Date: Tue, 13 Mar 2001 19:52:15 -0500 (EST) From: mi@aldan.algebra.com To: Will Andrews <will@physics.purdue.edu> Cc: Alexander Langer <alex@big.endian.de>, "Andrey A. Chernov" <ache@nagual.pp.ru>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: revising the package building (Re: cvs commit: ports/graphics/Image Magick Makefile pkg-descr pkg-p list) Message-ID: <200103140052.f2E0qGC13849@misha.privatelabs.com> In-Reply-To: <20010313182036.M61859@ohm.physics.purdue.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On 13 Mar, Will Andrews wrote: = On Tue, Mar 13, 2001 at 05:54:20PM -0500, mi@aldan.algebra.com wrote: = > A lot of ports have WITH and WITHOUT options. Honoring them all with = > a separate sub-port would explode the ports tree. Think of the = > amount of permutations of mod_php, ghostscript, or postgresql will = > have. Is NO_X -- an exception? May be... I'll do what my mentor = > tells me to. = = Will's principle: If it's not your port, it's not your problem. It's nobody's :-( ... = I dislike creating new directories just for certain options too, but = the way bento builds packages severely limits our options. There was a = huge discussion about this last September (see: "ports options paper" = I wrote), and according to Satoshi some years ago as well. Perhaps, bento's script needs to be modified. I'll even say it is broken, because it assumes the one-to-one relation between ports and packages. It can have a list of "interesting" options, and will try to produce the name of each package with and without this options (all permutations). If a new name is obtained -- the port is rebuilt with that option set and the new package is created -- from the same port. This way, the decision, which options are worthy is centralized, but a port's maintainer can turn the separate package build on and off by modifying the package name (typicly -- by altering the prefix and/or suffix). And there is now need to add another port per option. How is that? -mi (ducking, 'cause he did not read his mentor's paper) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200103140052.f2E0qGC13849>