[00:00] sin18, so how did you setup such VM for it to do that? [00:00] sergiusens : just installed ubutu on virtualbox as a guest (using ubuntu iso) [00:00] sin18, iirc you needed to manually add usb devices in virtualbox [00:01] sergiusens : i have noticed that vbox is smart where it just shows up under vm - usb devices [00:02] sin18, hmm, didn't work last I used vbox 2 years ago. [00:02] sin18, how's your ram and tmpfs? [00:02] that might be another problem [00:02] sergiusens : 8 GB, 4 of which allocated to VM [00:03] I still feel you should use your native environment [00:03] this is trusty I suppose? [00:03] sergiusens : yes though i tried from saucy too on VM [00:03] sergiusens: btw can i install ubuntu touch on to nexus 4 from fedora ? [00:04] sin18, yes, as I told you above [00:04] sin18, you just need go installed [00:04] adb and fastboot installed [00:16] sergiusens : was jsut surprised because i didnt realize that was an option. Will explore it [00:21] sin18, you can flash from a mac as well [00:22] sin18, you can problably even just use the binary in the deb package from fedora === chriadam|away is now known as chriadam === LarrySteeze|Away is now known as LarrySteeze [00:55] sergiusens : thanks, heading out will try it [02:03] hello? === Ubuntu is now known as Guest52095 [02:05] hello how do you instal ubuntu for nexus7 1 gen? [02:09] Guest52095, You can use multirom from google play to multiboot it with whatever is the nexus now [02:09] on* [02:10] The touch, saucy was the last desktop image [02:10] err raring === vying is now known as Guest1273 === 64MAACJZF is now known as frecel === chihchun_afk is now known as chihchun [05:04] add-apt-repository seems broken on #244, is this known? [05:04] http://pastebin.ubuntu.com/7118003/ [07:53] good morning === chriadam is now known as chriadam|away [08:49] hmm, i'd really like to have the app window that gets shown upon ctrl-r in qtcreator to remember its dimensions and position on the screen [08:50] [it always goes to my second monitor and i always have to resize it manually] [08:50] in a qt c++ app i'd know how to achieve that [08:50] but i'm a complete qml/ubuntu touch noob ;D [08:51] will i have to manually check what kind of device the app is running on and then use code to store/retrieve the dimensions/positions at shutdown/startup? [09:01] also: how to deploy the app to the phone and run it? [09:01] Good morning all; happy Chocolate Caramel Day! :-D [09:02] oh, found the menu entry [09:02] let's see if it does anything === Ursinha is now known as Ursinha-afk [09:04] mandel, ping? [09:05] mhr3, morning! [09:05] mhr3, how can I help :) [09:05] mandel, morning, could you tell me where do i find udm logs? [09:05] hmm, ok [09:06] enabled networking on the device and used the build -> ubuntu touch -> run on device menu entry [09:06] nothing :( [09:06] mhr3, sure, le me double check, will be a second [09:06] i get some errors in the "general messages" log [09:07] enabled networking on the device and used the build -> ubuntu touch -> run on device menu entry[10:06:32] bash: cannot set terminal process group (-1): Invalid argument [09:07] bash: no job control in this shell [09:07] oops [09:07] copy and paste error, but i guess it's still readable [09:07] oh well, will check this channel later. have to do some money earning work now ;) [09:09] mhr3, the session ones are in ~/.cache/ubuntu-download-manager/ the system ones are in /var/log/ubuntu-download-manager [09:09] mhr3, let me know if you need help reading them, what is the problem? [09:12] <_matti_> Hi, I'm trying to build Ubuntu Touch for Nexus 4. I successfully created the build environment, downloaded the sources via repo and got the Nexus 4 files by calling "breakfast". Now I get some error messages (some not existing files, missing permissions to /system/etc/firmware). Does the extrac-files script assume that I have installed CM on the device? Shouldn't it work with Android also? [09:12] <_matti_> Sorry, wanted to write: "Now I get some error messages when using extract-files.sh" === Ursinha-afk is now known as Ursinha === davidcalle_ is now known as davidcalle [09:16] mandel, mind looking at https://launchpadlibrarian.net/170005100/udm.log ? [09:16] mandel, search for "com (4)" and tell me why :) [09:16] mhr3, sure :) [09:17] mhr3, the file is already there (at least 3 times) and the download manager is adding the number before the extension, in this case, cim is the only part of the name before a . [09:18] mhr3, check if the files are there, remove all of them, try the installation again and let me know if after installation is present (it should not) [09:19] mandel, well the bigger problem is that the install fails with a space in the name [09:20] mhr3, yes, seems that the execution does not like those spaces :-/ [09:20] mandel, http://paste.ubuntu.com/7118794/ [09:20] mandel, whose bug? :) [09:20] you or click? [09:20] mhr3, lets first see if the files are left there after installation, they should not [09:20] mhr3, also, I think after the error the should be cleaned... [09:21] apparently they aren't ^^ [09:22] mhr3, let me check the code, give me a few mins [09:24] barry, known that system-image-cli can't upgrade? it tells me that it can't download channels.json [09:25] mhr3, hm... weird, we tested the umd and the si before landing them and had no problems, what img are you at, and device? [09:26] mhr3, can you let me know the version of system-image-common and ubuntu-download-manager?? [09:26] mandel, i dist-upgraded 30minutes ago [09:26] mako, trusty-proposed [09:26] mhr3, dist-upgrade... hm, I would not use that it breaks things [09:26] flashing now with ubuntu-device-flash... so can't tell you the versions anymore [09:27] mhr3, flahs it to version 246, I'll be testing with that on mako [09:28] mandel, fwiw http://paste.ubuntu.com/7118817/ [09:29] mhr3, that is an old bug we fixed some time ago where there is a race condition in system image [09:29] update* [09:29] mhr3, AFAIK it should be fix in the latests img [09:29] mandel, seems like it isn't [09:30] mandel, my base image was something between 242 and 244 [09:30] mhr3, those had broken si, it landed yesterday [09:32] mhr3, I'm testing with the latests img the udm to see if the files are not cleaned after installation [09:38] mhr3, I cannot reproduce your issues, I'm downloading files apps, installing them and removing them, I'm going to break pkcon and will let you know [09:41] mandel, hm, looks like the entire Downloads dir was deleted after flash [09:42] mhr3, ok, it indeed does not remove the package when there is an error.. we should talk about this, I think it is left behind in order to be able to debug [09:42] ah, no... i did s/local/cache/ :/ [09:42] mhr3, since, if we remove it, we don't know what was the issues (I did this my removing pkcon from my system) [09:43] mandel, but consider you just downloaded 500mb pkg and it fails to install cause you don't have space on the card [09:43] now you not only don't have space but lost 500mb [09:43] mhr3, hm, indeed, we need to talk with alecu and see what we want to do [09:44] mandel, hm... so i removed everything from Downloads and it's still failing to install [09:44] mhr3, hm, do you have the logs? [09:45] mhr3, that is very strange, could be an issue with that package [09:45] mhr3, I installed about 3, 4 before I told you it was ok [09:47] mandel, http://paste.ubuntu.com/7118892/ [09:47] mhr3, there are two changes we can do, one, that I remove the file (something that is currently done in the success case) or we trust the scope to do the right thing when it gets the error signal [09:47] mhr3, that seems to be a click issue, can you install other apps? [09:48] mhr3, lin 10 - 30 are the click stdout and stderr [09:49] mandel, same error with different apps [09:50] mhr3, wow, I don't have that issue at all! [09:50] mhr3, is that devel-propose? [09:50] line 30 is very interesting.. Permission denied [09:51] mandel, yes, 246 [09:51] mhr3, ok, reflashing to see if I can reproduce it from scratch [09:52] maybe i should do clean flash and dump my ~ [09:52] mhr3, if you have no problem in removing it, it would be a nice test, yest [09:52] yest* [09:53] gosh.. let me type yes! [09:55] ok, as an aside: as the deploying to phone doesn't work with QtCreator, what's the manual process? [10:08] mhr3, I just installed that same package with no problem.. it looks like it could be related to the state of your system :-/ [10:09] mhr3: what are the permissions on the directory /opt/click.ubuntu.com/.click/users/phablet ? [10:10] mhr3: and is it possible that you worked around an issue from last week or thereabouts by manually creating that directory? [10:11] mhr3: this sounds awfully like the situation I cautioned against in my comment #2 on https://bugs.launchpad.net/ubuntu/+source/click/+bug/1291192 [10:11] Ubuntu bug 1291192 in click (Ubuntu) "click hook run-user fails when booting image 233" [Critical,Fix released] [10:15] Mirv, hey, I’m reading your blog entry (congrats on the almost entirely smooth landing of Qt 5.2 btw!), and I’m interested in "xvfb used to run unit tests needed parameters updated in most packages because of OpenGL changes in Qt", can you point me to those changes? [10:18] Mirv, I mean the parameters updates to xvfb? I’ve seen similar problems with my oxide branch of webbrowser-app, where I started using xvfb, and I’d like to check that I solved it in a consistent way with what you’ve done [10:21] cjwatson, no, i never touched that dir [10:21] oSoMoN: sure. there were a bunch of "Could not initialize GLX" which were fixed by passing a "-s '-screen 0 640x480x24'" parameter like in https://code.launchpad.net/~timo-jyrinki/ubuntu/trusty/maliit-framework/fix_tests_under_qt52/+merge/203039 [10:22] some used QT_QPA_PLATFORM=minimal instead since that was the first glue found, but that should be better in allowing the emulated X tests to continue to run [10:22] mhr3: what does ls -ld /opt/click.ubuntu.com/.click/users/phablet say? [10:22] oSoMoN: and thanks :) [10:28] cjwatson, drwxr-xr-x 2 clickpkg clickpkg 4096 Mar 18 16:51 /opt/click.ubuntu.com/.click/users/phablet [10:28] mhr3: chown phablet:phablet /opt/click.ubuntu.com/.click/users/phablet [10:29] mhr3: maybe an artifact of some earlier bug if you didn't create that yourself [10:29] mhr3: was this a fresh install, or has it been upgraded? [10:29] I mean the user data [10:30] cjwatson, upgraded [10:32] ok. I'll worry about it if I get reports from fresh bootstraps, then [10:37] cjwatson, yea, that fixed it [10:42] Mirv, thanks, that’s exactly the fix I applied to my webbrowser-app branch, glad it’s consistent with your approach :) [10:43] oSoMoN: good! :) [10:44] how do I take screenshots on the device? [10:45] gcollura, http://paste.ubuntu.com/7073413/ [10:45] thanks ogra_ :) [10:45] (there is a fix for phablet-screenshot in the queue, but i think it did not land yet) [10:49] is there a stable (devel, not proposed) image yet with qt52? [10:49] t1mp, nope === vrruiz_ is now known as rvr [10:53] ogra_: ok [10:53] i guess we might have one by end of the week === netcurli_ is now known as netcurli === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [11:06] * janimo` is awed by 78 mountpoints existing on a fresh mako install [11:06] heh [11:07] this is a metric we sure have any competitor beat at [11:07] what scares you about it ? apart from getting a large mtab [11:08] ogra_, same things as about large codebases I guess. The implied complexity [11:08] well, the code that does the mounting is pretty simple [11:09] ogra_, which code does this? I don't even dream of documentation existing for this :) [11:10] the code in /usr/share/initramfs-tools/scripts/touch [11:10] it reads "writable-paths" from disk onboot and loops over it [11:10] ogra_, ok, that is code I had briefly looked at before, thanks [11:11] ogra_, /dev/mmcblk0p23 on /lib/ufw/user6.rules type ext4 (rw,relatime,discard,data=ordered) [11:11] not sure what to make of that. A file as a mountpoint? [11:12] well, effectively it is /userdata/system-data/ [11:13] imagine the bind mounts as links into that dir [11:14] Morning all === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [11:21] I have image 237, and the device boots (I can connect with adb), but the screen stays black. What can be the cause? [11:21] @nexus4 [11:21] missing patience ? [11:22] (if whoopsie/apport process crash files the boot can take ages) [11:22] I never experienced a boot process on that device yet that takes >5min [11:22] well, check top [11:23] t1mp: press the power button twice [11:23] I don't see weird stuff in top http://ubuntuone.com/5EZIq6O1ZpRKF3uWWT7AG1 [11:24] davmor2: doesn't do anything [11:25] t1mp: do ps aux | grep unity8 [11:28] davmor2: phablet 1892 0.1 1.3 234928 24956 ? Ssl 11:18 0:00 unity8 [11:28] interesting [11:28] so you should see something on screen [11:29] nope, fully black (no pixels, no backlight) [11:29] without adb I would guess the phone is off [11:31] when rebooting, I see the Google logo and then it turns black. [11:31] * t1mp reflashing to -proposed === MacSlow is now known as MacSlow|lunch [11:46] after flashing the device, screen is still black :( === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [11:46] how do I change the brightness of the screen? [11:47] with the indicator [11:47] (battery) [11:47] from command-line I meant, my screen is black so no indicator [11:47] oh I see the flash failed. I still have image 237 [11:51] the last file fails to be pushed - http://pastebin.ubuntu.com/7119371/ [11:52] I don't know where it is pushed, but there seems to be space available - http://pastebin.ubuntu.com/7119373/ [11:53] t1mp, are there any xz files in /cache/recovery? [11:53] t1mp, it pushes in recovery ... and it uses /cache [11:54] t1mp, list them to me and wipe them [11:54] 146M free isn't enough [11:54] s/6/3/ [11:55] sergiusens: https://pastebin.canonical.com/106765/ [11:55] I'll delete thoe [11:55] *those [11:55] geez [11:55] 2fa [11:56] my thought is the upgrader is supposed to clean that; I'll check it's code [11:56] ogra_: http://pastebin.ubuntu.com/7119385/ [11:56] ogra_: "pas" auto-completed to pastebin.canonical.com :) [11:56] :) [11:56] version-101.tar.xz [11:56] ??? [11:56] wow [11:57] * ogra_ wonders where that comes from [11:57] ogra_: that's probably from today. I did a "ubuntu-device-flash devel-proposed" (forgot the --channel) and it started flashing 101 before I cancelled [11:58] sergiusens, ^^^ that should rather produce an error i suppose [11:58] ogra_, it should; I have an MR somewhere around for that [11:58] heh [11:58] there is an MR for everything :) [12:01] thanks guys, I reflashed, and at least now I can see a spinning ubuntu logo :) === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [12:22] mhr3, any luck with the installation of click packages? [12:25] mandel, did the chown, removed everything from the downloads dir, now it works [12:27] mhr3, ok, nevertheless you did bring up two bugs, the one realted with the space in the name, which should be fixed by adding "" around it and is part of the click scope and what do do will the click package on error [12:27] mhr3, lets wait for alecu and we can decide how is the best way to fix those [12:33] mandel, opened https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1294631 [12:33] Ubuntu bug 1294631 in unity-scope-click (Ubuntu) "Click packages with spaces in file name fail to install" [Undecided,New] [12:33] mhr3, awesome, thx [12:34] mhr3, I think it is due to the installation command, and I don't thing that udm should have any knowledge about it [12:34] mandel: don't just add quotes like that, you'll create other bugs [12:34] cjwatson, true... meh, I hate the fact that udm does the installation [12:34] mandel: either switch to the packagekit dbus method, or use proper subprocess invocation that doesn't go through the shell [12:35] cjwatson, I'm using a correct process invocation, they send the shell like that so that the can call a dbus method in the scope [12:36] you're clearly not since you wouldn't have this bug if you were :) [12:36] correct subprocess invocation doesn't have quoting bugs [12:38] I guess there's no way to get udm to perform multiple commands in sequence other than that && hack? [12:38] cjwatson, that is the issue, is not that I'm not using the correct way to execute the process is that they send the command with && [12:39] cjwatson, I guess I can change the api to support a list of commands, where I execute them one after the other one as long as the previous is successful [12:39] cjwatson, but at that point, it feels that udm is doing waaaaay to much, I don't want to be a process dispatcher [12:40] I didn't mean udm was doing it wrong, I meant that the scope was [12:40] cjwatson, I was covering my ass hehehe :) [12:40] * mandel is emotional when he makes mistakes ;) [12:40] cjwatson, but you are indeed right in saying that is a hack and is dangerous and not very flexible [12:41] the safe fix with least structural rearrangement would be to have the scope do proper metacharacter escaping [12:41] http://doc.qt.digia.com/qtcreator-extending/utils-qtcprocess.html seems to have the right kinds of functions [12:41] oh, though that's in QtC, bah === chihchun is now known as chihchun_afk [12:44] cjwatson, well, honestly, I'm more than able to tell them to send a list and do the hard work on my side it just feels bad, but it might be a personal taste of mine [12:44] sed -r 's/([^a-zA-Z0-9+,./:=@_-])/\\\1/g' [12:44] an equivalent of that in QRegexp shouldn't be too hard [12:45] I certainly prefer list forms where possible :) [12:45] cjwatson, yeah, I'm not doing a regexp hehe [12:46] cjwatson, less code + problems ;) [12:46] wow, that regex looks like asciiart [12:46] ogra_, there is even a scary face in there :=@ [12:47] yeah :) [12:48] the safe punctuation symbols are in ascii order, that's not my problem :) [12:48] (apart from - because character class syntax) [12:51] mhr3, cjwatson I'll fix it allowing a list to be passed as per bug 1294634 and I'll send a patch to the click scope [12:51] bug 1294634 in ubuntu-download-manager "Only one post download command can be passed to the download via the metadata" [Medium,Triaged] https://launchpad.net/bugs/1294634 [12:56] cool === chihchun_afk is now known as chihchun [13:05] cjwatson, on other news, testing umd again with your branch, if everything goes ok (updates and click installation) I'll give a +1 to that silo === alan_g is now known as alan_g|lunch === chihchun is now known as chihchun_afk [13:11] mhr3, mandel did you identify a problem in the system update stack? [13:12] barry, I think it was mhr3 system state, it was in an older img version and did a dist upgrade and things went wrong [13:12] barry, AFAIK everything is working now as expected [13:12] mandel: ok, thanks [13:12] barry, 2 issues: wrong owner of the /opt/... dir + space in click file name [13:13] mhr3: ack === _salem is now known as salem_ [13:29] bzoltan: I've landed click 0.4.19, which should have the chroot fixes you wanted, so I think you should be able to create both 13.10 and 14.04 chroots now [13:30] cjwatson: nice! thank you. I am dogfooding the 0.1.19 since yesterday and it created the chroots [13:31] cjwatson: today when I have destroyed and recreated the chroots I have seen that many SDK packages are missing. [13:34] if I can get to the point of being able to just install ubuntu-sdk-libs-dev:armhf it'll be much easier to handle this [13:34] I'm getting there but currently blocked by https://code.launchpad.net/~cjwatson/libaccounts-glib/multiarch/+merge/211468 and https://code.launchpad.net/~cjwatson/dee/multiarch/+merge/211463 [13:36] cjwatson: OK [13:37] cjwatson: the QtCreator frontend is ready for the new frameworks and click stuff [13:38] bzoltan: which packages in particular are missing? [13:38] cjwatson: how can i query the valid frameworks for the click chroot command? [13:38] cjwatson: after a clean chroot creation all qt5declarative*plugin [13:39] asac: no interface for that yet [13:40] file a bug :) [13:40] or write an interface ;) [13:40] I'm fine with writing it [13:41] ogra_: I got that card too already :D [13:41] heh [13:41] well, I didn't ask for writing it, I asked for suggesting a sketch shape for the interface you needed, not quite the same thing [13:42] we need to keep asac's coding skills alive else he fully turns into a manager some day ;) [13:42] cjwatson: aye [13:42] bzoltan: right, I'm currently only installing the -dev bits and not the runtime. I think honestly it's best for me to fix this by focusing on making ubuntu-sdk-libs-dev:armhf installation [13:43] *installable [13:43] hardcoded lists in click chroot are a stupid hack [13:43] ogra_: :) [13:43] mhr3: any chance I could get https://code.launchpad.net/~cjwatson/dee/multiarch/+merge/211463 reviewed/landed? [13:43] cjwatson: If there is anything I can help please bug me [13:43] bzoltan: will do. not at present though, I need to get dee and libaccounts-glib fixed before I can see the next level in the stack (if any) [13:46] mardy: any news on the libaccounts-glib landing? [13:46] dbarth: ^ [13:46] dbarth: maybe we should get a silo just for that, to make sure it can land soon? [13:48] this should land with the webapps-oa silo; it's ready now [13:50] Elleo: hi! The peer picker UI you added to the content hub, in which process will it live? In the app or in the content-hub service? [13:51] mardy: for 14.04 it lives in the app, once there's the infrastructure in place to support it it'll live in a separate helper [13:52] I forget the name of the bit of infrastructure we need for that though [13:52] bfiller: ^ ? [13:53] mardy: lives in the app, will move to the hub process once trusted sessions are in place post 14.04 [13:54] trusted sessions, that's what I was thinking of :) [13:55] Elleo, bfiller: thanks [13:56] dednick: hi! So, it seems that for the trust session, we'll need to be able to add specific surfaces, not just pids [13:57] dednick: because both in the Online Accounts and Content Hub cases, the trusted helper is a single process (a D-Bus service) which will have more windows === alan_g|lunch is now known as alan_g [13:58] mardy: the trusted helper has surfaces? [14:00] kenvandine: ^ what's the final structure of the content-hub picker service likely to be like? I seem to remember you mentioning it possibly living outside of content-hub-service on its own? [14:00] hm, is OTA update possible without having a ubuntu one account? [14:00] dednick: yes: for Online Accounts it's the Online Accounts UI itself (or some other page in other cases), for the Content Hub it will be a screen to select the peers [14:00] oops, time for a meeting; back later [14:00] Elleo, mardy: i've been wanting to keep the part that has a surface outside of the hub itself [14:00] separate process [14:01] but still a trusted helper of some sort [14:01] kenvandine: OK, so the hub will spawn a separate process and talk to it via some IPC? [14:01] mardy, i think so [14:01] sil2100, can you take a look at cjwatson's dee branch? too much debian magic for me :) [14:01] i wanted to keep the hub service as simple as possible [14:02] and having another process that handles the surface as needed [14:02] short lived [14:02] mhr3: which branch? :) [14:02] bzoltan: I just tried installing ubuntu-sdk-libs:armhf (no -dev), fwiw, and that fails too for basically the same reasons - I think it'd be unacceptable to hardcode the huge pile of stuff in sdk-libs in click chroot, we'd lose track really quickly [14:02] sil2100, https://code.launchpad.net/~cjwatson/dee/multiarch/+merge/211463 [14:02] kenvandine: they'll have the same lifetime, won't they? === marcoceppi is now known as marcoceppi-mobil [14:03] the process that shows the picker component would only be running while it's needed [14:03] cjwatson, iirc gir didn't support multiarch last time we tried, did that change? [14:03] and i was thinking it would be multiple instance, so there could be multiple processes running handling different requests [14:03] kenvandine: ah, OK, your case is a bit different [14:03] mardy: as far as I'm aware, the trusted helper shouldn't really have any surfaces. It should just be managing the stacking of other surfaces (sessions). [14:03] kenvandine: yep [14:03] mardy, that isn't set in stone though :) [14:04] but i wanted it more of a fire and forget from the hub (not totally forget) [14:04] the hub will of course watch the transfer request [14:04] but not care about the state of the helper process [14:04] which would get torn down when the trusted session is done [14:04] mhr3: it doesn't reliably, no, but I didn't mark any gir packages as multiarch [14:04] mhr3: for exactly that reason [14:04] dednick: well, if that's a hard requirement, I can refactor how Online Accounts works, it should be doable to delegate the surfaces to other processes [14:05] dednick: but I believe you'll get this requirement when you start targeting the desktop won't you? There it's full of processes with multiple windows [14:05] cjwatson, so dee kinda is yet is isn't multiarch? === MacSlow|lunch is now known as MacSlow [14:08] mardy: I think it's going to have to be a hard requirement for the moment. [14:08] mardy, i'd prefer that for online accounts too [14:08] i think splitting that would be overall a good thing [14:09] mardy: at the moment, there's no way to reference a single mir session of a process let alone one of it's surfaces. [14:09] kenvandine: I don't think so... we certainly don't gain anything from splitting [14:09] mardy, then signon-ui would just be a service process that manages the helpers [14:10] mardy, right now if there is 2 requests for auth, it queues them right? [14:10] so you only see one [14:10] kenvandine: I already plan to merge signon-ui in the online-accounts-ui process [14:10] haven't we had problems with that? [14:11] kenvandine: no, they are not queued (unless they are for the same account, but that won't change anyways) [14:11] ah... i thought they were [14:11] kenvandine: they are queued *per client* [14:12] kenvandine: so, we cannot stack two signon-uis on the same client window [14:12] mardy, what's a client in this case? [14:12] an app that's requesting the auth? [14:12] kenvandine: yep [14:13] so if friends and the browser both need auth, it would show them both? === awafaa_ is now known as awafaa [14:13] i guess where that was biting me was when we had the script for adding the accounts [14:13] kenvandine: yes, unless it's the same account [14:13] it would add multiples and both get auth requests [14:16] dednick: do you have a vague idea of when you'll be able to address single surfaces? So that I know if I can work with pids for the time being (and have bugs on a few corner cases) and wait, or if I should refactor the code and work with pids forever [14:18] mardy: i'll need to check with tvoss about how this fits into current cycle requirements [14:18] * tvoss reads backlog [14:19] tvoss: ah. that's helpful :) [14:19] dednick, gimme a few for context switch [14:19] sure === chihchun_afk is now known as chihchun [14:21] mhr3: it can safely be multiarch at the runtime level, which is all that I need for my use case [14:21] mhr3: multiarch -dev packages are more complicated, but in a lot of cases that's a stretch goal anyway [14:23] mhr3: but the lack of a multiarch runtime here blocks doing a foreign-arch install (needed for click chroot) of anything involving friends, since that has Multi-Arch: foreign components (i.e. in practice installed from the native architecture but able to satisfy dependencies from other architectures) that depend on libdee, as well as having components from the foreign architecture that depend on libdee [14:24] mhr3: so I have to have the libdee runtime be Multi-Arch: same in order to make this work [14:24] mhr3: fortunately the packaging is already in the right shape for that, it's just missing metadata === vying is now known as Guest78258 === marcoceppi-mobil is now known as marcoceppi === marcoceppi is now known as marco-traveling [15:07] jono, on the api page there is a sort of block diagram of the qml api [15:07] http://developer.ubuntu.com/apps/qml/api/ [15:07] I'd like to put that into a presentation I am making [15:08] was there some app that I should install for all that? [15:11] rickspencer3, which block diagram? === chihchun is now known as chihchun_afk === alan_g is now known as alan_g|tea [15:41] cjwatson: would it be OK to install the click packages in an armhf ubuntu-sdk-14.03 click chroot and aa-clickquery the policy version there? [15:44] bzoltan: that's probably a question for jdstrand [15:44] cjwatson: we are discussing with him about it ... on the #sdk [15:45] cjwatson: as the click-review-tool does not recognize the -qml, -html frameworks [15:45] cjwatson: as they are not listed under /usr/share/click/frameworks [15:46] they should be, they're in ubuntu-sdk-libs-dev [15:46] but I can't cross-install that in click chroots yet for the reasons given earlier [15:49] cjwatson: so, as for now ... the locally installed ubuntu-sdk-libs-dev is the only usable source [15:49] right, I don't really see a more sensible approach than driving towards making that work [15:49] which I'm already doing, if I can persuade everyone to land my changes :) === alan_g|tea is now known as alan_g [15:50] cjwatson: yes... now I start to understand where we are :) [15:57] hmm, there's seven updates available [15:57] i just logged into my ubuntu account (since that seems to be nessecary (???)) [15:57] they all fail with "COMMAND ERROR" [sic] [15:58] is that expected behaviour at this point? [16:00] this is not using apt or anything, but the update manager shipped with ubuntu 14.04 (r237) [16:00] fps: yes. [16:01] fps: that will go away when we promote the next image [16:01] popey: ok, so updates with that image are broken [16:01] ? [16:01] yes, long story ☻ [16:01] bug [16:01] bug 1293797 [16:01] bug 1293797 in Software Center Agent "nexus 4 "COMMAND ERROR" when updating apps" [Undecided,Confirmed] https://launchpad.net/bugs/1293797 [16:01] fair enough :D [16:02] sorry about that ☹ [16:02] no problem at all.. [16:02] the issue is known [16:02] the bug is filed === chihchun_afk is now known as chihchun [16:02] what could go wrong? ;D [16:04] somewhat related: will i make this channel my enemy when asking why ubuntu touch doesn't just go with apt in a more vanilla way? [16:04] why all this system write only, and OTA update breakage if manually using apt? [16:04] no, not at all [16:04] it's a valid question [16:04] that sentence of mine was bad.. if it was undecipherable, let me know and i'll rephrase :D [16:04] but one that we've answered a few times, so it should be a faq [16:04] checking the FAQ [16:05] yeah, i dont think we've written it up anywhere === jhodapp is now known as jhodapp|afk [16:06] the write only part of it i can guess about [16:06] it's probably for security reasons [16:06] there's a number of facets, including security, yes [16:06] https://wiki.ubuntu.com/SecurityTeam/Specifications/ApplicationConfinement/Manifest [16:07] it also makes system image updates a lot simpler to apply, because you can do them as a big delta blob rather than as something with lots of moving parts that can go wrong [16:07] i see. [16:07] we don't want to run maintainer scripts as root [16:08]