Return-Path: X-Original-To: apmail-incubator-ooo-dev-archive@minotaur.apache.org Delivered-To: apmail-incubator-ooo-dev-archive@minotaur.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 130F57404 for ; Sat, 10 Sep 2011 14:57:44 +0000 (UTC) Received: (qmail 49528 invoked by uid 500); 10 Sep 2011 14:57:43 -0000 Delivered-To: apmail-incubator-ooo-dev-archive@incubator.apache.org Received: (qmail 49476 invoked by uid 500); 10 Sep 2011 14:57:43 -0000 Mailing-List: contact ooo-dev-help@incubator.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: ooo-dev@incubator.apache.org Delivered-To: mailing list ooo-dev@incubator.apache.org Received: (qmail 49468 invoked by uid 99); 10 Sep 2011 14:57:43 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Sat, 10 Sep 2011 14:57:43 +0000 X-ASF-Spam-Status: No, hits=0.0 required=5.0 tests=FREEMAIL_FROM,RCVD_IN_DNSWL_NONE,SPF_PASS,T_TO_NO_BRKTS_FREEMAIL X-Spam-Check-By: apache.org Received-SPF: pass (nike.apache.org: domain of Mathias_Bauer@gmx.net designates 213.165.64.23 as permitted sender) Received: from [213.165.64.23] (HELO mailout-de.gmx.net) (213.165.64.23) by apache.org (qpsmtpd/0.29) with SMTP; Sat, 10 Sep 2011 14:57:34 +0000 Received: (qmail invoked by alias); 10 Sep 2011 14:57:11 -0000 Received: from e176050156.adsl.alicedsl.de (EHLO [192.168.1.250]) [85.176.50.156] by mail.gmx.net (mp027) with SMTP; 10 Sep 2011 16:57:11 +0200 X-Authenticated: #17242763 X-Provags-ID: V01U2FsdGVkX19X7mYBkhpXejU8RZK271mibtbis8zcNwW8RiLvkv b4m7O2irK4BnvR Message-ID: <4E6B7AC6.6060507@gmx.net> Date: Sat, 10 Sep 2011 16:57:10 +0200 From: Mathias Bauer User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.8.1.21) Gecko/20090302 Thunderbird/2.0.0.21 Mnenhy/0.7.5.0 MIME-Version: 1.0 To: ooo-dev@incubator.apache.org Subject: Re: Fix me: Abnormalities during bootstrap References: <4E6B5B93.30304@gmx.net> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit X-Y-GMX-Trusted: 0 X-Virus-Checked: Checked by ClamAV on apache.org Am 10.09.2011 15:00, schrieb Joost Andrae: > Hi Mathias, > >>> After starting the build process on my Win7 64bit notebook (dual core >>> Intel 2.8 Ghz with 8 GB memory) it quite often stopped but rebuilding >>> went always fine (until now; I'm still compiling binfilter). I thought >>> it might be useful to implement a watchdog process watching >>> stderr/stdout that restarts dmake as soon as the build breaks but it >>> needs to interpret the error thrown. >> >> Probably you have been hit by the well-known bug in GNU Make 3.81 that >> lets it core-dump at times. This needs to be solved, the developers of >> Libre Office have gained some experience with that. > > GNU Make 3.8.1 is installed on my machine but I thought that dmake is > always used instead ? You really did not recognize our "new build environment" project? Parts of OOo are built with GNU Make, hopefully somewhere in the future everything. >>> Documentation at >>> http://ooo-wiki.apache.org/wiki/Documentation/Building_Guide/Building_on_Windows >>> was really helpful despite I had to disable ATL >> >> If you used VC Express: IIRC the documentation mentions that this is >> necessary in that case. > > Yes it is MSVC++ Express 2008 using Cygwin. > > btw. I wonder about the low CPU load during compile job (max. 50%). Is > there a way to allow dmake to parallelize compile jobs ? You can use "build -Pn" or "build -- -Pn" where n is the number of parallel build jobs you want to use. As a rule of thumb you will get the best performance with n = 2*number of (virtual) cores that you have. The first variant "build -Pn" starts n parallel processes of dmake, the second "build -- -Pn" tells dmake to do n parallel tasks. You also use "build -Pn -- -Pn" and divide your cores between both options. If you do a full build, I recommend "build -Pn -- -Pn" with n=square root of number of your cores (means 2 or 3 for 4 or 8 cores). This procedure is awkward and doesn't scale well. This was one of the reasons why we wanted a new build system. If everything was build with our new GNU Make based build system, the build would scale much better, nearly linear with the number of available cores. Regards, Mathias