Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!rphroy!caen!uwm.edu!ux1.cso.uiuc.edu!sirius!gerber
From: gerber@sirius.astro.uiuc.edu (Richard Gerber)
Newsgroups: comp.sys.amiga.programmer
Subject: YASG (Yet another SAS gripe)
Summary: my mistake or SAS's?
Keywords: SAS/C
Message-ID: <1991Apr9.141609.19375@ux1.cso.uiuc.edu>
Date: 9 Apr 91 14:16:09 GMT
Sender: usenet@ux1.cso.uiuc.edu (News)
Organization: University of Illinois at Urbana
Lines: 48

I'm not an expert C programmer, but is this my mistake, or a problem
with SAS/C 5.10a?

Here's the relevant code:

#define ISOTHERMAL 0
...
if(ISOTHERMAL)
{
	code A;
}
else
{
	code B;
}
...

In a program with 11 modules (I mean separate functions in different
files, if this is not the proper lingo), code A gets executed no matter
what value ISOTHERMAL takes. I even replaced 
if(ISOTHERMAL)
with
if(0)
and
if(1)

and it made no difference. code A was always executed. I even watched
it do it with cpr.

After mulling this over for a while, I switched to the "large data
module" option when compiling. Lo and behold the code then worked
properly.
Now I am perfectly ready to admit that I may have been at fault with
compiler settings. But even if I was, there should at least be some sort
of warning. This would be a major mistake with the code; the output from
the entire program would be worthless. It's only because I'm writing this
so as to be portable that I found the problem. When I ran the code on
a Sun 4, there was some error being produced in "code B". It was only when
I ran cpr to try to figure out what was going on that I found this
error with SAS/C.

I'm going to give SAS a call about this unless someone points out to me
that I have turned off a switch that would have produced a warning (or
some such thing).

Regards,
Richard
gerber@rigel.astro.uiuc.edu