[07:12] infinity, bdmurray slangasek we really need to the casper fix in bug 1561302, before our final images spin, can one of you look at that this week? [07:12] bug 1561302 in casper (Ubuntu) "gdm won't allow passwordless login" [Critical,Triaged] https://launchpad.net/bugs/1561302 === med_ is now known as Guest15381 === mapreri_ is now known as mapreri === marlinc_ is now known as marlinc [11:13] darkxst: Have you tested that does reasonable and sane thing? [11:13] s [11:14] infinity, i tested it, by dumping it in place on the live images [11:15] infinity, the flag does not get set because casper adds the user before accountsService starts [11:15] darkxst: I assume this works due to some PAM module kicking in when you run passwd -d? [11:16] infinity, not sure its a PAM module, but accountsservice picks up the passwd -d call [11:16] I wonder if it has an inotify watch on /etc/passwd or something equally disgusting. :P [11:17] Anyhow, looking. [11:19] infinity, likely, everything seems disgusting at this point in the hacks [11:20] including the pam code that detects if a user has no passwd [11:21] darkxst: Mind if I move it to the "if [ -d /root/etc/gdm3 ]" block, so you're the only one that suffers if it proves slightly wonky? :P [11:22] infinity, sure if you want, its a generic bug though, doesnt only affect us (maybe others just havent seen side-effects yet) [11:23] darkxst: Well, it's a subtle change (changing from a blank password to no password), and I'm not positive what the fallout there might be in other bits. [11:24] infinity, ok [11:24] wq [11:25] darkxst: http://paste.ubuntu.com/15891954/ <-- Okay having your name on this version? [11:26] infinity, ack [11:31] darkxst: TBF, accountservice is gloriously buggy if it needs to be running to process passwd *changes* to set flags. [11:32] darkxst: But what could be happening is that it only sets "it's null, yo" on accounts where the password is actually :: [11:32] darkxst: And we set the password to the crypt() equivalent of "blank". [11:32] darkxst: Which is quite different. [11:33] darkxst: Might be a simple patch in accountservice somewhere to mangle its startup parser to look for the crypted version of blank as well. [11:33] darkxst: Anyhow, casper hack is in for now. [11:34] infinity, k thanks, I am glouriously buggy also retreating to bed in attempt to sleep off flu [11:37] darkxst: Ick. No good. Get better! [11:40] sgclark: http://launchpadlibrarian.net/254213242/kcachegrind_4%3A15.12.3-0ubuntu1_4%3A15.12.3-0ubuntu2.diff.gz [11:40] sgclark: Another diff for you to pull into git. [11:40] infinity: yeah just heard, recently woke up, trying to injest some coffee [11:41] thats what happens when folks ignore VCS entries... [11:41] sgclark: Heh. Well, there it is all bundled up for you, I already uploaded. :P [11:41] thanks [11:41] sgclark: To be fair, ignoring VCS entries is entirely the right thing for us to do when we can't commit to the VCS in question. [11:41] fair enough. but we need to be informed [11:42] * infinity mutters something about "always debdiff before you upload". [11:42] anyway, we are moving back to launchpad, next release [11:42] * sgclark mutters huh? [12:02] infinity: pushed. thanks bunches. [12:10] Please approve the above xubuntu-default-settings. It sets our default Libreoffice theme, which is very much a part of our desktop experience. :) [12:10] * stgraber looks [12:11] bluesabre: translation update intentional? [12:11] it's not listed in the changelog [12:11] ah [12:12] meant to list that as well [12:12] ok, that's fine then [12:12] just wanted to make sure that those aren't broken/bad translations you had around on your machine :) [12:12] I've only uploaded bad translation once, and swore to never do it again :) [12:12] +s [12:13] :) [12:13] thanks stgraber [12:16] stgraber: I'll add some thanks too :) === s8321414_ is now known as s8321414 [12:22] any idea why ktp-kded-integration-module doesn't migrate? [12:23] trying: ktp-kded-integration-module [12:23] skipped: ktp-kded-integration-module (3 <- 3) [12:23] got: 139+0: a-41:a-15:a-16:i-15:p-19:p-18:s-15 [12:23] * amd64: kde-telepathy-integration-module-dbg [12:26] Looks like the dbg just needs to be removed. I'll do that. [12:26] ok [12:27] this one might be good to fix, used as a b-d for a bunch of other packages: https://launchpadlibrarian.net/253777080/buildlog_ubuntu-xenial-amd64.pkg-info-el_0.6-1_BUILDING.txt.gz [12:28] doko: Any idea what's up with the pip vs virtualenv autopkgtests? [12:28] Or does that need barry? [12:29] infinity, no, however there's a rc issue for unstable today [12:31] Oh, statsmodels and pandas are in a build-dep loop. Can probably bootstrap that on s390x. After I've beered with stgraber. [12:31] pandas just migrated [12:31] Yeah, but without s390x. [12:32] Just aiming for more coverage. [12:34] There's also the libsemanage that no one ever bothered to see through. [12:34] Probably too late to care now. [12:37] infinity, according to mdeslaur they don't want to look at updating the policy packages [12:38] so what we need to do is to build libsemange in -proposed after release to get the ruby2.3 module [12:43] doko: No need to do it after release. I can do it in a PPA, delete libsemanage from proposed, copy PPA version in, let it migrate, then copy the old one back. [12:43] works for me [12:43] doko: I'll put that on the post-beer TODO. [13:41] infinity, You online? I have an release critical issue with a potential fix. I need someone to sanity check the fix. [13:45] Hi, I just downloaded the Kubuntu ISO, burned it, and it did not start plasma. However, this ISO did boot. I can get a login prompt and type kubuntu and sudo poweroff just fine. It is just that plasma did not load. Is there some larger issue that is already known? [14:16] mparillo: No idea, perhaps ask yofel? [14:16] flexiondotorg: Point me at the issue/fix? [14:18] infinity: TY. I already asked in kubuntu-devel. I guess he is not on. I will try again with tomorrow's image. [14:18] infinity: qatracker stuff all done [14:19] stgraber: Guess I should uncron. [14:19] * infinity does that. [14:20] infinity: yeah, then update debian-cd and trigger a first spin so folks can do some smoketesting on "rc" images [14:20] stgraber: yep. [14:27] infinity, https://bugs.launchpad.net/ubuntu-mate/+bug/1571242 [14:27] Launchpad bug 1571242 in ubuntu-mate "Ubutnu MATE 16.04 Grub Install Fails in UEFI Mode" [Undecided,New] [14:27] The work around is to uninstall grub-pc from the live session before starting the installer. [14:29] flexiondotorg: That seems incorrect. It's in all the live tasks. If it's only failing for you, something weird is going on. [14:29] I've checked the other seeds, all flavours unclusing Ubuntu proper, and none have grub-pc in ship-live but Ubuntu MATE does. [14:29] *including [14:29] (base)adconrad@nosferatu:~$ apt-cache show grub-pc | grep ^Task [14:29] Task: ubuntu-live, kubuntu-live, edubuntu-live, xubuntu-live, mythbuntu-live, lubuntu-live, ubuntustudio-live, ubuntustudio-dvd-live, ubuntu-gnome-live, ubuntukylin-live, ubuntu-mate-live [14:29] infinity, What puts grub-pc in the live task? [14:30] Because in Ubuntu MATE live seed it is expressly included. [14:30] But not in the others. [14:30] grub2? [14:30] And a recommends of ubiquity. [14:31] So, you don't follow recommends, right? [14:31] Having it in your live task would be correct to fix that. [14:31] Correct, but nither does Lubuntu. [14:31] But not in ship-live. [14:31] lubuntu might be broken differently from you. :P [14:32] infinity, If I remember correctly it was added to live by cjwatson sometime back to fix a similar EFI issue. [14:32] OK, so removing from ship-live is OK? [14:32] But not actually going to fix anything? [14:32] Oh, no. lubuntu/live does follow recommends. [14:32] OK, but as far as I can tell Ubuntu MATE should too? [14:33] Should what? [14:33] Because I don't have anything in the seeds to say no-recommends for live. [14:33] Yes you do. [14:33] * Feature: no-follow-recommends [14:33] Shit [14:33] OK. Sec... [14:34] OK, that is the issue then. [14:34] Nah. [14:34] Shouldn't be. [14:34] That's why you include it explicitly, though. [14:34] I should follow recommends in live. [14:34] But you can drop it from ship-live. [14:34] OK. [14:34] If dropping it from ship-live magically fixes the bug, yay. But if not, we can investigate WTF is going wrong for you. [14:34] So the issue, if I'm reading it right is EFI install fails on EFI machine when the HD is zeroed. [14:35] Yeah. Easy enough to test. [14:35] But, if the HD has an existing legacy install on it, the install works. [14:35] I don't have a suitable machine. [14:35] But I know a man who does. [14:35] I'll get more details from them. [14:36] But, by comparision Ubuntu and Lubuntu are said to work under those condidtions. [14:38] And I do need to not follow recommends in the live seed, other I get most of Unity along for the ride. [14:42] Could someone please help us get a couple of changes in for ubiquity and ubiquity-slideshow-ubuntu. They are outdated for us, completely, and both merge requests were from before Final Freeze. Unfortunately we could be done sooner with the changes, since they were the last in a series of others. [14:42] Bug [14:43] 1569005 [14:43] Ahh, Bug 1569005 [14:43] bug 1569005 in ubiquity-slideshow-ubuntu (Ubuntu) "New ubiquity slideshow for Ubuntu Studio [UI Freeze Exception]" [Undecided,New] https://launchpad.net/bugs/1569005 [14:43] Bug 1568981 [14:43] bug 1568981 in ubiquity (Ubuntu) "new Ubuntu Studio wallpaper for ubiquity installer [UI FREEZE EXCEPTION]" [Undecided,New] https://launchpad.net/bugs/1568981 [14:43] infinity, OK, here is a to the point description of the problem. [14:43] infinity, "So, installing ubuntu MATE in UEFI mode kicks off an error at the end of the install because the grub-efi-amd64-signed package is not installed on the install image. And to install that grub-pc must be removed. The problem seems to only be appearing on ubuntu mate. Lubuntu and Ubuntu are fine." [14:46] flexiondotorg: Okay, well your manifest for installed packages matches ubuntu. [14:46] flexiondotorg: manifest-remove doesn't match, which is curious in itself. [14:50] infinity, Is this my weird no-recommends making a mess of things? [14:51] Ah-ha. [14:51] Yes. [14:51] queuebot and I will be offline for a few minutes, relocating to a new server [14:51] flexiondotorg: You're not pulling bootloaders recommended by linux-image-$(uname -r) ... We can hardcode those somewhere probably. [14:52] Maybe. [14:52] Kinda gross, though. [14:52] * infinity ponders. [14:52] infinity, Is this something I should add to the seed or something the build system can do? [14:53] flexiondotorg: In this case, it's not the seed, but the APT_OPTIONS="--yes --no-install-recommends" in livecd-rootfs, I think. [14:54] OK [14:54] Though, lubuntu should have this issue. [14:54] * infinity looks. [14:54] Yeah, lubuntu has the same manifest-remove issue. [14:55] Why do they not has the same bug? [14:55] flexiondotorg: I think I might need a full syslog from that failed install to unravel what's actually happening. English descriptions of what people *think* is happening aren't ideal. [14:55] What logs would you like? [14:56] Just syslog or anything else? [14:56] flexiondotorg: syslog should do. [14:56] Just for Ubuntu MATE? [14:56] For MATE and lubuntu. I see no reason they should differ. [14:57] infinity, For UBuntu MATE, you want the syslog upto the point the error occurs right? [14:58] You do not want to have grub-pc remove before hand, right? [14:58] flexiondotorg: The whole /var/log/syslog from the system, copied after the failure. [14:58] OK [14:59] infinity, Instructions dispatched to the man with the computer. [15:00] I'll got and try to fix a laptop and see if I can prepare some logs too. [15:00] I have a busted EFI laptop. [15:04] infinity, This is the Recommends in linux-image- you're referring to right? [15:04] Recommends: grub-pc | grub-efi-amd64 | grub-efi-ia32 | grub | lilo [15:06] infinity, I added APT_OPTIONS="--yes --no-install-recommends" in livecd-rootfs to also prevent pulling in Recommends. [15:06] Is APT_OPTIONS needed? [15:06] infinity: Any chance I could bother you with the ubiquity merges? [15:07] Just a change of artwork, really. But, it's from 14.04, so.. [15:07] I mean, the current artwork is from 14.04 [15:07] Would be nice to not have to release our ISO with that (though I know we are very, very late, and I have to ask people for favors) [15:08] * zequence hates asking people for favors, and can't spell today [15:16] infinity: ping [15:17] clivejo, don't ping, just write [15:17] Have another problem, with plasma-discover [15:18] basically, it was fixed and in our staging PPA - https://launchpad.net/~kubuntu-ppa/+archive/ubuntu/staging-plasma/ [15:18] plasma-discover (5.5.5-0ubuntu1~ubuntu16.04~ppa2) [15:18] yofel cherrypicked two upstream patches which fix a problem with the updater [15:19] but somewhere along the line, those patches didnt make it into the archive [15:19] The patches in question are 0001-Fix-refresh-on-the-notifier-plasmoid-for-QApt-backen.patch and 0002-Improve-how-we-react-to-failed-parses.patch [15:21] The one publsihed to the archive seems to be misisng them - https://launchpad.net/ubuntu/+source/plasma-discover/5.5.5-0ubuntu1 [15:21] published on the 14th [15:21] flexiondotorg: I'd rather not take shots in the dark, let's wait until we have a syslog to stare at. [15:22] cyphermox: You around to maybe help zequence with his merges? [15:23] clivejo: So get someone to upload a fixed -0ubuntu2 with the midding patches? [15:23] s/midding/missing/ [15:23] infinity: If I knew someone with uploading permission I would love to [15:24] sgclark is currently trying to get some rest [15:24] clivejo: Ahh. I'll try to look in a sec. [15:25] we are all a bit stressed, Im new to this all so I cant be must help on that front. But I know that version in the PPA works and has worked and been tested [15:26] I cant explain or even understand how it was ommited, but the patches arent in the main archive for some reason [15:32] infinity, Understood. Test are being conducted. [15:35] clivejo: Uploaded in yofel's name. [15:35] infinity: thanks, do you know how or why the patches got dropped? [15:36] clivejo: No idea, I didn't do the previous upload. [15:36] do you know who uploaded? [15:37] LP says it was Phil himself [15:37] clivejo, does the package come from the kubuntu staging ppa's? [15:37] clivejo: Phil's name in the changelog, Scarlet's sig. [15:37] they tend to override changes which go directly into the ubuntu archive [15:38] we follow a process, uploaded to staging, then landing, then to archive usually [15:38] clivejo: anyhow, this should fix it: https://launchpad.net/ubuntu/+source/plasma-discover/5.5.5-0ubuntu2 [15:38] but its gone from landing [15:39] infinity: thanks [15:48] infinity, syslogs attached - https://bugs.launchpad.net/ubuntu-mate/+bug/1571242 [15:48] Launchpad bug 1571242 in ubuntu-mate "Ubutnu MATE 16.04 Grub Install Fails in UEFI Mode" [Undecided,New] [15:53] flexiondotorg: Okay, you can see that both of them remove grub-pc before trying to install grub-efi-amd64-signed, so that's a red herring. [15:53] Yep. [15:53] Just looking now. [15:53] flexiondotorg: Figuring out why the grub-efi-amd64-signed install doesn't want to shouldn't be too tough. [15:55] infinity, I'm looking too. [15:55] Perhaps grub-efi-amd64 should be listed in ship-live? [15:55] I mean, grub-efi-amd64-bin [15:56] It's there. [15:56] See http://cdimage.ubuntu.com/ubuntu-mate/daily-live/20160417/xenial-desktop-amd64.list [15:56] Oh. [15:56] That could have just been a version mismatch. [15:56] Can he reproduce with a newer image? :P [15:57] Betting grub-signed and grub were just out of sync on that image build. [15:58] And we're chasing shadows. [15:58] Ha! [15:58] I'll check. [15:59] However, the original report is from the 15th and the tests were done this morning by Mike. [15:59] So maybe not that. [16:02] infinity, Running tests again for Ubuntu MATE using the current daily. [16:03] flexiondotorg: Can confirm that grub in /pool/ and livefs were out of sync for 20160416, they look fine for 20160417 [16:04] flexiondotorg: So that's probably what it was. [16:04] infinity, I really hope so :-) [16:17] infinity: not sure what's fixed what here but I did a rebuild for another reason here and bug 1570901 not apparent now [16:17] bug 1570901 in ubiquity (Ubuntu) "Cd menu not booting to ubiquity try/install menu but always to live session" [Undecided,Confirmed] https://launchpad.net/bugs/1570901 [16:18] flocculant: Spooky. [16:18] yup [16:19] just thought I'd mention something positive ;) [16:26] * infinity rebuilds the world. [16:26] (Don't worry, there's no way these are "final", don't sweat if your last minute stuff isn't in, this is just to get more testing) [16:27] * flocculant doesn't panic 5 days before :) [16:27] 4. [16:28] 3, if you consider that we like to be ready by Wed night. [16:28] :) [16:28] But no pressure. :P [16:28] ha ha ha [16:30] infinity: mmm so spookier - if I choose install from the first menu you could get to - I do land at the live desktop [16:31] That sounds delighfully broken. [16:31] :) [16:31] Might be a cyphermox bug. I think he was rooting around in gfxboot menus last week. [16:31] :) [16:37] infinity, Two different computers test in EFI. All Good! No issue! :-) [16:38] infinity, As always, thanks for your time helping investigate. [16:38] So the images from April 17th are not affected. [16:39] flexiondotorg: Righto. Final images obviously can't have this bug, as we take great care to keep the archive consistent while building during release week. [16:39] flexiondotorg: So go ahead and close the bug. [16:39] * flexiondotorg is closing the bugs now. [16:39] Please approve the above xubuntu-artwork package, fixes an issue where our wireless icons become very difficult to see in the network indicator [16:39] also, hi everyone :) [16:45] infinity, Am I right in saying we should only see release critical bugs being added to the archive from now on? [16:52] flexiondotorg: Ideally removed, not added. [16:54] * infinity is going to have a short jetlag nap. [17:34] pkg-info-el built with fixed emacs24 [18:20] cyphermox: I just got the UEFI secureboot debconf prompt on an upgrade within xenial (not at the time I upgraded to xenial). Will get more info and file a ug [18:20] or a bug [18:21] slangasek, fyi the mok secure boot disable does not appear to be working, i am looking at it [18:21] apw: ok [19:34] rbasak: ping [20:34] I'm not happy about the state of dbus in xenial. you see test related dbus failures in many packages: libnih, libsecret, nux, are the ones I touched today [20:47] infinity, shall we have unity on s390x in xenial? doko wants it to run openjdk tests, and IBM java team do too (so far directed them to use xfce, but doko is resisting) [20:47] https://launchpadlibrarian.net/254258618/nux.debdiff [20:47] heh [20:47] build stuff if we can ... [20:47] infinity, a noopt build guest nux going on s390x, i shall test unity build with that one in a ppa. [20:47] also looking at lz4 on s390x [20:47] which works with noopt again [20:48] doko, well if we have unity & nux, we may as well publish ubuntu-desktop. [20:48] looks like toolchain issues [20:48] lz4 did build on s390x in february -> https://launchpad.net/ubuntu/+source/lz4/0.0~r131-2/+build/9029716 [21:40] how did the cantor autopkg tests get fixed? [22:00] still don't know what to do about the wine mess ... === stgraber_ is now known as stgraber === rharper` is now known as rharper === debfx_ is now known as debfx