[03:45] Good morning [04:10] RAOF, Thanks! [04:18] tedg: No problem. [05:21] good morning [07:06] good morning desktopers [07:11] re seb128 [07:11] lut didrocks ;-) [07:50] happyaron, hey, how is the debugging going on bug #1380982? that's the most reported e.u.c issue on vivid [07:50] bug 1380982 in ibus (Ubuntu) "ibus-ui-gtk3 crashed with SIGABRT in g_assertion_message()" [High,Triaged] https://launchpad.net/bugs/1380982 [08:00] hey hey [08:01] morning Laney [08:01] hey Laney, how are you? [08:03] good thanks! [08:03] it's friiiiday [08:03] you? [08:05] I'm fine, thanks [08:13] ogra_, mvo_, do you know how/where is the includes.chroot livecd-rootfs used? [08:14] seb128, that are files that get copied into the chroot [08:14] before tarring it up [08:15] ogra_, do you know what copy them? would it work for ubuntu-desktop-next as well? [08:15] oh, seems to be lb_chroot_includes [08:15] yeah, some live build script [08:16] thanks [08:16] the current personnal image has no /etc/hosts content [08:16] so I guess we need to copy that dir [08:16] ogra_, mvo_, also you might be able to help with the users/groups hackery in livecd-rootfs [08:17] so the hooks have those /etc/groups|passwd expected content [08:18] does it mean users don't get added the correct way later on? [08:18] or the real issue/questions [08:18] the lightdm postinst is supposed to add the lightdm user with an userdir /var/lib/lightdm, that doesn't seem to happen though [08:18] the directory is missing on the image [08:18] do you have any idea why? [08:19] I guess it's because it thinks the user already exist, which might be due to the passwd hackery? [08:19] seb128: that sounds likely it depends on when exactly the scripts for the chroot run I don't know that without looking at the code [08:22] mvo_, was that a reply for the includes or the user? ;-) [08:23] seb128, hmm, that shouldnt be possible unless the postinst hardcodes a UID that already exists [08:23] seb128: that was for lightdm [08:23] ogra_, [08:23] # creating lightdm user if he isn't already there [08:23] if ! getent passwd lightdm >/dev/null; then [08:23] adduser --system --ingroup lightdm --home /var/lib/lightdm lightdm [08:24] hmm, wasnt there a --crate-home option ? that looks rather like the package ships the dir [08:24] (ICBW) [08:24] ? [08:25] no, the package doesn't ship the dir [08:25] the adduser call creates it [08:25] ah, sorry ... there is a --no-create-home ... [08:25] it creates the dir by default [08:25] right [08:25] (i think useradd was the one weher you need to call --create-home) === JMulholland_ is now known as JMulholland [08:28] seb128, and on the resulting image, do you see lightdm in the passwd file ? [08:29] seems that will only fail if a lightdm user exists already [08:29] s/fail/skip/ [08:29] ogra_, let me check, but I think so, it's in http://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/ubuntu-desktop-next/hooks/00-uid-gid-fix.chroot_early and otherwise the adduser would work [08:30] seb128, you are missing a mkdir in your live-build/ubuntu-desktop-next/hooks/00-uid-gid-fix.chroot_early [08:30] (compare to live-build/ubuntu-touch/hooks/00-uid-gid-fix.chroot_early) [08:31] ? [08:31] ogra_, no mkdir in http://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/ubuntu-core/hooks/00-uid-gid-fix.chroot_early [08:31] (i think when i added that i wrote a long bzr commit to explain that ... ) [08:31] seb128, thats where you need to look into touch ... core has no lightdm indeed [08:32] oh, touch [08:32] pre-added users whose homedirs are created from postinst scripts need to have these dirs added from the script [08:32] gotcha [08:32] ogra_, that's hackish :-/ [08:32] but thanks for the pointer ;-) [08:32] yes, it is ... but there is no way around such getent code [08:33] I don't understand why we need an hardcoded list of users/groups [08:33] at least not with the design we currently have for the passwd file handling [08:33] rather than just getting those from the chroot [08:33] seb128, because packages would dynamically add the IDs during install otherwise [08:33] and that's fine? [08:33] so your readonly passwd file would have changing uids [08:33] the install is in the chroot [08:33] and? [08:34] who cares what the uid is === mvo_ is now known as mvo [08:34] and your existing writable dirs would be owned by the UIDs that were existent when the image was first created [08:34] now if an OTA with a changed passwd file comes the ownership would be all wrong [08:34] I see [08:35] ogra_, thanks [08:35] the two options we have is either to loop over all writable dirs every update ... and potentially miss files ... (re-owning all of it) ... or make sure the passwd file never changes [08:36] * ogra_ would be really grateful if anyone has a less hackish idea for that ... we picked the option of the two that looked less hackish and less dangerous for the user [08:36] (fail at build instead of fail at OTS if something is wrong) [08:37] *OTA [08:41] ogra_, right [08:41] ogra_, mvo, do you know where is the writable-path list for snappy? ;-) [08:41] seb128: ubuntu-core-config [08:41] seb128: do you need to add stuff there? [08:41] mvo, yes, /var/lib/lightdm and /var/lib/lightdm-data [08:42] otherwise lightdm refuses to start [08:42] can't write info about the current sessions, users, etc [08:43] seb128: are you getting close to having it start? [08:43] Laney, it's starting working with the items I mentioned today in the backlog [08:43] like I hope that this livecd-rootfs upload + writable image is enough [08:44] it enough on my vm with manual changes [08:44] nice [08:44] seb128, is the livecd-rootfs side ready for upload (i just pushed my chjange to wily too and see you have a lot pending now) [08:44] ogra_, you pushed your change to wily mean? [08:44] yes [08:45] no cowboy upload right? [08:45] no, still UNRELEASED :) [08:45] good [08:45] I'm about to upload [08:45] I had dch -r locally [08:45] please pull then :) [08:45] let me pull your changes [08:45] good that we talked about it :) [08:45] oh, I'm going to remove 10-remove-documentation.binary as well [08:46] personnal can include manpages ;-) [08:46] yeah [08:46] that's also going to make systemd happier [08:46] * ogra_ would really love if that silly script didnt remove all changelogs ... painful on phones at times ... [08:46] systemd-tmpfile-setup.service is failing because it tries to set up /var/cache/man [08:46] but it likely buys us 500k diskspace or some such :P [08:47] that script is misnamed :p [08:47] echo "I: Removing /var/lib/apt/lists/*" [08:47] that's not documentation! [08:47] I'm going to keep the lines to clean apt dirs [08:47] complain to achiang :P [08:48] (oem team script .. ) [08:49] (even before oem was called pez) [08:49] seb128: haven't looked at it last week... [08:49] * happyaron O_O [08:49] mmm pez [08:52] happyaron, k [08:52] mvo, can I just upload ubuntu-core-config ? ;-) [08:53] seb128: sure [08:53] is there a vcs? [08:56] ogra_, mvo, what does "transition" mean? [08:56] seb128: data is copied from the uderlaying fs iirc [08:56] from the underlying RO parts [08:56] thanks [08:58] mvo, http://paste.ubuntu.com/11739310/ ok ? [09:00] looks fine [09:01] thanks [09:01] hopefully with that & livecd-rootfs we have a personal snappy that gives you a lightdm greeter [09:01] unsure about unity8 yet, I did testing in kvm and mir doesn't like software rendering so unity8 doesn't start [09:01] that's next ;-) [09:02] well, thats all in the homedir ... so at least you dont need to fiddle with writable bits anymore :) [09:02] (most likely at least) [09:04] right [09:30] seb128: heya, could you consider to bump libwps in wily/main to 0.4-4 from debian experimental for me? [09:32] Sweet5hark1, bump === sync? [09:32] hey btw [09:33] that's a transition [09:33] small but still [09:34] Laney, that's a nack? [09:34] no, more an fyi [09:34] k [09:34] if it's a small one I'm sure we can manage :-) [09:35] Sweet5hark1, right? ;-) [09:35] like maybe test build stuff before doing it and/or know that lo is going to come soon [09:35] Sweet5hark1, ^ please comment on that [09:36] seb128: lets just wait for LibreOffice 5.0 final upstream then. [09:37] Sweet5hark1, wfm [09:37] it might still be good to start test building other rdepends [09:37] like abiword&co [09:37] Laney, thanks for pointing it out :-) [09:37] this is not superurgent for me right now. [09:38] np! [09:38] (can of course sync into a PPA) [09:38] yeah, kudos to Laney for being on the watch ;) [09:38] * Laney the archive police [09:39] who watches the watchmen? [09:39] http://www.matiasvangsnes.com/wp-content/plugins/rss-poster/cache/ca4e2_20131210_Mark_Shuttleworth_LeWeb_001_610x416.jpg this guy [09:41] * Sweet5hark1 is feeling safe now. [09:42] lol [09:44] hum, what's going on with proposed migration, it's being slow today? [09:44] oh? [09:44] http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt also lists half the world with nettle, I hope it doesn't mean every include livecd-rootfs is blocked [09:44] there's a log in proposed-migration/log/ [09:45] k, no livecd-rootfs in there [09:45] maybe next one [09:45] it's weird that all the logs in http://people.canonical.com/~ubuntu-archive/proposed-migration/log/wily/2015-06-19/ are 2.7M [09:45] but the most recent one is 278k [09:45] oh, it's moving, ongoing I guess [09:46] it outputs it live [09:46] right [09:46] I guess you can ssh to snakefruit.canonical.com and tail -f it if you care ;-) [09:46] k, so let's go for some errands en lunch, things hopefully are in place for an image retry once I'm back [09:46] seb128: btw, i just walked through all our 'isnt MIRed yet' special casing vs. debian: http://anonscm.debian.org/cgit/pkg-openoffice/libreoffice.git/commit/?h=ubuntu-wily-5.0 -- This upstream release comes with surprisingly little new dependency madness. [09:47] good [09:47] Copying: livecd-rootfs/2.317 [09:48] Copying: ubuntu-core-config/0.6.22 [09:49] seb128: if this works out good we could MIR libgltf, coinmp, libetonyek -- which would leave use with only one difference (bundling liborcus) vs. debian. [09:49] seb128: not super high prio for me though -- lets see how it goes. [10:08] cyphermox: are you (or any chance you could) taking another look at https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/1436330 ? [10:08] Launchpad bug 1436330 in network-manager (Ubuntu Vivid) "Network Manager doesn't set metric for local networks any more, causing connection issues" [Critical,Fix committed] [10:09] it's annoying when moving between wired and wireless [10:09] and https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/1445134 is also annoying :) [10:09] Launchpad bug 1445134 in network-manager (Ubuntu) "Network manager never scanning for new access points" [High,Confirmed] [11:48] qengho, hey, are you looking at the chromium test issues? [11:49] Laney, did the iso.qa.u.c cdimage retry issue got resolved? I just tried to trigger a rebuild for desktop-next but it's not reflecting on the cdimage side [11:49] seb128: should be, I think stgraber said he tested it [11:50] it's a cron which runs every 5 minutes iirc [11:50] k [11:50] let's wait a bit [11:50] oh, yeah, worked ;-) [11:50] Laney, thanks [11:51] its there [11:51] nice! [11:51] go self service :-) [11:56] Sweet5hark, hey [11:56] Sweet5hark, dpm was looking for you ;-) [11:56] seb128: kk [11:56] dpm, ^ [11:56] dpm: Im working on it ;) [11:56] you are hard to find [11:57] Sweet5hark, do you have details about this libreoffice widget/toolkit? how is the rendering work? is it going to work on Mir? [11:57] thanks seb128, Sweet5hark :) [11:59] seb128: libreofficekit just renders into a bitmap. if your toolkit can display a bitmap (and do the offscreen-rendering/scrolling part) it will work with MIR. [11:59] k [12:00] seb128: the example we have in the source is plain gtk. [12:00] Sweet5hark, do you have an url to an example of the api/code? [12:01] seb128: speaking of gtk -- upstream LibreOffice might switch to using gtk3 by default in LibreOffice 5.0 already: http://nabble.documentfoundation.org/minutes-of-ESC-call-td4151949.html [12:02] nice [12:02] I dont think we want to vanguard that, better stay with gtk2 by default for wily IMHO unless it is reported to be exceptionally stable. [12:03] ^^ desrt: FYI [12:03] Sweet5hark, do you think you might be able to finish the example app some time today so that the core app devs can start looking at it? [12:03] seb128: example code is at https://github.com/LibreOffice/core/tree/master/libreofficekit [12:04] dpm: will try too, no guaratuees. But I should have something in the next days. [12:05] Sweet5hark, thanks [12:05] thanks === alan_g is now known as alan_g|lunch [12:16] Sweet5hark: interesting. [12:18] slow backup is slow. [12:19] or rather, I have become jaded and realize now that the 'storage is cheap, backup all the things!' strategy is bottlenecked by bandwidth. === MacSlow is now known as MacSlow| === MacSlow| is now known as MacSlow|lunch [12:40] larsu, Laney, other wily theming issue, in eog the overlay buttons for next/previous/rotation are barely visible [12:40] not sure if you noticed [12:40] indeed [12:40] thanks [12:41] yw! [12:41] want a bug report about that? [12:41] oh yeah, I've not seen those buttons for a while [12:41] on the theme or on eog? [12:41] * Laney usually uses keys [12:41] ye [12:41] start with theme I guess [12:41] k === ara is now known as Guest87687 === MacSlow|lunch is now known as MacSlow [13:18] grrrr, disk going crazy, doing crazy writings and locking up system suddenly [13:18] had to hard reboot [13:18] still the schroot issues? [13:19] don't really know, I still have one remounting, but not the mountloop crazyness === alan_g|lunch is now known as alan_g_ [13:28] hum, I have a bunch of: [13:28] juin 19 15:18:38 tidus nullmailer[2095]: Starting delivery: protocol: smtp host: mail file: 1428584426.15044 [13:28] juin 19 15:18:38 tidus nullmailer[2095]: Starting delivery, 129 message(s) in queue. [13:28] juin 19 15:18:38 tidus nullmailer[3051]: smtp: Failed: Connect failed [13:28] juin 19 15:18:38 tidus nullmailer[2095]: Sending failed: Host not found [13:30] it's anacron spamming me… [13:35] add a GPS app so it can find the host :) [13:57] oh great, libreoffice 5.0 fails with a flaky test on wily it seems :/ [14:12] great weekly update! [14:17] Laney, thanks ;-) [14:17] larsu, Laney, https://bugs.launchpad.net/ubuntu/+source/ubuntu-themes/+bug/1466890 [14:17] Launchpad bug 1466890 in ubuntu-themes (Ubuntu) "eog overlay buttons wrongly themed" [Undecided,New] [14:19] nice one [14:20] qengho, hey, are you around today? [14:24] chrisccoulson, hey, how are you? [14:25] chrisccoulson, question for you, I saw that bug about webbrowser/oxide not working on unity8 desktop because some env variable is not set anymore ... is the proper fix going to be land soon or should we consider setting to env back as a workaround? [14:39] seb128, I haven't managed to find time to look at that yet. It is milestoned on https://launchpad.net/oxide/+milestone/branch-1.9, but not assigned atm [14:39] chrisccoulson, would exporting the env you mentioned be enough to workaround it? [14:40] seb128, it would [14:40] chrisccoulson, seems like we should do it then ;-) [15:02] wooot [15:03] ubuntu personal boots to a lightdm greeter ;-) [15:03] for some reason it only has a "guest session" button and no user entry [15:03] but we are getting there! [15:08] seb128: Yep. What's up? [15:08] seb128: I have a test fix, yes. [15:08] My debdiffs have been strange, though. [15:09] qengho, it was about the test fix ... can we get that in wily? [15:09] how strange? [15:09] seb128, hey, sweet i just had a question about creating a snappy personal image :) (with lightdm) [15:10] seb128, did you just use ubuntu device flash to create the core image? Then push debs to it? [15:10] seb128: just larger than I expected. I will finish pruning it. [15:11] bschaefer, no, we are building an official snappy personal image [15:11] qengho, k [15:11] seb128, well i need to look at building a personal image with session upstart/lightdm/system compositor [15:11] seb128, not really sure how to create a personal image atm (cant find things online? my googling could be poor though) [15:11] bschaefer, why? that's what I'm working on for like 3 weeks ... that looks like duplication [15:12] bregma, ^ what's going on? [15:12] seb128, creating a kiosk snappy image is the goal [15:12] seb128, sooo a bare bone system with mir + qml/qt (SDK) apps [15:12] bschaefer, but you don't want unity8/standard apps? [15:12] which looks to require a dbus session (which we will need to create a session) [15:12] seb128, no unity8! [15:12] but we want the apps [15:12] shrug [15:13] its a bit confusing atm...but what i understand is i just need a dbus session to please the SDK [15:13] seb128, soo thats pretty much what you've made atm? [15:14] bschaefer, no, what I made is similar to desktop-next we had, just based on snappy [15:14] seb128, o right yeah... soo i was looking at doing something like that but much more minimal [15:15] just need the system compositor to run mir, then a session for dbus... from there just regular apps running on it [15:15] bschaefer, but it's not trivial "you can build locally", we added a new project to the launchpad cdimage build and livecd-rootfs, added a new device channel on the s-i server, we updated goget tools to know about this channel, etc [15:15] unsure if you need to go through all the same things for your image [15:16] can't you just make a framework on top of the core image? [15:16] seb128, to be honest im not even sure what i need, i know i need a personal image with a few more deps [15:16] installed into the core [15:16] seb128, not sure how an upstart session could become a framework [15:16] (doesnt it need to be started by lightdm?) [15:17] soo i would need lightdm in the framework? (which would need to be started by the system upstart/systemd) [15:18] bschaefer, I don't think framework allow to do what you need atm [15:18] seb128, yeaah which would mean i would have to install these deps as part of the core image [15:18] you might need to set up a full 3rd "kioske" image similar to core/personal then [15:19] which is what i thought personal image was [15:19] no [15:19] personal is a desktop image [15:19] it includes unity8 [15:19] sooo a personal image is exclusive to what you're doing? [15:20] bregma, might be able to explain a bit more... [15:20] it feels like naming overload [15:20] it's more like desktop-next atm [15:20] i know :( [15:21] I think we want to set up our own "personal basic" image, no Unity 8, no indicators, none of that [15:21] maybe eventually there will be some modular thing but that is not what this is now [15:21] personal ... yeah ^ [15:21] just the compositor, LightDM, and for now upstart for session management [15:21] bregma, I though the idea was to have 2 images, so now we want 3? [15:22] seb128, for now, yes, but this third one is not an official one [15:22] at least, not yet [15:22] k, unsure how you can build it [15:22] the tools are not really made for local hacks [15:22] for personal we had to get the tarballs in the s-i server as an official channel [15:22] yeah, I see that as a problem, but maybe I'm just old fashioned [15:22] there is code tweaking the images there on the server [15:23] also the command line tools don't understand local tarballs [15:23] they just speak to the server [15:23] right [15:23] it would be better if it was otherwise [15:23] yes [15:23] that seems problematic hmm [15:23] I'm just saying that you are going to need to improve the tools [15:24] * bregma looks at golang like a disease === alan_g_ is now known as alan_g === alan_g is now known as alan_g|EOW === dpm is now known as dpm-afk === qengho is now known as CardinalFang === CardinalFang is now known as qengho