From owner-freebsd-standards Fri Feb 7 14:33:41 2003 Delivered-To: freebsd-standards@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 44CB737B401 for ; Fri, 7 Feb 2003 14:33:40 -0800 (PST) Received: from fafoe.dyndns.org (chello212186121237.14.vie.surfer.at [212.186.121.237]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9027B43FA3 for ; Fri, 7 Feb 2003 14:33:39 -0800 (PST) (envelope-from stefan@fafoe.dyndns.org) Received: from frog.fafoe (frog.fafoe [192.168.2.101]) by fafoe.dyndns.org (Postfix) with ESMTP id A10C54066; Fri, 7 Feb 2003 23:33:37 +0100 (CET) Received: by frog.fafoe (Postfix, from userid 1001) id 695A8253; Fri, 7 Feb 2003 23:33:36 +0100 (CET) Date: Fri, 7 Feb 2003 23:33:36 +0100 From: Stefan Farfeleder To: "M. Warner Losh" Cc: standards@FreeBSD.ORG Subject: Re: fpclassify() for review Message-ID: <20030207223336.GB620@frog.fafoe> Mail-Followup-To: "M. Warner Losh" , standards@FreeBSD.ORG References: <20030206224455.B78590@espresso.q9media.com> <200302071825.h17IP4s7081758@khavrinen.lcs.mit.edu> <20030207.151652.38309440.imp@bsdimp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030207.151652.38309440.imp@bsdimp.com> User-Agent: Mutt/1.5.3i Sender: owner-freebsd-standards@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Fri, Feb 07, 2003 at 03:16:52PM -0700, M. Warner Losh wrote: > In message: <200302071825.h17IP4s7081758@khavrinen.lcs.mit.edu> > Garrett Wollman writes: > : I think that, if we had some help from the compiler, it ought to be > : possible to implement FLT_EVAL_METHOD == 2 for i386, and that's > : probably the closest to the actual hardware implementation. Something > : similar to Kahan's PARANOIA ought to be able to distinguish the > : various methods and thereby serve as a regression test. > > What's FLT_EVAL_METHOD == 2 mean? > > I'd love to make the long doubles actually have their full range by > default. 5.2.4.2.2 Characteristics of floating types 7 The values of operations with floating operands and values subject to the usual arithmetic conversions and of floating constants are evaluated to a format whose range and precision may be greater than required by the type. The use of evaluation formats is characterized by the implementation-defined value of FLT_EVAL_METHOD:19) -1 indeterminable; 0 evaluate all operations and constants just to the range and precision of the type; 1 evaluate operations and constants of type float and double to the range and precision of the double type, evaluate long double operations and constants to the range and precision of the long double type; 2 evaluate all operations and constants to the range and precision of the long double type. Regards, Stefan Farfeleder To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-standards" in the body of the message