From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 32572 invoked by alias); 10 May 2006 20:17:42 -0000 Received: (qmail 32516 invoked by uid 48); 10 May 2006 20:17:35 -0000 Date: Wed, 10 May 2006 20:17:00 -0000 Message-ID: <20060510201735.32515.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug middle-end/27528] compiling linux kernels 2.6.16.14/15 2.6.17-rc3 on powerpc (7450) get error on long exixting code In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "malitzke at metronets dot com" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2006-05/txt/msg01030.txt.bz2 List-Id: ------- Comment #8 from malitzke at metronets dot com 2006-05-10 20:17 ------- Well Fellas: Either have the Steering Committee revise the Invitation to participate in testing; quoted iselectively below. Or,have a member from the Steering Committe ask me to refrain from further participation. I, for one, am no longer willing to be at the receiving end of snide comments from people who can not admit that, volunteering or not, are talking through their hats. I am over 70 years of age and have come to my expertise (which is not in compilers) the hard way. I started with plugboards and worked my way up to real time assembly systems programming (telephone Central Office switches and dispatching systems). The proof of any programming effort is for the hardware to generate the right output to control other hardware or to be comprehensible by human beings. Now to the specifics: > This page describes regular efforts to test GCC thoroughly, > plus ideas for additional testing by volunteers who have > machine cycles to spare. You might not believe it, but I have a router-firewall, a MAC (dual 800 Mhz G4's), a Pentium 3 server (four 550 Mhz Pentium 3 with 2Gbyte error correcting memory an 130 Giga byes of SCSI) assorted other machines all running with software entirely compiled with gcc-4.1 and gcc-4.2. I only report to gcc.gnu.org what I consider important problems. In this specific case I just eliminated the __FUNCTION__ part from bug.h (asm-powerpc) and I am writing this with kernel-2.6.17-rc3 as compiled by the current gcc-4.2 on the MAC. > Perform regular builds and testing of current GCC sources > that are not already being reported regularly. Compiling everything but Ada and running the full test suite now takes 8.5 hours on a ~800 Mhz pentium 3. There is realy no publicly available Ada source. I, personally do not care for Java, but some source packages require it and sun is apparently no coming out with new releases for the powerpc series (competition for server sales) I check the test suite output to see if that particular gcc is to be my current production compiler for either pentium3 or powerpc. > If the operating system kernel you use is normally compiled > with GCC, try building it with the current sources; > such as a release branch, use the newly built kernel for > running further GCC tests. I am a user of compilers (not only gcc) and not a compiler builder. The four or five problems I reported (and caused changes in gcc) sofar were not evidenced by the test suite. > Build and test applications that are important to you; > investigate and report any problems you find. > Build and test packages that are normally available on > your platform and for which you have access to source.a I have about 40 Gigabyte of source code (no games, no IRC, no themes or their engines, no music or downloading software, as little Java as possible, and no Pascal). This is really software for workstation use. Regarding __func__ C99 declares it a "predefined identifier" and "Like keywords, predefined identifiers must no be defined by programmers." I am not asking that it becomes part of GCC. however it should made clear that it and certainly __FUNCTION__ are no longer supported. Regarding __LINE__, __FILE__, __DATE__, etc are required as per Standard C. Again, the GCC community can state that it is no honoring that particular requiremnt. However, the GCC comunity can not unilaterally abrogate that requirement. The trigraphs come to mind. These are probably matters for the Steering Committee. -- malitzke at metronets dot com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |UNCONFIRMED Resolution|INVALID | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27528