Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!asuvax!ncar!elroy.jpl.nasa.gov!swrinde!zaphod.mps.ohio-state.edu!caen!news.cs.indiana.edu!bronze!cica!ogre!emo
From: emo@ogre.cica.indiana.edu (Eric Ost)
Newsgroups: comp.sys.sgi
Subject: Re: Help, porting AT&T C++ to SGI 4D220
Summary: porting C++ to SGI
Keywords: AT&T, C++
Message-ID: <1991Feb26.221821.26298@cica.indiana.edu>
Date: 26 Feb 91 22:18:21 GMT
References: <52372@cornell.UUCP>
Sender: emo@ogre.cica.indiana.edu
Reply-To: emo@ogre.cica.indiana.edu (Eric Ost)
Organization: Center for Innovative Computing Applications
Lines: 22
Nntp-Posting-Host: ogre.cica.indiana.edu


I have successfully ported AT&T C++ version 2.1 to a couple of SGI platforms.
You should select SYS=SYSV and OS=svr3 in all significant makefiles.
You should also pay particular attention to the #include files which are
constructed by the makefile in ./incl-master.  The only proto header file
which required modification was  math.h.  In particular, several function
prototypes are inconsistent with the definitions contained in the distributed 
/usr/include/math.h.   The easiest fix for these differences is to look
at the proto header file in ./incl-master and add lines similar in form
to those which were added to compensate for differences in the Alliant
architecture.  This permits you to then use the same set of proto header
files to build C++ on several difference architectures thereby only having to
modify the makefiles.  Compare the prototypes presented in the SGI version 
of math.h, i.e. /usr/include/math.h, with those presented in AT&T's proto 
header file in order to construct the corrected instances.

If you run into problems, feel free to send me Email and I will make an
attempt to respond as soon as my schedule permits.

Good luck.

eric


Brought to you by Super Global Mega Corp .com