From owner-freebsd-gnome@FreeBSD.ORG Sun Jan 3 03:52:51 2010 Return-Path: Delivered-To: gnome@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 32C1E1065698 for ; Sun, 3 Jan 2010 03:52:51 +0000 (UTC) (envelope-from rfg@tristatelogic.com) Received: from outgoing.tristatelogic.com (segfault.tristatelogic.com [69.62.255.118]) by mx1.freebsd.org (Postfix) with ESMTP id E5BFE8FC1E for ; Sun, 3 Jan 2010 03:52:50 +0000 (UTC) Received: from segfault-nmh-helo.tristatelogic.com (localhost [127.0.0.1]) by segfault.tristatelogic.com (Postfix) with ESMTP id 2143CBDC43; Sat, 2 Jan 2010 19:52:45 -0800 (PST) To: Joe Marcus Clarke In-Reply-To: <1262486170.16768.56.camel@shumai.marcuscom.com> Date: Sat, 02 Jan 2010 19:52:45 -0800 Message-ID: <17006.1262490765@tristatelogic.com> From: "Ronald F. Guilmette" Cc: gnome@FreeBSD.org Subject: Re: ports/140563: net/avahi-app doesn't build - depends on missing "gnome-config" X-BeenThere: freebsd-gnome@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: GNOME for FreeBSD -- porting and maintaining List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 03 Jan 2010 03:52:51 -0000 In message <1262486170.16768.56.camel@shumai.marcuscom.com>, you wrote: >> If not, then may I hunbly suggest fixing the simple (and now obvious) goof >> in the avahi-app Makefile? > >No. If you want to fight this battle, then do so with the Avahi author. Now that I know the problem is in the avahi sources, I will do that too, however... As the maintainer of the FreeBSD port, you _could_ write a FreeBSD-specific patch (along the lines of the one I supplied) which would be applied against the "standard" avahi-app distribution tarball prior to configure/build. I'm not saying that that is the Right Thing or even the Best Thing, but I think you should at least consider it before dismissing the idea out of hand. I understand your point that the FreeBSD port maintainers cannot possibly test all ports with all possible combinations of environment variables, however it seems to me that ports should build with any & all ``reasonable'' sets of environment variables... and I don't consider POSIXLY_CORRECT at all unreasonable. (Take it from me... The vast majority of ports _do_ build with this set. I have hundreds and hundreds that I've built & installed with POSIXLY_CORRECT set.) But your point is well taken that this isn't in any sense FreeBSD's fault. Nontheless, a certainly accomodative frlendliness towards wayward souls such as myself (who use FreeBSD and who have a health respect for standards) would not, I think, be unreasonable. A somewhat different suggestion: I can't see that it would hurt anything to, for example, create a net/avahi-app/scripts/configure script for the port that would explicitly "unset POSIXLY_CORRECT" just prior to invocation of avahi's own configure script. In addition to avoiding having other people like me fall into the trap, at least that would serve to _document_ the fact that this particular port uses non-POSIXly-correct command invocations (unlike the vast majority of other FreeBSD ports). >> P.S. I see that you have closed this PR (ports/140563). I disagree that >> it should be closed, and I request that you re-open the PR. > >No. I have identified the problem as not being a FreeBSD-specific bug. Why didn't you say that in your closing message in this PR then, rather that chalking it up to pilot error, as you seem to have done? (Call me overly sensitive if you like, but this was _not_ a case of pilot error. *I* did not screw up. Whoever wrote configure.ac did, i.e. by breaking a rule relating to UNIX command invocation that has been THE rule since the beginning of time.) I'd appreciate it if you would amend your closing annotation to indicate that yes, this is a _real_ bug... just not one that belongs to FreeBSD exclusively. (I have no objection to you closing the PR as long at it is _with_ that clarification.) >If you really want to get this fixed, then you need to open a bug at >http://avahi.org. Thanks. Shall do. >But be aware that other ports may do exactly the same >thing, and you may have to fight this battle in a lot of other places. Thanks for the friendly caution, but I am not at all worried. As I've said, I have built hundreds and hundreds of FreeBSD ports... over a span of more than ten years. In all that time, I've only seen a grand total of -two- (count 'em, two) instances of this kind of command invocation faux pas... one of them being this new one in avahi-app. (The other one, as I said, was in some port required by the nmh mail port... and that one got fixed already, I think.) >Additionally, POSIXLY_CORRECT is a non-default environment variable, and >one that we do not test with. OK. Sounds reasonable. Are these two important facts documented anywhere? (One hopes so. If not, they should be.) >We do not have the resources to make sure >that all ports build and operate properly with it set. Roger wilco. Understood. (Anyway, you don't have to worry about that. *I* have already, informally, done the test builds on hundreds and hundreds of ports, and trust me... hardly anybody messes up so bad that they put non-option arguments before option arguments. Like I said, most people learn the Right Way to order command line things on Day One of UNIX school.) >I strongly suggest that if you want a more problem-free experience moving >forward, you unset this. Thank you for the friendly suggestion. I wil take it under advisement. (Certainly, you are correct that my life is likely to be more problem-free if I follow your advice in this regard. But I have this unfortunate and lifelong habit of swimming against the tide, whenever the tide doesn't seem Right to me.) Regards, rfg