Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!spool.mu.edu!uunet!ns-mx!ccad.uiowa.edu!cadsi
From: cadsi@ccad.uiowa.edu (CADSI)
Newsgroups: comp.os.msdos.programmer
Subject: Re: Problem with BC++ and non-prototyped functions
Message-ID: <1991Apr9.175029.9197@ccad.uiowa.edu>
Date: 9 Apr 91 17:50:29 GMT
References: <Ze54Z1w164w@turbo.atl.ga.us>
Distribution: na
Organization: CAD-Research, U. of Iowa, Iowa City, Iowa
Lines: 23

From article <Ze54Z1w164w@turbo.atl.ga.us>, by greg@turbo.atl.ga.us (Greg Montgomery):
> I'm having a little problem with BC++ and functions that I didn't use
> the include files. For instance, I noticed using the fabs() function
> would actually increase the variable I was getting the abs of by 90
> until I added a #include <math.h> in the top of the .C file.
> Why would this happen? I can't figure out why having a declaration (I
> think that's the right word??) would affect how the function actually
> works. I've noticed this a couple times tonight. I think the same thing
> happened with atof(). Any ideas on this?? Is this a normal C happening
> or a bug that needs to be reported to Borland?

This is a standard C beginners problem.  C assumes integer if not
told otherwise.  Thus, without the prototype in the math.h header
file, atof is assumed to pass back an integer, not a double.  In
addition, without the prototype, the compiler will allow you to pass
an integer to the function.  Even uglier, since atof will read a
double off the stack starting from where the integer was placed.

|----------------------------------------------------------------------------|
|Tom Hite					|  The views expressed by me |
|Manager, Product development			|  are mine, not necessarily |
|CADSI (Computer Aided Design Software Inc.	|  the views of CADSI.       |
|----------------------------------------------------------------------------|