From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 8090 invoked by alias); 18 Oct 2002 14:33:55 -0000 Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org Received: (qmail 8075 invoked from network); 18 Oct 2002 14:33:52 -0000 Received: from unknown (HELO nile.gnat.com) (205.232.38.5) by sources.redhat.com with SMTP; 18 Oct 2002 14:33:52 -0000 Received: by nile.gnat.com (Postfix, from userid 338) id CE144F28CC; Fri, 18 Oct 2002 10:33:51 -0400 (EDT) To: rth@redhat.com, steve@moshier.net Subject: Re: real.c fails floating point tests Cc: gcc@gcc.gnu.org Message-Id: <20021018143351.CE144F28CC@nile.gnat.com> Date: Fri, 18 Oct 2002 09:11:00 -0000 From: dewar@gnat.com (Robert Dewar) X-SW-Source: 2002-10/txt/msg01114.txt.bz2 > Previously the REAL_VALUE_TYPE could not contain auxiliary > information, because often it was actually just emulating a host computer > double. If you have removed that restriction then you can put various > other items into the data structure and then it looks like it would be > simple to make the 160-bit precision follow some arithmetic rules that > are easier to analyze and test. I definitely agree with this analysis. High precision is never an adequate susbtitute for well defined axiomatic properties when it comes to fpt.