From owner-freebsd-current Sun Nov 24 16:20:48 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id QAA22236 for current-outgoing; Sun, 24 Nov 1996 16:20:48 -0800 (PST) Received: from vector.jhs.no_domain (slip139-92-4-91.mu.de.ibm.net [139.92.4.91]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id QAA22117; Sun, 24 Nov 1996 16:19:50 -0800 (PST) Received: from vector.jhs.no_domain (localhost [127.0.0.1]) by vector.jhs.no_domain (8.7.5/8.6.9) with ESMTP id BAA14499; Mon, 25 Nov 1996 01:12:17 +0100 (MET) Message-Id: <199611250012.BAA14499@vector.jhs.no_domain> To: Poul-Henning Kamp , joerg_wunsch@uriah.heep.sax.de, John Hay , Stephen McKay cc: current@freebsd.org Subject: Re: ctm-cvs-cur dealta 2716 broken From: "Julian H. Stacey" Reply-To: "Julian H. Stacey" X-Organization: Vector Systems Ltd. X-Mailer: EXMH 1.6.7, PGP available X-Address: Holz Strasse 27d, 80469 Munich, Germany X-Phone: +49.89.268616 X-Fax: +49.89.2608126 X-ISDN: +49.89.26023276 X-Web: http://www.freebsd.org/~jhs/ In-reply-to: Your message of "Fri, 22 Nov 1996 08:18:19 +0100." <8083.848647099@critter.tfs.com> Date: Mon, 25 Nov 1996 01:12:16 +0100 Sender: owner-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi Poul-Henning, J"org, John Hay, Stephen McKay CC current Reference: > From: Poul-Henning Kamp > > Hi Julian, > > If you had subscribed to "ctm-announce" like you should have if you > receive ctm deltas from the freebsd project, I have been subscribed ages :-) > you would have know > this already. Yes, Thanks, that's where I read your announcement of problem, & manual work-around. [ I didn't reply on ctm-announce@ as I thought readers might not want a mere question on that low traffic announcement list, however, I didn't make the change of list clear in my post, so sorry for confusion. ] Reference: > Reply-to: joerg_wunsch@uriah.heep.sax.de (Joerg Wunsch) > > > Will a subsequent or re-issued cvs-cur.[] self correct ( I hope ) ? > > Nope. Think about it, it's impossible. Once cvs-cur.XXXX didn't > apply, no other subsequent delta will apply. My subsequent cvs-cur have all applied (without -F (Force)), which puzzles, as Poul-Henning reported: > Martin Cracauer writes: > >ctm cvs-cur.2716.gz > > FN: CVSROOT/val-tags md5 mismatch. I don't know why mine didn't break too, (I'm rebuilding my CVS tree again from 1500, to try again). Reference: > From: John Hay > > But you can send a second (fixed) one with the same number and on most > systems it will just overwrite the previous one and everything will be > working. :-) :-) (Not that I'm suggesting that it be done.) Reference: > Stephen McKay > > Regenerating the deltas is pretty much impossible, since most people have > managed to get past the problem. However, I have faked up a CTM delta that > can be inserted between 2715 and 2716 to fix the problem for those people > (like me) who were missing CVSROOT/val-tags, and might want to start over > from scratch (due to bad disks or sudden finger trouble). > > I call it 'cvs-cur.2716+val-tags.hack' so that it sorts between > cvs-cur.2715.gz and cvs-cur.2716.gz. Any similar sorting name will do. > > Voila! Your (A) scenario is fulfilled. > > --snip-- > CTM_BEGIN 2.0 cvs-cur 2716 19961118210000Z . > CTMFM CVSROOT/val-tags 633 552 664 8f1e5fc47ec0d2efa49234192d1cf9c9 52 > RELENG_2_1_0 y > HEAD y > JULIAN_HACK y > SCSI y > ALLMAN y > > CTM_END 4da2de1d94fa81145a224934bec69d49 > --snip-- Copying Stephen's bit above into ftp:freebsd.org/b/ftp/pub/CTM/cvs-cur/cvs-cur.2716+val-tags.hack seems a nice fix, to allow for occasional rebuilds. Would you be willing to install it in the archives Poul-Henning ? Thanks Stephen I'm rebuilding from a 1500 base, & will test your delta tomorrow, when ctm gets to it. Julian -- Julian H. Stacey jhs@freebsd.org http://www.freebsd.org/~jhs/ ================================================================================ ANNEX Copy of my original post ================================================================================ > Hi Poul-Henning > CC Current > > Reference: > > From: Poul-Henning Kamp > > Subject: Re: ctm-cvs-cur dealta 2716 broken > > Date: Tue, 19 Nov 1996 11:50:31 +0100 > > > > In message <9611190816.AA13849@wavehh.hanse.de>, Martin Cracauer writes: > > > > > >ctm cvs-cur.2716.gz > > > FN: CVSROOT/val-tags md5 mismatch. > > > FN: CVSROOT/val-tags edit returned 1. > > >Exit(73) > > > > There is an unfortunate error on delta 2716. > > > > Please run these commands before applying the delta: > > > > echo 'RELENG_2_1_0 y' > $CVSROOT/CVSROOT/val-tags > > echo 'HEAD y' >> $CVSROOT/CVSROOT/val-tags > > echo 'JULIAN_HACK y' >> $CVSROOT/CVSROOT/val-tags > > echo 'SCSI y' >> $CVSROOT/CVSROOT/val-tags > > echo 'ALLMAN y' >> $CVSROOT/CVSROOT/val-tags > > > Will a subsequent or re-issued cvs-cur.[] self correct ( I hope ) ? > Or do CTM-CVS recipients need to make careful note of this in our private > local directory archives of received ctm diffs ? > > IE when people blow away their local cvs tree & try to rebuild, > will this (example) be sufficient : > (A) > ctm ......./cvs-cur.1500A.gz > ctm ......./xx/cvs-cur.1[5-9]* > ctm ......./xx/cvs-cur.2[0-9]* > > or will this be necessary: > (B) > ctm ......./cvs-cur.1500A.gz > ctm ......./xx/cvs-cur.1[5-9]*.gz > ctm ......./xx/cvs-cur.2[0-6]*.gz > ctm ......./xx/cvs-cur.270*.gz > ctm ......./xx/cvs-cur.271[1-5].gz > echo 'RELENG_2_1_0 y' > $CVSROOT/CVSROOT/val-tags > echo 'HEAD y' >> $CVSROOT/CVSROOT/val-tags > echo 'JULIAN_HACK y' >> $CVSROOT/CVSROOT/val-tags > echo 'SCSI y' >> $CVSROOT/CVSROOT/val-tags > echo 'ALLMAN y' >> $CVSROOT/CVSROOT/val-tags > ctm ......./xx/cvs-cur.271[6-9].gz > ctm ......./xx/cvs-cur.27[2-9]*.gz > ctm ......./xx/cvs-cur.2[8-9]*.gz > ctm ......./xx/cvs-cur.3*.gz > > Hoping (A) is sufficient :-) > > Julian ================================================================================