Path: utzoo!telly!attcan!dptcdc!jarvis.csri.toronto.edu!mailrus!ames!lll-winken!vette!brooks
From: brooks@vette.llnl.gov (Eugene Brooks)
Newsgroups: gnu.gcc.bug
Subject: Re: inline math.h
Message-ID: <22469@lll-winken.LLNL.GOV>
Date: 24 Mar 89 17:37:47 GMT
References: <GRUNWALD.89Mar24084502@guitar.cs.uiuc.edu>
Sender: usenet@lll-winken.LLNL.GOV
Reply-To: brooks@maddog.llnl.gov (Eugene Brooks)
Distribution: gnu
Organization: Lawrence Livermore National Laboratory
Lines: 19

In article <GRUNWALD.89Mar24084502@guitar.cs.uiuc.edu> grunwald@guitar.cs.uiuc.edu writes:
>according to man pages, these should really be:
>extern double
>  frexp (double x, int *exp), modf (double x, double *ip);
>
>and
>
>__inline static double modf (double x, double *ip)
>
>Do others agree? This is needed to include math-68881.h in G++.
According to the Aug. 1987 DRAFT of the "proposed" ANSI C standard
the second argument should indeed be (double *).  There is a rationale
for this, a double can have an integer part much bigger than a 32 bit
int.  Anyone got a copy of what was actually recently voted on to
make sure it was not changed with regard to this?



brooks@maddog.llnl.gov, brooks@maddog.uucp, .../uunet!maddog.llnl.gov!brooks