From owner-freebsd-ports Thu May 8 02:48:06 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id CAA18626 for ports-outgoing; Thu, 8 May 1997 02:48:06 -0700 (PDT) Received: from silvia.HIP.Berkeley.EDU (ala-ca26-32.ix.netcom.com [207.93.42.96]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id CAA18621 for ; Thu, 8 May 1997 02:48:03 -0700 (PDT) Received: (from asami@localhost) by silvia.HIP.Berkeley.EDU (8.8.5/8.6.9) id CAA18182; Thu, 8 May 1997 02:47:46 -0700 (PDT) Date: Thu, 8 May 1997 02:47:46 -0700 (PDT) Message-Id: <199705080947.CAA18182@silvia.HIP.Berkeley.EDU> To: james@nexis.net CC: chuckr@mat.net, freebsd-ports@FreeBSD.ORG In-reply-to: (message from James FitzGibbon on Sun, 4 May 1997 20:00:58 -0400 (EDT)) Subject: Re: Suggested change to apache port From: asami@vader.cs.berkeley.edu (Satoshi Asami) Sender: owner-ports@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk * This works, but in the specific case of p5-Apache, I required two things * that the patch didn't cover: * * - The ability to specify a double target ("clean patch") so that I know * I'm working off of clean source I don't think that's what we really want to do. The ports' directories are supposed to have states, that only go forward unless you explicitly type "make clean". I don't know what's going to happen, for instance, if you type "make install package" in ports/www. Actually I do know what's going to happen; apache will be built and installed, p5-Apache will come and clean apache/work, build and install itself, apache will be installed (again) and packaged (will p5-Apache build in apache/work? if that's so, this second installation will be that of p5-Apache, not vanilla apache), then p5-Apache will be packaged. As you can see, this opens up a whole new can of worms. * - Putting it in DEPENDS instead of *_DEPENDS. Yeah, as I said, that's only a temporary patch. I didn't want to duplicate code before I knew it was working. ;) * For now, I kludged it in to cd into ${PORTSDIR}/www/apache and do a make * clean patch rather than keep the requestor of the port waiting. Let me repeat my original suggestion, that is to use as many of the port files under apache as possible but build by yourself under p5-Apache. You can't really use the depend mechanism unless you are (1) only moving forward, (i.e., if you specify "patch", you can use the "build" or "install"ed state too), and (2) not changing anything under the depended port's work/ subdir. * Still, as Chuck mentioned, these changes will make life a lot easier for * future ports (and some existing ones that do really ugly things right * now). Let me know more about the ugliness. Satoshi