Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!lll-winken!ncis.tis.llnl.gov!helios.ee.lbl.gov!ux1.lbl.gov!beard
From: beard@ux1.lbl.gov (Patrick C Beard)
Newsgroups: comp.sys.mac.programmer
Subject: Re: MPW C & registers
Summary: Take Address of variables to force stack allocation.
Message-ID: <4404@helios.ee.lbl.gov>
Date: 7 Dec 89 21:10:25 GMT
References: <1989Dec6.211341.762@ux1.cso.uiuc.edu>
Sender: usenet@helios.ee.lbl.gov
Reply-To: beard@ux1.lbl.gov (Patrick C Beard)
Organization: Lawrence Berkeley Laboratory, Berkeley
Lines: 28
X-Local-Date: 7 Dec 89 13:10:25 PST

In article <1989Dec6.211341.762@ux1.cso.uiuc.edu> dorner@pequod.cso.uiuc.edu (Steve Dorner) writes:
>SADE is capable of examining variables in previous stack frames, but not if
>those variables are in registers.  Therefore, I'd like to be able to tell the
>MPW C compiler to not put variables in registers when I'm compiling test
>versions of things.
>

I suppose a #pragma directive might be a nice way to do it as well, but you
can keep variables out of registers by taking their addresses:  i.e. statements
like

foo()
{
	int x, y, z;
	&x, &y, &z;	// keep these suckers out of registers.
	...
}

will keep them on the stack.  

After looking in my MPW C manual, I also see that the keyword volatile can
be used to keep variables on the stack.  See the manual, pg. 48.


-------------------------------------------------------------------------------
-  Patrick Beard, Macintosh Programmer                        (beard@lbl.gov) -
-  Berkeley Systems, Inc.  ".......<dead air>.......Good day!" - Paul Harvey  -
-------------------------------------------------------------------------------