del.o and main.o
	as well as in free().  Has anyone looked into this at all?

	Our thanks in advance.

					Glenn Little

					little@sdcsvax