Path: utzoo!utgpu!water!watmath!clyde!bellcore!rutgers!tut.cis.ohio-state.edu!cwjcc!gatech!gitpyr!edp
From: edp@pyr.gatech.EDU (Warren Tucker)
Newsgroups: comp.unix.xenix
Subject: Re: pcomm 1.1 on SCO Xenix
Message-ID: <6527@pyr.gatech.EDU>
Date: 8 Oct 88 19:40:41 GMT
References: <157@libove.UUCP> <402@hawkmoon.MN.ORG>
Reply-To: edp@pyr.UUCP (Warren Tucker)
Organization: Georgia Institute of Technology
Lines: 20

In article <402@hawkmoon.MN.ORG> det@hawkmoon.MN.ORG (Derek E. Terveer) writes:
>above.  I suspect that use of -M2 turns on some sort of different inclusion
>mechanism for include files.  For example, i frequently find param.h (i think
>it was param.h) to generate errors just like the above.  Much as i hated to do
> [stuff deleted]
>Now, a question.  Just how much difference is there between running a -M2
>program and a plain old 8086 (or whatever the default code generation is -M0?)
>program in SCO xenix (i have 2.1.3).
>
-M2 causes the compiler to do the equivalent of a -DM_I80286
(#define M_I80286). -M0 does -DM_I8086; param.h and other standard
include files have conditional compilation depending on this switch.
(please no flames on the exactness of the identifiers; I'm doing this
from memory).  Also, -M1 or -M2 causes the compiler to use the 80186
and beyond ENTER and LEAVE instructions, which speeds up subroutine
linkage quite a bit.  So -M2 does improve execution speed.

-- 
Warren Tucker
...!{akgua,allegra,amd,hplabs,ihnp4,seismo,ut-ngp}!gatech!gitpyr!edp