Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!hao!hull
From: hull@hao.ucar.edu (Howard Hull)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Why won't VT100 2.8 compile???
Summary: Reserved Function Call
Keywords: ConsoleDevice vt100
Message-ID: <1703@ncar.ucar.edu>
Date: 1 Apr 89 01:34:42 GMT
References: <424a49e8.1285f@maize.engin.umich.edu> <11G0Ye8rZ=10101gm52@amdahl.uts.amdahl.com>
Sender: news@ncar.ucar.edu
Reply-To: hull@hao.UCAR.EDU (Howard Hull)
Distribution: na
Organization: High Altitude Observatory/NCAR, Boulder CO
Lines: 42

In article <11G0Ye8rZ=10101gm52@amdahl.uts.amdahl.com> lfk@tde.uts.amdahl.com (Lynn Kerby) writes:
>In article <424a49e8.1285f@maize.engin.umich.edu> beh@caen.engin.umich.edu (Bob Hruska) writes:
> >
> >I've got Manx 3.6a and I'm trying to compile the source to vt100 2.8,
> >but it chokes every time on vt100.c and init.c with the following error:
> >
> >init.c:24: ERROR 49: multiply defined symbol: ConsoleDevice
> >
> >(the same for vt100.c)
> >
>[other stuff deleted] 
> >Bob Hruska   University of Michigan Computer Aided Engineering Network (CAEN)
> >
> >INTERNET: beh@caen.engin.umich.edu
> >    UUCP: {umix|ucbvax|uunet}!caen.engin.umich.edu!beh
>
>The problem appears to be a bad definition in the MANX 3.6 functions.h file.
>I went through all the RKM stuff and the Autodocs and cannot find a function
>called ConsoleDevice.  I just commented out that line in functions.h and 
>everything worked fine.
>

I sympathize, but I never like to touch the original includes.  For instance,
just because you can't find a referenced function, you cannot necessarily
conclude that there won't be one by that name in some future software release.
Since you have the Autodocs, you also have the CBM Native Developer include
files and library tags in machine readable form; did you check those as well
to see if there was any reference to ConsoleDevice anywhere in there?  I
gather that you did.  However, placing the name of a reserved function in a
type list is a good way to keep people from assigning that name to one of
their globally-linked functions.  Manx may have done this for some reason of
their own making as well.  Some sweet day you may discover the reason...
[In checking some of the subsequent replies to this article, I did see a note
by Tony Sumerall explaining how Manx had used this tag in connection with the
RawKeyConvert implementation they provide with the Aztec C 3.6a release.]  Is
this a reserved item, S. Walton or CATS?   Perhaps a better way to handle this
is to copy the sys:include/functions.h file to your local source directory and
change your #include <functions.h> references to #include "functions.h" in this
program (only) so that you can then edit the local functions.h to your heart's
content...
						Howard Hull
						hull@hao.ucar.edu