From owner-freebsd-ports  Thu May 30 09:07:29 1996
Return-Path: owner-ports
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id JAA18851
          for ports-outgoing; Thu, 30 May 1996 09:07:29 -0700 (PDT)
Received: from time.cdrom.com (time.cdrom.com [204.216.27.226])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id JAA18835
          for <ports@freebsd.org>; Thu, 30 May 1996 09:07:25 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by time.cdrom.com (8.7.5/8.6.9) with SMTP id JAA20697; Thu, 30 May 1996 09:06:55 -0700 (PDT)
To: asami@cs.berkeley.edu (Satoshi Asami)
cc: ports@freebsd.org
Subject: Re: Request for feedback: REQUIRES_OS_VERSION feature. 
In-reply-to: Your message of "Thu, 30 May 1996 02:41:15 PDT."
             <199605300941.CAA02402@silvia.HIP.Berkeley.EDU> 
Date: Thu, 30 May 1996 09:06:55 -0700
Message-ID: <20695.833472415@time.cdrom.com>
From: "Jordan K. Hubbard" <jkh@time.cdrom.com>
Sender: owner-ports@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> First, disabling reinstall based on NO_INSTALL is not necessary, as
> reinstall calls install and that will be disabled by NO_INSTALL.

I know, I just did that for orthogonality with the package and
repackage rules - think of it as an optimization. :-)

> Next, check sitting between configure and build is simuly bogus, if
> it's not suitable for the user's system, it shouldn't attempt to do
> anything.  Having a missing dependency (configure -> check) and
> calling configure from check is a pretty gross hack too.

I know, this was all of 10 minutes work here and I took the path of
least resistance (this was the implementation of an _idea_, not so
much the final implementation of whatever we adopt - that's why I
posted it here rather than simply commit it).  If you can come up
with a better way of causing ports who's checks fail to skip
absolutely everything, that'll be fine by me.

While you're in there, why not add an option to make clean's "chain"
to dependent ports?  It's always annoyed me that you can do a build in
port A and have it build ports B, C and D as part of the dependencies
but when you then do a make clean from A you still wind up with B, C &
D fully populated.  Basically, the clean target is next to useless
unless you always run it from the top right now.

> What we should probably do is to add it to the beginning before
> fetch, so it can also be used for other things (security/ssh's
> USA_RESIDENT check comes to mind).  Let me look into this.

Find with me!  Just so long as it's not a fatal error to fail the
check.

					Jordan