Date: Tue, 29 Apr 2003 04:24:00 +0700 From: Alexey Dokuchaev <danfe@nsu.ru> To: Peter Wemm <peter@wemm.org> Cc: freebsd-arch@freebsd.org Subject: Re: [FreeBSD-rc] Re: RFC: Removal of the old rc system from -current Message-ID: <20030428212400.GC91050@regency.nsu.ru> In-Reply-To: <20030428211331.ABDF12A7EA@canning.wemm.org> References: <3EAD83F5.7030302@btc.adaptec.com> <20030428211331.ABDF12A7EA@canning.wemm.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Apr 28, 2003 at 02:13:31PM -0700, Peter Wemm wrote: > Scott Long wrote: > > > I think that you and others have convinced me that there isn't an > > overwhelming need to deprecate rcOG. > > My take on this whole thing is that having two implementations doesn't > encourage a clean break. Suppose we have two third party vendors. One > supplies a rcOG hook because they were too lazy to convert it, and put in > their instructions "Be sure to set rcng=NO in /etc/rc.conf". > > Then, you get another component from another vendor, who only supplies > a rcNG startup module. The user now has two conflicting sets of startup > hooks to reconcile and will be forced to get their hands dirty and translate > one of them to the other. > > IMHO, make a clean break and get it over and done with. Get everybody on > the same page. Making the clean break also means that we will find anything > that has been missed (eg: /etc/netstart as referenced later in this thread) > sooner rather than later. I tend to agree: vendors that are too concerned with keeping their software compatible with RELENG_4 could spend a few human/minutes for compatibility with rcOG; others should just go for rnNG as de facto standard. Dropping rcOG for 5.1(2?) seems to help to avoid future problems rather than creating them. Just my $.02. ./danfe
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030428212400.GC91050>