From owner-freebsd-bugs@FreeBSD.ORG Wed Dec 19 05:08:23 2007 Return-Path: Delivered-To: freebsd-bugs@FreeBSD.ORG Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 0352616A41A for ; Wed, 19 Dec 2007 05:08:23 +0000 (UTC) (envelope-from turutani@scphys.kyoto-u.ac.jp) Received: from fb1.aams.jp (fb1.aams.jp [202.189.147.105]) by mx1.freebsd.org (Postfix) with ESMTP id C018113C459 for ; Wed, 19 Dec 2007 05:08:22 +0000 (UTC) (envelope-from turutani@scphys.kyoto-u.ac.jp) Received: from proxy2.aams.jp (proxy2.aams.jp [202.189.147.98]) by fb1.aams.jp (8.13.8/8.13.4/AAMS SPECIAL) with ESMTP id lBJ2b12D005311 for ; Wed, 19 Dec 2007 11:37:01 +0900 Received: from [192.168.3.2] (softbank218183189199.bbtec.net [218.183.189.199]) (authenticated bits=0) by proxy2.aams.jp (Switch-3.2.7/Switch-3.1.7) with ESMTP id lBJ2au5K017643 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 19 Dec 2007 11:36:57 +0900 Message-ID: <476883C3.4090406@scphys.kyoto-u.ac.jp> Date: Wed, 19 Dec 2007 11:36:51 +0900 From: Tsurutani Naoki User-Agent: Thunderbird 2.0.0.9 (Windows/20071031) MIME-Version: 1.0 To: Tsurutani Naoki , freebsd-bugs@FreeBSD.ORG References: <20071217075522.GA8703@VARK.MIT.EDU> In-Reply-To: <20071217075522.GA8703@VARK.MIT.EDU> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-AAMS-Virus-Status: clean Cc: Subject: Re: kern/102424: [libc] printf(3) prints ill result. X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 19 Dec 2007 05:08:23 -0000 David Schultz wrote: > Thanks for investigating this. It's interesting that the problem > only happens with -fgcse, and then only on amd64 machines in > 32-bit mode. Does this happen in a more recent version of FreeBSD? > > I don't have an amd64 machine available for testing, and the > vendor's dtoa code is a little hard to read, especially if we're > looking for a compiler bug here. Some ideas: > > - See whether contrib/gdtoa/dmisc.c or contrib/gdtoa/dtoa.c is the > problem. (That is, compile only one of those with -O2 and see > what happens.) > > - Compile libc with -g in your CFLAGS, then run the program with a > breakpoint on the 'quorem' function. See what value it returns. > It should never return more than 9. Thank you for your reply. But I do not know how to execute you wrote... Would you tell me ? --- Tsurutani Naoki turutani@scphys.kyoto-u.ac.jp