[00:07] yes RobbyF i wanted advice from someone into the project as we will be making a hackaton in few days and advice/recommendations are very appreciated ;) === chriadam|away is now known as chriadam [00:35] hi everyone [00:35] I have a question: [00:37] how does Ubuntu Touch know when the phone it is running on is connected to an external monitor and to switch to desktop mode? [00:37] magic, I assume [00:37] Does anyone know if https://code.launchpad.net/~mterry/lightdm/socket-junk/+merge/186415 currently blocking lightdm on Ubuntu touch? [00:44] mterry, hey [00:45] mterry, sorry to bug you post eod [00:46] mterry, short question, is https://code.launchpad.net/~mterry/lightdm/socket-junk/+merge/186415 blocking lightdm on touch and is that the last remaining issue? [00:47] what screen resolutions does Ubuntu Touch support when in desktop mode? [00:48] robert_ancell, that is unrelated to lightdm on touch, it's just related to lightdm + mir [00:48] mterry, i.e. unity8 running in u-s-c? [00:48] robert_ancell, right [00:48] mterry, which is not a high priority at the moment? [00:49] robert_ancell, but lightdm on touch will start with surfaceflinger, not blocking. right, not high priority [00:49] mterry, so OK if I do the MP but it's not released immediately? [00:49] robert_ancell, sure [00:49] mterry, thanks [01:17] sergiusens: do we have a standard webserver anywhere for phablet stuff? i plan to add a --no-network flag to demo-setup to download a tar from somewhere on the host machine, but unsure where to put it [02:06] is ubuntu touch coming for galaxy s4 (i9500)?? [02:22] hey everyone, I was wondering, if I have a phone running Ubuntu Touch connected to an 11.6" external monitor with a resolution of 1920 x 1080, will everything on screen appear too small? [02:41] boiko: hey, you guys wouldn't happen to have a history.sqlite that you use to test the history-service that we could use as fake conversations do you? === chihchun_afk is now known as chihchun [03:26] ok, I'm up on the tablet [03:27] looks like the touch screen might have the axes flipped [03:27] and or inverted, i think the scale is off too as little gestures in the corner seem to move things pretty far [03:27] how is the touch screen loaded and configured? === Namidairo`bnc is now known as Namidairo [04:04] stgraber: all the jobs have been regenerated now since landing that change - we shouldn't have any more references to daily-proposed now [04:13] plars: cool, I'll kill the daily-* channels tomorrow morning then [04:18] is there any way to join a real session from adb shell? [04:43] nik90: Great if that app is useful :) pmcgowan started it and I decided not to create my own similar version, so added the network, storage and location tabs. [06:05] bzoltan: will qtlocation in your code also work for geoIP as source? [06:05] bzoltan: I tried it on my desktop and sometimes it found the location coordinates but sometimes it didnt [06:05] nik90: That is what happens [06:06] nik90: I do not know much about the sources ... [06:06] bzoltan: you mean you also found that it sometimes found it and other times didnt? [06:07] nik90: Yes [06:07] nik90: it is unreliable ... But I tested on phone only [06:07] bzoltan: currently I used a xml list model to retrieve the current location from geoip.ubuntu.com/lookup directly [06:08] bzoltan: however dpm asked me to look into using qtlocation [06:08] bzoltan: still unsure which one to use [06:08] nik90: My objective with that app was simple to see if the qml api works ... I am not the one to debug the backends :) [06:09] bzoltan: :) [06:10] dpm: any suggestions what can done ^^ === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [06:21] with image 20130924, I need to login to U1 to install an app, is it by design or a bug? it is really annoying [06:29] nik90, tvoss created the Qt Location backend, but I believe he is away for the week. In that case, let's use direct lookup for now, and switch to Qt Location when we know more. Would you mind sending an e-mail to the phone and core apps mailing lists to see if someone else can help with using Qt Location in the meantime? [06:35] dpm: definitely [06:36] dpm: as in I will send it asap [06:36] * dpm hugs nik90 [06:38] jdstrand: https://bugs.launchpad.net/unity-mir/+bug/1230091 [06:38] Ubuntu bug 1230091 in unity-mir "[feature] Window reparenting" [Undecided,New] [06:46] nik90, we had a discussion last week with the calendar developers about https://bugs.launchpad.net/ubuntu-weather-app/+bug/1226834 - which API/service do you use in clock to retrieve city locations? Do you have any issues with it or would you recommend to use the same for weather? [06:46] Ubuntu bug 1226834 in Ubuntu Weather App "Identically named cities are indistinguishable in weather-app" [High,Triaged] [07:09] good morning [07:15] morning Saviq. Looking at https://bugs.launchpad.net/music-app/+bug/1228345 - is this something you think is related to how unity or the music scope launch the music app? I'm not sure we can fix it in the music app itself [07:15] Ubuntu bug 1228345 in Ubuntu Music App "Launching track from dash new instance of app, not re-use of existing one " [Undecided,New] [07:23] I recently installed the latest ubuntu-touch non preview onto my P4WIFI and its just a black screen I am able to adb root adb shell ubuntu_chroot shell just fine [07:24] I piped dmesg into a text file is there anywhere else I could look to solve this issue [07:42] Is it possible to conect ubuntu phone to the keyboard? | http://askubuntu.com/q/349808 === VDVsx_ is now known as VDVsx [08:39] dpm, yeah, it involves work in a few areas, it's being worked on === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [09:29] hi everyone [09:30] it might be not a new question but will there be a ubuntu-touch version for nokia n9 phones? [09:32] hmm, waiting for more than 1min for an answer might have helped === b0bben_ is now known as b0bben === vrruiz_ is now known as rvr === gusch is now known as gusch|lunch [10:04] Morning all [10:18] mornin davmor2 [10:28] does the web-brower app need some extra css styling, to style buttons and such | http://askubuntu.com/q/349845 [10:29] ogra_ I have been in here for over an hour and still waiting lol [10:30] DJJeff: looks like you're using the old non-flipped image? [10:31] popey, how do I check? [10:31] oh you mean of cyan mod ? [10:31] No, the fact that you mentioned ubuntu_chroot makes me think you're using an old version of the image [10:32] I downloaded both http://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/current/saucy-preinstalled-touch-armhf.zip [10:33] and http://ubuntuone.com/3eqlvC2YW3xO0suhkLvmVi which is cyan mod (2013-07-14) [10:34] ( cm-10.1-20130714-UNOFFICIAL-p4wifi.zip ) [10:35] looks like I need to wait for this cm to be updated to work with the newer ubuntu touch? [10:39] most likely, yes [10:42] I will output dmesg and pastebin it so MAYBE just MAYBE p4wifi maintainer (patrickseemann) can see it [10:42] anything else I should pass on besides dmesg output? [10:43] popey: ping === smartboyhw_ is now known as smartboyhw [10:44] oSoMoN: did you saw http://askubuntu.com/questions/349845/does-the-web-brower-app-need-some-extra-css-styling-to-style-buttons-and-such ? [10:45] daker: no, reading now [10:46] daker: huh, can alert boxes be CSS-styled in other browsers? [10:46] oSoMoN: it's about the default css for elements (checkbox, buttons, textinput, etc...) they look different/ugly from the native ones [10:47] tinti_: pong [10:47] DJJeff: sorry, i dont know, not done porting myself [10:47] what's the command to take a screenshot again? [10:48] daker: the way the question is formulated, I understand it differently [10:48] im not asking about porting, im asking about what log files to look in [10:48] to see how to fix the black screen [10:48] IE ( /var/log/???????? ) [10:48] oSoMoN: "As you can see from the image below, the Show alert box, looks differently, then in firefox, IE, chrome." [10:49] popey: hi. Will ubuntu traditional ui be developed or it is stopped in flavor of ubuntu touch ui, for tablets? [10:49] oSoMoN: he means that even on the desktop they don't look native like in FF & chrome [10:49] daker: ah [10:49] tinti_: by 14.10 it'll be the same codebase. convergence! [10:49] I mean for tablets will we have in a near future ubuntu desktop + ubuntu touch? [10:50] ok, but ubuntu touch more close to an Android rather a Linux desktop right? [10:50] daker: well that’s kind of on purpose I would say, the browser uses the Ubuntu UITK, so its native dialogs have the default Ubuntu SDK look’n’feel [10:51] oSoMoN: no no look at the buttons on the background [10:51] daker: I don’t see the point of applying the standard GTK theme to all SDK widgets on desktops [10:51] oSoMoN: do they look native on the phone ? [10:51] tinti_: we use the same UI on phone and tablet [10:51] tinti_: I don't know what you mean "close to an android" [10:52] daker: no, they don’t look native [10:52] oSoMoN: that's what he is talking about ;) [10:52] daker: not really clear… but ok. We would need to implement a custom renderer for webkit, that uses the SDK widgets. Not on the roadmap for sure [10:53] oSoMoN: ok :) [10:53] I mean compared to the last release, the 13.04 image === MacSlow is now known as MacSlow|lunch [10:53] daker: will you answer the question, or do you want me to do it? [10:53] oSoMoN: just do it [10:54] daker: ok [10:55] popey: this article helps a lot to understand some key points of ubuntu touch [10:55] http://www.xda-developers.com/android/ubuntu-touch-next-generation-os-or-just-another-skin/ [10:55] tinti_: thats old [10:56] and inaccurate [10:56] still the "cyanogenmod chroot valid" [10:56] * tinti_ take more care of what you read [10:57] nope [10:57] we use some android drivers which are confined to an lxc container [10:57] hum, is it armel or armhf? [10:58] for compatibility with the blobs right? you have a small android subsystem to use as HAL [10:58] armhf [10:58] well, our bits are armhf [10:59] ok, but will ubuntu desktop and ubuntu touch ui converge? [11:00] remove java from android and all you have is linux :) [11:00] tinti_: yup [11:00] put ubuntu touch on top and presto a bad ass phone/tablet [11:01] haha [11:01] * ogra_ wishes it would be that easy [11:01] there was a talk at DEFCON 20 about removing java from android [11:01] * tinti_ QML rocks! [11:01] well, we only use a really tiny bit of the linux part of android even [11:01] DEF CON 20 Hacking Conference Presentation By John Floren - Hellaphone Replacing the Java in Android [11:01] so better get used to the new UI :) [11:02] DJJeff: +1 === greyback is now known as greyback|lunch [11:05] tinti_: chris doesn't agree https://twitter.com/chrisccoulson/status/382820387748188160 :) [11:06] daker: why ... [11:07] tinti_: https://twitter.com/chrisccoulson/status/382802413444268032 https://twitter.com/chrisccoulson/status/382807488493408257 [11:09] daker: as far as I remember there is no QShortcut too [11:10] but I still think it is a nice way to generate a UI [11:10] the StateMachine and the Transition system saves a lot of time and logic [11:11] does anyone know if the demo-assets-* packages have changed over time? [11:11] I'm trying to flash a phone with some history... [11:11] https://launchpad.net/ubuntu/+source//+publishinghistory [11:11] apt-cache madison demo-assets-contacts, for example shows only one version in the repository. Is that expected? [11:12] Published archives typically only contain about one version per series at a time [11:12] But the history is in Launchpad [11:12] aha - thank you [11:13] is there a boot animation/splash feature being worked on ? [11:15] I Haven't tested for myself but plymouth may work. [11:16] FastCode, apparently not yet used: https://lists.launchpad.net/ubuntu-phone/msg00803.html [11:17] that answers my question even if it's a somewhat old post [11:17] rsalveti, do you know of any updates ^ ? [11:18] It didn't work and my phone got really weird. [11:20] removing plymouth gave me 'Writing Boot Image /dev/disk/by-name/boot' [11:20] Am I done? [11:20] as in dead? [11:25] I would expect that plymouth is still required for its console multiplexing even if its splash screen feature isn't used. [11:25] (That's certainly true on the desktop.) [11:25] Is anyone else running build 61 on a Nexus 4 and unable to check for updates? I just get the spinner and "Checking for updates..." [11:26] yes, plymouth was installed.installing solar and ubuntu-logo themes made no difference [11:26] plymouth is disabled because it hangs the device and nobody's debugged it yet. [11:27] By which I mean, the plymouth splash screen is disabled. plymouth itself (the console multiplexer) is still in use I believe [11:27] If build 61 is for about a week ago, then yes.I've had this issue. [11:27] cjwatson, is there a LP bug for that plymouth hang? [11:27] Not sure, ogra_ might know [11:28] Well, that's what /etc/ubuntu-build says anyway. I updated yesterday, not quite sure what time. [11:29] I don't have an 'ubuntu-build' what does it mean? [11:31] janimo, yeah, there definitely is, but i filed it over a year ago ... for the nexus7 images [11:31] * ogra_ cant find it [11:32] janimo, but plymouth is completely disabled on out images, so that shouldnt be an issue [11:32] FastCode, I'm not sure. Are you using a CDimage build or Image Based Upgrade? [11:33] (it cant talk to surfaceflinger and cant talk to Mir either yet i think (i might be wrong on the latter after the Mir sprint) [11:34] Me neither, I did phablet-flash, and used apt ever since. what does it make me? [11:35] FastCode, system-image-cli -i .... === thostr_ is now known as thostr_lunch [11:35] popey: I'm gonna hunt you down and make you listen to beiber, I only just finished flashing 61 and you go and update to 62 ;) [11:35] if that returns a 0 for "current build number" you are on a cdimage image [11:36] bulid-number is 0 [11:36] right [11:36] channel: daily [11:36] well, thats moot [11:36] seb128, Laney: got any time to look at https://code.launchpad.net/~mardy/ubuntu-system-settings/testability/+merge/187011 ? [11:36] HAHA, I forgot again [11:36] if build number is 0 you are on an unupgradeable cdimage image [11:36] let me look right now [11:36] Laney: I guessed :-) [11:36] Laney: thanks :-) [11:37] Current build number: 61 Channel: devel-proposed [11:37] (you can only use apt with that, not the update mechanism) [11:37] JamesTait, thats a system-image image [11:37] That means I have to reflash if I want to use GUI-based updates? [11:37] JamesTait, you need to update the a new channel name though [11:37] hmm, or not [11:38] I switched to the system-image method last week, I think. Last update, yesterday at 17:17. [11:38] well, there should be 62 and 63 images [11:39] ogra_, I'm unable to check for updates at the moment, I just get the spinner. [11:39] This is getting very confusing for me, do I lose anything if I use apt? [11:39] popey: 62 hmmm my box just downloaded 63 [11:39] * FastCode loves command line [11:39] mardy: I can't tell if that's right or not [11:39] what does it do? [11:39] davmor2, i tested 62 (and now run 63) ... dont worry :) [11:40] FastCode, on what device are you ? [11:40] nexus 4 [11:40] FastCode, well, then you do ... [11:40] Laney: you can test it this way: if you start "system-settings -testability" without the patch, and then run "autopilot vis", autopilot won't be able to inspect the system settings app [11:40] Laney: if you run the same commands with the patch applied, then within "autopilot vis" you'll be able to navigate the system settings UI [11:41] Laney: it's a hack I copied from the camera-app [11:41] what? [11:41] FastCode, supported devices work best with the system-image ... we also dont test the cdimage images anymore, these are largely for porters [11:41] (since we cant offer system-image images for them yet) [11:41] Laney: it's needed because while QApplication properly loads the module, QGuiApplication doesn't [11:41] Is that a bug? [11:41] Laney: good question [11:42] mzanetti: hi! Do you know if there's a Qt bug about ^ ? [11:42] FastCode, if you want to do local builds of stuff etc the cdimage image might be better though, but yoou should test yoour stuff against the rreadonly system-image [11:43] mardy: I don't think there is... But I can fix it in Qt propose a merge [11:43] nice [11:43] mardy: altough I'm not sure if its even a bug or it is wanted this way [11:43] mzanetti: that would be excellent [11:44] Well, it'd be good to investigate and fix it there if possible [11:44] in the meantime I'll approve this [11:44] mzanetti: you could ask to the ML first [11:44] I get it now, If I go with cdimage then I'm on my own and my applications may not work as intended on other systems.not as bad as it first looked like. [11:44] true [11:44] Laney: thanks [11:44] ok. will do [11:44] mzanetti: and thanks :-) [11:44] FastCode, thats an appropriate summary [11:44] ogra_: thanks for the explanations. [11:44] mardy: thanks [11:45] FastCode, you also have to stick to apt for upgrades, the updater wont work (neither gui nor cmdline) [11:45] Hm, something's failing to download, it seems; the signatures, I think. [11:46] blacklist.tar.xz.asc is taking a while, then eventually says "No blacklist found" [11:46] that's ok with me, I didn't even know system settings had update functionality until a few days ago. [11:47] another question, what is with partition 22(mmcblk0p22) [11:47] ? === greyback|lunch is now known as greyback [11:48] It seems to be mounted on /dev which is really weird. [11:48] 512 MB wasted [11:50] http://paste.debian.net/45977/ [11:51] Then channels.json.asc is also taking a while and ultimately timing out. I get a backtrace. :( I wonder if it's my dodgy network [11:57] seb128: is the battery screen on system settings supposed to work? [11:57] gema, yes [11:57] seb128: ok [11:57] wfm [11:57] it doesn't for you? [11:57] seb128: I've managed to get the wifi togle from system settings and the network indicator out of sync [11:58] seb128: I am going to restart and retrace my steps [11:58] yes, the wifi toggle is not updated [11:58] see if I can reproduce [11:58] seb128: ok, so that's known [11:58] that's a known lack of feature [11:58] yes [11:58] and the brightness not changing the actual britghness? [11:59] that's fixed in trunk [11:59] ack [11:59] you should ask asac for an update to land [11:59] its really had to slide the battery gauge up/down [11:59] well, left/right [11:59] as you do so, if you deviate from the centre line, the page scrolls and you lose control of the slider [11:59] toolkit issue? [12:00] what is the state of keyboard going nuts when opening the indicator sliders?It really hurts when working with terminal. [12:00] is it even a reported bug? [12:00] does it happen on the readonly image ? [12:00] :) [12:01] seb128: also the battery charge level seems to be 100% on the system settings but the indicator shows 78% [12:01] cjwatson: regarding click backports, I'm finding a new delta that it requires libglib2.0-dev >= 2.34 that's not in precise. I wonder if it'd work with 2.32? [12:01] ogra_: are you talking to me? It was like this from day one on cdimage [12:01] seb128: and when you unplug the device, last full charge shows as N/A === hikiko is now known as hikiko|lunch [12:02] FastCode, right, just wanted to point out that bugs on cdimage might not be on system image [12:02] the settings for sleep when idle don't stick either [12:02] I didn't have this issue like 3 month ago, It happened this time I flashed. === MacSlow|lunch is now known as MacSlow [12:02] ok, so I have a fully non-functional battery screen :D [12:02] FastCode, and we dont really focus on cdimage bugs unless they also affect system-image [12:02] seb128: will wait for new version [12:02] seb128: is the fix for background selection in? [12:03] you're really getting me to reflash [12:03] Mirv: Hm, I'll have added that because of a particular API I was using, let me check [12:03] cjwatson: ok [12:03] didrocks: so it seems to me that if I press download a third time while it's downloading, I dont get to see the "reboot to install" button in system-settings OS updates [12:04] didrocks: if you could confirm this when you next see it, that would be nice [12:04] didrocks: I can file it as a bug in doubt [12:04] i think I saw it twice [12:04] I can see how it might happen is the previous button press was synchronously blocked on a handler [12:04] hmm maybe that goes away with d/l service, not sure [12:05] Mirv: So, it was for g_spawn_check_exit_status, which I can't easily drop; but actually I arranged things a while back so that the entire PackageKit plugin isn't built on precise [12:05] Mirv: I suggest you just drop the libglib2.0-dev, libjson-glib-dev, and libpackagekit-glib2-dev build-deps entirely for the backport === alan_g is now known as alan_g|lunch === chihchun is now known as chihchun_afk [12:11] cjwatson: ok, sounds good, I'll do that and test [12:11] gema, was the device ever fully charged? [12:11] seb128: yes, many times [12:11] asac, yes [12:11] gema, could be an upower issue then, we get the datas from it [12:11] seb128: but I will charge it fully again just to make sure [12:12] dbarth_: so unity-webapps-qml is in as the whole stack was run, but without its testsuite [12:13] dbarth_: FYI, https://code.launchpad.net/~zaspire/address-book-service/GaleraManagerEngine_implement_saveContact_few_methods/+merge/181438 is "missing commit message" but it also misses a human review of course [12:13] seb128: ack, btw, whenever you get into the set date and time screen with the pickers , theres no way back, you have to kill the app [12:13] seb128: the lower edge bar doesn't come up again [12:13] dbarth_: also, you link to https://bugs.launchpad.net/cordova-ubuntu/+bug/1227427 but that bug is Invalid? [12:13] Ubuntu bug 1227427 in address-book-service "Mobile Spec Suite crashes: Symbol lookup error in libqtcontacts" [Undecided,New] [12:13] dbarth_: and the branch isn't linked to it [12:15] gema, that's fixed in trunk, ask to asac to let the fixes in so you can confirm it works ;-) [12:16] asac: man, release all the fixes [12:16] asac: you are slowing us down! [12:17] lool: checking [12:17] gema: you can sneak test from daily-build and help us grow confidence in those changes [12:17] thats alwayus helpful [12:17] the binaries are all there [12:17] asac: last time I checked that wasn't my job anymore ;) [12:17] and ready to install on top of a phone boot [12:18] depends [12:18] lool: the bug is invalid for cordova, the issue is in the service [12:18] asac: anyway, plenty of other things to do, just be aware that we are eagerly waiting for that ! [12:18] right, but we have an open system [12:19] you can see whats coming, so if you see something coming it very likely will be there soon [12:19] lool: as for the review, the mp is there for a while; we've asked for a review already [12:19] asac: what do you mean we can see what's coming, where do I look for that? === _salem is now known as salem_ [12:20] gema: here is the stuff that is staged and ready for publishing to archive/image in case we can validate that it doesnt regress the image [12:20] http://people.canonical.com/~platform/cu2d/results [12:21] * gema looking [12:21] in optimisitic case those things will come [12:21] asac: ack, that's useful, thanks [12:21] and you can take sneak previews pretty easily [12:22] lool: i'll ask bill for an urgent review in a few [12:22] seb128: ok, so the time pickers are also landing [12:23] lool: while we're at it, apps have all been re-verified yesterday; sergio should be able to pre-install them today [12:23] gema, right [12:23] seb128, asac: thanks! [12:26] mhr3: Hey man o/ Is there a way to add a new scope from QML? Even directly from DashMusic.qml? I want to overrideResults() anyway, so I wonder if there's a quicker way than writing a dummy scope in cpp [12:27] mhr3: Why the question - it's hard to hardcode a view within ScopeListView (under the header, but above other scopes). So I might need to take the fake scope route. [12:28] karni, if you just want a new scope that doesn't do anything but provides categories that can all be done with a simple .scope file, no cpp needed [12:28] doanac: sergiusens: i should have an MR for you guys soon to make phablet-demo-setup work on r/o images :) [12:29] cwayne, great [12:29] karni, so just look in /usr/share/unity/scopes/ and just do cp music.scope foo.scope [12:29] cwayne, btw, did you see pitti's comment on the timezone MR? [12:29] ogra_: can you confirm (or not) bug 1230247 [12:30] sergiusens: hm, no i didn't, let me check it out [12:30] karni, or just edit the .scope file right away [12:30] up to you [12:30] mhr3: perfect! thank you :) [12:30] exactly what I needed [12:30] popey: you killed the bug bot [12:31] karni, you need to `pkill -f unity-scope-home` to the changes to be picked up [12:31] mhr3: great! [12:32] sergiusens, cwayne, note that pitts change requires systemd changes, so it will be most likely stuck in proposed until the beta freeze is lifted [12:33] ogra_: ACK [12:33] although his way is much better than mine, heh [12:34] cwayne, sergiusens https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1227520 for background [12:34] * cwayne never thought changing a timezone would ever be so damn complicated [12:34] cjwatson: could the configure.ac in click somehow be updated to auto-detect precise and disable needed parts? it seems it'll need manual changing as well to drop dependencies. [12:34] Ubuntu bug 1227520 in systemd (Ubuntu) "Timezone changes are not working due to ro /etc and bind mounts" [High,Fix committed] [12:35] dbarth_: cool [12:35] Mirv: No need, debian/rules already takes care of that. [12:35] Mirv: See debian/packagekit-check [12:35] cjwatson: ok, I'm just seeing some error about PKPLUGIN https://launchpadlibrarian.net/151432678/buildlog_ubuntu-precise-amd64.click_0.4.8~precise1~test2_FAILEDTOBUILD.txt.gz [12:36] ogra_, sort of goes against asac's target of having all of /etc writable [12:36] hello ogra_ :) can you take a quick look at my init.find5.rc file, i am not sure if the way find5 mounts system is correct for ubuntu : https://bitbucket.org/anders3408/ubuntu_device_oppo_find5/src/cacb1b4e927b7f567909cb1a1113fdeeef793fb3/configs/init.find5.rc?at=ubuntu#cl-21 [12:36] sergiusens: my target? [12:36] :) [12:36] then a few lines later : # mount_all ./fstab.find5 [12:36] Mirv: Ah, I suspect this will work if you insert an explicit build-dep on pkg-config so that autoreconf can expand that properly [12:37] * asac tries to remember [12:37] asac, you made a comment in Oakland about that ;-) [12:37] oakland? [12:37] Mirv: click ought to have such a build-dep, so that's partially my fault [12:37] sergiusens: i think i made a comment on that i believe the config system we traditionally use [12:38] munges package defaults and system defaults in one place [12:38] asac, yeah, you said anything that is not meant to be changed should go into /usr/share and anything that is meant to be dynamic in /etc [12:38] which makes it a bit hard to transition nicely without having user interaction dealing with conflicts [12:38] right === chriadam is now known as chriadam|away [12:42] popey, do I gather from the change log for 61 that there is a bunch of new SDK stuff in it? [12:43] cjwatson: ok, seems to build locally in chroot when pkg-config is installed, so probably good now. [12:43] Mirv: OK, I'll add that build-dep to trunk for clarity [12:43] cjwatson, back to the arch thing.... for arch specific, would you prefer that we do arch specific but leave the arch field unassigned? [12:44] rickspencer3: we're on 62 now .. [12:44] ooops! [12:44] but yeah, lots of changes [12:44] so we are :) [12:44] sergiusens: I'm still thinking about what's best - leave it as it is for now [12:44] popey, but 61 had a bunch of qtdeclarative stuff [12:44] today new ui-toolkit went into saucy archives, so it'll included in the next build [12:45] cjwatson, ack [12:45] so I think the next one after 63 which already built [12:45] sergiusens: I'm pretty clear that a package's architecture should be in its metadata somewhere [12:45] architecture(s) === alan_g|lunch is now known as alan_g [12:54] rickspencer3: this is diff from ubuntu=20130923 to ubuntu=20130924 which corresponds to image 60 to 61 http://people.canonical.com/~ogra/touch-image-stats/20130924.changes [12:54] thanks lool [12:55] rickspencer3: are you on 62 yet? [12:55] lool, yes [12:55] I update immediately whenever I receive popey's avengers mail :) [12:55] ☻ [12:55] Avengers, unite! === oreneeshy_ is now known as oreneeshy [12:56] rickspencer3: so opening music files works for me from home and music scope on 62 (only home scope for thostr and for me in 61) [12:56] Saviq, kgunn_ https://bugs.launchpad.net/unity-mir/+bug/1230091 [12:56] Ubuntu bug 1230091 in unity-mir (Ubuntu Saucy) "[feature] Window reparenting" [High,New] [12:56] popey: did you find some logs for your issue? [12:57] lool, well, I opened a music file (with a space!) from the music scope [12:57] it opened the music app, but didn't bring it to the front [12:57] but I can't actually hear the music! [12:57] which is, you know, not cool [12:57] * rickspencer3 tries home scope [12:58] lool, same deal with home scope [12:58] didn't bring it to front: hmm, currently it should still open a *copy* of the music-app, and then this should be in front and play [12:58] rickspencer3: are you using system-image or apt to update? [12:59] did the database fix for the music -app already make it in ? [12:59] system-image of course [12:59] I am an avenger [12:59] :) [12:59] hmpf [12:59] 4 [12:59] grr [12:59] ogra_: yes [13:00] lool, I'm a bit concerned that I can't actually hear the music [13:00] rickspencer3: is that mako? [13:00] yes [13:00] rickspencer3: and is that a mp3? [13:00] yes [13:00] rickspencer3: and is your volume >> 0? :-) [13:00] yes [13:00] I hear lool asking "is the device turned on?" hehehe [13:01] olli, thanks, will keep an eye on it [13:01] cjohnston: :-) === gusch|lunch is now known as gusch [13:01] lool, is it working for you? [13:02] greyback, can you please have a look at bug #1230091 and update, where applicable? [13:02] bug 1230091 in unity-mir (Ubuntu Saucy) "[feature] Window reparenting (required for appstore app trust model)" [High,New] https://launchpad.net/bugs/1230091 [13:02] popey: goto the home lens, type in adele, find skyfall, click on the amazon link, does it randomly open settings and then forward you to amazon? [13:02] Saviq: looking === hikiko|lunch is now known as hikiko [13:02] rickspencer3: Yes, I've just posted it to G+ just for you [13:02] stgraber: meeting? [13:03] rickspencer3: but a recent music-app bug lead me to read the startup code and I wonder whether it might be an issue with older settings [13:05] rickspencer3: https://plus.google.com/109334143378931299553/posts/QpfrBumCbp2 [13:05] lool: just curious, have you seen the music app test results from today? [13:07] cjohnston: no, I hope they are good [13:08] heh [13:08] http://reports.qa.ubuntu.com/smokeng/saucy/touch_ro/4439/music-app-autopilot/ [13:08] lool: ^ [13:10] davmor2: "randomly open settings" I think is a common problem, it's opening the last app you had open [13:12] rickspencer3: ah! just hit the no audio bug [13:12] popey: I had no apps open [13:12] popey: this is the first thing I touched :) [13:13] odd [13:14] davmor2: goes directly to music app for me, but doesn't play [13:14] and it launches multiple music apps, which I filed a bug for already [13:15] to be precise, it seems to play, but muted [13:15] (here at lesat) [13:16] not here [13:16] "No songs queued" at the bottom [13:18] cant have enough music :) [13:18] There is a graphics bug in browser that i think is worth mentioning [13:18] It only happened because my connection was laggy [13:19] I Had Terminal Open, put it in the background, Opened the browser [13:19] the ubuntu page showed up but there was a problem [13:20] Images hadn't loaded completely and i could see the terminal through them [13:20] merhaba [13:20] I even scrolled to bottom of page and this behavior didn't change [13:20] Halo [13:20] مرحبا [13:20] one x için ubntu touch varmı? [13:20] and I'm not an arab [13:21] english please [13:21] bug 1229905 lool [13:21] bug 1229905 in Ubuntu Music App "Launch song from dash doesn't play the song" [Critical,Triaged] https://launchpad.net/bugs/1229905 [13:22] oke :) I am from Turkey .. For one x kurabilirmiyiz touch ubuntu? [13:22] popey: do you know of another place I can play sounds? I tried opening a mp3 in the browser, but it plays muted [13:23] asac, ogra_, cwayne: lightdm landed last night. The session-manager-touch branch could be landed again. Third times the charm.. :) [13:23] animal farm app [13:23] Ohh, ic. I don't understand turkish well. and I'm guessing you're asking if ubuntu touch works on htc one x. [13:23] or jono's sleeptime app [13:23] mterry, it only landed in proposed [13:23] mterry, desktop beta freeze [13:23] lool, rad.io always works for me [13:24] ogra_, sure. we could land session-manager-touch in proposed too [13:24] Though I guess that doesn't buy the images much [13:24] Where and how to download and install? I could not find so I researched ... I looked at xda. [13:24] https://wiki.ubuntu.com/Touch/Install [13:24] demirbas: ^^ [13:24] mterry: ok that needs thorough testing :) [13:24] thanks [13:24] mterry, we could but then if there comes someything i want to land before in the session stuff which doesnt depend on a blocked package we would be blocked [13:25] mterry, merging and uploading is the smallest part :) [13:25] ogra_, fair [13:25] I think there's one missing for x or beklemelimiyim kurmalımıyım now little more than just the idea:? [13:25] asac, agreed. I tested it on maguro and mako yesterday. cwayne tested on maguro. Please test more [13:25] mterry, making sure all tests still pass after polkit is on is more important [13:28] demirbas: AFAIK these are the only supported devices: http://paste.debian.net/46009/ [13:28] I may be wrong. [13:29] ogra_, can we do that without putting the session-manager-touch bits in the image? [13:30] thnk [13:30] mterry, no, for testing you would have to install a deb built with your changes indeed [13:30] but I do not want it enough data yet: ( [13:31] ogra_, I could emulate the daily image tests on my device? [13:31] There are no data [13:32] mterry, yeah, by using phablet-test-run [13:33] boiko: morning [13:33] demirbas: what do you mean, There are no data? there are 20 phone/tablet model numbers that ubuntu touch works on [13:33] cwayne: morning! [13:34] boiko: It'n night here but good morning anyway. [13:34] FastCode: :) [13:34] mandel: ping [13:34] cwayne: so, did the import tool work? [13:34] oh, that was a reply. [13:35] boiko: nope, I couldn't get it to do anything [13:35] cwayne: ok, so you probably don't have the old telepathy logger history [13:35] cwayne: let me give you a prepopulated db [13:36] sorry for my english is bad. I would say that one x does not have a satisfactory for the device. https://wiki.ubuntu.com/Touch/Devices [13:36] boiko: that'd be amazing! is it just the same stuff that was in demo-assets? [13:36] cwayne: hey, I have a branch that allows customizing the homepage (read-only settings stored in ~/.config/webbrowser-app/settings.conf), if you wanna have a look it’s there: https://code.launchpad.net/~osomon/webbrowser-app/qsettings/+merge/187353 [13:37] cwayne: nope, it has lots of random messages I was sending/receiving on the desktop to try it out :) [13:37] boiko: as long as you don't mind those being used as demo content that works for me :) [13:39] oSoMoN: i'll try it out, thanks! [13:39] oSoMoN: did you notice it seemed to fail autolanding? [13:39] cwayne: yeah, twice, some tests seem to be flaky (the failures are unrelated to this MR) [13:40] demirbas: I'm really sorry, I don't understand much of what you say.what do you mean by "satisfactory"? [13:40] oSoMoN: ah, alright [13:40] popey, ogra_: Rad.io doens't work for me; does it work for you guys? [13:42] it worked for me once in the past [13:43] lool, havent tried today, but it definitely worked on yesterdays image (if i find a station that actually provides the right playlists) [13:43] lool: yes, search for 977, choose .977 The 90's Channel [13:43] * ogra_ installs it [13:43] get loud music [13:45] oSoMoN: so theoretically when would this land in the image? [13:46] ogra_: about my reboot issue, when i reboot lxc-container doesnt mount the system partitions correctly somehow, what if i make an init.d script that runs in the android parts that remounts all partitions on bootup will that work or ? [13:46] cwayne: it should be merged today in trunk, and if everything goes well there will be a new release tomorrow, so at the latest in Friday’s image, possibly earlier with a bit of luck === anders3408|afk is now known as anders3408 [13:47] oSoMoN: wondeful, thanks so much :) [13:47] wonderful even [13:47] cwayne: you’re welcome, it turned out to be quick to implement so I JFDI :) [13:48] oSoMoN: i love it === thostr_lunch is now known as thostr_ [13:53] popey: lool: do you have that track on you devices if so pick something else that was just an example :) [13:54] tmoenicke: hey, is there a bug report to track the fact that the return key on the OSK doesn’t validate the current form in the browser? === chihchun_afk is now known as chihchun [13:57] how do I tell what version of the image is on my device? [13:57] does anybody know if SDK apps automatically use the system proxy settings? [13:58] https://plus.google.com/u/0/100568396676405064966/posts/WrfB21vK1x5 [13:58] davmor2: dunno what you're referring to [13:59] mhall119: just to be sure , you havent send me an email yet ? :) [14:00] popey: it doens't play for me either [14:00] anders3408: ok, I'm putting the box on my desk, right in front of me, so I don't forget it :) [14:00] and the French radios used to work [14:00] * lool reboots [14:00] anders3408: he lies, he doesn't have a desk! [14:00] ok, it's a kitchen table, the point is I work here [14:01] ☻ [14:01] mhall119: it was just to confirm you didnt send the mail to a wrong person :) [14:01] Don't make a cake with it! [14:01] mhall119: I can give you anders3408 address, he's lodging with me :þ [14:01] seb128, system-image-cli -i on the device [14:01] popey: what ? :) [14:01] kidding [14:02] popey: you working on ubuntu touch ? [14:02] popey: nice try [14:03] popey: rebooted, and it worked -- couldn't pause though -- and music-app worked again... [14:03] popey: he has my adress :) [14:03] jibel, thanks [14:05] lool, just fyi i'm listening to radio with rad.io === dandrader is now known as dandrader|lunch [14:06] on maguro ... where even the volume buttons work :) [14:06] ogra_: and how does music-app work for you? [14:06] ogra_: clicking a song from the music-scope or home-scope for instnace [14:07] i dont have mp3's around ... let me try coying one of the alsa test wavs into Music [14:07] ugh.... /system/bin exsist, but /proc/{pid}/root/system/ is empty..... [14:07] anders3408: i do [14:09] ogra_: just grab one [14:10] hello [14:10] ogra_: http://samplemedia.linaro.org/Audio/ [14:11] does anyone know how dialer-app/telephony-service/telepathy-ofono actually initiates a call? [14:11] it apparently does not happen via ofono's Dial() D-BUS method [14:11] but I grepped all three projects and can't figure out how it actually happens [14:11] lool, Front_Center.wav plays fine [14:11] pitti, awe certainly knows the stack [14:11] ofono only sends out signals, no methods are invoked on it at all [14:11] seemingly in an endless loop though [14:12] ChickenCutlass: right, but he's not online, so I was hoping someone else would know as well [14:12] argh ! [14:12] ah, pause workss :) [14:12] haha :) just made a symlink to /system /proc/415/root/ and then tried to run android-chroot which uses sh from the lxc container , it gave kinda funny output : chroot: failed to run command 'sh': Too many levels of symbolic links [14:12] pitti: maybe boiko can help? [14:12] boiko: hello [14:13] ogra_, can the new system images still be downloaded the old way (e.g rsync/adb push/adb reboot recovery)? [14:13] pitti: the one that does create the call is the telepathy-ofono [14:13] seb128, sadly not :( [14:13] at least i didnt dig up a manual way yet [14:14] ogra_, ok, so I've to use the phablet tool? [14:14] seb128, yeah, thats best ... [14:14] boiko: and how does it do that? [14:14] ogra_, hum, k, thanks [14:14] sudo phablet-flash ubuntu-system --channel saucy-proposed -d maguro --no-backu [14:14] +p [14:14] ogra_, danke [14:14] thats what i use here for testing === Namidairo is now known as Namidairo`bnc [14:14] pitti: salem_ implemented this, maybe he can give you more details [14:14] (-d mako for the N4 indeed) [14:15] boiko: thanks [14:15] ogra_: in music-app? cool [14:15] pitti: I'm checking the code in the mean time [14:15] lool, yeah, so wav definitely works [14:15] salem_: how does telepathy-ofono initiate a call? It doesn't seem to call ofono's Dial() method (which appears to be the standard API for that) [14:15] ogra_: saucy-proposed? devel-propose surely ;) [14:15] davmor2, nah, boring, i prefer it saucy [14:15] hi! can someone enlighten me about backporting apparmor? [14:15] (davthey are the same, one is an alias) [14:16] ogra_: yeah so is daily-proposed :D === Namidairo`bnc is now known as Namidairo [14:16] mamenyaka, jjohansen surely can lend you a hand if he finds time ... there should also be links to the patches in the porting wikipage [14:16] davmor2, thats going away though [14:16] ogra_: can you explain me this : /proc/{pid}/root is a symlink to / , but /proc/{pid}/root/system is empty but /system/ is not empty.... why does the symlink not work 2nd time i bootup [14:16] pitti: so it uses ofono-qt, and calls the dial method there [14:17] boiko: ah, I wasn't aware of ofono-qt, thanks (yay, a sixth layer in the stack :) ) [14:17] anders3408, no, stop using android_chroot, it doesnt do what you think it does (and i'll remove it before release) [14:17] ogra_: you guys just don't want the same command use two days in a row do you [14:17] pitti: :) [14:18] anders3408, something in your android must be wrong wrt /system mounting (or probably unmounting ? so there is some entry in mtab or some such which makes it not mount) [14:18] ogra_: im not using android_chroot , im just looking in the folders where the lxc-conatiner mounts the android parts [14:18] boiko: you guys are working on making the call stop if the app gets killed, right? [14:18] davmor2, we dont want to restrict sergiusens creativitiy in inventing new options for phablet-tools :) [14:18] gema: nope, the call should not be killed when the app gets killed [14:18] boiko: or on a way to make the user able to hang up somehow [14:19] mardy: https://codereview.qt-project.org/#change,66513 [14:19] boiko: so how do I hang up a call when the app goes away? [14:19] gema: well, yes and no :/ [14:19] ogra_, thank you, I have a 3.4 kernel for sony and a kernel 3.0 for samsung, any suggestions on git checkout ? [14:19] boiko: I am tired of having to switch off my phone to end calls [14:19] barry, pong [14:19] boiko: it seems a bit cumbersome [14:19] gema: design has planned for an indicator to show there is a call and to make it possible for the user to go back to the call, but that (as designed) won't be ready for 13.10 [14:20] pitti, sorry, I will read the backlog, just a sec [14:20] gema: you don't need to switch off the phone, just launch the dialer-app again and hangup the call [14:20] boiko: ok, will try that next time [14:20] gema: we are discussing alternatives to that, but no conclusion yet [14:20] boiko: this should have a huge release note , btw [14:20] boiko: in flashing lights x) [14:20] gema: but there will probably be something to go back to the call [14:21] boiko: ack [14:21] boiko, salem_: so http://paste.ubuntu.com/6154648/ is what happens to/with ofono when I do a call; i. e. just signals, no method calls [14:21] mamenyaka, not beyond whats on the porting wiki, sorry [14:22] pitti: let me check what ofono-qt does internally [14:22] ogra_, okay, I understand [14:22] boiko: OfonoVoiceCallManager::dial does call Dial() [14:22] pitti: yep [14:22] boiko: just curious that I don't see that in dbus-monitor === mzanetti is now known as mzanetti|food [14:22] ogra_, last thing: any info on why suspend/resume isn't working? [14:22] mamenyaka, there should be links for the different kernel versions [14:22] davmor2, ogra_ that's a backend change and stgraber sent an email 2 weeks ago for the channel name change [14:22] ogra_: i have a init.qcom.post_fs.sh script that remounts system partition on bootup... can it be that which conflicting something ? [14:23] sergiusens, we're just joking :) [14:23] pitti, yep, it isn't shown in dbus-monitor for some reason. I dont know why to be honest. but I am pretty sure it calls Dial() [14:23] anders3408, yeah, try to make it a no-op for a start ... see what happens [14:23] no-op ? [14:24] no operation [14:24] pitti: do you have an apparmor denial in /var/log/syslog? there's a chance that apparmor dbus mediation could be blocking the message from dbus-monitor [14:24] ogra_, after the blockade is lifted I am coming up with new options though ;-) [14:25] ogra_: but thats a part of starting the modem ... === alan_g is now known as alan_g|tea [14:25] bfiller: are you responsible for the camera-app? [14:26] sergiusens, i hoped so ! :) [14:26] anders3408, so commment only the mount bits [14:26] jdstrand: my team is yes [14:28] bfiller: so, audio has an out of process daemon that everything goes through (puselaudio), does the camera? [14:28] it should mount /system as -rw now by default, and i have commented those remount parts out. lets see how it goes [14:28] bfiller: hi! Do you know if there are working ubuntu-keyboard AP tests already, or it's still in progress for now? [14:29] bfiller: perhaps that is a question for ChickenCutlass... [14:29] Mirv,didrocks: the process listing from https://launchpad.net/~ubuntu-unity/+archive/daily-build/+build/5048682 is https://pastebin.canonical.com/98017/ [14:30] Mirv,didrocks: if that's enough to diagnose the problem, I suggest cancelling the build [14:30] tyhicks: ah, that would explain why I don't even see them with "sudo dbus-monitor"? [14:30] tyhicks: no, I don't [14:31] jdstrand: not sure, gusch may know [14:31] tyhicks: nor apparmor denials in sudo dmesg [14:31] pitti: hrm... you should definitely get a denial in the syslog if it is due to dbus mediation [14:32] sil2100: as of rev 43 the autopilot tests should be working, talk to veebers about how to run them [14:32] sil2100: you need to stop unity8 and stop maliit-server then run maliit-server -testability [14:32] veebers: ping [14:32] pitti: since there's no denial, I don't think apparmor is blocking the message [14:32] sergiusens: no rush obviously, but heres an MR for you guys :) https://code.launchpad.net/~cwayne18/phablet-tools/demo-setup-fixed/+merge/187523 [14:32] bfiller: ah, thanks :) [14:33] I'll try that [14:33] tyhicks: yeah, I tried with /e/i/a teardown, and still no methods there [14:33] jdstrand bfiller not sure, but camera uses the android media-server for all audio/video (and that does not use pulseaudio) [14:34] pitti: tried here, I don't see the method call either [14:34] bfiller: hah! Seem to be working! At least things are 'happening' now, thanks [14:34] xnox, so i'm looking at https://code.launchpad.net/~xnox/ubuntu-seeds/ubuntu-touch.saucy/+merge/186803 ... these removals from the touch seed dont look right (and i know a bunch of apps which will break without these bits) [14:34] boiko: ok, thanks for the heads-up; so Dial() method should be used, I just don't see it [14:34] that helps [14:34] gusch: but we aren't going to rely on binder forever for camera, correct? [14:35] gusch: I thought the plan was for 14.04 to not use binder for camera [14:35] tedg: should the messaging indicator change when there is a message arrived? [14:36] jdstrand: for details of that, you should ask jhodapp|afk [14:36] davmor2, Yes, I think we're waiting on a unity branch to land there. [14:36] gusch, i think rsalveti made some changes to the camera service when pulse landed [14:36] tedg: ah great thanks for the info :) [14:36] * jdstrand steps into a meeting === alan_g|tea is now known as alan_g [14:38] cjwatson: ogra_: do note that ubuntu-touch meta package becomes dependant on ubuntu-sdk-libs meta package, or maybe that dependency should be declared in the lifebuild instead. That branch should be viewed together with debian/control changes in the ubuntu-touch-meta package, see https://code.launchpad.net/~xnox/ubuntu/saucy/ubuntu-touch-meta/xnox/+merge/186802 [14:38] jdstrand: then spends 10 minutes trying to wipe the meeting of the bottom of his shoe [14:38] ogra_: plus I need to catch up on the recent seed changes ( i think i am a couple of revisions behind) [14:39] yeah [14:39] we had and still have a lot of them currently [14:40] xooh, yeah, that looks fine then [14:40] ogra_: at the moment ubuntu-touch & ubuntu-sdk are disconnect, now i'm bringing in guarantees that runtimes match across sdk & actual touch images =) [14:41] jdstrand: yeah, for 13.10 we'll still be using the camera-service, goal for 14.04 is to just use the HAL directly [14:41] xonyeah, i was just missing that last commit [14:42] xnox, ^^ [14:44] mterry: any idea on our welcome screen issue? [14:44] sergiusens: updated the mr to use fileutils, also it seems that the CI failed from some bzr error [14:46] cwayne, yeah, glitch... it will retried with your new commit [14:46] sergiusens: ack, thanks [14:49] hum [14:49] is changing gsettings's com.canonical.powerd activity-timeout supposed to change the time before the device screen power off? [14:51] ogra_: i have a modded powerd.deb i need to use , its made by sforshee, that enables support for backlight from leds folder instead... but how can i add that so people doesnt need to install it manually ? if i let it push to /data , and in a init.d script runs dpkg -i /data/powerd.deb on bootup will that do ? [14:51] sounds not like a good ide btw :) [14:53] cwayne, right... what does this say: dbus-send --system --print-reply --dest=org.freedesktop.Accounts /org/freedesktop/Accounts/User32011 org.freedesktop.DBus.Properties.Get string:org.freedesktop.Accounts.User string:BackgroundFile [14:53] cwayne, also, I was thinking, for future proofing, you may want to use /org/freedesktop/Accounts/User$UID in your script [14:54] mterry: ah, that's a good idea [14:56] seb128: yes, in theory. But it doesn't work in the read-only images because the new setting can't be written. [14:56] doanac, hey, wrt to your aa-click rule for phablet-config, should I wait for autopilot to land? [14:56] anders3408: I've proposed a merge for those changes, but no one has reviewed and approved them [14:57] ogra_, when trying to git clone git://kernel.ubuntu.com/jj/ubuntu-saucy.git, I get remote: error: refs/heads/master-next does not point to a valid object! [14:57] ahh :) i hope it will be approved :) [14:57] sergiusens: i forgot i had an MP out for that :) === dandrader|lunch is now known as dandrader [14:57] sergiusens: I suppose we should, just so we don't give someone the wrong impression [14:57] rsalveti: does "HAL directly" direct access to devices? I'm asking because that is somewhat problematic for enumerating those devices, but we have a plan for that in lxc-android-config for shipping apparmor policy snippets. But more importantly, we will need to be able to give a contextual runtime prompt for recording (like we want o with online accounts) [14:57] rsalveti: and I'm not sure how that would work with direct hardware access [14:58] sforshee, what new setting? [14:58] rsalveti: eg "Foo wants to access the camera. y|n", then cache the result appropriately [14:58] sergiusens: i did just discover a dumb bug I caused last week in phablet-test-run: https://code.launchpad.net/~doanac/phablet-tools/test-run-collect-error/+merge/187531 [14:59] jdstrand: yeah, need to think more about that one, as we don't want more than one user using it at the same time as well [14:59] sforshee, oh, I see, we write to the phablet user settings and powerd runs as root :/ [14:59] rsalveti: might also want a visual notification when recording in the background so users know they are being recorded [14:59] jdstrand: it would probably be better to bring the service itself, so we can protect it [14:59] right [14:59] seb128: yeah, I think gsettings wasn't the right way to go for that [14:59] will think a bit more about the camera architecture and let you know [15:00] sforshee, what would be the right way? [15:00] seb128: unity is supposed to handle the timeout, but I don't think that's going to land this cycle [15:00] sforshee, so basically that option is not going to work this cycle? [15:00] rsalveti: I tend to agree. if there is some performance benefit to giving an app direct access, the service could pass an fd and that would work via apparmor delegation. if you want to go that route-- would want to talk to jjohansen [15:00] right, sure [15:00] thanks [15:01] seb128: I don't know. I guess there could be a dbus api for the setting, but powerd has nowhere to "remember" it so it would need to be set each boot [15:01] rsalveti: do you mind if I create a bug on this? if that is ok, against what package/project [15:01] seb128: but then that api will have to go away when unity if finally handling the timeout [15:01] seb128: maybe ricmm can better say what might actually land this cycle [15:01] sforshee, ok... is there a bug handling the fact that unity is going to take over that setting? [15:02] ricmm, hey [15:02] sforshee, thanks [15:02] jdstrand: please open it against https://launchpad.net/touch-preview-images for now, and assign it to me [15:03] rsalveti: awesome, thanks! :) [15:03] anyone? error when trying to clone git://kernel.ubuntu.com/jj/ubuntu-saucy.git [15:04] seb128: sup [15:05] ricmm, do you know when is unity8 going to take over the "turn off/lock screen after "? [15:05] seb128: theres no bug, unity will take over but its not scheduled yet [15:05] ricmm, ok, let me open a bug about that [15:06] Saviq: ^ === mzanetti|food is now known as mzanetti [15:10] ricmm, Saviq, sforshee: https://bugs.launchpad.net/ubuntu-system-settings/+bug/1230345 [15:10] Ubuntu bug 1230345 in Unity 8 "Unity8 should control the display "lock after idle" (with a way to configure the delay)" [Undecided,New] [15:12] lool, Did you notice my comments on the fdo-application branch? Does that address your concerns? [15:12] looking [15:16] tedg: I dont understand the dbus design of this I'm afraid, mind reexplaining this to me? I was indeed expecting we'd add have a confinment update to allow app id com.xyz.foo to listen on that dbus name [15:16] hum [15:17] tedg: indeed I see now that the code uses org.freedesktop.Application (headdesk), but I don't see how just using the app name is unique? [15:18] stgraber, pitti, ogra_: charge history is buggy on the device because upower can't write to /var/lib/upower ... is that a lxc-android-config tweak (e.g adding the directory there)? [15:18] seb128, perferably the single files [15:18] ogra_: the names are random and the directory is empty by default, so in that case, the whole directory needs to be made writable [15:18] adding a whole dir a) makes the whole dir writable and b) slows down the boot even more [15:18] stgraber, ah, thanks [15:19] seb128, yeah [15:19] lool, It's not unique, as we're not using a bus name. We're always sending to the unique name on the bus. [15:19] (to lxc-android-config) [15:19] lool, app name is only used for the object path. [15:20] lool, We don't want confined apps to register names because then apps can determine what other apps are running. [15:20] seb128: so you want "/var/lib/upower auto persistent none none" in writable-paths (lxc-android-config) [15:20] tedg: well we could firewall them with apparmor [15:20] stgraber, ok, I'm going to mp that, thanks [15:21] tedg: I mean protect access to listing the bus [15:21] lool, Not really, because they need to be able to request all the names on the bus to get to the services. So we'd have to add support for sending a different list depending on who asks. [15:21] tedg: hmm why do they need to request all names? [15:23] tedg: also, I did read the part about generating the object path wrong [15:23] lool, Hmm, I guess they don't *need* to as the could always ask if a name had an owner, but we'd have to verify every lib handled that correctly. [15:23] tedg: but it still feels wrong that this is declarative [15:24] tedg: could we actually express the object path thing in apparmor? [15:24] tedg: maybe that's enough [15:24] stgraber, is there a vcs for lxc-android-config ? [15:25] tedg: that is, don't allow Open on anything else than the object path that corresponds to your app id on org.freedesktop.Application [15:25] lool, The path? Why would we need to express the object path in app armor? [15:25] seb128, no [15:25] lool, The app is exporting that, so there's no limitation on unconfined stuff talking to confined stuff. [15:25] seb128, use UDD and assign to me, i have other stuff in there [15:25] tedg: right, but I want to stop a confined app from pretending to be another app [15:25] ogra_, thanks [15:26] tedg: if we can stop other apps from connecting there, that would be enough [15:26] ogra_: can you review sforshee changes to powerd ? :) [15:26] lool, They could all have an app "foo", right? Why is that an issue? [15:26] cjwatson: sure, please go for cancelling [15:26] lool, If it's just an object path on their connection [15:27] didrocks: I think you should be able to do it [15:27] tedg: can we go a quick hangout? I'm sure either I don't explain myself right or I don't understand you right [15:28] lool, K [15:28] Pff and G+ wont load [15:28] seb128: ubuntu:lxc-android-config [15:29] tedg: see calendar [15:30] stgraber, ogra_: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/saucy/lxc-android-config/saucy/revision/103 [15:31] cjwatson: indeed, done [15:31] seb128, thx === dednick is now known as dednick|lunch [15:31] gema, ^ that's the fix for the last charge = n/a but you mentioned earlier [15:35] cwayne, so what was the output of the dbus-send command? [15:35] xnox, w-flo ported to a yaffs device i think, that might probably help you [15:36] (he doesnt seem to be around atm) [15:36] mterry: ah sorry, i was reflashing, i'd need to reinstall your lightdm right? or is it in today's daily-proposed? [15:36] cwayne, it's only in saucy-proposed, so probably not [15:36] cwayne, but you can download from saucy-proposed instead of building yourself [15:39] mterry: would i need just lightdm or also ubuntu-touch-session? [15:41] cwayne, both [15:41] cwayne, but you have to build u-t-s [15:41] ogra_: firstboot : EXT4-fs (mmcblk0p19): mounted filesystem with ordered data mode. Opts: noauto_da_alloc but 2nd boot : EXT4-fs (mmcblk0p19): mounted filesystem with ordered data mode. Opts: errors=remount-ro what is causing the diff ? [15:41] mterry: cool beans, branch? [15:42] mmm beans [15:42] aw man now i want beans [15:42] scratch that, i want a burrito [15:42] is it true [15:42] that ... [15:42] first bootup shows that its init.find.rc that mounts /system but 2nd bootup it looks like init.find5.rc doesnt mount /system becasue it already mounted ... [15:42] * Laney creates suspense for a boring question [15:43] anders3408, hmm, looks like the first one uses some android options [15:43] is it true that the nexus 7 runs loads slower when it's not on power? [15:43] building stuff seems to take forever when it's not plugged in [15:44] yes ogra_ its from the ramdisk from the android part, but 2nd bootup it mounts from /etc/fstab/ which somehow fails [15:44] cwayne, lp:~mterry/session-manager-touch/lightdm-switch [15:44] cwayne, sorry for delay, am shaving while chatting :) [15:45] ah i already have that branch [15:45] mterry: lol [15:46] Hello hello? [15:48] oh darm.... i see the issue now ogra_ [15:48] rsalveti: fyi, bug #1230366 [15:48] bug 1230366 in touch-preview-images "Please provide Ubuntu camera service that integrates with trust-store" [Undecided,New] https://launchpad.net/bugs/1230366 [15:48] Anyone can help me? [15:52] Hello? :D [15:54] ElectroPug, if you have a question, just ask it [15:55] jdstrand: thanks [15:56] * cjwatson finally gets https://click.readthedocs.org/en/latest/ updating again (used to be click-package.readthedocs.org, and I had to fix readthedocs.org not to crash on bzr branches with tags ...) === dednick|lunch is now known as dednick [15:56] oh alright, well on our device (hTC one S c2) somebody managed to compile Ubuntu touch, the problem is that the screen doesn't work because of some chroot issue, the developer says that he can't get chroot access. Does anyone know how he could accomplish that? [15:58] The system works according to him it's just the screen === gatox is now known as gatox_lunch [15:59] tedg: which channel you said? [15:59] ogra_: is it var/lib/lxc/android/pre-start.sh that is responsible for mounting partitions in lxc-container correct ? and can i make some device specific changes in the lxc-container stuff ? [15:59] sergiusens: sorry to be annoying, but this would be super-helpful for our team right now: https://code.launchpad.net/~cwayne18/phablet-tools/demo-setup-fixed/+merge/187523 [15:59] like how it should mount or something ? [16:00] lool, I usually start at #ubuntu-security because I can remember it, but it redirects to #ubuntu-hardened [16:01] right [16:01] lool, I do find it rather ironic that the security channel does a redirection [16:01] cwayne, done [16:01] sergiusens: thank you sir [16:02] * cwayne adds 1 to the countless beers he owes sergiusens [16:02] np [16:04] tedg, upstart app launch sends an event directly to zeitgeist, right? we don't need the datahub to pick up the gvfs signal? [16:05] well on our device (hTC one S c2) somebody managed to compile Ubuntu touch, the problem is that the screen doesn't work because of some chroot issue, the developer says that he can't get chroot access. Does anyone know how he could accomplish that? [16:06] System boots but screen doesn't work [16:06] mhr3, For start/stop, yes [16:06] dpm: ping [16:06] mhr3, And I don't think there'll be a gvfs signal after the upstart-app-launch stuff lands. === psivaa is now known as psivaa-afk-bbl [16:09] seb128: excellent thanks! [16:09] ricmm: seems we actually do need a SF solution as we'll have SF on maguro [16:09] ricmm: can you directly discuss this with didrocks/asac? [16:10] thostr_, just make the Mir guys fix maguro ;) [16:10] ricmm: in a meeting, happy to discuss and have the details about it just afterwards [16:10] ogra_: also a solution. point is we need A SOLUTION right now [16:10] yeah [16:11] ricmm, didrocks: please discuss this [16:12] ogra_: what's the quickest way to modify the touch script? or shall automate unpackging, updating, repacking with abootimg? [16:12] thostr_: discuss what? [16:12] thostr_: URL on SF ? [16:12] ricmm: yes [16:13] ricmm: that we need a SF solution in addition to mir code [16:13] what exactly do I need to get from didrocks/asac for this? [16:13] ricmm: basically what you described before as throwaway code [16:13] xnox, for fiddling with it in emu ? [16:13] thostr_: ah, none of that is related to them if thats what you are thinking [16:14] ogra_: yeah. [16:14] its non-existant code right now, it would require developing the same solution over in the android app manager (SF) [16:14] well on our device (hTC one S c2) somebody managed to compile Ubuntu touch, the problem is that the screen doesn't work because of some chroot issue, the developer says that he can't get chroot access. Does anyone know how he could accomplish that? [16:14] ricmm: correct. that's why you said you prefer not to do that and instead focus on the mir solution [16:15] ricmm: so, now it seem we need a solution working with mir and one for SF [16:15] ricmm: or do I get that wrong? [16:15] thostr_: who said that we will ship nexus with SF ? [16:16] ricmm: didrocks/asac that's why I said you should talk to them [16:16] tedg: :-) [16:17] ogra_: my interractive shell is about to become adding statements to the touch script and rebooting. [16:20] loicm, Hey, do you have a link to your qtubuntu plugin so I can play with it some? [16:23] tedg: you mean the stuff I'm doing or the current qtubuntu branch? [16:24] loicm, The stuff you're doing [16:24] loicm, For the URL passing. [16:24] sergiusens: how often does phablet-tools get released? [16:27] cwayne, as often as it's added to the release list [16:27] ogra_, can you add a phablet-tools release request? [16:27] tedg: so there's nothing to play with yet, I only got a hackish standalone Qt DBus based app and a custom build of fdo-application-open.c to test that the communication works [16:27] sergiusens, sure ... after the meeting i'm currently in [16:28] tedg: but I still have to add support for a custom type for the platform_data arg "a{sv}" [16:28] tegdg and the communication works btw [16:28] tedg: then exposing all that as a library + qml binding is easy [16:29] tedg: I4ll ping you when it will be ready to play with [16:29] ogra_, great [16:29] ogra_, also, if you get the seeds in, give me a ping please [16:29] sergiusens, will do === chihchun is now known as chihchun_afk [16:31] loicm, Great, thanks! === Namidairo is now known as Namidairo`bnc [16:34] lool / loicm: can you make the morning (for me) sync meeting tomorrow? I have a few topics to cover [16:37] doanac: ping [16:37] cwayne: hey [16:38] xnox, did you update the seee merge already ? [16:38] doanac: hey, were you the one working on the customization test suite CI stuff? [16:38] *seed [16:38] i.e. getting it run automatically [16:38] cwayne: yes. i have a branch that can do it [16:38] cwayne: was waiting to here if it had been approved to add to our daily smoke results [16:40] didrocks: I've added system-image to the "ask" spreadsheet now; AFAIU there are ubuntu-system-settings changes that go with this but don't have to land at the same time, is that your understanding too? === gusch is now known as gusch|dinner [16:41] doanac: who's that waiting for? asac? [16:42] cwayne: yes. i think there was supposed to be a discussion with him [16:42] slangasek: I'm unsure about I. I don't know of any change on ubuntu-system-settings for the update panel. barry: any hint? [16:43] didrocks: barry's afk at the moment, which is why I asked you :) [16:43] slangasek: FYI, we have the latest system-settings already [16:43] any one know of any issues with the touch screen that would keep it from working? [16:43] cwayne, since we landed the tz data stuff new landings to phablet-tools are blocked [16:43] (apart from rev 379 which is about testing) [16:43] * didrocks looks at proposed MP [16:43] didrocks: does that already have the support for download progress etc.? [16:44] slangasek: for more than a month yeah ;) [16:44] jdstrand, ping [16:44] sergiusens: agh, ok [16:44] didrocks: ok, so we just need to land the system-image side, got it ;) [16:44] slangasek, you didnt use it yet ? [16:44] salem_: yes? [16:44] slangasek: right, just send me the signal! :) [16:44] ogra_: download *progress* [16:44] cwayne, we need to either revert that or something... did you already talk to stgraber? [16:44] slangasek, oooh ! [16:44] * ogra_ wants !!! [16:44] cwayne, you can also request landings for our broken process on #ubuntu-ci-eng [16:45] didrocks: the dbus signal? :) [16:45] sergiusens, well, pitti is the one who wants it to work differently, not stgraber [16:45] slangasek: yeah, updateProgress(int) ;) [16:45] ogra_: no, no seed update. will look at it later tonight, i need to leave for volleyball on time. [16:45] ok :) [16:45] sergiusens, ^^^ [16:45] ogra_, I didn't name names ;-) [16:45] slangasek: so, ok, is it ready to launch now? we have images for barry or you to tests if needed [16:45] (I guess in the proposed channel) [16:46] sergiusens, so seed change will wait until tomorrow [16:46] jdstrand, hi, I am reading a bug you reported: 1226844, and my question is what is the problem of allowing in apparmor access to org.ofono.NetworkRegistration.GetProperties ? [16:46] didrocks: he's running through some final testing AIUI [16:46] sergiusens, we changed to much since the MP was made :( [16:46] didrocks: there's a non-user-affecting bug in the download-manager that makes the testsuite run very slowly [16:46] jdstrand, it contains no private information [16:46] slangasek: ok, can you please update as soon as the test finish? we can publish it tomorrow [16:47] jdstrand, also, I think you forgot to attach the qml example file. [16:47] didrocks: "please update" - the spreadsheet? or upload to -proposed? [16:47] sergiusens: i guess we can revert it, or could we have it push those files to /etc/writable/[timezone,localtime] instead? [16:47] slangasek: if it's tested, update to proposed tonight is fine (just not tomorrow morning, we'll try to have one slot for a big change alone) [16:47] cwayne, no, since thats still in discussion between pitti and stgraber [16:47] didrocks: ok [16:48] ugh [16:48] didrocks: though, the whole point of -proposed is to let devs upload straight there, and you guys can decide when to let things through to the images... [16:48] who'd have thought something as simple as changing a timezone would be such a PITA [16:48] cwayne, i didnt revert your three files yet ... but the whole thing will likely change again so the question is if it makes sense for you to land what you have right now [16:48] cwayne now you know why I am not pushing changes in so quickly [16:48] sergiusens: yeah, fair enough :) [16:49] ogra_, I say we push [16:49] sergiusens, ok [16:49] sergiusens: ogra_: i'm happy for us to revert the changes if it helps get other (not-quite-as-broken) stuff in [16:49] ogra_, if it's reverted, cwayne will have a fix ;-) [16:49] yeah, thats what i thought [16:49] slangasek: well, we had a long discussion about it, but it needs more discussion I guess ;) [16:49] :P [16:49] salem_: I'll need to regenerate that qml and apparmor profile. [16:50] didrocks: how come? [16:50] sergiusens, cwayne btw ... are we by chance talking about https://launchpad.net/ubuntu/saucy/+source/phablet-tools/1.0+13.10.20130925.1-0ubuntu1? [16:50] seems that landed [16:50] slangasek: let's discuss when I won't have 100 pings (and with asac and all other guys) please ;) [16:50] slangasek: added your slot for upload [16:51] ok [16:51] slangasek: so feel free to test with the ui and upload [16:51] ogra_, cwayne yes we are [16:51] slangasek: the UI works with the mock for a month without changing the code, so it should be ready [16:51] lol [16:51] so much about the spreadsheet :P [16:51] ogra_, cwayne then the request is just to land phablet-demo-setup fixes for read only images [16:51] heh, yeah [16:51] ogra_, not sure who triggered, but whatever [16:52] lol [16:52] looks like Mirv did on lool request [16:52] at least thats what the spreadsheet says [16:52] well, saves us one :) [16:52] * ogra_ likes that [16:56] sergiusens: Hey ! the docviewer is ready to be click packaged, but I don't fully understand the script you just gave me (http://paste.ubuntu.com/6145304/). The target path must be the path of the current docviewer source code ? [16:57] salem_: but that interface actually does expose more information than the app needs. all the app really wants to now is 'am I online'-- that shouldn't give the app the serial number of the device [16:57] s/now/know/ [16:57] salem_: dbus-send --system --print-reply --dest=org.ofono /ril_0 org.ofono.Modem.GetProperties [16:58] or the revision, etc [16:58] oh, that was the wrong dbus command [16:58] Chocanto, I don't understand your question [16:59] Chocanto, just tell me where the deb module lives and remind me where the source was [16:59] sergiusens, i think that needs a few seed changes first [16:59] jdstrand, I see. that's not good indeed [16:59] salem_: right, dbus-send --system --print-reply --dest=org.ofono /ril_0 org.ofono.NetworkRegistration.GetProperties gives away the carrier, contry code, cellid, etc, [16:59] sergiusens: The deb modules are on the ubuntu-core-apps PPA [17:00] sergiusens, or going along with it at least [17:00] sergiusens: If I understand your question === alan_g is now known as alan_g|EOD [17:01] ogra_, ? [17:01] ogra_, for the doc viewer? oh, for poppler [17:01] ogra_: im trying to get some more debug information from the lxc-container , and i have editted lxc-android-config.conf and lxc-instance.conf, added -l DEBUG to the 2 lines that contains lxc-start , but i dont see any new loggings .... [17:01] sergiusens, docviewer need the pdf plugin and stuff [17:01] yeah [17:02] anders3408, logcat might be more informative [17:02] ogra_: Yes, it needs two modules === gatox_lunch is now known as gatox [17:02] ogra_: do you mind including robru in the phablet-tools and other landing like that? [17:02] Chocanto, what's the DEBs name in the ubuntu-core-apps PPA? [17:02] didrocks, you mean on the spreadsheet ? [17:03] sergiusens: qtdeclarative5-poppler-qml-plugin and file-qml-plugin [17:03] ogra_: yeah, spreadsheet + publishing stuff [17:03] Chocanto, both of those are in that PPA? [17:03] sergiusens: Yes, both [17:03] stgraber: are you interested in a phone not starting after an apt-get install + shutdown -r now [17:03] ricmm: do you have time now to discuss? [17:03] stgraber: if yes I have one [17:04] didrocks, publishing stuff ? (i have no control over phablet-tools i think) [17:04] stgraber: it's stuck with blackscreen (I installed test packages, but shouldn't affect boot) [17:04] ogra_: yeah, just sync with him to get phablet-tools landed on dailies [17:04] stgraber: I see sensorservice using a lot of CPU [17:04] lool: can you still get into adb? [17:04] stgraber: yes [17:04] apt-get install foo [17:04] didrocks, pahblet-tools doesnt land on images ;) [17:04] apt-get: error while loading shared libraries: libapt-pkg.so.4.12: cannot open shared object file: No such file or directory [17:04] ogra_: right, but aren't you the one adding it to the landing plan? [17:04] didrocks, but thanks, will ping him for a buiold [17:05] didrocks, i am ... [17:05] didrocks: I do [17:05] ogra_: that's why I pinged you, just if you want that early, ping him, it's an easy one! ;) [17:05] lool: ok, so pretty clear sign of fs corruption... the annoying part is that it's almost impossible to debug since any log that may be useful will also be missing [17:05] didrocks, definitely, and thanks ! [17:06] ricmm: https://plus.google.com/hangouts/_/8e573ae15f745c78a78ac5ab65bed5ecb3b73118?hl=fr [17:06] ogra_: thanks to you! [17:06] lool: I guess I'll have to go through the init scripts to see if anything in there bypasses the standard shutdown sequence and if not, what's preventing a clean unmount of all partitions [17:06] stgraber: Ok; thanks [17:07] ricmmmmmmmm ;) [17:07] lool: in theory, the container should first be killed, then all bind mounts and android partition unmounted, leaving only / and /userdata, after that / should be remounted ro, same goes for /userdata and then the system should reboot [17:07] lool: /userdata won't be unmountable since / is coming from it and / can't ever be unmounted [17:08] stgraber: maybe we will see some warnings that it's not happening on the console on shutdown [17:08] ogra_: the thing is /system is mounted correctly , but lxc.rootfs/system doesnt mount correctly and i cant trace it to why [17:08] lool: the clean fix for that is slangasek's repartitioning work which will split / and /userdata to two separate partitions [17:08] right [17:08] anders3408, because lxc/rootfs is not what you think [17:08] lool: yeah, if only we could see the console, because we sure won't have the logs... [17:08] stgraber: I gave a heads up to Steve about this issue, but I don't think we can land this in the next couple of weeks [17:08] anders3408, thats just the place to unpack the initrd, nothing gets moounted there [17:09] lool: no and it won't fix the issue on grouper and the ports [17:09] rue [17:09] true [17:09] lool: one way around this would be to have a shutdown job running before the reboot which attemps to remount / ro, /userdata ro and ends with a sync call [17:09] stgraber, reboot into recovery without powering off and you have the console output in /proc/last_kmsg [17:09] ogra_: that'd be useful if sysvinit was logging on there [17:09] it gets kept in the ramconsole as long as power is there [17:10] jdstrand, I am going through the network manager api and I am able to get through dbus if the active connection is wifi or gsm for example. [17:10] stgraber: right... we need to try to solve it for all devices, just not the "good" ones that don't need loopback [17:10] stgraber, we should make it log there on touch ;) [17:11] mfisch: I could do half an hour later, but not at the current time [17:11] slangasek: do you happen to know anything about unmountfs and how it's supposed to behave when / depends on another mount? My hope would be that it'd try to force / read-only, do the same with any outstanding mount, then call sync. But we appear to have evidence that it doesn't [17:11] mfisch: did you see the package autolanded BTW? [17:11] ogra_: so where does the lxc-contanter starts those init.find5.rc scripts from ? [17:12] salem_: ok, I attached the files needed to reproduce [17:12] jdstrand, just wondering if we couldn't use that somehow. I wanted to avoid changing ofono-qt as we are not upstream. [17:12] salem_: well, so the network manager api isn't any better. if gives up all kinds of things [17:12] mfisch: so essentially your changes will go into archive when we publish the "misc" stack of packages; we will see if any change needs publishing, but if you need something you need to give Landing team a heads up; this is all taken from trunk, so upstream merger is your friend for this part (top approving mp to get them landed etc.) [17:13] salem_: what is interesting is that the libs can figure out if they are online without network manager on a desktop system or on the nexus 7 [17:13] anders3408, the content of /rootfs gets copied to the proc space you already know ... then ./init gets executed ... thats all lxc does, eveything else that happens on container boot is then handlled by androids init [17:13] salem_: (ie, when ofono isn't around) [17:14] salem_: even without the denied access to network manager. it just pokes around in /proc and /sys and handles the NM denial fine [17:14] jdstrand, you mean reading direct from /proc and /sys? [17:14] jdstrand, ah ok [17:15] salem_: yes. but, doing the same thing on the nexus 4 (ie, ofono is in the mix), all of a sudden /proc and /sys aren't enough, and an error is displayed [17:15] but in /rootfs the system folder is empty, and also is the folder in /proc , so the android init will fail when ever it should use any files from /system, but from / then /system folder is as it should be [17:16] salem_: I just now verified this. nexus 7: works. nexus 4: doesn't work [17:17] anders3408, androids init is the one caring for the mounting too [17:18] lool: okay, I'll move the meeting [17:18] mhr3: Any idea why would MusicFilterGrid.qml never show more than 3 items from the model? (and if that's how it works, what's the trick to add that tick and make it unwind its content? :) ) [17:19] Chocanto, can you see if this works for you? https://code.launchpad.net/~sergiusens/ubuntu-docviewer-app/wrapper--/+merge/187571 [17:20] Cimi, poke! If you haven't started with the dynamic-language change for the welcome wizard, I have time to look at it [17:20] mterry: pinagling [17:20] cwayne, hello! [17:20] cwayne, sorry I was offline, had to head to starbucks again [17:20] mterry: so, i did that dbus-send you told me to do, and it returns the right value, but the welcome screen background is still unchanged [17:21] mterry: no worries! i had plenty of people to bother while you were gone :P [17:21] stgraber: by 'umountfs' you mean the init script? If it tried to force / read-only, it would certainly fail [17:21] cwayne, hmmm... So that means bug in unity8 itself [17:21] cwayne, it was working last time I tested it [17:21] mhall119, as well for you https://code.launchpad.net/~sergiusens/ubuntu-docviewer-app/wrapper--/+merge/187571 [17:22] mhr3: well.. I think expandable: true might do the trick, but I've upset my phone with multiple unity8 shell replaces, I have to reboot it ;) [17:22] mterry: last time i tested it, i was able to at least do it manually with that dbus-send, but i couldnt from upstart [17:22] now i can run it from upstart, but it changes nothing [17:22] mhall119, that is a better option than doing the wrapper script [17:22] mterry: but hey,t he good news is this counts as further testing of the ubuntu-touch-session package right? :P [17:22] cwayne, last time you tested it, you could see background change, right? [17:22] cwayne, :) [17:22] mterry: right [17:22] last time as in a couple weeks ago, pre lightdm [17:23] cwayne, ok, good. That's confirmation that we used to work. lightdm shouldn't affect our ability to read the value [17:23] jdstrand, ok, reading the code now I see that this issue isn't ofono-qt related, as the qofonowrapper.cpp in qtdeclarative5-systeminfo-plugin handles all the ofono dbus communication internally. [17:23] robru, bzoltan: any word on that qtaudioengine issue? [17:23] salem_: cool, feel free to reassign-- I took my best guess :) [17:23] mterry: so who can we bug to fix this then? [17:24] cwayne, me....... [17:24] slangasek: right, the init script [17:24] salem_: I guess it needs to do its error handling similar to how it does error handling with NM [17:24] mterry: PING I FOUND A BUG [17:24] * mterry quits [17:24] lol [17:24] mterry: this is all cus of that time you didn't invite me to your birthday party. [17:24] cwayne, guh! [17:25] I had no idea of the cost! [17:25] jdstrand, hehe ok, np. I will discuss this issue with the other networking guys. thanks for the report. [17:25] stgraber: so yeah, I don't think the umountfs script is that clever... the code predates /proc/mountinfo ;) [17:25] cwayne, what is severity of this problem for ya? [17:26] ogra_: does lxc-android-boot.conf runs before or after the container is started ? [17:26] anders3408, before i think ... but it doesnt do much after first boot [17:26] mterry: it's not SUPER-URGENT but it'd be good to have I think [17:27] greyback, heyo! About OSK in welcome-wizard... So I added the print() that you suggested to the surface manager, and I never see it print anything [17:27] greyback, so looks like maliit-server isn't creating the surface or we aren't noticing [17:27] greyback, I tried delaying the start of maliit-server, but no luck [17:27] ogra_: i just wanted to try and add these lines somewhere before it starts the container but after the /proc/pid/root is made [17:27] containerpid="$(lxc-info -n android|grep pid|sed 's/^pid:.* //')" [17:27] cp -r /system/* /proc/$containerpid/root/system/ [17:27] greyback, is maliit-server hardcoded to only talk to unity8 or some such similar thing? [17:28] anders3408, that wont work [17:28] slangasek: so I think the easiest way to fix touch is to ship an extra umountdata init script which does the same as umountroot but for /userdata [17:28] anders3408, find out whats wrong with your android [17:28] instead of adding hacks on the ubuntu side [17:28] slangasek: I'll do some tests here, but I'm pretty confident that umountfs should take care of all fs except / and /userdata, umountroot should take care of / and we just need something to remount /userdata ro before the reboot to make sure everything is flushed to flash [17:29] salem_: np, thanks for looking at it [17:29] i simply cant see what should be wrong in the android side.... [17:29] salem_: have you seen this? https://bugs.launchpad.net/ubuntu/+source/messaging-app/+bug/1230420 [17:29] Ubuntu bug 1230420 in messaging-app (Ubuntu) "When sending a message to a new phone number, the number is inserted at the beginning of the message" [Undecided,New] [17:29] anders3408, so read the logs [17:29] stgraber: ok. [17:29] anders3408, logcat should have errors that help you to track down the issue [17:29] those just say that the android init cant find /system/bin/* just as its not correctly mounted in the lxc-container === psivaa-afk-bbl is now known as psivaa [17:30] cwayne, ouch, never seen this. let me test [17:30] logcat just says sensorservice and surfaceflinger are not started, and they are not started becasue android init fails to find sh and sensorservice and all those in system/bin/* [17:30] because under /proc/pid/root/system/ its empty [17:33] stgraber: so... trying to update from build 56 to build 62, I'm consistently getting timeouts. "Upgrade path is 57:60:61:62" [17:33] cwayne, I am unable to reproduce it [17:33] cwayne, which phone? [17:33] mterry: sorry, shame on me, but no ... let's ping Mirv tomorrow morning and check with him [17:33] stgraber: is it just timing out because it takes longer than n minutes to download it all? [17:34] salem_: mako, latest devel-proposed [17:34] salem_: happens to me every time [17:34] cwayne, ok, will install [17:34] slangasek: that'd be a question for barry [17:34] ok [17:34] barry: ^^ why am I timing out? [17:34] anders3408, you only look at the fallout, there mst be mount errors and the like [17:34] barry: am I hitting the 10m dbus limit? and does the download-manager migration fix this? [17:34] ogra_: in syslog + [17:34] ? [17:35] in logcat [17:35] syslog is ubuntu [17:35] it's possible! maybe you can bump up the logging and tail the log file. [17:35] logcat nope [17:35] you want to debug the android side [17:35] slangasek: ^^ and if so crank up the timeout [17:35] barry: oh, there's a logfile? where? [17:35] slangasek: /var/log/system-image/client.log [17:35] but i think it only logs at debug level by default [17:35] barry: so how do I turn up the logging? [17:35] ogra_: logcat only contains 2 diffrent lines... [17:35] I/ServiceManager( 1177): Waiting for service SurfaceFlinger... [17:35] I/ServiceManager( 1257): Waiting for service sensorservice... [17:35] (until s-i 1.6) [17:35] I was trying to debug by calling system-image-dbus -v by hand [17:35] slangasek: /etc/system-image/client.ini [17:36] anders3408, well, there should be a lot more [17:36] mterry: sorry, was on call [17:36] there isint [17:36] slangasek: you can also system-image-cli --dbus [17:36] anders3408, well, init starting prints a ton of stuff [17:36] mterry: hard-coded to unity8, certainly not [17:36] but dmesg fails big time [17:36] including some cyanogenmod ascii art [17:37] barry: so if I *am* hitting that timeout, does that get resolved with download-manager? [17:37] mterry: check maliit's output, it should print a "creating surface..." message [17:37] ogra_: thats only if android is really started, and it doesnt as all android services is not starting at all becasue of the mount thing [17:38] anders3408, if init wouldnt have started the services wouldnt try to start [17:38] how are you getting your logcat ? [17:38] slangasek: maybe ;). i think u-d-m is faster overall for downloading than the built-in downloader, but if the bottleneck is network or server, then no, it won't help. we could probably crank the timeout to 1h or so by default [17:39] slangasek: oh [17:39] ogra : i do ./system/bin/logcat [17:39] try with -d [17:40] ogra_: see this is all : http://logcat.scheffsblend.com/view?id=407001 [17:40] also try the different buffers [17:40] slangasek: heh, it looks like the default timeout might be 1m [17:41] slangasek: if that's the case, edit [system]timeout value in client.ini [17:41] anders3408, yeah, do it right after boot, logcat uses a ringbuffer ... so the actual boot messages seem to have been lost since the other stuff already printed to much [17:41] (i'm changing it in my dev branch right now ;) [17:41] anders3408, add something like "/system/bin/logcat -d >/var/log/foo.log" to /etc/rc.local [17:42] that should get you a dump right after boot [17:42] ogra_: the log you just saw was directly after boot [17:43] does anyone else have a dock for their nexus 7? [17:43] anders3408, thats definitely not init output [17:48] ogra_: /system/bin/logcat -d > /var/log/logcat.log from rc.local points out the same lines only as i showd you :) [17:48] ogra_: all the android parts is not started up at all [17:49] barry: well, it's clearly not reasonable to expect the downloads to finish in a minute... or even 10. Why would we want the dbus service to time out at all while a download is in progress? [17:49] well, it wouldnt print that stuff if itr wasnt [17:49] phones just sits on the boot logo [17:49] first bootup shows all the lines it should just like it was a normal aosp rom that booted up [17:50] ogra_: try and see the dmesg.: http://paste.ubuntu.com/6155437/ [17:51] so yes , a part of android is booting up or else no dmesg, but all the services that android's init starts fails [17:51] cwayne, just tested here and couldn't reproduce. is it easy for you to record a screencast? [17:51] ogra_: hi, do you plan to switch to cm 10.2 for ubuntu touch base? [17:51] salem_: not sure, is there a documented way to do it on the phone? [17:51] Tassadar, ask rsalveti :) [17:51] cwayne, I dont know :) [17:51] slangasek: i think it was a failsafe so the process wouldn't hang around forever. i don't particularly like the timeout, but i guess it makes some sense. [17:52] slangasek: so let's s/m/h/ :) [17:52] salem_: last i checked there wasnt :/ === Zuardi_ is now known as zuardi [17:52] rsalveti: do you plan to switch to cm10.2 base? I'm asking because I _think_ flo's binary blobs need 4.3 :/ [17:52] cwayne, what about record a video using another phone? [17:53] salem_: that i could probably try :) [17:53] ogra_: if you see line 635 and 638 and 647 , system and data and cache is mounted corrctly, but already in line 679 things starts to break when it trying to start things up [17:53] cwayne, ok... for myself, it seems to work [17:53] cwayne, cool thanks. this bug sounds critical to me. but neither me or boiko could reproduce it. [17:54] mterry: on a maguro? [17:54] cwayne, I added "dbus-send --print-reply --system --dest=org.freedesktop.Accounts /org/freede [17:54] sktop/Accounts/User32011 org.freedesktop.Accounts.User.SetBackgroundFile string: [17:54] /usr/share/unity8/graphics/tablet_background.jpg" to unity8.conf right before we call unity [17:54] nevermind about the dock question [17:54] ick, that formatted bad [17:54] cwayne, uh, no on mako [17:54] cwayne, you are saying this is maguro specific? [17:54] with the newest touch-session and lightdm? [17:54] mterry: im not sure if it is, ive only tested maguro so far [17:55] cwayne, hm. this is on mako with latest lightdm/touch-session [17:55] mterry: could you try it with my upstart job and see if that works? [17:55] cwayne, sure, pastebin me? [17:55] anders3408, 635 is inside the initrd [17:55] mterry: http://paste.ubuntu.com/6155484/ [17:56] mterry: note youll need a picture as /custom/usr/share/backgrounds/welcome.png [17:56] anders3408, 640 - 642 are more intresting, this is your containers init [17:56] cwayne, k [17:57] salem_: i'll have to get it for you after my dr's appt [17:58] cwayne, is it at least picking up the background on a reboot? [17:58] ogra_: here is the diff between the mouting from dmesg on first bootup and 2nd bootup : http://paste.ubuntu.com/6155493/ [17:58] mterry: no [17:58] cwayne, no problem [17:59] cwayne, fascinating [17:59] mterry: ? [18:00] ogra_: 640 - 642, the service qcamer thing is just a duplicated line in that file, and the setprop is something i was testing for simcard, not related to the mouting :) [18:00] cwayne, fascinating that it doesn't pick it up on reboot either. Which means it's not just a problem with notification of change but reading in the first place [18:01] cwayne, so it worked for me, with your job (well, I changed url of /custom/* but same idea) [18:01] mterry: i wonder if it is a maguro thing then, or if i'm just missing something (entirely possible) [18:03] cwayne, I don't know why it would be maguro-specific, but I've learned not to trust my assumptions about what is device-specific [18:03] mterry: lol, i'd wager that I probably am missing something on my maguro [18:03] mterry: which image are you on? [18:03] cwayne, it's clearly being set in AS, which means that policykit and all that goodness is working. So lightdm side is covered. unity8 just can't ask for the file from AS for some reason [18:03] cwayne, how do I find that out? [18:03] cwayne, I may have apt-get upgraded at some point too [18:03] cwayne, something from last week I think [18:03] mterry: cat /etc/system-image/channel.ini [18:04] mterry: ah, i did mine on the latest devel-proposed [18:04] maybe that has something to do wit it [18:04] cwayne, 57 [18:04] with* [18:04] either way, i'll try it on my mako later when im back from my dr's appt [18:04] cwayne, I don't want to wipe this guy yet, I'm in middle of other stuff [18:04] cwayne, I'd be curious about your mako experience [18:04] mterry: no worries at all, wouldn't expect you to drop everything and wipe a device for this :) [18:05] i'll shoot you a ping when i get to test it on a mako [18:05] k [18:05] anders3408, what happens if yu comment /system in fstab, did you try that yet ? [18:05] mterry: if it works on my mako i may not really care anyway AFAIK none of the customization stuff has been very heavily tested on maguro [18:08] mterry: also, so locale stuff should be good to go now,r ight? [18:08] cwayne, in theory. Also, I'm working now on the welcome wizard which asks the user for their language [18:08] ogra_: fstab.find5 is not being called from init.find5.rc at all [18:08] If that's relevant to you [18:09] all is mounted from init.find5.rc [18:09] anders3408, on the ubuntu side i mean [18:12] ogra_: same thing happens if i look in dmesg [18:13] slangasek, lool: so we appear to have a few problems on touch currently. At the time umountfs is called we still have open write fds to /var/log and /var/lib, if the device is writable, we also have a few more fds open. [18:14] ogra_: also lxc-conatiner fails to start if i do that [18:14] slangasek, lool: upower and rsyslog are the main problem, followed by upstart with adbd (fixable with a console none in the job file) [18:14] once I get rid of all of those, umountfs unmounts pretty much everything, umountroot remounts / read-only and I can remount /userdata read-only too [18:15] ogra_: wait a sec, ill reflash :) [18:15] anders3408, well, then start by adding sleeps to the lxc-android-config upstart job ... right before lxc-start [18:15] reflashing will just get you a working system again [18:16] ogra_: adding waits to lxc-android-config.conf ? or pre-start.sh ? [18:16] to the upstart job [18:16] ehh ? [18:16] and which file would that be ? [18:17] upstart jobs live in /etc/init [18:18] stgraber: rsyslog is certainly meant to be shut down before then; is this not happening? [18:18] ogra_: so that would be /etc/init/lxc-android-config.conf ?? [18:18] yes [18:19] add something like: [18:19] pre-start script [18:19] sleep 5 [18:19] end script [18:19] and then reboot and see if that changes anything [18:19] (will indeed make the boot 5 sec slower) [18:19] and add it just after the post-start script ? [18:20] haha :) [18:20] yeah, just below the end script [18:21] slangasek: it sure isn't, that's what I get after running S20sendsigs stop: http://paste.ubuntu.com/6155595/ [18:21] stgraber: sendsigs has nothing to do with it, rsyslog is an upstart job [18:22] and is 'stop on runlevel [06]' [18:22] ogra_: is it possible that you can look at sforshee changes in powerd and possible approve it ? :) i need it for find5 brigtness [18:23] slangasek: gah, yeah, good point, need to rework my test a bit to get a more realistic view of the last second of the system [18:27] ogra_: sadly it didnt do anything diffrently [18:38] barry: so, this timeout is something other than the dbus service timeout... the UI is returning a "timed out" / "retry" while the dbus service is still running [18:39] ogra_: when is lxc-instance.conf being used ? [18:39] slangasek: any clues in the log file? [18:40] barry: hadn't adjusted the log level. and does ubuntu-system-settings log there? [18:40] slangasek: nope, just s-i [18:40] ok [18:43] stgraber: \o/ awesome, thanks a lot for looking into this [18:46] slangasek: so, just to confirm I'm not going mad, reboot is supposed to be the equivalent to telinit 6, which is supposed to emit the matching runlevel event in upstart which in turns will trigger /etc/init.d/rc, right? [18:47] slangasek: I'm asking because I noticed that at least lxc-android-config has a broken post-stop that should make our devices take >1min to reboot, and well, they don't which leads me to believe that this never gets executed. I also added some debugging to /etc/rc6.d/ and it doesn't appear to run (I'm logging to the kernel ring buffer but never see those messages post-boot in last_kmsg) [18:48] stgraber: post-stop or pre-stop? The version I see here only has the pre-stop [18:49] stgraber: but yes, I can't think of any relevant differences between 'telinit 6' and 'reboot' [18:49] slangasek: oh yeah, pre-stop, anyway that lxc-stop is going to take a minute to run, so clearly it's not running on reboot [18:51] hmm, ok [18:51] why would it take a minute? [18:51] because it's waiting for a clean shutdown from android [18:52] which won't happen as android doesn't know to cleanly shutdown (in response to SIGPWR) [18:52] instead it should be lxc-stop -n android -k [18:52] to send a sigterm+sigkill to all the processes [18:54] hmm, strange [18:54] ok then [18:55] stgraber: but I have certainly run 'stop lxc-android-config' here and watched the container shut down [18:58] Hey guys. Any idea why would a DashPreview not want to close itself? I see there have been on changes to DashPreview.qml for a while, so perhaps I'm doing something wrong. [19:03] slangasek: gah, the more I look at those upstart script the more issues I find... [19:04] stgraber: well, currently I get this gem: [19:04] # status lxc-android-config [19:04] lxc-android-config stop/post-start, (post-start) process 497 [19:04] the post-start script is hanging in the 'initctl emit android', and I can't see why [19:05] hmm, my phone is eating cpu on 102 root -51 0 0 0 0 D 26.1 0.0 0:18.50 irq/330-wcd9xxx [19:05] slangasek: right, I found the reason for that one [19:05] wossat? [19:05] slangasek: it's the new mtp stuff [19:05] stgraber: really? mtp-server-bootup is 'start/running' [19:05] slangasek: do "stop mtp-server-bootup" and that'll get lxc-android-config into start/running [19:05] slangasek: yeah, it's start/running but without a pid, so it's blocking the initctl from lxc-android-config [19:06] oh, but it's 'task' [19:06] stgraber: it probably needs to not be 'task' [19:07] slangasek: hmm, actually it may be fine as a task if it would do its job and not get stuck somewhere half way through its script :) [19:07] it's not spawning any long lasting process in theory [19:07] unless something depends on that android_usb-device-changed and gets stuck there (as it's a blocking initctl call) [19:08] stgraber: 'task' would imply that it would immediately run the post-stop script [19:08] slangasek: oh yeah, good point, so definitely shouldn't be a task, though I'm a bit confused as to why it's not running the post-stop immediately at the moment then :) [19:08] stgraber: right, so dropping the 'task' unblocks the 'emit android'... which means that now, lxc-android-config is stopping instead [19:09] someone has done something very bad to the handling of the container [19:10] Tassadar: ogra_: not until 13.10 is released [19:10] we're just working on bugfixing at this point [19:10] slangasek: http://paste.ubuntu.com/6155757/ that's my suggestion for lxc-android-config [19:10] stgraber: any idea what's killing the container? [19:11] slangasek: also, look at the exec line of the mtp.conf job ;) [19:11] stgraber: I think the 'initctl emit android' should probably be 'nowait' [19:11] stgraber: I don't have an 'exec' line in this version [19:12] maybe if I could update the phone I would [19:12] exec /usr/bin/mtp-server & [19:12] I don't have that here [19:12] oh, that's in the user session [19:12] ok [19:12] yeah [19:13] rsalveti: okay, thanks, so I guess I'll have to wait (or update it myself). flo throws errors about proprietary .so libraries not having certain symbols, so I guess there is nothing else to do with it [19:13] expect fork + backgrounding in the exec line is pretty fun, I think I'll fix that while I'm at it :) [19:13] stgraber: are you sure android init needs that -k? Because I know I've seen the container shut down quickly and cleanly in the past [19:14] slangasek: lxc-stop in the past was defaulting to -k [19:14] ah [19:14] slangasek: or rather, we had lxc-shutdown, lxc-halt and lxc-stop all doing different things, we merged those in 1.0 [19:14] stgraber: eh, why are you changing it to a post-stop? that's got to be wrong [19:14] Tassadar: you could try rebasing our patches on top of cm 10.2, but that might not be that trivial [19:15] exactly) [19:15] slangasek: hmm, actually, why am I keeping the lxc-stop at all, upstart will send a SIGKILL anyway won't it? [19:16] so long as upstart tracks the pid properly, it'll kill lxc-start which will make the whole tree die properly, no need for lxc-stop at all [19:16] stgraber: upstart sending SIGKILL is certainly not graceful [19:17] stgraber: what does lxc-start do when killed with SIGTERM? That's the more important question [19:18] slangasek: forwards it to the container's pid 1 which usually ignores it [19:19] stgraber: then how is that an appropriate shutdown? Doesn't something need to kill init and clean up the container behind? [19:19] I'm assuming that lxc-stop does some clean-up [19:20] stgraber: anyway, are you seeing this same issue with lxc-android going into stop/post-start? because that's a VERY serious problem [19:20] slangasek: lxc-stop doesn't really do any cleanup, it just sends SIGPWR to PID1 in the container and after x seconds sends it SIGKILL, that kills all the processes, empties the namespaces which in turn gets destroyed by the kernel [19:20] and needs to be fixed before we can realistically change any of the rest of this [19:20] slangasek: yep, I do [19:20] stgraber: ok, so SIGKILL to lxc-stop would have the same effect - in which case I think we do want to use lxc-stop -n android -k in the pre-stop, so that we don't have to wait for upstart to send SIGKILL on its own [19:21] slangasek: anyway, I guess we can keep the lxc-stop in pre-stop for now (but using -k) so that it's easy to change if we ever support clean shutdown within android, I'll update my diff accordingly [19:21] ok, I'll do that then [19:21] stgraber: ok [19:21] hello [19:21] but again, we have a serious problem with lxc-android-config being 'stop/post-start' on boot [19:21] the blocked event is the only thing keeping the container from shutting down! [19:22] stgraber: is lxc-start forking? [19:22] slangasek: ok, so http://paste.ubuntu.com/6155797/ should do the trick for LXC, then we need http://paste.ubuntu.com/6155801/ which should fix mtp and unblock lxc-android-config [19:22] i'm trying to install ubuntu n a spare nexus 4.. its downloaded the image but its not able to boot into the recovery image [19:23] slangasek: not unless you pass it -d [19:23] it gets stuck at the blank screen [19:23] stgraber: and at what point does it fork if you pass it -d? Should we be using -d + expect fork? [19:24] slangasek: I believe it does it after it's done parsing the config, so -d + expect fork may be fine, I just have to check whether we fork or double fork [19:25] stgraber: appears to be a double fork (proper daemonize) [19:26] ok, so I can remove the console none switch to expect daemon and use -d [19:27] stgraber: ok. also, 'pre-stop script lxc-stop -n android -k end script' is written 'pre-stop exec lxc-stop -n android -k' :) [19:28] ok, let's see what happens on my device with those changes [19:28] stgraber: why do you want 'console none', anyway? the android container should also be shut down before we try to remount / [19:28] slangasek: doh, can't use expect daemon [19:28] oh? [19:29] lxc spawns a monitoring daemon at startup time and upstart picks up that one instead of the main lxc-start [19:29] ok [19:29] then drop the -d, the post-start script polls anyway [19:29] so I'll drop -d and re-add console none (unless we want to spam /var/log/syslog with /dev/console output from android) [19:30] you mean /var/log/upstart/lxc-android-config.log, surely? [19:30] yeah, that one [19:30] anyway, if the reason to kill it is because it's too spammy, then yeah [19:30] makes sense to me [19:30] yep, that was the only reason to put a console none in there [19:31] ok [19:31] fwiw, there are some interesting errors in my log when running without -d [19:31] maybe we want to make sure we know what those are before silencing [19:31] alright, so the only odd one left is: ofono start/pre-start, process 633 [19:32] http://paste.ubuntu.com/6155830/ [19:32] ofono is start/running for me here [19:32] slangasek: most relevant stuff would probably also be in /var/log/lxc/android.log [19:32] slangasek: you're testing on mako, right? [19:32] yes [19:32] slangasek: I'm testing on grouper at the moment here [19:32] confirmed, the log is a duplicate [19:32] which doesn't have a modem [19:32] right [19:33] anyway, made the lxc changes locally here and it looks solid [19:33] the ofono job is sane too, it just waits in pre-start for a while, then exits cleanly [19:34] so that's lxc-android-config: http://paste.ubuntu.com/6155848/ [19:34] and mtp: http://paste.ubuntu.com/6155850/ [19:34] yes, looks right to me [19:34] hmm, mtp-server isn't running here with those changes... [19:36] oh, I see, apparently you need to unlock the screen for it to start? anyway, I have it running now [19:41] that still doesn't tell me why my rc6.d scripts don't get run, but I'll get MP for those two packages sent out before I dig any further... === salem_ is now known as _salem [19:44] ogra_: https://code.launchpad.net/~stgraber/ubuntu/saucy/lxc-android-config/upstart-fixes/+merge/187589 and https://code.launchpad.net/~stgraber/mtp/upstart-fixes/+merge/187587 [19:45] ogra_: let me know when I get those uploaded to the archive (I can commit to the first but can't to the latter which seems wrong to me since I'm a coredev...) [19:45] mterry: installing mako now [19:47] slangasek: any clue? http://paste.ubuntu.com/6155893/ [19:48] stgraber: did you have a clean shutdown last time? [19:48] slangasek: it really looks to me like something else is making the device reboot (that or /dev/kmsg magically stops working when in the middle of a reboot) [19:48] hmm [19:48] dunno [19:48] slangasek: the script I wrote to know whether I'd get a clean shutdown doesn't get run, so I don't know ;) [19:48] well, no idea, sorry [19:49] stgraber, why console none ? the output in the log is pretty helpful for debugging [19:50] specifically for ports [19:51] ogra_: because the output is already in /var/log/lxc/android.log and as I'm dropping -d, you'll now get every single message printed to /dev/console in the upstart log (which can then get pretty big and use persistent space) [19:51] ok [19:52] if something goes wrong, I'd expect porters to just use adb shell and run lxc-start -n android -l debug -o /tmp/debug so they can really see what's going on, but I doubt it's a good thing to do by default :) [19:53] true ... but its something that should be on the Porting wikipage ;) [19:58] stgraber: i just tried and added -l debug -o /tmp/debug to the lxc-android-config.conf where it runs lxc-start but it gives me absolut nothing in /tmp regarding lxc no new debug file or anything :) [20:00] anders3408: where exactly did you add that to the exec line? [20:00] anders3408: can you paste me the resulting exec line? [20:00] added exec lxc-start -n android -d -- /init -l debug -o /tmp/debug to /etc/init/lxc-android-config.conf [20:01] and i have issues getting it to bootup 2nd time, looks like the lxc-container fails to mount /system [20:01] balloons: heya! would you think you could review https://code.launchpad.net/~lool/music-app/singleton/+merge/187583 ? [20:02] stgraber: /system is mounted , but /proc/containerpid/root/system is empty [20:02] lool, heh, big merge :-) sure [20:02] balloons: (or perhaps redirect me to usic-app developers) [20:02] balloons: ;-) [20:02] anders3408: right, you added that at the wrong place, that's why it doesn't work :) [20:03] but its only system that fails, as the rest in containerpid/root is fine [20:03] anders3408: exec lxc-start -n android -l debug -o /tmp/debug -d -- /init [20:03] balloons: I dont really know who to ping, I tend to ping Alan Pope or David Planella [20:03] balloons: thanks a lot [20:03] thanks :D ill try [20:03] lool, you are approved. [20:04] ogra_: could it be that last_kmsg is just completely useless on grouper? :) [20:04] lool, I normally end up talking to the music folks via email as I never find them on irc [20:04] ogra_: I'm writing stuff to /dev/kmsg, reboot and can't find it in /proc/last_kmsg [20:04] balloons: you didn't top approve it [20:04] balloons: but I can mail them [20:04] balloons: can you msg-ed me a list? [20:04] */msg-me a list [20:06] can someone tell Jenkins it's wrong wrt https://code.launchpad.net/~stgraber/mtp/upstart-fixes/+merge/187587 [20:06] cwayne, sorry, I'm back. Any luck with mako? [20:07] sergiusens, lool, plars: just a quick heads up, the daily and daily-proposed channels have now been removed from system-image.u.c [20:07] barry: here's my client.log: http://paste.ubuntu.com/6155968/ [20:07] mterry: no dice here [20:07] barry: so what's going on here, and how to I fix it? [20:07] mterry: although now sometimes it shows up as just black [20:07] so thats cool [20:08] barry: and why is a traceback about a TCP timeout showing up in the client log? [20:08] stgraber: lxc_start - '/init' started with pid '426' but then it says : lxc_start - invalid pid for SIGCHLD [20:08] cwayne, heh [20:08] anders3408: yep, that one can be ignored [20:08] bfiller, is https://bugs.launchpad.net/webbrowser-app/+bug/1231126 on the plan to fix? [20:08] Ubuntu bug 1231126 in webbrowser-app "When rotated to landscape the toolbar doesn't come out on the bottom" [Undecided,New] [20:08] cwayne, well... it's good that it's not device-specific [20:09] slangasek: great question. i wonder if it's a server or network problem? [20:09] slangasek: it's trying to read 4k bytes and getting a low-level socket timeout [20:09] mterry: i suppose so :) [20:09] jono: not sure, needs the shell guys to support changing edge gestures depending on orientation [20:10] stgraber: see : INFO lxc_conf - looking at .25 18 179:19 / /system ro,noatime,nodiratime - ext4 /dev/mmcblk0p19 ro,data=ordered [20:10] that looks fine, but still /proc/containerpid/root/system is empty ? :( [20:10] jono: i.e. a bottom swipe is being interpreted as right edge swipe in landscape mode [20:10] bfiller, gotcha [20:10] jono: that should really be a dupe of another bug that I can find on unity8 [20:10] so this is dependent on unity8 fixes? [20:10] yes [20:10] ok cool [20:11] Saviq, ^ [20:11] cwayne, maybe you are using a newer unity8 than me. Next time I flash, I will retest [20:11] mterry: cool beans, thanks [20:13] anders3408: do you have something in /android/system outside the container? [20:13] ehh ? i dont have a folder named /android ? [20:13] if you mean from root /android/ [20:13] barry: I have no problems with my network reliability; I can reach the server and download individual files from it. I suppose the switch to download-manager may make it more robust... [20:14] slangasek: ah, you're in standard flipped, hmm, let me check, I haven't use that in months [20:14] barry: and I guess I'll bite the bullet and reflash to get updated to that version :P [20:14] stgraber: what do you mean, "standard flipped"? [20:14] slangasek: it'll definitely change the behavior :) [20:14] slangasek: I mean "old flipped" :) [20:14] stgraber: what's "old" about it? [20:14] pre-system-image [20:15] stgraber: absolutely not [20:15] slangasek: i'm going to try upgrading my own n4 to see what it does [20:15] where do you get the impression that this is pre-system-image? [20:15] slangasek: because I didn't mean to talk to you, my fingers just typed s by reflex instead of a :) [20:15] ehh stgraber : saucy-preinstalled-touch-armhf.zip is that the old flipped ? [20:15] stgraber: oh, haha [20:15] slangasek: what channel are you on? [20:16] barry: "the default"? [20:16] anders3408: yep, that's old flipped which indeed doesn't have /android and directly mounts the partitions from Android [20:16] barry: it says "channel: devel" [20:16] so im using the wrong prebuilts ? :O [20:16] anders3408: depends, what device is that? [20:16] slangasek: k [20:16] oppo find 5, very simlar to lge mako device [20:17] so no, you're using the right thing on that device since we only have system-image for mako,grouper,maguro and manta [20:17] but unfortunately you're talking to the system-image guy who hasn't been using one of those images since mid-May so I doubt I'll be able to help you very much there :( [20:18] (well, I'm working on getting system-image to play nice with ports, so once that's done, I should be able to help people since it'll be a setup I know about) [20:18] ugh :( sounds like no one can.... [20:19] slangasek: well, i went from build 4->62 np, but i guess i was running old enough that whatever problem you were seeing i missed [20:19] anders3408: if the container appears to start but /system isn't mounted, that's probably because android couldn't find your system partition, now that can be a broken fstab, broken ueventd, ... [20:20] stgraber: the thing is , on first bootup it boots fine, no issues at all , but on 2nd bootup , i see /system mounted correctly but /proc/containerpid/root/system is all empty , also is /var/lib/lxc/android/system [20:21] barry: well, I had different issues with the older versions and had to manually update to this version of system-image with apt to be able to update at all... [20:21] anders3408: and by bootup, do you mean rebooting the whole device or just the lxc container? [20:21] rebooting the whole device :) [20:21] barry: so how do things look with the download-manager testing? [20:21] anders3408: and what do you have to do to get it to work again? [20:22] wipe all , and reflash is the only way to get it bootup again [20:22] anders3408: and when you reflash, do you also have to reflash the original system partition or just the ubuntu bits? [20:23] everything , i reflash the compiled zip , then the prebuilt then a zip with the device specific conf and rules in [20:24] all i am missing atm is adding the audio parts , so calling will work, and adding apparmor to the kernel, other than that, its going fine, having 3g and wifi to work, camera and music [20:24] anders3408: could you pastebin /proc//mounts on one of those times where /system is empty? [20:25] slangasek: i'm still seeing some weird things with the d/l service in my dbus tests. it's very hard to debug because there's currently no logging in the u-d-m so when i'm not getting a signal i expect, i'm not sure why. mandel did say he was going to add some syslogging to help with this (and the slow tear down problems), so that will be useful. as it is, all of my tests pass except the tests of s-i's dbus api (though some of those [20:25] do pass) [20:26] stgraber, depends if grouper actually powers off on reboot, kmsg/last_kmsg should persist as long as the ram is powered , but probably the grouper kernel misses the ram console option or some such [20:26] stgraber: just a sec :) [20:26] mhall119, Is there someone I can get to ack that they're okay with this URL format from calendar app? bug 1231136 [20:27] bug 1231136 in URL Dispatcher "Calendar needs to be invocable with a URL" [Undecided,New] https://launchpad.net/bugs/1231136 [20:27] stgraber: http://paste.ubuntu.com/6156029/ [20:27] stgraber: ok thanks [20:27] there you have from 1st bootup and 2.nd bootup [20:27] lool, So I'm having second thoughts on using the full triplet for the object path. [20:28] lool, The reason being that if you have something like GEdit, and you wanted to make a click package, even thought it is "org.gnome" the app wouldn't do that. [20:28] lool, But it would be wired to use application name already. [20:28] anders3408: can you try "mount /dev/mmcblk0p19 /mnt" outside the container? [20:29] anders3408: to check whether the partition is still mountable at all [20:29] stgraber: mount: /dev/mmcblk0p19 already mounted or /mnt busy mount: according to mtab, /dev/mmcblk0p19 is mounted on /system [20:31] tedg: kunal is usually the one to go to for calendar, but I know he has some help now [20:31] popey: who's working on calendar these days? [20:31] tedg: we could patch it to support both [20:32] lool, ? How? Send two messages? [20:32] tedg: doesn't fdo encourage all apps to move to namespace + app name anyway? [20:32] jono: for bug #1226183, does "not displayed" mean that "More Suggestions" shows up, but is empty? [20:32] bug 1226183 in Unity Click Scope "More Suggestions not displayed with mobile data connection" [Undecided,Confirmed] https://launchpad.net/bugs/1226183 [20:32] tedg: have two objects? [20:32] anders3408: so what happens if you umount /system and then start the container? [20:32] tedg: also, imagine a desktop with gedit.deb installed a gedit.click, what happens when you want to tell gedit to open an url? [20:33] s/a/and/ [20:33] lool, Kinda, I mean, the namespace/app might not align. They're encouraging the desktop file name to match that. But we're using that as the app name. [20:33] lool, So is an app id it could end up as: org.gnome_org.gnome.gedit_123 [20:33] tedg: 123? [20:33] lool, version [20:33] stgraber: just tried and outcommited /system in /etc/fstab , and then the container started, but still some issues like E/linker ( 1927): linker.c:1095| ERROR: Library '/system/lib/libubuntu_application_api.so' not found so it ends up rebooting after some time [20:33] barry: are the failing s-i dbus api tests bugs in system-image? in u-d-m? We're kind of at the end of our rope here [20:34] tedg: I didn't want the version in there, sorry should have been clearer in the mp [20:34] tedg: and I thought _ would be split away [20:34] lool, So you're just thinking /pkg/app ? [20:34] tedg: /parts/of/the/name/space/AppName [20:34] anders3408: right, so you're basically in a race between Ubuntu and Android mounting /system with the first boot case changing the timing enough to make it boot [20:35] com/ubuntu/stock-ticker-mobile [20:35] com/ubuntu/dropping-letters [20:35] dobey, yes, I think so [20:35] dobey, although I think some times it didn't show the section [20:35] yes that sounds correct :) stgraber [20:35] tedg: there are other conflicts between the two anyway, like lower case vs. mixed case [20:35] Gedit vs. gedit [20:35] lool, Perhaps we can just check to see if the pkg is the first part of the app, and ignore in that case. So if org.gnome is there twice, we don't put it there twice. [20:35] anders3408: that was one of my concerns with the flipped model which I addressed in the system images by having Ubuntu always mount everything in /android and then have lxc bind-mount those into the container (and disabling android's mount management code) [20:36] ssweeny: https://bugs.launchpad.net/touch-preview-images/+bug/1231140 [20:36] Ubuntu bug 1231140 in touch-preview-images "wifi password box clears when you check "show password", type something, then uncheck it" [Undecided,New] [20:36] tedg: I dont understand the thing coming in twice [20:36] lool, True, and that's what is nice about just matching the desktop file. It's known by everyone clearly. [20:36] lool, Case and all. [20:36] lool, If we do /pkg/app then app could have a namespace in it. [20:36] anders3408: so the good news is that your device will likely just work once there's documentation for how to setup a ports system-image server, the bad news is that this probably won't happen for another couple of weeks, so until then your best hope is to add a bunch of sleeps at random places and see if that helps... [20:36] stgraber: ugh.... so i cant do anything about it :( [20:37] lool, pkg = "org.gnome" ; app = "org.gnome.gedit" [20:37] jono: ok, thanks. [20:37] tedg: I dont get it, if we take sudoku out of com.ubuntu.sudoku_sudoku_0.4.3.desktop it's not distinguishable from sudo from net.launchpad.sudoku_sudoku_0.4.3.desktop [20:37] tedg: why would we ever get pkg = "org.gnome" ; app = "org.gnome.gedit" [20:38] lool, That's not an issue, we don't care if they export the same object name. [20:38] anders3408: maybe ogra_ can comment on what exactly is the done on first boot which isn't on subsequent ones which may help you figure out where to add a delay [20:38] stgraber: where should i add some sleeps ? right now it looks like if i can get android to mount /system first then ubuntu mounting /system after that , looks like a win win, just like it does on first bootup :) [20:38] concretely with "legacy" / .deb we have non-namespaced names, and they might be namespaced, and with click we have namedspaced [20:38] tedg: it's an issue because we're matching the name [20:38] lool, Because FD.o is encouraging apps to change their desktop files to include namespace. So it'd be "org.gnome.gedit.desktop" [20:38] lool, We're not matching the name. [20:38] tedg: so that's fine, and then it will be org/gnome/gedit [20:38] lool, We're just sending a message to that object on the bus. [20:39] I fail to see the difference [20:39] we're going to send the message to all sudokus on the bus [20:39] If two apps both export /sudoku and /sudoku it doesn't matter. [20:39] No, we're not. [20:39] well there's just the pid check [20:39] "just" ? [20:39] We're sending it to only a single unique name. [20:39] tedg: I don't find "sudoku" unique enough [20:40] There's no chance that it'll get to more than one connection there. [20:40] well you're saying we're stopping at the first one [20:40] it might still be the wrong one :-) [20:40] I don't see anyway it could be the wrong one. [20:40] tedg: also, you want to implement a non-namespaced object path when FDO pushes for namespaced object path? [20:41] It pushes for it to match the name of the desktop file. And for the desktop file to be namespaced. [20:41] The result is a namespaced path, but they're independent. [20:41] yes, and so we happen to have namespaces [20:41] ogra_: do you have some ideas on what stgraber said ? :) [20:42] and you don't want to use them :-) [20:42] Because I want to use the desktop file name. [20:42] Which could or could not be namespaced. [20:42] the desktop file name is com.ubuntu.sudoku_sudoku_0.4.3.desktop and is namespaced! :-) [20:43] tedg: I think we should use the app_id/desktop filename but strip the version and transform not only . but also _ in / [20:44] it guarantees uniqueness, and it matches the way we identify the application on the system [20:44] I think if we're going to do that we should use the DBus name [20:44] but I can pull in an independent expert, my wife, and convince here to vote for my option [20:45] *her [20:45] We've got the name that we're getting from nih_dbus_path() elsewhere. [20:45] (hud, etc.) [20:45] And just do /$(DBUS_NAME) [20:45] anders3408: tracking number should be in your inbox [20:45] anders3408: once again, fantastic work you did on the port [20:45] mhall119: nice :D [20:46] thanks mhall119 :) [20:46] I know many Find 5 owners will be looking forward to continued improvements to it to :) [20:47] tedg: ok, I'm happy to check how this looks like, but it will have to be tomorrow [20:47] stgraber: what if i add noauto,users, to the fstab file, where it mounts system, then in lxc-android-config.conf after it has run lxc-start, let it mount /system ? [20:47] tedg: the other thing we could think about is whether it's something we can easily change or not [20:47] if it's only done in qtubuntu, then certainly, but if individual apps do it, it's bad [20:47] gtg [20:47] super tired [20:48] Non-SDK apps would do it themselves. [20:48] But SDK apps would just be using QtUbuntu. [20:48] jdstrand: guess you forgot to attach the example qml/profile files at bug 1226844 [20:48] bug 1226844 in qtsystems-opensource-src (Ubuntu Saucy) "QtSystemInfo fails when using ofono and there are DBus denials to ofono" [High,New] https://launchpad.net/bugs/1226844 [20:48] It'd be nice if we could keep GApplication working. [20:48] you did attach the click package, but I'd prefer using the qml one to reproduce the issue [20:48] rsalveti: I did, but later attached a click package with instructions [20:49] right, let me give the click a try then [20:49] rsalveti: see comment #2 [20:49] fginther: not sure you saw https://code.launchpad.net/~stgraber/mtp/upstart-fixes/+merge/187587 failed upstream merging due to bazaar connection issue? [20:50] lool, So I'll use nih_dbus_path() for now and we'll go from there? [20:50] jdstrand: as we connect to 3g with nm anyway, I believe we just need to fix the ofono related error [20:50] cool [20:52] tedg: I'll read on that tomorrow; write me some pointers to it and I'll check it out tomorrow [20:52] updating last things I had to do before bed and going off [20:53] 'night lool [20:54] mhall119, popey, so I'm happy to take other options as a URL format there, we just need something that we can start coding to on the other sides. [20:55] mhall119, popey, It seems relatively arbitrary as I couldn't find any other prior art. [20:55] tedg: what is the url meant to do? [20:55] mhall119, Click on an appointment in datetime and open the calendar to the right view. [20:57] stgraber: ^ [21:00] ricmm: just getting ducks in a row to land kdub's flicker fix, need to bump so name again [21:00] ricmm: https://code.launchpad.net/~kgunn72/platform-api/bump-mir-dep12/+merge/187603 [21:00] tedg: http://iphonedevwiki.net/index.php/NSURL has calshow:// for urls to specific events already in the calendar [21:00] though it's not public [21:00] ricmm: didrocks said these build dep bumps should be top approved first before mir...altho, i suppose they need to be done in quick succession [21:01] kgunn: I'll approve it now, if thats ok [21:01] ricmm: i won't top approve until we've got mir on trunk ready w/ a bump [21:01] cool [21:02] kgunn: let didier do the top approving, as he usually does the succession [21:02] ricmm: thanks (i'll quote you :) [21:02] lool, no, I did not see the https://code.launchpad.net/~stgraber/mtp/upstart-fixes/+merge/187587 failure, will restart [21:03] stgraber: 6 [21:03] stgraber: ^ [21:06] mhall119, Yeah, but that was private and seems it was deprecated in iOS5 [21:07] ok [21:07] mhall119, So, no "useful prior art" :-) [21:07] tedg: something that points to a specific event would be better, I think, than just pointing to a time [21:07] mhall119, Why? Is there an event view? [21:07] tedg: I was close, it was "prior useful art" :) [21:07] tedg: there is, yes [21:08] and a better one being worked on currently as well [21:09] charles, Is there an event ID we could use? ^ [21:09] Something from EDS that is unique [21:11] tedg, yes, see e_cal_component_gen_uid() [21:11] that's the unique ID that the calendar uses [21:12] charles, Would it make sense to do a "calendar:///?eventId=$(foo)" ? [21:12] It'd be nice for calendar (on desktop) to be able to click on a day. [21:13] But it seems for events that'd imply the start/end time. [21:13] ted, right. in addition to startdate and enddate, calendar takes a handful of other arguments [21:13] source-uid, comp-uid, comp-rid [21:14] charles, K, that works for me. [21:15] charles, Add the useful ones to the bug? [21:15] tedg, sure [21:15] looking through the code right now to make sure one doesn't require the other two [21:19] charles, Think you forgot to bzr add the vapi for the indicator-sound branch. [21:19] teg, right you are [21:20] ...chasing the comp-uid, comp-rid handler from evolution to libecal, where it's turning into a dbus call, still digging :) [21:21] bfiller, hey, any idea who is working on the gmail web app? [21:22] noticed a bug where links are not loaded in the external browser [21:22] jono: the gmail webapp would be alex-abreu, I'm working on getting a good user-agent string so it renders nicely [21:23] thanks bfiller [21:23] bfiller, any idea where the gmail web app lives on LP? [21:23] so I can file a bug [21:24] jono: good question, just file it under browser for now and I'll make sure it gets to the right person [21:24] thanks bfiller [21:30] bfiller, https://bugs.launchpad.net/webbrowser-app/+bug/1231156 [21:30] Ubuntu bug 1231156 in webbrowser-app "GMail Web App: links don't open up in system web browser" [Undecided,New] [21:31] jono: thanks [21:31] fginther, lool: failed again: https://code.launchpad.net/~stgraber/mtp/upstart-fixes/+merge/187587 [21:32] meant, https://jenkins.qa.ubuntu.com/job/mtp-saucy-i386-autolanding/5/console [21:32] in the middle of the auto-landing this time... [21:37] stgraber, argh! These network failures will be the death of me. Will retry for now (an attempt at a fix is in the works) === wilee-nilee is now known as Guest60768 === sporkeee is now known as wilee-nilee [21:45] jono: how many contacts do you have on your phone? [21:47] popey, about 88 [21:47] ok go to them and scroll to the bottom [21:47] time it [21:49] on iphone it takes ~7 swipes to get from A to Z, on Touch it takes 7 swipes to get from A-G [21:50] * popey discovers you can tap the letter to do it quicker [21:57] mterry: so is the u-touch-session in the landing plan? [21:58] cwayne, oh, I guess I should re-do it. It is a twice-graduate, I figured it was still on the radar [21:58] cwayne, lightdm is stuck in proposed until the beta freeze gets lifted [21:59] earliest is tomorrow evening i guess [21:59] jdstrand: did you test your example click app after disabling ofono? [21:59] cwayne, looks like it got updated yeah. thanks ogra_(?) [21:59] even after I disable ofono, it still gives me status unknown [22:00] rsalveti: I tested on grouper vs mako [22:00] cwayne, I still could use confirmation that it fixes the daily tests. If you have free testing time... [22:00] rsalveti: you want me to disable ofono on mako now? [22:00] mterry, yeah, i keep an eye on this one [22:00] jdstrand: please, just stop ofono and give it a try [22:01] argh, something might be wrong locally, even the qml directly failed now [22:02] mterry: how can i help? i've never really touched/seen the daily tests [22:02] but i've got devices, am happy to help get stuff in :) [22:05] hm, stopped ofono, restarted nm, and still showing status unknown even when opening the qml file directly [22:05] rsalveti: I stopped ofono and now it gave an error [22:05] (both confined and unconfined) [22:06] cwayne, ogra_ tells me you have to use phablet-test-run [22:06] I noticed telepathy-ofono was running, so killed it too, no difference [22:06] cwayne, but I don't know the details [22:06] cwayne, this is to reproduce the same tests that the daily image tests do [22:06] yeah, we might have another bug around as well [22:07] rsalveti: restarted network-manager, same thing [22:07] jdstrand: works fine if you start ofono again, unconfined at least [22:08] I forced isOfonoAvailable to return false, and also gives me that error [22:08] rsalveti: yeah, I can confirm that (unconfined) [22:08] thanks for confirming it, will investigate a bit further [22:10] rsalveti: thanks. I tried uninstalled libofono-qt1 too, but that didn't help [22:10] uninstalling* [22:24] tedg: hey, gonna link to nih-dbus and use nih_dbus_path() too to get the object path