From owner-cvs-all@FreeBSD.ORG Sun Feb 19 16:45:28 2012 Return-Path: Delivered-To: cvs-all@FreeBSD.org Received: by hub.freebsd.org (Postfix, from userid 1033) id E584C106566C; Sun, 19 Feb 2012 16:45:28 +0000 (UTC) Date: Sun, 19 Feb 2012 16:45:28 +0000 From: Alexey Dokuchaev To: cvs-ports@FreeBSD.org Message-ID: <20120219164528.GA48166@FreeBSD.org> References: <201202182356.q1INuU7V061378@repoman.freebsd.org> <20120219060053.GA45762@FreeBSD.org> <20120219150943.GA6673@magic.hamla.org> Mime-Version: 1.0 Content-Type: text/plain; charset=koi8-r Content-Disposition: inline In-Reply-To: <20120219150943.GA6673@magic.hamla.org> User-Agent: Mutt/1.4.2.1i Cc: Eitan Adler , cvs-all@FreeBSD.org, ports-committers@FreeBSD.org Subject: Re: cvs commit: ports/devel/adime Makefile ports/x11-wm/icewm Makefile ports/graphics/scr2png Makefile ports/x11/xbindkeys Makefile X-BeenThere: cvs-all@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: **OBSOLETE** CVS commit messages for the entire tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 19 Feb 2012 16:45:29 -0000 On Sun, Feb 19, 2012 at 10:09:44AM -0500, Sahil Tandon wrote: > Alexey, I appreciate your perspective on this, but I want to make sure > you are aware of the related discussion (and opposing viewpoints) on > freebsd-ports: Yes, I've heard that reasoning a number of times since I started to hack on ports, which is a relatively long period. Technically, since there is no supporting infrastructure to ensure if some app really requires particular shlib version or not (and I personally do not see how it can be implemented automatically for any arbitrary port out there), the whole issue right now is little more than a matter of personal preference. I agree that most of the times it does not matter if you specify shlib version in LIB_DEPENDS or not; there're currently no ports of the same origin that install different ABI shlibs, and our framework does not support it either, AFAIK. I want to make it clear that I don't know of any technical problems of not specifying shlib version, as it is hard to verify and we cannot be sure that upstream maintainers bump it correctly in 100% of cases anyways. However, I know what most ports specify it, and it makes sence to me for the reasons I state above: it makes easier to spot when some port was not updated during the build phase on any tinderbox. Without that and forgotten port revision bump, users would find it out themselves that their not updated package does not work because of ABI changes. With shlib tracking, such port would be broken on tinderbox and likely be fixed before it will be installed by our end users. Bottom line: unless we have reliably working solution to ensure if a port really wants a particular ABI or not, it's better to be on the safe side. ./danfe