[00:00] bah...forgot to init submodules [00:15] iBelieve: seems to have built ok...but running err's on https://pastebin.canonical.com/122949/ [00:28] anyone here have any experience with the nexus 5 [00:33] dcbartlett: I've got a bit of experience. I haven't been using it for long though. [00:35] were you able to make phone calls with it? [00:36] Yup [00:36] hum, mine isn't making any calls with tmobile [00:36] I've only been called though. Where did it go wrong? [00:38] i dial the number and hit the green call button [00:38] it sits there for a bit [00:38] And then nothing? [00:38] and then it says "call failed \n calling" [00:38] and then drops back to the dialpad [00:38] kgunn: I don’t have access to the pastebin [00:39] Well... It's midnight over here so I can't help you any further. I can't call people really. [00:41] ok === _salem is now known as salem_ === duflu_ is now known as duflu === salem_ is now known as _salem [02:14] hey will https://wiki.ubuntu.com/Touch/Devices/i9505 for the galaxy s4 [02:15] work on a SGH - 1337m [02:15] android version 4.4.7 [02:15] android version 4.4.2 sorry [02:15] baseband version 1337MVLUFLN12 [02:24] Hello [02:25] any body here? [02:25] am asking about the recommended phone device for ubuntu tocu [02:25] any body could help? [05:25] https://www.youtube.com/watch?v=cmoGanNI-fw [05:25] one step closer to multiboot [05:29] if I could only read french [05:46] peppe le phew === ecloud is now known as ecloud_wfh [06:29] is there a reason this doesn't work? ... [06:29] http://cdimage.ubuntu.com/ubuntu-touch-preview/daily-preinstalled/current/quantal-preinstalled-phablet-armhf.zip [06:29] i was trying to install ubuntu touch on a tf101 [06:30] !devices | jrg [06:30] jrg: You can find the full list of devices, official images, community images, and works in progress at https://wiki.ubuntu.com/Touch/Devices [06:30] https://wiki.ubuntu.com/Touch/Devices/tf101#link_to_images [06:30] that is the page directing me to the broken link above [06:31] lotuspsychje: yah. that page directs me to the wiki page i listed above which directs me to a site which does't have the file necessary for installation [06:32] i was just curious if this is simply unsupported... it says "work in progress" but the link to get the second file is broken [06:34] i'm guessing that quantel was replaced with saucy? [06:35] http://cdimage.ubuntu.com/ubuntu-touch-preview/daily-preinstalled/current/saucy-preinstalled-phablet-armhf.zip [06:36] that file is there which looks similar but i am not sure if the kernel works [06:36] maybe check the XDA forums mate [06:36] what brand is your device? [06:36] asus tf101/transfrmer [06:36] transformer [06:36] just figured it would be safer to use the ubuntu wiki vs xda [06:36] since there is a device page for it [06:37] i guess i can try it and if it breaks i can just re-install and recover android from recovery [06:37] i dug this thing out of a drawer after being in there for a year and a half anyways [06:38] seems like that page needs to be updated tho heh [06:41] well i only ested for nexus7 so dont know for your asus [06:43] np heh [06:45] ah well... let me try it out heh [07:05] well... that didn't work heh [07:05] it just went into a reboot loop [07:06] think the XDA forums is your best bet [07:06] other devices then nexus need to be ported [07:06] yeah they had other methods but i was hoping to use the actual offical ubuntu wiki page methods [07:06] and i only found methods for full ubuntu installs and those didn't work out too well [07:07] i bought me a nexus7 specially for touch [07:07] because its most supported devices [07:07] lotuspsychje: yah but if you try to ebay a nexus they are hyperinflated [07:07] because it supports ubuntu ;) [07:07] i'm not paying $300 for a nexus4 [07:07] well the 2013 wifi nexus cost 269 [07:07] so its not a bad deal [07:08] and BQ ubuntu phones are comming soon also [07:08] see what i mean? in reality if that thing only ran android it would probably cost $100 or less [07:08] yah [07:08] i'll wait on that [07:08] they are supposed to have a dock as well [07:08] aren't they? [07:08] not sure how it will turn out [07:08] i wouldn't mind that [07:08] but im sure it will influence the world on ubuntu touch [07:08] other companys will follow [07:10] if it's even anything close to maemo it is win ;) [07:10] maemo was great [07:11] i miss maemo.. probably the closest the world ever came to a pocket linux phonelike device :/ [07:22] hey, I just ran through the process installing on my nexus 10 [07:22] it reboots to ubuntu loading screen, spins for a while, then reboots to android loading screen and sticks [07:22] same as http://askubuntu.com/questions/540244/while-intalling-ubuntu-on-nexus-10-recovery-rom-boots-back-into-android-not-ubun [07:23] no solution was provided [07:23] I ran udbflash --bootstrap=true --channel="trusty" [07:23] 2015/01/08 02:18:21 Rebooting into recovery to flash [07:23] I'm not really getting anywhere with this [07:23] any thoughts? [07:26] Nirgali42: I find sometimes it's more reliable holding the power button till it's fully off, and then starting in fastboot mode (turn on while holding both vol up and down) [07:29] okay but in fastboot mode what next? [07:29] I already went through the flash process... I thought [07:30] duflu: how do I make it boot to ubuntu? [07:30] I can get to recovery mode, which is ubuntu recovery [07:30] but not sure where to go from there [07:30] Nirgali42: The flash process will work more reliably if you start in fastboot. If it has worked at all, it will enter Ubuntu itself [07:31] udbflash --bootstrap=true --channel="trusty" [07:31] is what I ran [07:31] from fastboot [07:31] it does Rebooting into recovery to flash [07:32] then reboots to ubuntu spin loading screen [07:32] Nirgali42: Oh yeah, I think there are some USB cable reliability issues that make it mysteriously less reliable on some devices (Nexus 10, Nexus 7). Try a different cable or port [07:32] sits there for a while, then reboots again to android [07:33] Nirgali42: Does the recovery screen show device is "unlocked"? [07:33] yes [07:33] Done pushing /Users/dmart/.cache/ubuntuimages/ubuntu-touch/trusty/manta/version-303.tar.xz.asc to device [07:33] et cetera [07:33] it does all that, but I try now agian with better cable [07:34] Nirgali42: Don't know then. It sounds like your cabling is fine [07:35] okay it goes "2015/01/08 02:35:15 Rebooting into recovery to flash" [07:35] reboots [07:35] shows spinning ubuntu [07:36] always have bad luck [07:37] yup, same thing [07:37] bed time [08:06] good morning === ara is now known as Guest15321 === dholbach_ is now known as dholbach [09:34] Good morning all; happy Joy Germ Day! :-D [09:34] JamesTait, do you know enough about the rnr server to answer setup questions about it? rgallina isn't online [09:34] and nor is noodles [09:35] aquarius, I've never installed it, but I might still be able to help. fgallina won't be around for about 3 hours, and noodles is in Oz now, so won't be around for about 10. [09:37] JamesTait, http://bazaar.launchpad.net/~rnr-developers/rnr-server/trunk/view/head:/README says that I should enable the PPA and then install rnr-server-{developer-,}dependencies. There is no rnr-server-developer-dependencies afaict, and the PPA (https://launchpad.net/~ubuntuone/+archive/ubuntu/rnr-dependencies) only contains precise files? [09:38] ah, no, I lie, there is a developer dependencies deb [09:38] but they're all only for precise. Are you deployed on precise still? [09:38] aquarius, AFAIK, rnr is still running on Precise (because it's the old Reviews service from the old Software Centre). [09:38] k [09:38] So, LXC ftw. :) [09:39] I think the debs in the PPA are just virtual packages [09:39] so I might see if it runs on 14.04 [09:40] hm, many dependencies [09:40] maybe I do need an lxc after all :) [09:42] aquarius, that tends to be the way we do it. I currently have CPI and PKGME running on my host, and U1 Servers and SCA running in LXCs, plus an LXC for Snappy stuff, all on my laptop. [09:42] It keeps me nice and warm. ;) [09:43] I bet :) [09:46] JamesTait, can I create an i386 lxc container on an amd64 machine? [09:47] aquarius, I'm not 100% sure, but I think the answer is no. [09:47] hm [09:47] so: the packages in the rnr PPA are i386. [09:47] my machine is amd64. [09:47] That might not be a problem with multiarch. [09:48] aaaaah [09:48] that'd be the answer then [09:48] It might need soe config, I'm not sure, but I know I have i386 packages installed on my amd64 machine. [09:48] right, creating an amd64 lxc :) [09:52] I love how https://help.ubuntu.com/lts/serverguide/lxc.html doesn't actually explain how to start the lxc container :) [09:52] aquarius: that's left as an exercise for the reader. :) [09:53] aquarius, you got it though, right? [09:53] no. :( [09:53] $ lxc-start -n rnr [09:53] lxc_container: call to cgmanager_create_sync failed: invalid request [09:53] lxc_container: Failed to create hugetlb:rnr [09:53] lxc_container: Error creating cgroup hugetlb:rnr [09:53] lxc_container: failed creating cgroups [09:53] lxc_container: failed to spawn 'rnr' [09:53] lxc_container: The container failed to start. [09:53] lxc_container: Additional information can be obtained by setting the --logfile and --logpriority options. [09:53] aquarius, sudo lxc-start -n $CONTAINER_NAME [09:54] aquarius, I usually start it detatched (-d) and ssh into it, but that's entirely optional. [09:54] problem is https://github.com/lxc/lxc/issues/206 afaict, but stgraber said he fixed it on 14.04 9 months ago :( [09:55] stgraber, ping. :) [10:13] stgraber, unping. Got it to work after reading a bunch of bugs and mailig list posts. [10:17] Hello [10:18] JamesTait, am I thick? what's the user account in my newly created lxc? [10:19] aquarius, I think it depends how you created it. Did you use the option to copy your host user and bind mount the home? [10:19] nope [10:19] https://help.ubuntu.com/lts/serverguide/lxc.html doesn't mention anything about that. [10:19] unless I missed something obvious. [10:19] I just created an unprivileged contanier. [10:20] Is Ubuntu Touch compatible with Onda v989 ? [10:21] !devices | iMhack [10:21] iMhack: You can find the full list of devices, official images, community images, and works in progress at https://wiki.ubuntu.com/Touch/Devices [10:21] aquarius, You can add it after the fact, if that's what you want to do. [10:21] JamesTait, that's what I don't know how to do :) [10:21] aquarius, I just need to remember exacty how - there are a couple of lines to add to the lxc config, IIRC. [10:22] aquarius, just checking for you, 1 mo. [10:23] !devices [10:23] You can find the full list of devices, official images, community images, and works in progress at https://wiki.ubuntu.com/Touch/Devices [10:25] aquarius, /var/lib/lxc/rnr/fstab needs to contain something like: ```/home/aquarius home/aquarius none bind 0 0``` and then you just need to copy your passwd line into /var/lib/lxc/rnr/rootfs/etc/passwd and the same for /etc/group I think. [10:30] Onda v989 is not in the list. But it have an Allwinner A80T processor, so no UEFI problem. [10:32] aquarius, any luck? Also, I noticed the "-a amd64" parameter in the comments in /var/lib/lxc/foo/config so maybe i386 is possible. [10:35] Is there a place to ask for some ubuntu touch port ? [10:36] aquarius, possibly with `lxc-create -n rnr -- -a i386` [10:38] iMhack: the XDA forums might be able to help [10:39] lotuspsychje: Thanks, I'll check it. [10:49] JamesTait, cor, that's a faff and no mistake [10:49] I would expect lxc-attach to give me a shell in the container but it doesn't work :( [10:50] aquarius, what error do you get? [10:51] $ lxc-attach -n rnr [10:51] lxc_container: call to cgmanager_move_pid_abs_sync failed: invalid request [10:51] lxc_container: Failed to enter group /aquarius/rnr [10:51] lxc_container: error communicating with child process === vrruiz_ is now known as rvr [10:51] aquarius, WFM with sudo (I get a root shell), but that's not helpful to you. ;) [10:52] * aquarius grins [10:52] you've set up privileged containers [10:52] I hate doing things with sudo. So I have user container [10:53] aquarius, ah, well, there's your problem - too smart for your own good. :-P [10:56] JamesTait, heh :) [10:57] maybe I can get stgraber's attention and ask for help; he'll know :) [10:57] bzoltan: hi! About bug 1393697 [10:58] bug 1393697 in click (Ubuntu) "Cross qmake to the chroots" [High,Confirmed] https://launchpad.net/bugs/1393697 [10:58] bzoltan: all seems to work fine, but the "lupdate" tools won't run: [10:58] mardy: what fails there? [10:59] bzoltan: lupdate: could not find a Qt installation of '' [10:59] bzoltan: am I missing some package in the chroot? [10:59] bzoltan: If I run "QT_SELECT=5 lupdate" in the chroot, then I see the normal output of the tool [11:00] mardy: most likely the intltool ... why do not you create a fresh chroot? [11:00] bzoltan: it's a fresh chroot [11:00] mardy: 15.04? [11:00] bzoltan: yep, from the PPA (but on Trusty) [11:00] mardy: the click is the same on T [11:01] bzoltan: if I comment out the lupdate target from the qmake file, then the project builds fine [11:05] zbenjamin: ^ can you say something? [11:07] bzoltan: something :) [11:07] zbenjamin: hkhm ... thank you, thank you a lot [11:07] bzoltan: lupdate is the qt specific translation tool [11:08] bzoltan: seems like the qt_chooser fails [11:12] mardy: if your run lupdate without te QT_SELECT set i suppose you get the same? [11:13] zbenjamin: yep [11:14] zbenjamin: qttools-dev5-tools is installed in the chroot (the amd64 version not the armhf one) === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === _salem is now known as salem_ === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [12:18] bzoltan: in a qml touch app, is there anywhere you can get the application's id from? [12:18] bzoltan: the misnamed "applicationName" has the package name, not the application name nor the application id [12:22] Chipaca: you set it in the MainView [12:22] kalikiana_: no, you set applicationName, not the application id [12:22] kalikiana_: unless i've got something wrong :) [12:22] at do you mean by that exactly? [12:22] Chipaca: The APP_ID variable is obsolete for long time [12:23] https://wiki.ubuntu.com/AppStore/Interfaces/ApplicationId [12:23] the application id; either package_appname or package_version_appname would do [12:24] Chipaca: why do you need that? [12:24] bzoltan: for push [12:24] mmmmm overloading terms is not helpful; APP_ID never included the version [12:25] i might've put version in the wrong place, above :) anyway [12:25] Chipaca: until now every single use case involving an id for an app needed APP_ID, if anything else was needed apparmor, mir or whatever would figure out the rest if needed [12:26] is it really intended to change this? [12:26] i'm not sure what you mean [12:27] it's the first time there seems to be a need to know that string in an app [12:27] i don't know what's in APP_ID; i guess because it was obsolete nobody mentioned it when we asked about application ids way back [12:27] that I'm aware of anyway [12:28] kalikiana_: right now a developer needs to include it by hand in the push qml thing, but it's errorprone (and devs don't really understand them) so i'd rather avoid that if we have it already elsewhere [12:29] Chipaca: why the app_id is required for push? Who is doing the other side? I guess they can tell where to get the ID what they need. [12:30] bzoltan: which "other side"? [12:30] bzoltan: you mean the c++ side? [12:30] Chipaca: do you have some docs about that? [12:30] Chipaca: for example yes ... I mean whoever said that you need an ID should be able ti tell you where to get it from... I guess :) [12:32] zbenjamin: yes. It got moved, give me a minute (ralsina_ is digging) [12:32] https://developer.ubuntu.com/en/apps/platform/guides/push-notifications-client-guide/ [12:33] zbenjamin: search for Ubuntu.PushNotifications in there [12:33] ralsina_: i had it, i was just searching wrong (moar coffee maybe?) [12:33] ralsina_: thanks :) [12:33] :-) [12:34] bzoltan: um. I own the whole stack of this, and I don't know where to get it from from within qml, which is why i'm asking you [12:34] bzoltan: the point being, if it already is in qml, i don't need to go digging through click to figure it out === MacSlow is now known as MacSlow|lunch [12:34] but even digging through click would be better than asking the user to type it out, as we do now [12:35] in fact I'd suggest you do something similar for MainView.applicationName too [12:35] it's just asking for pain, forcing the user to specify the same information over and over [12:36] even some of our finest minds, and aquarius, get these things out of sync -- and then things almost-but-not-quite work, and debugging is a complete pain, and there's practically no warnings about it anywhere AFAICT [12:37] it's nice to be in the same sentence as fine minds. [12:38] I can only add my (considerable) weight to Chipaca's point here, bzoltan -- if i have to specify an appid (whether com.canonical.foo or com.canonical.foo_Foo for multi-app click packages), all that can happen is that I get it wrong and then things subtly weirdly don't work. FInd out some way to derive that information and do it in the OS: don't make me type in a thing which you ought to already know. [12:40] https://developer.ubuntu.com/en/apps/platform/guides/push-notifications-client-guide/ mentions appId: "com.ubuntu.developer.push.hello_hello" that's really just applicationName with a "_hello" added [12:40] kalikiana_: yes, that's ${packagename}_${appname} [12:40] aquarius: just to be clear ... _I_ do not make you type anything [12:40] * aquarius grins [12:40] bzoltan: not even MainView.applicationName? [12:40] bzoltan: i figured you own that one :) [12:41] * Chipaca is probably wrong, and used to it [12:41] bzoltan, that's synecdoche, unless it's syllepsis. You're the SDK; you can hit people on the head until they do things right :) [12:41] Chipaca: I am the right dude, but I do not set the rules for push [12:42] aquarius: if applicationName can ever be automatic it'll be after we have mandatory systemd or upstart everywhere we have UITK [12:42] bzoltan: right. push is me, and I'm going to make the appid optional one way or another. But the reasons why i'm going to do that apply as well to applicationName :) [12:43] kalikiana_, we do. You may say "no no you can run uitk apps on the desktop", to which I will say: that ability is meaningless, because you can't *give* someone a uitk app to run on the desktop, because click packages won't install on the desktop. [12:43] aquarius: they do :-) [12:43] only if you've got a mir desktop :) [12:43] no [12:43] I got them on unity7 [12:44] brb, do-system-upgrade over ssh is now telling me to restart the server [12:44] um. I was told in no uncertain terms that the security of a click package required Mir, and until that happened, no click packages on the desktop for anybody [12:44] is that incorrect, then? Can I now tell people that they can install my apps on their Ubuntu 14.04 desktop? [12:45] aquarius: the security drawbacks have not changed [12:45] but apps run, our uitk is aweseome [12:45] kalikiana_, how does an ordinary Ubuntu 14.04 desktop user install things from the click store? [12:46] I'm more than happy to add those instructions to my website. [12:46] they don't [12:46] * Chipaca back [12:46] did i miss anything? :) [12:47] kalikiana_, hence my point. They don't. So we have upstart running click packages everywhere meaningful there are click packages. :) [12:48] bzoltan: anyway, i guess the conclusion about the application id is "nope, no way"? [12:48] bzoltan: that is, there isn't any easy way from within qml to get the application id [12:49] aquarius: are you suggesting screw non-unity8 and therefore everyone running a stable de? [12:50] that seems a little bit premature at this point [12:50] Chipaca: please correct me if I am wrong, but there is not such as app id in QML level .. we use the Mainview.applicationName and it is $appname.$developer [12:50] Chipaca: and yes, that is set by the application and yes it should be in sync with the manifest [12:51] bzoltan: i *think* that was my question to you :) [12:51] Chipaca: there is no id outside of a click running within upstart - by extension it would be impossible to qmlscene foo.qml without setting it [12:52] kalikiana_, I am suggesting that you already *have* screwed those people! I can't install click packaged apps on my stable Ubuntu 14.04 desktop. Therefore, some super-techie-only way that it's possible for Ubuntu core developers to do that anyway should not be an important thing blocking the SDK working out appId for me. [12:52] Chipaca: Just I am a bit confused :) the application name is something what the developer sets .. it is not something what the store or the sdk or the shell assignes to the app... you the developer must set it [12:52] kalikiana_: when you say "there is no id", what do you mean? Because there clearly is, as per the wiki page [12:52] Chipaca: So I wonder if it makes sense to have an API to read a property what is actually set by the app [12:53] bzoltan: ah, MainView.applicationName does not have to be the package name? [12:53] Chipaca: the wiki pages describes a specification. same as a dbus service name. it won't exist until it is implemented somewhere :-) [12:53] Chipaca: not anymore [12:53] bzoltan: ah, ok then [12:54] Chipaca: cool :) Sorry for the confusion [12:55] kalikiana_: AFAIK click, url dispatcher, indicator-messages and push at least all use this specification [12:55] the indicator a little indirectly tho [13:00] Chipaca: yes. *use*. think where it's coming from. [13:01] kalikiana_: it's coming from the click package, if i understand your question [13:01] kalikiana_: also i forgot to mention ubuntu-app-launch in that list :) [13:01] * Chipaca purposely did not include upstart [13:02] Chipaca: exactly. that is the one I meant by "implemented". [13:02] now think how we run stuff like "qmlscene foo.qml" or "qmltestrunner tst_bar.qml" all over the place [13:03] kalikiana_: where do we do that? [13:03] Chipaca: development mainly. autopilot. unit tests. [13:03] kalikiana_: development of the platform, or of applications themselves? [13:04] Chipaca: both. === alan_g is now known as alan_g|lunch [13:05] well, you can develop apps by exclusively testing them within a device but that kinda kills convergience [13:07] hmm [13:07] wouldn't qmltestrunner run after the click package has built? [13:07] that is, don't you build the click package, install it, and run the tests? [13:10] oh, for apps, yes. I guess I was accidentally mostly thinking of platform dev then [13:10] toolkit, unity, etc don't run tests in click [13:11] ^^ qml unit tests [13:15] kalikiana_: ok. I think i have a way forward; thank you. [13:17] \o/ === MacSlow|lunch is now known as MacSlow [13:32] zbenjamin: about the lupdate issue, is it fine if I file a bug about it? [13:33] mardy: yeah , i do not know yet how fix it but feel free to do so [13:33] zbenjamin: cool, will do [13:33] zbenjamin: against "click"? === alan_g|lunch is now known as alan_g === oSoMoN_ is now known as oSoMoN [14:04] Chipaca: i'd expect qmltestrunner to be unit tests and run during build. autopilot tests i'd expect to run the built/installed app and something you run after building/installing the click [14:15] dobey: that's what *she* said! [14:18] she == nessita? because it sounds like something she would say too :) [14:21] hi all, trying to use mirscreencast on my device, but just getting a static image of the apps scope and none of the action. Can you tell me which socket I should point mirscreencast at? Is it /run/mir_socket? [14:22] dobey, does payui have a .desktop file? wondering how it is delivered as click but does not show up as an installed app [14:22] willcooke: yes [14:22] popey, thx. I wonder why I'm not getting any moving pictures then. [14:22] * willcooke tries again [14:22] pmcgowan: yes, but it is not an "app" it's a "pay-ui" [14:23] dobey, how does it prevent being seen as an app? can we do the same for mediaplayer? [14:23] pmcgowan: it's not something that users should launch directly [14:23] right [14:23] pmcgowan: it's a special hook for payment UIs; no you can't use the same method for media player [14:23] ah ok [14:37] What is the up to date "usable" image on Nexus 4 nowadays? [14:38] !devices [14:38] You can find the full list of devices, official images, community images, and works in progress at https://wiki.ubuntu.com/Touch/Devices [14:41] kenvandine, poke about https://code.launchpad.net/~mterry/ubuntu-system-settings/wizard-introspection-rtm/+merge/243852 -- rhuddie would like to see that wind it's way to landing [14:42] ybon: "stable" channel [14:42] more up to date than rtm? [14:43] mterry, we'll take a look [14:43] kenvandine, thanks! [14:51] ybon: stable is rtm [14:51] oh ok, so I guess I'm still up to date, thanks :) [14:51] (I'm running rtm) === salem_ is now known as _salem [15:23] jdstrand: ping === _salem is now known as salem_ [15:34] mardy: hey [15:42] jdstrand: so, I'm playing with the account plugin confinement, and I found an issue which I didn't think of: [15:44] jdstrand: the OAuth-based plugins open a webview, and set a dataLocation of ~/.cache/online-accounts-ui/id-/ [15:44] jdstrand: that's for the cookies [15:45] jdstrand: I wouldn't like to open up the whole ~/.cache/online-accounts-ui/ for the plugins, because they could steal each other's cookies [15:46] jdstrand: is it possible to tell apparmor that mkdir is allowed in that directory, and then any read/write access is allowed in the newly created dir? [15:47] you mean: [15:47] owner @{HOME}/.cache/online-accounts-ui/id-/ w, [15:47] owner @{HOME}/.cache/online-accounts-ui/id-/ rw, [15:47] ? [15:49] the answer is 'yes we can allow mkdir ('w' on a dir/)' and then other access to files in that dir or elsewhere [15:49] ' [15:50] wich version are you guys working so hard on? RTM? [15:50] jdstrand: more like "owner @{HOME}/.cache/online-accounts-ui/ w", because the account plugin must create the "id-" directory in there [15:51] jdstrand: but the problem is that only the account plugin knows the value of "" in "id-" [15:51] mardy: apparmor works differently than you expect here [15:52] mardy: with apparmor, you give write access to the actual directory being created, not the parent [15:53] Tassadar: hi. looks like hammerhead builds are behind now. build failling maybe? [15:53] mardy: DAC (ie traditional unix permissions) handles the parent dir [15:53] mandel, did you make progress on that updates pause fix? [15:53] morning [15:53] mardy: 'r' would be needed on the parent though [15:54] mardy: so we could do: [15:54] owner @{HOME}/.cache/online-accounts-ui/ r, [15:54] owner @{HOME}/.cache/online-accounts-ui/id- rw, [15:54] anyone else get the ubuntu start spinning for long time immediately after flash+reboot, then to have it finally reboot to permanent android boot logo? [15:54] jdstrand: if we knew what is, right? :-) [15:55] mardy: yes [15:55] trying trusty on nexus 10 [15:55] let me check something actually [15:55] dobey: wrong flavour in config, give it an hour or two, it will catch up [15:55] fixed it just now [15:55] Nirgali42: try using "stable" instead of "trusty" [15:55] Hi, [15:55] One month ago, I installed ubuntu touch from the testing (?) source (to try orientation). [15:55] It was working fine. I have update a few time with apt-get update/dist-upgrade. [15:55] Now, a silly issue appears : [15:55] On my nexus 11, I cannot login because the "password" field appears, misplaced, on top of the screen, [15:55] and the keybord does not show up to enter the 1111. [15:55] Tassadar: thanks [15:55] Any idea? [15:56] dobey: okay [15:56] mardy: I mispoke-- the 'r' on the parent directory is not needed [15:56] mardy: DAC (traditional unix permissions) handles that [15:56] slvn_: reflash and don't use apt to upgrade? :) [15:56] jdstrand: an option could be to let the plugin store the cookies in whatever directory it has access to, and then online-accounts-service (which is unconfined) could copy the whole tree back where it needs it, but it's hacky and assumes many things [15:56] mardy: so just need the one rule [15:57]