Date: Wed, 16 Nov 2016 01:15:54 +0100 From: "Julian H. Stacey" <jhs@berklix.com> Cc: "Montgomery-Smith, Stephen" <stephen@missouri.edu>, "ctm-users@freebsd.org" <ctm-users@freebsd.org> Subject: Re: current ctm reports: Fatal error: Probably not a CTM-patch at all. Message-ID: <201611160016.uAG0Fsne015388@fire.js.berklix.net> In-Reply-To: Your message "Mon, 07 Nov 2016 16:20:41 %2B0100." <201611071520.uA7FKfYQ007164@fire.js.berklix.net>
next in thread | previous in thread | raw e-mail | index | archive | help
"Julian H. Stacey" wrote: Mon, 07 Nov 2016 16:20:41 +0100 > "Montgomery-Smith, Stephen" wrote: > > On 11/05/2016 10:39 AM, Julian H. Stacey wrote: > > > Hi ctm-users@freebsd.org > > > Anyone else seeing similar strangeness ? ... > > > No problem running on 9.2-RELEASE, but on today's current : > > > ( /usr/src .svn_revision 308280 .ctm_status src-cur 12727 ) > > > after an installworld & kernel ) > > > > > > cd /usr/svn; cat .ctm_status > > > svn-cur 5081 > > > ctm -q /pub/FreeBSD/development/CTM/svn-cur/svn-cur.05[0-9][0-9][0-9].xz > > > /pub/FreeBSD/development/CTM/svn-cur/svn-cur.05000.xz \ > > > Fatal error: Probably not a CTM-patch at all. > > > # Yes I know I dont need earlier deltas, but its a cut & paste > > > # suggestion from my standard script > > > # The point is, ctm always used to just skip past deltas it didnt need, > > > # including those it now doesnt recognise as deltas. > > > # ctm also fails on delta it does need, see below: > > > # Both these fail: > > > ctm /pub/FreeBSD/development/CTM/svn-cur/svn-cur.05082.xz > > > ctm -v /pub/FreeBSD/development/CTM/svn-cur/svn-cur.05082.xz > > > Fatal error: Probably not a CTM-patch at all. > > > ctm: exit(65) > > > > > > Delta is not corrupt: > > > > > > cd /pub/FreeBSD/development/CTM/svn-cur > > > sha256 svn-cur.05082.xz > > > SHA256 (svn-cur.05082.xz) = d2b20f44265cca77a811eb074abdbe58466bf72d8f5f3a39113a7b82a5a0b1b8 > > > > > > ssh ctm.berklix.org (or as no public login, ftp & sha256 localy) > > > cd /pub/FreeBSD/development/CTM/svn-cur > > > sha256 svn-cur.05082.xz > > > SHA256 (svn-cur.05082.xz) = d2b20f44265cca77a811eb074abdbe58466bf72d8f5f3a39113a7b82a5a0b1b8 > > > > > > Proof that its not delta at fault, but ctm command: > > > > > > cd /pub/FreeBSD/development/CTM/svn-cur; pwd > > > /.amd_mnt/fire/0s4/ftp/pub/FreeBSD/development/CTM/svn-cur > > > cd /usr/svn;pwd > > > /data/usr/ftp/pri/FreeBSD/development/FreeBSD-SVN > > > > > > ssh fire # 9.2 host > > > cd /host/lapr/data/usr/ftp/pri/FreeBSD/development/FreeBSD-SVN > > > ctm /0s4/ftp/pub/FreeBSD/development/CTM/svn-cur/svn-cur.05082.xz > > > ctm: .ctm_status: Permission denied > > > # OK, i just auto apply deltas from /etc/aliases, hence > > > # it complaining about ownerships, but not about delta content. > > > > > > ssh -l mailnull-csh localhost # still on 9.2 host > > > cd /host/lapr/data/usr/ftp/pri/FreeBSD/development/FreeBSD-SVN > > > ctm /0s4/ftp/pub/FreeBSD/development/CTM/svn-cur/svn-cur.05082.xz > > > > FS .ctm_status > > > > SV > > > svnadmin: E000045: Can't get exclusive lock on file 'base/db/txn-current-lock': Operation not supported > > > Broken pipe > > > # OK that's some NFS problem applying delta, but at least old > > > # 9.2 ctm command accepts delta to apply, unlike current ctm > > > > > > Curiously I am not seeing same problem running on 12 current file system > > > & 12 current bin/ctm, applying deltas to src-cur/ > > > cat .ctm_status > > > src-cur 12727 > > > ctm -q /pub/FreeBSD/development/CTM/src-cur/src-*.1[0-9][0-9][0-9][0-9].gz > > > cat .ctm_status > > > src-cur 12729 > > > > > > Cheers, > > > Julian > > > > > > > Did you remember to apply the patch to ctm after applying the src-cur > > deltas? (I forget to do this all the time.) > > I started from a make world & kernel built from generic current src/ > uname -a > FreeBSD lapr.js.berklix.net 12.0-CURRENT FreeBSD 12.0-CURRENT > #12727: Sat Nov 5 13:42:38 CET 2016 > jhs@lapr.js.berklix.net:/data/release/12.0-CURRENT/usr/src/sys/\ > amd64/compile/LAPR.small amd64 > > I assumed all CTM patches were committed to current, as years old: > ssh ctm.berklix.org > cd /pub/FreeBSD/development/CTM/svn-cur ; ls -l | grep patch| grep -v .sig > drwxr-xr-x 2 jhs staff 1024 Nov 7 11:41 old-patches/ > -rw-r--r-- 1 jhs staff 16221 Nov 15 2014 patch-for-ctm-v09 > -rw-r--r-- 1 jhs staff 878 Sep 24 2012 patch-for-ctm_rmail > -rw-r--r-- 1 jhs staff 583 Dec 25 2011 patch-for-ctm_smail > -rw-r--r-- 1 jhs staff 8655 Nov 15 2014 patch-for-mkctm-v05 > > cd /usr/src/usr.sbin/ctm > # Current .svn_revision 308337 .ctm_status src-cur 12729 > patch < /pub/FreeBSD/development/CTM/svn-cur/patch-for-ctm-v09 > Reversed (or previously applied) patch detected! Assume -R? [y] > n > n > # OK, that patch seems to be in current. > > patch < /pub/FreeBSD/development/CTM/svn-cur/patch-for-ctm_rmail > Hunk #1 succeeded at 152 (offset 2 lines). > Hunk #2 succeeded at 201 with fuzz 1 (offset 2 lines). > # NOT IN CURRENT ! > > patch </pub/FreeBSD/development/CTM/svn-cur/patch-for-ctm_smail > Reversed (or previously applied) patch detected! Assume -R? [y] n > n > n > # OK in current. > > patch </pub/FreeBSD/development/CTM/svn-cur/patch-for-mkctm-v05 > Hunk #1 succeeded at 185 (offset 4 lines). > Hunk #2 succeeded at 230 (offset 4 lines). > Hunk #3 succeeded at 306 (offset 4 lines). > Hunk #4 succeeded at 338 (offset 4 lines). > Hunk #5 succeeded at 634 with fuzz 1 (offset 128 lines). > Hunk #6 failed at 888. > 1 out of 6 hunks failed--saving rejects to mkCTM/mkctm.c.rej > # NOT IN CURRENT ! > > Should patches not in current be commited ? Who is/are author[s] ? > > ctm_rmail.c: Seems a small patch for mutex, lack of application > probably not causing "Fatal error: Probably not a CTM-patch at all" > mkctm.c: lack of application of patches on receive end, probably not the > answer to "Fatal error: Probably not a CTM-patch at all" Hi CTM people, Belated apologies ! It looks like the original phenomena I described may have been caused by bad blocks on disk. Still not sure, but after an installworld & kernel & rebuild of svn tree it seems to have gone away. However the question on diffs above still applies. Cheers, Julian -- Julian Stacey, BSD Linux Unix Sys Eng Consultant Munich Reply below, Prefix '> '. Plain text, No .doc, base64, HTML, quoted-printable. http://berklix.eu/brexit/#stolen_votes
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201611160016.uAG0Fsne015388>