From owner-freebsd-ports Sun Dec 15 13:30:06 1996 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.4/8.8.4) id NAA01486 for ports-outgoing; Sun, 15 Dec 1996 13:30:06 -0800 (PST) Received: from apolo.biblos.unal.edu.co ([168.176.37.75]) by freefall.freebsd.org (8.8.4/8.8.4) with ESMTP id NAA01438 for ; Sun, 15 Dec 1996 13:30:00 -0800 (PST) Received: from unalmodem.usc.unal.edu.co ([168.176.3.39]) by apolo.biblos.unal.edu.co (8.8.2/8.8.2) with SMTP id QAA04737 for ; Sun, 15 Dec 1996 16:31:50 -0500 (EST) Message-ID: <32B496C4.4A12@fps.biblos.unal.edu.co> Date: Sun, 15 Dec 1996 16:24:36 -0800 From: "Pedro Giffuni S." Organization: Universidad Nacional de Colombia X-Mailer: Mozilla 3.0 (Win16; I) MIME-Version: 1.0 To: ports@freebsd.org Subject: FP exception on a port. Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-ports@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi, I am preparing a port for Felt ( a CAD program), but it exhibits a problem (using fbsd 2.1.5) pointed out by Martin Cracauer in his page: ___________quote__________ An example is IEEE floating point. FreeBSD triggers an FP exception when dividing by zero. All my other computers set it to NaN and that is what many programs rely on (that means, it is supposed to be changed to work with FreeBSD). For example, the xlock program from the FreeBSD ports collection sometimes dumps core, leaving the display open. Just unacceptable. I had some mail with Bruce Evans, and the reason for FreeBSD's behaviour is that there are some bugs in floating-point handling in FreeBSD's gcc related to handling of IEEE exceptional numbers. Therefore, Bruce felt FreeBSD should default to terminate an application that isn't aware of division by zero etc. issues. I strongly disagree here. ___________end_of_quote____ Was this problem resolved in fbsd 2.2? I canīt find the place where the division by zero occurs. Pedro. BTW, thanks Martin for pointing out the problem!