From owner-freebsd-hackers Thu Dec 17 04:01:38 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id EAA14365 for freebsd-hackers-outgoing; Thu, 17 Dec 1998 04:01:38 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from mail.ruhrgebiet.individual.net (in-ruhr.ruhr.de [141.39.224.38]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id EAA14355 for ; Thu, 17 Dec 1998 04:01:23 -0800 (PST) (envelope-from bs@adimus.de) Received: (from admin@localhost) by mail.ruhrgebiet.individual.net (8.8.5-r-beta/8.8.5) with UUCP id NAA12819 for hackers@freebsd.org; Thu, 17 Dec 1998 13:00:33 +0100 (MET) Received: from mail by mx.adimus.de with local (Exim 1.92 #1) for hackers@freebsd.org id 0zqbaG-0000ei-00; Thu, 17 Dec 1998 12:27:08 +0100 Received: from det.adimus.de(192.168.0.1) via SMTP by adimus.de, id smtpdax1542; Thu Dec 17 12:27:03 1998 Received: from bs by det.adimus.de with local (Exim 1.92 #1) for hackers@FreeBSD.ORG id 0zqba9-000337-00; Thu, 17 Dec 1998 12:27:01 +0100 To: hackers@FreeBSD.ORG Subject: Minimal base systems (was Re: Fortran in the base system (was Re: sysinstall)) References: Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit From: Benedikt Stockebrand Date: 17 Dec 1998 12:27:01 +0100 In-Reply-To: Chuck Robey's message of "Wed, 16 Dec 1998 15:17:05 -0500 (EST)" Message-ID: Lines: 49 X-Mailer: Gnus v5.5/XEmacs 20.4 - "Emerald" Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Chuck Robey writes: > Steve, if you can't reply without getting personal, please take it to > usenet. Just like Bud Dodson you might as well blame me for getting personal in the start. Well, at least some of the misunderstandings (apparently I wasn't the only one to mistake Steve Kargls posting) with this issue seem to be solved. Anyway, from fiddling with assorted free Un*xen I've learned some important things: - You need a reliable, no-junk base system as a starting point for all other work. This is where Linux with its kernel-distribution splitup really loses (details on demand, but I don't want this to be misunderstood as Linux-bashing). - Changes to the base system need to be thoroughly tested. Unless there's good reason to change things you better don't. As a consequence it's actually better to keep old versions in the base system. - Things like perl5, gcc-2.8.2 (with proper c++ exception handling) or a rdist-6.1.4 are nice to have. However, updating them in the base system is trouble-prone so they better be added to the ports collection, at least until they have been properly tested. And if they're not needed for base functionality it's usually a *win* to put them in a package/port/whatever so people can choose what version to use---the fancy new one or the old one they've been building their system around. - The bigger a base system gets, the harder it is to keep it consistent. Both inherent complexity and turnaround times increase noticeably with every bit added to the base system. - Admitting that ones pet toy isn't really important enough to go into the base system is bl**dy fscking hard. Ask me about rdist... So long, Ben -- Benedikt Stockebrand Adimus Beratungsgesellschaft für System- System Administration & Design, und Netzwerkadministration mbH & Co KG IT Security, Remote System Mgmt Universitätsstr. 142, 44799 Bochum Opinions presented are my own. Tel. (02 34) 971 971 -2, Fax -9 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message