From owner-freebsd-x11@FreeBSD.ORG Thu Jun 28 08:25:47 2012 Return-Path: Delivered-To: x11@freebsd.org Received: by hub.freebsd.org (Postfix, from userid 1033) id 178401065901; Thu, 28 Jun 2012 08:25:47 +0000 (UTC) Date: Thu, 28 Jun 2012 08:25:47 +0000 From: Alexey Dokuchaev To: Konstantin Belousov Message-ID: <20120628082546.GA46785@FreeBSD.org> References: <20120606131623.GA82148@ithaqua.etoilebsd.net> <20120628042051.GA8166@FreeBSD.org> <20120628080012.GT2337@deviant.kiev.zoral.com.ua> Mime-Version: 1.0 Content-Type: text/plain; charset=koi8-r Content-Disposition: inline In-Reply-To: <20120628080012.GT2337@deviant.kiev.zoral.com.ua> User-Agent: Mutt/1.4.2.1i Cc: x11@freebsd.org Subject: Re: Proposition to make nvidia driver stop overwritting files X-BeenThere: freebsd-x11@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: X11 on FreeBSD -- maintaining and support List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 28 Jun 2012 08:25:47 -0000 On Thu, Jun 28, 2012 at 11:00:12AM +0300, Konstantin Belousov wrote: > On Thu, Jun 28, 2012 at 04:20:51AM +0000, Alexey Dokuchaev wrote: > > Thanks for not touching 8.x. This is probably the last known branch that > > currently more or less reliably works (the trend was broken Januaryish). > > It seems that this statement makes a carte blanche to claim that current > situation with nvidia driver port just demonstrates complete lack of > care from the maintainer, for quite lot a time? If by "complete lack of care" you mean having fully working four branches of the driver accurately covering broad range of FreeBSD source base, then I'd have to say yes, but are we talking in the same language? All I was saying that there is a sad tendency that X+1 branch will less likely work than X on the same hardware. > > Include-dances in libmap.conf are standard way in freenix world to solve > > problems like this one, but right now I see this scenario happening on a > > developer's box rather that on user's. User would probably just want > > single libGL+friends implementation that best suits his/her gfx card. > > So you completely ignore/do not care about users of hybrid machines. > Nice. No, just didn't think about it. Good point, thanks for bringing it up. Now it's pretty clear that we need more complicated, but more functional solution. ./danfe