From owner-freebsd-chat Tue Aug 26 09:38:42 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id JAA20245 for chat-outgoing; Tue, 26 Aug 1997 09:38:42 -0700 (PDT) Received: from csemail.cropsci.ncsu.edu (csemail.cropsci.ncsu.edu [152.1.88.3]) by hub.freebsd.org (8.8.7/8.8.7) with SMTP id JAA20239 for ; Tue, 26 Aug 1997 09:38:37 -0700 (PDT) From: rdkeys@csemail.cropsci.ncsu.edu Received: by csemail.cropsci.ncsu.edu (5.61-AIX-1.2/1.0) id AA133423 (for freebsd-chat@freebsd.org, from rdkeys/rdkeys); Tue, 26 Aug 97 11:56:27 -0400 Received: by csemail.cropsci.ncsu.edu (5.61-AIX-1.2/1.0) id AA133414 (for rdkeys, from rdkeys/rdkeys); Tue, 26 Aug 97 11:55:53 -0400 Message-Id: <9708261555.AA133414@csemail.cropsci.ncsu.edu> Subject: Re: Suggestions from a unix dummy..... To: jamie@itribe.net (Jamie Bowden) Date: Tue, 26 Aug 1997 11:55:50 -0400 (EDT) Cc: rdkeys@csemail.cropsci.ncsu.edu () In-Reply-To: <199708261409.KAA23704@gatekeeper.itribe.net> from "Jamie Bowden" at Aug 26, 97 10:13:15 am X-Mailer: ELM [version 2.4 PL24] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-chat@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > On Tue, 26 Aug 1997 rdkeys@csemail.cropsci.ncsu.edu wrote: > > > > > > > On Mon, 25 Aug 1997, Tim Vanderhoek wrote: > > > > > > > Officially ports are only supported on -stable. This is to keep > > > > the amount of work manageable. However, if you find a port that > > > > doesn't work on -current or any other release (PROVIDED you are > > > > using an up-to-date bsd.port.mk and friends), you are encouraged > > > > to send patches in a pr. > > > > > > This is wrong, IMHO. Ports should build on the latest official release. > > > > > > Jamie Bowden > > > > From the philosophical point of view, why is it not acceptable that > ports > > should make on ANY FreeBSD box? Assuming that one starts from fresh > > archive sources (and it seems to this unix dummy that they do) unless one > > has more than trivial patches, a make is a make is a make. I have ported > > a lot of minor things to my boxes (aix/minix/linux/FreeBSD). Aside from the > > quirks of my early aix 1.2 and my minix not being ``GCC compliant'', I find > > that essentially everything makes out of the box on all the machines. > > Porting is not that difficult anymore (99% of the time). Most sources make > > out of the box with any sort of half-intelligent configure script. That > > would make me still think a common ports tree across all FreeBSD platforms > > would be an ideal goal, and not that difficult to attain. Then, again, I > > am just the unix dummy..... > > > > Bob Keys > > rdkeys@csemail.cropsci.ncsu.edu > > > > There are significant changes from 1.x to 2.x, and from 2.1.x to 2.2.x. I > am sure there are significant differences as well from 2.2.x to what will > be 3.x > > Jamie Bowden > > System Administrator, iTRiBE.net But how should that have any affect on ports. Ports are not part of the OS but are brought in from external sources. A well-made port should make on any standard unix box. FreeBSD is not that unstandard. The ports mechanism should only be a convenience overlayer on the normal external sources. If not, then the way ports are made on FreeBSD might need some rethinking. I can bring in GCC, 2.7.2.2 for example and install it on the 3.0 snap, the 2.2.x boxes and the 2.1.7.1 boxes and they all install fine, no patches, no hacks, just a proper configure, make, make install. I did the same thing with TeX and Groff and everything else that I consider ``sacred'' sources for graphics applications, programming, and other things that I need to do. I can't see any sane reason that a TeX port on 2.1.7.1, 2.0.5, 1.x or 3.0 should not do exactly the same thing for all versions of the OS. If any OS particular revision level patches are required (and they should only be rare on FreeBSD), the ports build should be sufficiently aware of the OS revision level to take that into consideration. Still thinking, wishfully..... Bob Keys