Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!agate!eos!ptolemy!fineman
From: fineman@ptolemy.arc.nasa.gov (Charles Fineman)
Newsgroups: gnu.g++
Subject: How about g++ on a 386i (running 4.0.1) [TAKE 2]
Message-ID: <1358@ptolemy.arc.nasa.gov>
Date: 13 Jul 89 23:40:46 GMT
Organization: NASA-Ames Research Center, Moffett Field, CA
Lines: 29

[                                                                ]
[   Sorry, something wierd happend with postnews so I don't know ]
[ if this made it to the outside world.)                         ]
[                                                                ]

I've been trying to get g++ up on a SUN386i (with 4.0.1 SunOS)
for the past few days and ran into three problems:

	1) Some file (newld.c I think) uses a macro called _buffend()
	  on the LHS but the macro is never defined.

	2) The SARMAG macro value does no get defined in ar.h (I tried
	 setting the PORTAR flag so it *would* get defined but that
	 caused an error when I ran the program when it tried to read
	 in a archive file).

	3)  The structure "struct exec" never gets defined with the
	  options given in the release of g++. 

One other apparent problem: Although there is support in the source files
for the 386i, there are no clauses in the Makefile to set up the options
for this machine (where it sets up the options for compiling the loader).
If I am supposed to be using the 386igas instead of sun386, then I have
to have the a.out.encap.h file which I could not find.

	Any clues?

		Chuck Fineman
		fineman@pluto.arc.nasa.gov