Date: Sun, 24 Nov 1996 16:04:54 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: "John M. Aldrich" <fighteer AT cs DOT com>
Cc: Panos Platon Tsapralis <panosts AT hiway DOT gr>, djgpp AT delorie DOT com
Subject: Re: PDCurses Compiling Problems.
In-Reply-To: <3297F5FE.1E44@cs.com>
Message-Id: <Pine.SUN.3.91.961124155754.29805B-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Sat, 23 Nov 1996, John M. Aldrich wrote:

> Panos Platon Tsapralis wrote:
> > 
> >         I try to compile the PDCurses package under DJGPP, Ver.2.01. I observe
> > that the function 'ungetch(int)' is defined twice : a) in <conio.h> and b)
> > in <curses.h>. What should I do to overcome the problem ?
> 
> I'm not sure about this, but you may want to comment out the #include
> <conio.h> that has the "old" definition.  If this causes other problems,
> then you can probably safely ignore the warning.
                        ^^^^^^^^^^^^^^^^^^^^^^^^^
No, no, no!  I guess the warning doesn't come from the compiler (it
doesn't care if you declare the prototype twice, unless the prototypes are
different--and in this case they aren't).  The warning probably comes from
the linker, and you should never ignore those.  (Would the original poster
please post the exact warning message, if any, that gcc prints?)

The *real* problem here is that you should never use conio and curses to 
gether, as they will most likely conflict with each other.  PDCurses have 
everything you should need, so just don't use conio functions.