/srv/irclogs.ubuntu.com/2014/08/07/#ubuntu-devel.txt

=== _salem is now known as salem_
hallynsarnold: jdstrand sadly, i'm getting the qcow corruption you do.  since i'm on a buggy thinkpad i haven't disproven that it's a thinkpad-only bug02:38
sarnoldhallyn: yay?02:38
sarnoldhallyn: sorry it hit you too, but hopefully that'll make recreating it easier..02:39
sarnoldI'd really hope qcow2 is pretty boring and not care about the hardware it's running on; it doesn't feel like a special-enough snowflake to care, you know?02:39
jdstrandhallyn: I would hope it wasn't thinkpad specific like you said, but yeah, 'yay'?02:43
jdstrandhallyn: is this on utopic or trusty?02:43
jdstrandhallyn: I'm on utopic now, but haven't tried 2.1 since I've been needing my VMs lately02:43
hallyni'm on utopic02:48
hallynunfortunately it's interfering with my attempts to test qemu live migration through libvirt02:48
hallyni wasn't (yet) looking to reproduce it :)02:48
=== timrc is now known as timrc-afk
=== timrc-afk is now known as timrc
=== salem_ is now known as _salem
=== ryanneufeld_ is now known as ryanneufeld
=== ivoks_ is now known as ivoks
Unit193After the upstart upgrade, the following files are considered obsolete: /etc/X11/Xsession.d/00upstart /etc/X11/Xsession.d/99upstart /etc/cron.daily/upstart /etc/bash_completion.d/upstart /etc/upstart-xsessions05:51
pittixnox: I did an initial cleanup of the "uncategorized" ones in http://pad.ubuntu.com/missing-systemd-units05:55
pittislangase`, xnox: ^ what Unit193 said: that's what I meant by "needs dpkg-maintscript for migration" (same for conffile questions)05:56
=== sunweave1 is now known as sunweaver
=== RAOF_ is now known as RAOF
=== stub` is now known as stub
=== buxy_bak is now known as buxy
=== yofel_ is now known as yofel
=== slangase` is now known as slangasek
dholbachgood morning06:51
=== tvoss is now known as tvoss|test
=== tvoss|test is now known as tvoss
=== arrrghhh is now known as arrrghhhAWAY
=== doko_ is now known as doko
dokotvoss, please don't close MIR issues in changelog entries07:54
dokotvoss: lp: #1338587 is incomplete07:54
ubottuLaunchpad bug 1338587 in qtquick1-opensource-src (Ubuntu) "[MIR] trust-store" [Undecided,Incomplete] https://launchpad.net/bugs/133858707:54
tvossdoko, I can get rid of gcovr for the package07:56
tvossdoko, and I thought all of qt is in main?07:57
dokotkamppeter, pitti: foo2zjs build-depends on lcms which is in universe. only lcms2 is in main, you need to port to lcms2 if you want to have  foo2zjs built07:57
dokotvoss, wishfull thinking =)07:57
dokoMirv, ^^^07:57
Mirvtvoss: um, why is trust-store using qtquick1-5-dev, nothing we use should be using it?08:09
tvossMirv, just patching it out08:10
Mirvtvoss: maybe an accidental build dependency?08:10
Mirvtvoss: ok, thanks!08:10
MirvQt upstream indeed has some funky naming that can cause confusion08:10
tvossMirv, yup08:10
Mirv("qtquick1" source package provides "libqtdeclarative" library, while "qtdeclarative" source package provides "libqtquick" and "libqtqml")08:11
Mirvthe explanation is something along the lines that Qt4 also had declarative, so for the version 1 they stuck with that source package name, even though Qt 5 has the new version 2 declarative too08:12
Mirvs/source package name/library name/. see, the confusion.08:12
tvossMirv, https://code.launchpad.net/~thomas-voss/trust-store/remove-unneeded-build-dependencies/+merge/22990708:13
tvossdoko, ^08:13
=== Riddelll is now known as Riddell
darkxstpitti, so 214 seems to be working well with both upstart and systemd boot, although I am getting very occasional, races on upstart boot where gdm/X fails to start (pretty sure that has happend on 208 as well though08:32
pittidarkxst: ah, great to hear!08:41
=== tvoss is now known as tvoss|afk
tkamppeterdoko, pitti, the lcms1 build dependency in foo2zjs is bogus, it actually does not depend on any lcms library (any more). I will remove it.08:57
tkamppeterdoko, pitti, OdYX, has freshly introduced this dependency to the obsolete library, as there was some utility program copied from lcms1 in the source code of foo2zjs. I will ask him whether there is a way he can solve it with lcms2. If we do not get a quick solution I will put up an out-of-sync vesion simply taking the newest upstream source but not changing anything else.09:03
=== tvoss|afk is now known as tvoss
xnoxjdstrand: ping bug #1341083 will you upload that, or shall i?09:14
ubottubug 1341083 in ufw (Ubuntu) "ufw needs systemd unit or init.d script" [Medium,In progress] https://launchpad.net/bugs/134108309:14
xnoxogra_`: do you agree that ubuntu-defaults-nexus7 can be removed? bug #135392209:22
ubottubug 1353922 in ubuntu-defaults-nexus7 (Ubuntu) "RM: ubuntu-defaults-nexus7 obsolete product" [Medium,Confirmed] https://launchpad.net/bugs/135392209:22
=== pete-woods1 is now known as pete-woods
mlankhorstxnox: I can't even get it to work on raring or later09:23
ogra_`xnox, yeah,, burn it with fire09:23
* mlankhorst tried09:23
xnoxexcellent, mark it as "ported to systemd by means of removal"09:24
pittiand *swoosh* gone it is, updated bug09:32
=== bluesabre_ is now known as bluesabre
=== MacSlow is now known as MacSlow|lunch
pittixnox: very nice, systemd-sysv is now installable in an utopic-proposed chroot (and upstart can be removed)10:48
pittixnox: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt still complains about systemd-sysv, so I figure before britney is happy we'll need to fix *all* dependencies to upstart (and we don't actually want that..)10:49
=== Pici` is now known as Pici
=== ogra_` is now known as ogra_
xnoxpitti: hm. but systemd-sysv does declare conflicts and replaces upstart right?11:06
xnoxsomething is odd, we do know they are not co-installable. and plenty of other packages/providers act the same way.11:06
xnoxpitti: latest upload 208-7ubuntu4 ?11:06
pittixnox: no, not replaces: (that would be wrong), just conflicts:11:07
pittiskipped: systemd (20 <- 270)11:07
pitti    got: 40+0: i-4011:07
pitti    * i386: systemd-sysv11:07
pittiit's not very verbose, but I take it it's complaining about it being uninstallable11:08
pittixnox: or perhaps it didn't yet take the new sysvinit into account (that's not yet in _output, as mysql-5.6 is still running)11:08
xnoxwait and see i guess.11:10
=== Wellark_ is now known as Wellark
cjwatsonpitti: that'll mean it's uninstallable in isolation11:40
cjwatsonwhen considering release pocket + whatever set of uploads it's considering at that point11:40
pitticjwatson: "in isolation" means "just on a debootstrapped env", i. e. just essential+required packages?11:42
xnoxi thought it was just recursive depends11:42
pittianyway, seems it's happy now11:42
pittihttp://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt doesn't have systemd-sysv any more, so it was just lagging behind the sysvinit publication11:42
pittifinal: systemd,sysvinit,trust-store11:43
pittiall good now and accepted11:43
xnox\o/11:46
cjwatsonpitti: this is pure dependency analysis, no debootstrap or anything like that11:49
cjwatsonpitti: take release's Packages files, take the set of sources that are being considered, remove all old entries corresponding to those sources, add new entries ditto, see if anything new is uninstallable11:50
cjwatsonor strictly, see if the uninstallable count increases11:50
mptev, do you remember where the usb-creator a.k.a. Startup Disk Creator design spec is? I’ve gone through a dozen wiki pages and haven’t found it12:01
mpthttps://wiki.ubuntu.com/USBInstallationImages has early Ascii art, but I remember doing proper sketches afterward :)12:02
=== pete-woods is now known as pete-woods|lunch
=== MacSlow|lunch is now known as MacSlow
evmpt: not sure - I'll dig through my email12:25
tvossdoko, build-deps are patched out of trust-store12:30
dokotvoss, great! any update on net-cpp?12:32
tvossdoko, next on my list12:32
tvossdoko, but likely not today12:32
xubincs #ubuntu-mir12:34
=== _salem is now known as salem_
=== pete-woods|lunch is now known as pete-woods
jdstrandpitti: hey, is there a wiki page on how to change init systems? (testing the ufw change)13:33
jdstrandxnox: ^13:33
jdstrandwiki page or similar13:33
pittijdstrand: boot with init=/bin/systemd in the grub shell13:34
pittijdstrand: there's https://wiki.ubuntu.com/systemd with that too (but slightly out of date)13:35
xnoxjdstrand: init=/lib/systemd/systemd (init=/bin/systemd requires new enough initramfs, thus may fail)13:36
xnoxon kernel boot params.13:36
pittiright, init=/lib/systemd/systemd has worked since utopic's early days13:36
jdstrandI have an up to date utopic vm13:38
pittijdstrand: yep, that'll do; once booted, check with "systemctl" that you get a lot of output, instead of an error message that you can't connect13:40
pittijdstrand: in particular, "systemctl status ufw"13:41
pittijdstrand: should look something like http://paste.ubuntu.com/7979840/13:41
tedgNow that I'm using libmir-client I can't build on powerpc or ppc64el. What's the recommend way of showing that?13:42
tedgJust list the arch in the control file?13:42
tedgCan I say "everything that mir builds on" ?13:43
pittitedg: kind of, if you build-dep on mir, it will just fail to build on ppc*13:43
pittiwhich is fine13:43
xnoxtedg: no need for arch restrictions, just build-deps on things you need. FTBFS are not blockers, regressions in FTBFS across arches are but can be cleaned up by archive admins when intentional.13:48
jdstrandpitti: oh heh, I guess I need to install systemd :)13:48
xnox(e.g. previous version build everywhere, new one depends on libmir-client and hence need binaries removed to migrate in britney)13:48
pittijdstrand: it's pretty hard to not have it installed..13:49
xnoxjdstrand: =)13:49
pittitedg: right, we need to remove the ppc binaries of the previous version, then it'll propagate13:49
pittijdstrand: is that a minimal server-like VM or so? on a desktop it's a very firm dependency (since saucy)13:49
jdstrandpitti: no, full desktop. been upgrading for a while. I was guessing it didn't have it installed-- it didn't come up. still diagnosing13:51
pittijdstrand: verify in /proc/cmdline?13:51
jdstrandI didn't get that far13:52
jdstrandpitti: I'm just updating /etc/default/grub to have:13:52
jdstrandGRUB_CMDLINE_LINUX_DEFAULT="init=/lib/systemd/systemd"13:52
jdstrandthen doing 'sudo update-grub'13:53
tedgpitti, xnox, okay, cool. So how do I request the binaries to be deleted?13:53
jdstrandhuh, weird, it came up this time13:53
pittijdstrand: hm, that sounds right; what happened then?13:53
jdstrandI did install systemd-ui before rebooting13:53
pittijdstrand: ah, I never tried -ui13:54
jdstrandpitti: I missed the early boot-- it was just a blck screen13:54
jdstrandwhen I looked13:54
xnoxtedg: you don't have to, unless it's already stuck in proposed-migration.13:54
jdstrandmight be unrelated and a kvm issue13:54
xnoxtedg: what packages are you talking about?13:54
tedgxnox, pay-service, my issue is that ci-train doesn't think it's built, figure it's checking the same thing.13:54
jdstrandpitti: it appears to be working :)13:54
jdstrandpitti, xnox: thanks for your help!13:55
xnoxtedg: i don't know what ci-train looks for. I'm only talking about general ubuntu archive rules.13:55
jdstrandpitti: I need to do a little more testing, then will just sponsor your changes13:55
jdstrandpitti: thanks again for your work on this!13:56
xnoxrobru:  sil2100: pay-service will be regressing in supported architectures due to new build-dependencies, does ci-train handle that sensibly? or is something needed to be done there.13:56
pittijdstrand: great, you're welcome!13:56
sil2100xnox: you mean, it will now not build for some architectures?13:57
xnoxsil2100: correct.13:57
xnoxsil2100: well, it will rather FTBFS on some.13:57
sil2100xnox: so, sadly CI Train might have some problems with that, as long as the archive still provides binaries for the packages in the archive13:58
xnoxsil2100: well, dep-wait on libmirclient-dev.13:58
sil2100But if those will be dropped from the archive, CI Train will work ok13:58
xnoxsil2100: horum.13:58
sil2100I guess we could also just ignore the failing architectures, which is possible but a bit dirrrty, as build failures will have to be ignored13:59
xnoxtedg: so you need an archive-admin to review the diff and request removal of binaries by an archive admin.13:59
* tedg looks for an archive admin to delete powerpc14:00
tedg;-)14:00
pittitedg: i. e. all powerpc binaries of the "mir" source?14:00
pittiah no, that already doesn't exist on ppc14:01
tedgHeh, no I was joking about the architecture14:01
pittitedg: what does "I" expand to?14:01
pittiin "Now that I'm using libmir-client I can't build on powerpc"14:01
tedgProbably easiest way to see the diff is the one that is generated by the PPA: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-00814:01
tedgpitti, pay-service source package14:02
tedgpitti, Generates libpay* pay-service14:02
pittiunity-scope-click and unity8 also build-dep on libpay2-dev, so these would need to go as well14:03
pittiah, they don't build on ppc, so that's fine14:03
tedgAre they there? I believe the click scope is not.14:03
tedgNot having Mir on PPC is resulting a large set of packages not building there.14:04
pittitedg: so to be completely clear, you are going to upload a new pay-service source which is going to build-dep on Mir?14:06
pittibecause, if you are not going to, then the next upload of pay-service will just bring them back14:07
tedgpitti, Yes, I'd like to publish silo 8, which depends on mir.14:07
pittitedg: usually I'd only remove the binaries from utopic if there was a new version waiting in -proposed, but I suppose the CI airline doesn't work that way14:07
tedgYou can rest assured that *I* won't upload anything :-)14:08
* pitti turns the LP crank, go faster!14:08
pittitedg, sil2100: go wild: http://paste.ubuntu.com/7980028/14:09
tedgpitti, awesome, thanks pitti14:09
sil2100\o/14:09
cjwatsontedg: Can the package in question build without mir?  If so, the correct answer is mir [amd64 arm64 armhf i386].  If not, removing the binaries was correct.14:57
tedgcjwatson, No, it needs Mir now.14:57
cjwatsonRight.14:57
tedgThe trusted prompt support specifically.14:57
bdmurraypitti: is there something wrong with ddebs.ubuntu.com? all the -security pockets seem empty15:01
bdmurrayhttp://ddebs.ubuntu.com/dists/saucy-security/15:02
xnoxbdmurray: better that than publishing them before USN notices & unembargo go out15:10
pittibdmurray: yes, I removed them as they were horribly outdated15:10
pittibdmurray: they aren't really needed either15:10
pittipeople were complaining about hash sum mismatches, etc.15:10
pitti-security is just a subset of -updates these days anyway; I haven't yet taught ddeb-retriever about building -security properly (that'll again double all the special cases of emulating soyuz)15:11
bdmurraypitti: some notification about that would have been nice as the retracers have been failing since that happened15:12
pittibdmurray: oh, if they use them, how did they not fail before?15:12
pittias their Packages could never be authenticated through the Release files15:12
pittibdmurray: anyway, it's probably possible to build them properly, but it'll again take a while to sort that out15:13
pittijdstrand: traincon? ufw is on the touch images?15:14
bdmurraypitti: http://pastebin.ubuntu.com/7980482/15:14
jdstrandpitti: yes, it is and its part of the smoke image tests15:14
pittibdmurray: argh, sorry about that15:14
slangasekpitti: do you have a handle on this, or do you want me to look into it?  I'm not sure if it just needs recreating the files?15:15
pittislangasek: we can copy the old (and broken) files, if that helps any15:15
pittibut frankly it might be just better to drop the -security ddeb apt sources from the retracers15:15
bdmurraypitti: well I can just disable the -security pocket in the sources.list files for the retracers but it would have helped to do that before they were removed15:15
pittibdmurray: well, want me to put the old ones back?15:16
pittiwell, that still wouldn't update the Release files, but apparently that didn't stop the retracers15:16
bdmurraypitti: how long would it take to put the old ones back?15:17
slangasekpitti: I think the point here is that ddebs.u.c, for all its bodgery, is a critical dependency of errors.u.c; so if you need to make changes to it, however innocuous, could you please keep bdmurray in the loop so he doesn't have to root-cause from the other end?15:17
pittislangasek: ack, will do next time15:17
pittibdmurray: a few minutes15:18
bdmurraypitti: that is probably faster than submitting an RT and finding a webop15:18
pittibdmurray: done15:19
bdmurraypitti: thanks!15:19
bdmurraypitti: so I should still remove -security though?15:19
pittibdmurray: yeah, please; they are useless and not updated anyway15:20
dokobarry, any update on the MIR fron?15:20
dokofront even15:20
barrydoko: not much.  been busy with other things. :(15:20
=== psivaa is now known as psivaa-afk-bbl
rbasakI'm flipping some binary packages from being generated by mysql-5.5 to being generated by mysql.5.6.15:37
rbasakJust to check, I don't need to do anything special, do I, except to upload both to migrate from utopic-proposed together?15:38
rbasakThey should be ABI compatible, and will have a higher packaging version number.15:38
rbasakThis will create some component mismatches too, but I presume that this will just be flipped afterwards by archive admins?15:39
rbasak(to demote 5.5 and promote 5.6)15:39
pittirbasak: binary overrides won't change due to an upload, so yes, it'll appear in component-mismatches afterwards16:00
pittiwould be nice to fix mysql-5.6's autopkgtest if it wants to become the new default version (5.5's work)16:00
rbasakpitti: yeah, I get dep8 failures here. I've asked upstream to take a look.16:14
=== chiluk` is now known as chiluk
rbasakinfinity: there's some confusion about HWE EOL prompts before 14.04.1 is recommended by update-manager, which I think is fair.16:19
rbasakinfinity: users are asking if the block is for some reason that might mean they should delay forcing an update.16:19
rbasakinfinity: http://irclogs.ubuntu.com/2014/08/07/%23ubuntu-server.html#t15:2416:19
rbasakinfinity: do we have an answer for that question please, until 14.04.1 is recommended?16:20
=== arrrghhhAWAY is now known as arrrghhh
=== charles_ is now known as charles
hallynzul: so yeah.  feh!  systemd-shim is now being used by libvirtd, and not quite doing the right thing bc it only expected logind to talk to it17:38
=== roadmr is now known as roadmr_afk
=== desrt_ is now known as desrt
=== chuck_ is now known as zul_
hallynSUCCESS19:23
hallynzul: I'm going to push http://paste.ubuntu.com/7982242/ to utopic to fix standard migration;  then I will probably push http://paste.ubuntu.com/7982241/ to support migration from precise, as well as push Alex Bligh's qemu patch (also required)19:28
=== roadmr_afk is now known as roadmr
Elv1313Hello, we have build issues on our PPA since a few day because of an upstream package change that seem to conflict with its own dependencies. Can someone take a quick look and revert that chnage? https://launchpadlibrarian.net/181543532/buildlog_ubuntu-utopic-amd64.sflphone-daemon_1.4.1-rc20140804~ppa1~utopic_FAILEDTOBUILD.txt.gz19:42
=== timrc is now known as timrc-afk
hallynjust running qrt first...20:05
=== salem_ is now known as _salem
hallynxnox: :(23:40
hallyn(#server)23:40
sarnold:(23:40
xnoxi've kept -devel, -desktop (cause seb128 is funny when he is trolling), -dmb, -installer, -motu, -uk23:41
xnoxand #upstart, #lxcontainers and #cgmanager.23:41
xnoxundecided about a few others.23:41
xnoxhallyn: sarnold: such is life =)23:42
hallyn\o see you at plumbers :)23:43
xnoxhallyn: yeah, need to register and organize travel there.23:43
sarnoldxnox: heh, if you're at plumbers and at debconf, I might actually see you more -after- you've left ...23:43
xnoxsarnold: *giggle* true23:44
hallynzul: ok, if you could take a look at the qemu and libvirt pkgs in ppa:serge-hallyn/virt, each has 2 extra packages in it, i intend to push them to utopic tomorrow23:44
sarnold(i'm not actually doing much for debconf, but some of the debian folks interested in apparmor want to get together, sounds like fun...)23:44
hallynnever been to debconf myself23:44
xnoxsarnold: oh, that does sound like fun!23:44
xnoxsarnold: hallyn: this one looks busy. Best one was in Nicaragua. Excellent location, cheap bear, 8 talks per day across 2 rooms, breakfast / lunch / dinner / hack labs, 1 full day sightseeing day trip.23:46
xnoxit's very laidback.23:46
* hallyn jealous23:47
shadeslayer_xnox: did you have a look at the ubiquity bug I pinged you about?23:47
xnoxshadeslayer_: nope, been testing/fighting to push out 12.04.5. Let me look at it.23:48
* shadeslayer_ is pondering about coming to conf week23:48
xnoxshadeslayer_: conf week? DebConf that is?23:49
shadeslayer_no, the large conference week at Düsseldorf23:51
shadeslayer_with linux plumbers conf and couple of other ones23:53
stgraberI know quite a few Ubuntu folks will be there, we'll be a bunch from Foundations and there will be all the kernel team too I think23:54
shadeslayer_Embedded Linux Conference was the one I was thinking about23:55
shadeslayer_there's also Cloud Open Europe and the KVM forum that happens the same week23:56
shadeslayer_mental ^_^23:56
shadeslayer_xnox: re ubiquity, its something to do with privilege escalation23:58
xnoxshadeslayer_: hm.23:59
shadeslayer_it complains about not being able to run commands as a different user, and works fine when using sudo, which sounds like a privilege escalation issue to me23:59

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!