[05:51] hi [08:20] Morning hikiko [08:21] morning flexiondotorg [08:23] hey flexiondotorg, hikiko [08:23] Morning didrocks [08:25] hi didrocks [08:38] morning all [08:40] morning Sweetshark [08:41] :)p [08:55] morning all [08:59] hey hey [08:59] ahoy Laney [08:59] Beautiful day today [09:01] hey willcooke [09:01] it is! === davmor2_ is now known as davmor2 [09:04] https://youtu.be/4OESp5Sts4s?t=14s [09:04] :) [09:04] * Laney used to think that was "drinks and gravy in the park" [09:04] ahh, bisto [09:05] Morning all [09:05] hi davmor2 [09:06] good morning Sweetshark willcooke Laney [09:06] Laney, seb128, willcooke: was there a new version of software updater pushed out seems the details view is only one line thick again [09:06] I want to buy an attachment for the hoover for hard floors. Dyson want 35 quid for a bit of plastic. True story. [09:06] davmor2, what distro serie? [09:07] willcooke, they know how to make money :-/ [09:07] willcooke: 3d printer! [09:07] seb128: zesty it was fixed to 200px iirc with flexiondotorg patch that barry landed but now it is down to one line again [09:08] davmor2, dpkg -l | grep aptdaemon? [09:08] davmor2, that version is stil in zesty-proposed it seems [09:08] did out test out of proposed before and reinstalled without that pocket? [09:09] libsecret autopkg issues blocking it [09:10] seb128: ah I might of tested the fixed package that might be the reason [09:10] which probably has nothing to do with the change [09:10] I pinged barry about it [09:10] libsecret isn't blocking aptdaemon, it's aptdaemon itself [09:10] good morning everyone! [09:11] hey seb128! [09:11] hey desrt! [09:11] hi laney :) [09:11] hey Laney :-) [09:11] what's up [09:11] ups [09:11] seb128: hi! [09:11] I hit the wrong line [09:11] Laney is right [09:12] Laney: not much. looking forward to a relatively relaxing day. think i will stay home tonight, even [09:12] still probably not due to the change [09:12] but some look real issue, maybe packagekit version changed or bindings did [09:12] barry should sort it out... [09:22] desrt: nice [09:22] we're supposed to be singing at the folk club tonight [09:27] Laney has a busy life :-) [09:27] pub quiz yesterday, climbing and singing tonight [09:27] yep [09:28] won't be doing very much for the rest of the week though (H) [09:28] Laney, seb128 - what are your thoughts on this bug: https://bugs.launchpad.net/oem-priority/+bug/1667222 [09:28] Ubuntu bug 1667222 in OEM Priority Project "set password in "user accounts panel' won't remove user from "nopasswdlogin" in all cases" [Critical,Confirmed] [09:28] low priority ^ [09:30] willcooke, it seems a bit annoying ... it's somewhat similar/has to do with the one Laney fixed no? [09:30] I'm sure there are some other problems there [09:30] or at least the "user not removed from nopasswd" [09:30] did we SRU that fix to xenial btw? [09:30] the group thing is okay I think, but switching between the modes is a bit ropey [09:30] yes [09:30] great [09:31] might be waiting verification still [09:31] * seb128 looks [09:31] yeah, still in proposed [09:31] so you think that setting a new password should indeed remove the user from the nopasswd group? [09:32] well if you have a password then you are not *no*passwd... [09:32] I think the testcase I gave in the SRU bug that I fixed is fixed [09:32] I don't claim that all bugs around nopasswdlogin are gone [09:32] right [09:33] that's what is writing on the SRU bug comments [09:33] yeah, the original problem is fixed [09:33] they made it verification-done and opened the other one as a follow up [09:33] like, [09:33] but the appdata one failed verification [09:33] be in nopasswdlogin, open u-c-c, weird shit happens [09:33] which is what is blocking the SRU :-/ [09:33] I can believe that [09:35] seb128, while we're waiting to see what is going on with livepatch, would you be able to see if we can get that sorted out? ^ [09:35] well maybe Robert can have a look at his fix and decide if it should go in anyway or if he can fix it up or if he needs to revert it [09:36] robert_ancell, ^ [09:36] woah [09:36] * Laney isn't used to that guy being on IRC when he is [09:36] yeah, I was surprised to see his nick on the channel :p [09:37] willcooke, "that" being the stucked SRU or the new problem? [09:38] seb128, I guess both really :) [09:38] willcooke, but yeah, I'm going to have a look to that new one, it's not an option used often I think and not really a security issue but it could mislead users to think their account is locked when it's not which is not nice [09:38] let's see what robert_ancell says about the SRU [09:38] thanks chaps [09:38] thx [09:39] yw! [09:39] I think I worked out whats wrong with my printer [09:39] seb128: all those code paths are a bit annoying btw [09:39] there is a [09:39] lemme know if you want me to review your fix :-/ [09:39] Laney, thanks, reviews are always welcome! [09:39] also I think upstream just removed the nopasswd stuff [09:39] * robert_ancell forgot to close XChat... [09:39] lol [09:39] run [09:40] night robert_ancell [09:40] :-) [09:40] *an electric motor inside which seems to drive some belt connected to the big transfer drum. It seems to be jammed and getting very hot, so I think it's just broken. A new part > cost of a new printer. [09:40] night robert_ancell :) [09:41] willcooke: can you get it out? [09:41] Laney, what was my fix? [09:41] appdata for u-c-c [09:42] the bug got verification-failed [09:42] Laney, no, it's really annoying - the mounting seems to be moulded in and so getting just the motor out would break the plastic supporting the rest of the gubbins. [09:42] oh, I was looking at the nopasswd bug [09:42] heh [09:43] night all! [09:46] willcooke: just buy an Air Ram mk2 if you need to vac ceilings get a multi too and be done ;) [10:12] * Sweetshark tried kvm/virt-manager and feels stupid about having VirtualBox around still. [10:23] Sweetshark, what's wrong with virtualbox? [10:23] Sweetshark, oh btw how did that ppa upload from yesterday go? build is done? [10:26] seb128: 5.2.6/yakkety is smoketested and moved to the libreoffice ppa, 5.3.1/zesty is uploaded and currently building in a personal ppa. 5.3.1 snap was also updated, succeeded on amd64, broke on armhf due to infra, broke on i386 in a test, but maybe a heisenbug. retried armhf, i386. [10:27] k [10:27] Sweetshark, let me know if/when you need sponsoring for something [10:30] seb128: willdo. [10:31] seb128: compared to vbox, kvm is just much smoother in desktop/video emulation for me for starters. installs were also really quick, so IO might be better too. but that might also just be my new hardware. [10:31] I should give it a try again [10:34] seb128: also thorsten from #libreoffice-dev showed me he can boot into his native windows dual-boot install from the running linux, which is quite cool. [10:35] that sounds useful [10:36] seb128: He said the only tricky thing is not to miss the grub menu, because booting into the linux install that you are currently running causes some confusion between host and guest. [10:37] "Yo dawg, We heard you like linux, so we put your linux in your linux ..." === hikiko is now known as hikiko|ln === dandrader is now known as dandrader|afk === hikiko|ln is now known as hikiko === dandrader|afk is now known as dandrader [13:43] Laney: So I have a silo building (failing on PPC) that has UAL without Upstart, but it's still blocked on s390x because of other deps. [13:44] Laney: Do you want my to try to get those in the same silo or do you want to do "archive magic" to do rebuilds later? [13:44] tedg, what silo? =) can i look? [13:45] xnox: No, it's for Laney only. [13:45] xnox: https://bileto.ubuntu.com/#/ticket/2577 [13:45] hi tedg [13:45] * xnox is *sad* [13:45] Nice one [13:47] tedg: It comes down to u-s-s? [13:47] Laney: Yeah, I've pinged kenvandine about that one and he's looking into it. [13:47] tedg, powerpc should be dropped as an architecture anyway; and s390x should not have ual as we don't support clicks/desktop there at all [13:48] Stop it please [13:48] tedg: Thanks, ok, sounds good [13:48] You can upload the things into the silo to see if they work [13:48] xnox: Well, I'm not going to argue those points. But from an academic perspective, my stuff shouldn't break there either. :-) [13:48] on my todo list for this morning [13:49] tedg, right, imho then powerpc/s390x builds should be libertine-less as we don't support graphical desktops on those arches anymore. [13:49] xnox: Also, FWIW, this silo drops click support as well ;-) [13:49] or at least we have for sure not ported libertine, nor tested it on powerpc & s390x [13:49] xnox: We don't add arch-specific hacks all over the archive [13:50] Can you please leave it? [13:50] We're close to getting this fixed now [13:50] For the record, if someone wanted to get me an s390x rack I could use as a desktop, I'd be happy to dogfood it :-) [13:50] lol, ok. [13:50] *happily* :-D [13:50] Laney, ok [13:51] thx [13:52] tedg: kenvandine: I'll be off tomorrow and Friday, so if you could check that you can get gtk built against this then please feel free to publish [13:52] sure [13:52] after uss is fixed I would hope that libertine and then ual should get builds automatically [13:52] assuming there's no other problem [13:54] tedg, does this silo also fix the issues renatu was talking about with app starting when dbus-user-session is installed? [13:54] kenvandine: Yes [13:55] woot [13:55] wow \o/ [13:55] kenvandine: When it's not installed though, right? [13:55] application starting is sort of important :) [13:55] the issue was when dbus-user-session was installed [13:56] tedg, yes after install dbus-user-session the apps does not start from content-hub anymore. [13:56] Okay, perhaps I'm confused then. It fixes the issue of dbus-user-session not being installed, which ual needs. [13:56] yeah, that's not the issue i was talking about [13:56] tedg, did you remember that you told me that the env vars could be wrong [13:56] Yeah, but that's a content-hub issue. [13:57] tedg, but only happens after install dbus-user-session [13:57] Yes, it happens when content-hub is using systemd instead of upstart. [13:57] Upstart has the environment of the full graphic session because it is once per session, where systemd is once per user. [13:58] humm I tought that was transparent for content-hub [13:58] So we have to include more environment. [13:58] to be clear, content-hub is using UAL [13:58] we don't use systemd directly [13:58] That environment is getting taken from the launching process. [13:58] or is it the exec-tool? [13:58] Which *should* be the same for everyone. But somehow must be different. [13:59] kenvandine: Wait, are we talking "applicaiton launching" or "untrusted helper launching" ? [13:59] i think it's both [13:59] renatu, right? [13:59] renatu, it's an issue with both importing and exporting? [13:59] kenvandine, let me try. I am not sure [14:00] Laney, https://code.launchpad.net/~ken-vandine/ubuntu-system-settings/no_upstart/+merge/319926 [14:00] Laney, perhaps you can review that [14:01] the upstart build depends was synthetic, looks like Mirv added that just to keep it from building on those arches [14:04] kenvandine, tedg: works to pick a file (dbus-user-session) [14:04] ok [14:04] so the untrusted helper launching does work [14:04] kenvandine, ted: but does not work to export [14:04] it's just the app launching that doesn't [14:05] kenvandine, what is the diff? [14:05] export just uses UAL to launch the app [14:05] importing uses the untrusted helper to setup a trust session and launch the app [14:05] it's a little different [14:05] so the trust session works, so woot [14:05] So this branch changes that :-) [14:05] simple app launching doesn't... [14:06] It makes the helpers use systemd as well. [14:06] tedg, so i wonder if you are breaking content-hub :) [14:06] i'll test that with the silo [14:06] Cool. [14:06] tedg, i'm going to build uss in the silo, want me to rebuild ual too? [14:06] kenvandine: Give me two minutes... [14:07] ok [14:07] tedg, i'll let you kick the build tne [14:07] then [14:07] K, trying to find a fix for these powerpc failures. [14:24] willcooke: do you know when the new slide artworks is likely to land? Still got the Yak on installer slides [14:24] davmor2, I just branched the code [14:24] willcooke: awesome thanks [14:25] in anticipation of the graphics hitting my inbox in the next few mins [14:25] davmor2, should be the next hour or so [14:25] willcooke: so should be in tomorrows iso right? [14:26] davmor2, depends on how charitable xnox is feeling :) [14:27] willcooke: between you saying please and me shouting DO IT NOW DAMN YOU! should be good ;) [14:27] willcooke, did you vote remain? [14:27] xnox: did you? [14:27] of course [14:27] * willcooke is wearing a Union Jack as a shirt and drinking lager at 2:30pm [14:28] willcooke, with or without scottish flag union jack? [14:28] * willcooke <- not a xenophobe [14:28] * xnox wants dragon on top of union jack [14:29] willcooke: I think you'll find that is a Union Flag unless it is on a ship :P [14:29] davmor2, damn it, I nearly put "flag" [14:29] I knew someone would point that out :) [14:29] xnox: don't give George another target damn you I like dragons [14:30] willcooke: I like the title pedant :D [14:34] xnox: https://www.youtube.com/watch?v=86MO-JWTNjo I think that is Mays official response right? [14:37] more like this: https://www.facebook.com/999BABF/posts/1417849411607460:0 [14:44] kenvandine: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/2577/+build/12129615 [14:44] kenvandine: Can we drop U1 auth as well? [14:44] kenvandine: I think it doesn't work... [14:45] tedg, in uss? [14:48] kenvandine: I thought all the u1auth stuff was now macaroons. [14:48] i haven't looked at any of that [14:48] I might be confused, but I thought dobey was trying to move out of libu1auth everywhere. [14:48] i think we only use it in the updates panel [14:48] Yeah, you need to talk to snapd for that now. [14:48] right [14:48] Well and apt. [14:49] so all of that can be removed for now [14:49] jgdx, ^^ [14:49] tedg, what about system image? [14:49] kenvandine: snapd [14:49] There is no system image [14:49] snapd ate it. [14:49] right, i'm thinking we just remove the updates panel completely [14:50] for now [14:50] Well, I imagine you'll want the same QML, just a different model. [14:50] yeah, but until we have time to port that [14:50] we can just disable building it [14:50] The updates made it look really nice, would hate to lose that :-) [14:50] kenvandine, yes to that [14:51] porting that panel isn't on our backlog [14:51] we want to use some qml bits from it in the apt-based one, but other than that… [14:51] but it should be [14:53] do what [14:54] kenvandine: yeah you should just get rid of the updates panel entirely now [14:54] i'll prepare a branch that does that [14:56] i don't think we want to spend time writing a bunch of code to have apt updates in u8 system settings do we? [14:56] dobey, i don't think we know right now [14:57] current update-manager should work under unity8 i think; might need a little tweaking to ensure it's getting started in that session though [14:57] whatever does the popping up of it and such, that is [15:16] ouch. $600+ for 64GB DDR4 :-/ [15:24] tedg: you dont have to care about ppc; it's going away and britney already ignores it [15:26] Laney: I think I fixed it :-) [15:26] good, if it indicated a real bug :P === dandrader is now known as dandrader|afk [15:46] Laney, willcooke: can we just make gtk+ not build the mir back-end on s390x? [15:47] "can" yes, "should" no [15:47] these upstart dependencies have to be sorted out anyway [15:48] yes, but support mir/unity8 stuff on s390x is an oxymoron [15:48] support upstart into the future becomes one too [15:49] i'm not talking about supporting upstart [15:49] adding technical debt into the gtk+3.0 package doesn't seem like fun either [15:49] doing what you want just kicks the can down the road [15:49] you are going to have to unwind the upstart stuff at some point [15:50] upstart is irrelevant to what i'm saying [15:51] it is the reason you are asking me to insert a hack into the gtk3 package [15:51] no [15:51] what is causing the problem then [15:51] the reason i'm asking is because mir/unity8 on certain architectures makes absolutely no sense [15:52] that's not how the archive works [15:53] what you're finding out now is that adding a pile of hacks to work around an upstart bug on s390x is a painful experience to undo [15:54] and I am to blame too, because I added the one to ubuntu-system-settings [15:54] no, what i'm finding out is that we care more about building on architectures that don't make sense, than being efficient [15:56] You're wrong, and you don't want to listen to what I'm saying. [15:57] I'll upload ubuntuone-credentials. [15:58] upload it to what? [15:58] no i am not wrong. and you don't want to listen to what i'm saying. [15:59] why would you upload ubuntuone-credentials? [15:59] we want to remove it from the archive, not add more architectures to it [15:59] it needs to die [16:00] Sounds good to me, feel free to make its dependencies not be depdencies [16:01] system-settings is getting rid of it per previous discussion [16:05] imho we actually do need gtk+3.0 on s390x with support for xvfb and remote-forwarded X11 server [16:05] xnox: yeah, but mir? [16:05] but we are not commiting to support it; or run either unity7 or 8 on it. [16:05] dobey, but mir is build/available on s390x [16:06] built [16:06] yeah, but it's a pointless build [16:06] ideally one wants similar gtk3+ to other arches, to lower any maintaince costs [16:06] that [16:07] and why not having it if it's not extra cost? it might be useful to some people some day [16:07] it is extra cost [16:07] which one? [16:08] having s390x builders to idle a bit less? [16:08] how many developers have s390x or ppc64 boxes sitting on there desk? where's the hardware to run our CI on them? [16:09] in the datacenter [16:09] as for other archs? [16:09] no [16:09] dobey, most of our mainframe is not utilised; and actually s390x works as a great canary as it finishes builds before anything else most of the time. [16:09] not launchpad builders, CI for developers [16:09] dobey, add a worker to your jenkinsaas environment? [16:10] dobey, you are speaking for the mir team there? [16:10] where are my jenkins nodes that are s390x and ppc64 [16:10] dobey, ppc64 are in scalingstack & s390x are coming. [16:10] why do i have to care if my software keeps building on s390x, when it will NEVER be run on that architecture? [16:10] dobey, e.g. juju teams; hwe; server teams are all running slaves in their jenkins on both ppc64 and s390x [16:11] dobey, because we have reverse-dependcy chains to build (eventually) source packages that are useful in their command-line equivalent. [16:11] for example we had to build qt, to build cmake, to build a bunch of command-line utilities and server daemons that use cmake. [16:11] ditto with gtk+3.0, we need it because there are useful command-line tools that link with gtk+3.0 [16:11] yeah, but choosing to add a new dependency to something that doesn't build on an archtecture where you normally build, doesn't mean the thing you depend on needs to change. it means you need to make it optional [16:11] ditto e.g. mir -> it's in the stack for u-a-l which can be used to launch s390x snaps from command line [16:12] so when is the mir team going to provide remote terminal support for mir so i can run it as a thin client? or forwarding of mir over ssh? [16:12] dobey, we've tried that, but the dependency cycles are weird and instead of redoing the chains of optional builds; it's cheaper and easier to just build everything and use the parts that are useful. [16:13] dobey, and acutally e.g. hwe; certification; java; toolchain teams have been asking for unity7 or unity8 support on s390x for java test-suite validation, etc. [16:13] such that e.g. if mir forwarding over ssh works; we may would actually want to have that supported on s390x for niche certification stuff. [16:14] so far i have been pushing back on that. [16:14] because we are clearly not in the position to support that with trivial amount of cost. but we should not be slashing our wrists over it as well by making parts of the archive not build at all, artificially [16:39] desrt, hey! Would you be able take a look at these MPs and sanity check for GMenuModel? [16:40] https://code.launchpad.net/~aacid/qtubuntu/aboutToShow/+merge/319080 [16:40] https://code.launchpad.net/~aacid/qmenumodel/aboutToShow/+merge/319078 [16:41] willcooke: sure. [16:42] desrt, no immediate hurry, merci === dandrader|afk is now known as dandrader [17:57] davmor2, xnox - squeek squeek, the mouse is here! [17:58] https://code.launchpad.net/~willcooke/ubiquity-slideshow-ubuntu/ubiquity-slideshow-ubuntu-zesty/+merge/319962 [18:01] BYEEEEEEEEEEEEEEEEEEEEEEEEEEEEE PALS [18:02] have a nice weekend Laney [18:02] thx, will do! [18:02] * Laney is going to be So Mindful === dandrader is now known as dandrader|afk [18:22] Enjoy Laney [18:22] willcooke: \o/ [18:23] have a nice evening desktopers, Laney, enjoy the long w.e! [18:24] I'm off too, night all === dandrader|afk is now known as dandrader [18:47] libreoffice 5.3.1~rc2 compiled and build successfully to private ppa, running autopkgtests in a local VM right now. [18:49] Sweetshark: do you have good visibility into how many items the clusterfuzzer has found, how many cpu hours they've spent fuzzing, etc? [18:49] (in a shell on desktop, not exactly an adt-testbed, but it should be good enough) [18:51] sarnold: I should be able to find out items found, cpu hours spent would require me to do some digging (if that is even knowable outside GOOG) [18:52] Sweetshark: ah, don't bother then; it's an idle curiosity, not worth spending time on it [18:52] Sweetshark: thanks [19:23] subsequentcheck and manual test succeeded for 5.3.1 build === ayan is now known as Guest18912