From mboxrd@z Thu Jan 1 00:00:00 1970 Message-Id: <6.0.0.22.0.20031026203433.04e13770@pop.noos.fr> To: 9fans@cse.psu.edu From: Philippe Anel Subject: Re: [9fans] pcf config file ... In-Reply-To: <200310261832.h9QIWJl21300@augusta.math.psu.edu> References: <200310261832.h9QIWJl21300@augusta.math.psu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Date: Sun, 26 Oct 2003 20:37:49 +0100 Topicbox-Message-UUID: 7995e4aa-eacc-11e9-9e20-41e7f4b1d025 At 19:32 26/10/03, you wrote: >"Russ Cox" writes: > > pull is hard-coded to run "9fs kfs". really you should be using > > dan cross's /dist/replica/dist, but i never remember how to do that. > > instead you can just create a /srv/kfs that pull will be able > > to mount (it doesn't see one so it starts kfs). > > > > echo 1 >/srv/kfs > > > should do the trick. > >It's called `inst', and while it's on sources, it never made it to the >base distribution. It is, however, pretty easy to use: > > term% mount -c /srv/boot /n/inst > term% 9fs sources > term% cp /n/sources/plan9/dist/replica/inst /dist/replica > term% bind /n/sources/plan9 /n/dist > term% replica/pull -v /dist/replica/inst It worked ! Thank you. > After doing so, however, I'm a little dismayed >about how much still depends on kfs, particularly the replica scripts >under /dist/replica (and in individual package replica scripts). It >seems to me that if fossil is the way to go moving forward, it would >behoove us to update the other parts of the system to treat it with a >little more respect. To that end, I like Andrey's modification that >Ron just posted; it'll also work on fileservers and the like. In >short, we should be updating the system to be more general, not trying >to fool things into looking vaguely like kfs so old tools will continue >working. I agree with you. Philippe,