From owner-freebsd-current Sun Oct 4 16:19:39 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA16772 for freebsd-current-outgoing; Sun, 4 Oct 1998 16:19:39 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from mail.scsn.net (scsn.net [206.25.246.12]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA16759 for ; Sun, 4 Oct 1998 16:19:29 -0700 (PDT) (envelope-from dmaddox@scsn.net) Received: from rhiannon.scsn.net ([209.12.57.9]) by mail.scsn.net (Post.Office MTA v3.1.2 release (PO205-101c) ID# 0-41950U6000L1100S0) with ESMTP id AAA106; Sun, 4 Oct 1998 19:10:21 -0400 Received: (from root@localhost) by rhiannon.scsn.net (8.9.1/8.9.1) id TAA00538; Sun, 4 Oct 1998 19:19:44 GMT (envelope-from root) Message-ID: <19981004191943.A422@scsn.net> Date: Sun, 4 Oct 1998 19:19:43 +0000 From: dmaddox@scsn.net (Donald J. Maddox) To: Mike Smith Cc: current@FreeBSD.ORG Subject: Re: Shouldn't 'make includes' install stand.h? Reply-To: dmaddox@scsn.net References: <19981004133239.A309@scsn.net> <199810042108.OAA06658@dingo.cdrom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.93.2i In-Reply-To: <199810042108.OAA06658@dingo.cdrom.com>; from Mike Smith on Sun, Oct 04, 1998 at 02:08:42PM -0700 Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Sun, Oct 04, 1998 at 02:08:42PM -0700, Mike Smith wrote: > > > > Shouldn't a 'make -DCLOBBER includes' result in a _complete_ set > > > > of includes? Are there other includes than stand.h that don't > > > > get installed by 'make includes'? > > > > > > No. "Make includes" installs random header files. libstand.h is > > > installed at the same time libstand is; if you install just the former, > > > you're going to die in the link phase when you can't find the latter. > > > > 'Random header files'? Ok, if you say so, but so far, the only "standard" > > component of /usr/include I've managed to identify as _not_ installed by > > 'make includes' is stand.h. This seems counterintuitive to me... > > libstand.h is not a "standard" component of /usr/include, any more than > eg. zlib.h. > > > Your point about installing the header without the lib is valid, but it > > seems to me that this is applicable to just about all of the includes, in one > > way or another. Maybe there shouldn't be a 'make includes' target at all? > > 'make includes' theoretically exists as a catchall to install headers > not associated with any particular item. It's commonly abused to avoid > the chicken-and-egg problem that occurs when you try to build a tool > that consumes an interface to something else that hasn't been built yet. > > > I'm not trying to be combative here; this is not a religious issue to > > me... The current behavior just seems to me to violate POLA. > > Only if you have misapprehensions about what 'make includes' does. It > seems mostly to trip up people with such misapprehensions. Well, I can only plead guilty as charged. If I want to rebuild everything in /usr/bin, I can cd to /usr/src/bin, type 'make all install', and watch everything in /usr/bin be rebuilt. The same applies to /bin, /sbin, /usr/sbin, /usr/share, etc., etc., etc. I thought the 'make includes' target was intended to rebuild /usr/include in the same way; it appears, however, that the only way to repopulate the /usr/include dir _completely_ is 'make world'. Oh well... To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message