Path: utzoo!attcan!uunet!super!udel!gatech!bloom-beacon!mit-eddie!uw-beaver!cornell!batcomputer!riley
From: riley@batcomputer.tn.cornell.edu (Daniel S. Riley)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Lattice Compilation under ARP vs. AmigaDOS
Message-ID: <6249@batcomputer.tn.cornell.edu>
Date: 9 Sep 88 01:00:25 GMT
References: <3885@bsu-cs.UUCP>
Reply-To: riley@tcgould.tn.cornell.edu (Daniel S. Riley)
Organization: Cornell Theory Center, Cornell University, Ithaca NY
Lines: 23

In article <3885@bsu-cs.UUCP> cfchiesa@bsu-cs.UUCP (Christopher Chiesa) writes:
>   Well, I've finally figured out what the problem has been all this time I've
>been having trouble with Lattice C (v3.03).  Namely, compiling-and-linking 
>under ARP produces binaries that work ONLY under AmigaDOS, and NOT under ARP!
>Compiling-and-linking under AmigaDOS, on the other hand, produces binaries
>that work under BOTH AmigaDOS AND ARP.

The ARP startup routine does a fair bit of stuff that ARP expects to work,
that is quite dependent on the compiler version.  In particular, the tags
ARP examines for the resident stuff assumes some features of Lattice
4.00 or later that simply aren't there in 3.10 or 3.03.  What I would
guess is happening is that ARP sees that tag, and then assume that your
program supports some ARP features that it actually doesn't, because
you aren't using it with the supported version of the compiler.

It should be possible to adapt the startup code back to Lattice 3.10;
3.03 support would be hard.

I believe (but can't check at the moment) that the documentation clearly
specifies the compiler versions that the startup code is written for.

-dan riley (dsr@lns61.tn.cornell.edu, dsr@crnlns.bitnet)
-wilson lab, cornell u.