*BSD News Article 10388


Return to BSD News archive

Received: by minnie.vk1xwt.ampr.org with NNTP
	id AA39 ; Wed, 27 Jan 93 13:11:52 EST
Newsgroups: comp.unix.bsd
Path: sserve!manuel.anu.edu.au!munnari.oz.au!hp9000.csc.cuhk.hk!saimiri.primate.wisc.edu!zaphod.mps.ohio-state.edu!rpi!gatech!destroyer!fmsrl7!lynx.unm.edu!zia.aoc.nrao.edu!laphroaig!cflatter
From: cflatter@nrao.edu (Chris Flatters)
Subject: Re: Correct values for float.h, compiling gcc-
Message-ID: <1993Jan26.003133.20730@zia.aoc.nrao.edu>
Sender: news@zia.aoc.nrao.edu
Reply-To: cflatter@nrao.edu
Organization: NRAO
References: <1993Jan25.152028.10017@yrloc.ipsa.reuter.COM>
Date: Tue, 26 Jan 93 00:31:33 GMT
Lines: 27

In article 10017@yrloc.ipsa.reuter.COM, bsd@yrloc.ipsa.reuter.COM (Brian de Alwis) writes:
>What exactly are the correct values for <float.h>?
>
>This problem has come up while attempting to compile gcc-2.3.3 (are
>there any patches out there? I couldn't find any, with my limited
>resources here). It seems that the build proceeds with no problems
>until enquire.c is compiled with the new xgcc. On line 2307, it
>craps out with a `floating point constant out of range'. I haven't
>had too much time to look at any further.
>
>When I tried compiling enquire with the `stock' gcc-1.39, it manages
>to compile properly, but dumps core with a precision exception when
>commencing the double tests. I think it may be my implementation of
>strtod (from the Tcl distribution).

The values that need to be corrected are DBL_MAX and DBL_MIN.

#define DBL_MIN         2.2250738585072014E-308
#define DBL_MAX         1.7976931348623157E+308

(DBL_MAX in /usr/include/float.h is one digit too short and rounded high).
There are also problems with HUGEVAL being defined to be 1E500 in math.h ---
fixing this will require some changes to libm.

	Chris Flatters
	cflatter@nrao.edu