From: kagel AT quasar DOT bloomberg DOT com
Date: Tue, 18 Feb 1997 16:01:28 -0500
Message-Id: <9702182101.AA02135@quasar.bloomberg.com >
To: nikki AT gameboutique DOT co
Cc: djgpp AT delorie DOT com
In-Reply-To: <5e4dji$99q@flex.uunet.pipex.com> (nikki@gameboutique.co)
Subject: Re: weird gcc thing
Reply-To: kagel AT dg1 DOT bloomberg DOT com

   Errors-To: postmaster AT ns1
   From: nikki AT gameboutique DOT co (nikki)
   Newsgroups: comp.os.msdos.djgpp
   Date: 15 Feb 1997 13:27:14 GMT
   Organization: GameBoutique Ltd.
   Lines: 28
   Nntp-Posting-Host: www.gameboutique.com
   Mime-Version: 1.0
   X-Newsreader: knews 0.9.8
   Dj-Gateway: from newsgroup comp.os.msdos.djgpp
   Content-Type: text/plain; charset=us-ascii
   Content-Length: 647

   is there any particular reason why :

   double wib;

   wib= (double) atof(argv[2]);

   or

   float wib;
   int a;

   wib= (float) a;

   should both give wrong values under djgpp? only these both work fine in
   a program i've been porting from unix, yet the atof one will give wildly
   wrong answers. eg if i made the argument '8.0' it gets evaluated as 
   wib=456246264624525.00 or something ridiculous. removing that casting makes
   it work of course (and yes i know the casting is pointless here but it was
   in the program i was converting :)

   just wondered why this casting should cause a problem.

Do you have a prototype for atof() in your source or did you include stdlib.h?

Without the prototype atof is asssumed to return an int hence the screwed up
result when you explicitly cast that int to a double.  Try adding either an 
explicit prototype of atof() [double atof( const char * )] or including 
stdlib.h.  If your UNIX compiler is pre-ANSI that would explain why it works 
there, too.

-- 
Art S. Kagel, kagel AT quasar DOT bloomberg DOT com

A proverb is no proverb to you 'till life has illustrated it.  -- John Keats