From owner-freebsd-ports  Tue Sep 24 01:59:16 1996
Return-Path: owner-ports
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id BAA27231
          for ports-outgoing; Tue, 24 Sep 1996 01:59:16 -0700 (PDT)
Received: from silvia.HIP.Berkeley.EDU (ala-ca8-50.ix.netcom.com [207.93.141.178])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id BAA27194;
          Tue, 24 Sep 1996 01:59:09 -0700 (PDT)
Received: (from asami@localhost) by silvia.HIP.Berkeley.EDU (8.7.6/8.6.9) id BAA18585; Tue, 24 Sep 1996 01:59:01 -0700 (PDT)
Date: Tue, 24 Sep 1996 01:59:01 -0700 (PDT)
Message-Id: <199609240859.BAA18585@silvia.HIP.Berkeley.EDU>
To: jkh@time.cdrom.com
CC: andreas@freebsd.org, ports@freebsd.org
In-reply-to: <3226.843546061@time.cdrom.com> (jkh@time.cdrom.com)
Subject: Re: cvs commit: ports/graphics/xv Makefile
From: asami@freebsd.org (Satoshi Asami)
Sender: owner-ports@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

 * Yeah, I've commented on this in the past - our clean rule should
 * follow the same dependencies which would have been traversed in
 * building the port.  Right now, you can do:
 * 
 * 	make all install && make clean
 * 
 * In many ports and wind up with far more bits on your disk than when
 * you started due to depended ports being built but not cleaned again.
 * I'd say it's a "bug" for bsd.port.mk to fix and not something the
 * ports should be doing on their own, definitely.

Well, I wouldn't call that a "bug" though.  The problem here is that
we don't really know if the dependency was built just for this port or 
not.

For instance, if someone is trying to fix a bug in the jpeg port, and
a "make clean" in xv deletes jpeg's work directory with all his
patches and stuff, we'll have one very unhappy hacker. ;)

I can change bsd.port.mk to keep the list of ports actually built
because of dependencies, but is this really worth it?

Satoshi