From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nivedita Singhvi Subject: Re: [PATCH] tools: Fix network-bridge to work withGentoo (version 2) Date: Wed, 05 Oct 2005 13:55:38 -0700 Message-ID: <43443DCA.4080200@us.ibm.com> References: <1128545051.21137.61.camel@bree.local.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1128545051.21137.61.camel@bree.local.net> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jeremy Katz Cc: Ian Pratt , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org Jeremy Katz wrote: > On Wed, 2005-10-05 at 21:25 +0100, Ian Pratt wrote: > >>>Thinking aloud -- maybe just setting a variable so we know >>>we're in the 'make dist' case and using that to install all >>>the distro specific stuff into DESTDIR and then letting >>>install.sh also pick out the right distro would work. Then >>>for the more normal install case, just the proper things for >>>the distro would get installed. Does this seem reasonable? >>>I'll try to cook up something today >> >>This would be great, but does have the danger that we end up having a >>swathe of very similar scripts that have to be maintained together. > > > Yeah, this is the downside. After spending a while earlier looking for > how other projects handle it, it looks like most have gone the route of > "don't provide anything at all and then distros ship their own". Which > kind of sucks. :/ It's much like the current problem of not having the > packaging scripts for each distro. > > The only exception tend to be a few projects which detect your > distribution at configure time and then install an appropriate > "backend". We've been thinking a bit about this too, and I think honestly it won't be that bad - once we have stuff working (and allow distros to maintain their separate segments), there won't be that much churn in the underlying xen scripts. Your approach above is what we'd like to have. >>It was annoying that gentoo chose not to use ifup/down. I guess with the >>various udev/hotplug versions we'll need vendor-specific ways of >>addressing them too. > > > Yep, I'm having to run with Chris Bainbridge's udev rules (with a few > minor tweaks) since Fedora devel is very current udev as the hotplug > handler. That's what really got me thinking down this path. thanks, Nivedita