Date: Sun, 20 Aug 1995 12:45:32 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Emil Wataralov <emilw AT msil DOT sps DOT mot DOT com>
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu
Subject: Re: Problem with double on large program


On Sun, 13 Aug 1995, Emil Wataralov wrote:

> Be aware for following problem with djgpp's printf command.
> _doprnt engine, crashes program with Floating point exception,
> if trying to print FP NaNs (Not A Numbers). This happens with
> all printf, sprintf, fprintf, vsprintf...
> 
> Rajiv, I've fixed this by checking FP numbers for NaNs before calling 
> printf. Mail me if you need the code or try V2.0, maybe there _doprnt
> works fine.

The DJGPP FAQ list mentions a program which can diddle with the 80387 
control word to change this behavior:

11.7  Q: I have a program which works with FP emulation, but dies with
         ``Numeric Exception'' when run on a machine with a
         co-processor.  It also runs OK when compiled with Microsoft C.
         Can't you people make your floating-point code right?
      A: This might be still a problem with your program.  Under DJGPP,
         the 80x87 control word is set up so that it generates an
         exception when your program feeds it with a NaN (Not a Number),
         while the emulator doesn't have this behavior.  You should check
         that your program doesn't generate NaNs, or set 80x87 control
         word to a different value.  There is a program called ctrl87.c
         which you can find by anonymous ftp to omnigate.clarkson.edu,
         directory pub/msdos/djgpp/pub, which enables this kind of
         diddling the 80x87 control word; check it out.

(The latest version of the FAQ is available as faq102.zip from the same 
place you get DJGPP.)