From owner-freebsd-questions Mon Oct 30 1:44:12 2000 Delivered-To: freebsd-questions@freebsd.org Received: from smtp6.mindspring.com (smtp6.mindspring.com [207.69.200.110]) by hub.freebsd.org (Postfix) with ESMTP id 50CD437B479 for ; Mon, 30 Oct 2000 01:44:08 -0800 (PST) Received: from jayk3 (user-2ini89m.dialup.mindspring.com [165.121.33.54]) by smtp6.mindspring.com (8.9.3/8.8.5) with SMTP id EAA27118; Mon, 30 Oct 2000 04:43:59 -0500 (EST) From: jay.krell@cornell.edu Message-ID: <00c201c04255$b3bfaab0$8001a8c0@jayk3.jaykhome> Reply-To: To: "Mike Meyer" Cc: Subject: Re: make install multiprocess safe? Date: Mon, 30 Oct 2000 01:42:20 -0800 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.3612.1700 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3719.2500 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Imho it should be more safe. Even concurrent builds in the same directory should work better -- one should notice the other and either wait or abort. ..Jay -----Original Message----- From: Mike Meyer To: jay.krell@cornell.edu Cc: questions@freebsd.org Date: Sunday, October 29, 2000 11:31 AM Subject: Re: make install multiprocess safe? >jay.krell@cornell.edu writes: >> This has been bugging me a while. I've always just avoided it. >> Is it safe to >> cd /usr/ports/1/2 >> make install & >> cd /usr/ports/3/4 >> make install >> ? > >Mostly it's safe. > >> if both go to like register the package at about the same time, will the >> package database stay not corrupted? > >The "packages database" is a collection of flat text files. You only >get into problems if you try writing to the same one at the same time. > >> I'm assuming both have all their dependents installed or they share no >> dependents -- to avoid the question of building in the same directory at the >> sam etime. > >*That's* the real problem: dependencies. If both ports depend on the >same third package, and they both start writing on the +REQUIRED_BY >file at the same time, it could mess up that file. On the other hand - >that's not a major breakage; it just means that you don't get warned >about all the dependencies when you remove the third package. As >opposed to what happens when the two makes start trying to build the >same package, which tends to break that build. > >If no package in system is directly required by more than one unbuilt >package in the tree of dependents, you will be safe. Given that all >the dependents of the two ports are built, this means that no package >is directly required by both ports. > >