Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!think.com!paperboy!hsdndev!cmcl2!adm!smoke!gwyn
From: gwyn@smoke.brl.mil (Doug Gwyn)
Newsgroups: comp.lang.c
Subject: Re: Here's a challenge for floating point lovers.
Message-ID: <14938@smoke.brl.mil>
Date: 23 Jan 91 22:36:10 GMT
References: <1991Jan21.215629.8393@watdragon.waterloo.edu> <1991Jan22.180927.29232@zoo.toronto.edu>
Organization: U.S. Army Ballistic Research Laboratory, APG, MD.
Lines: 14

In article <1991Jan22.180927.29232@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes:
>>Does ANSI X3.159 require the ability to define such a constant?
>No.  Nobody's come up with a good way of doing this that I know of,
>especially when portability is desired.

Presumably portability is not an issue for this application.  What some
implementors do is to initialize what amounts to a union type with the
proper bits then refer to the floating-point member of that union to get
the value interpreted properly with the proper floating type.  The
standard does NOT require that the values of the <float.h> macros be
CONSTANT expressions; in particular they may access external storage at
run time (contents of such a union, for example).  Note also that this
means that FLT_RADIX and other <float.h> macros must not be used in #if
expressions in strictly conforming programs.