=== fmasi is now known as Guest6550 [00:23] this is only a idea/suggestion. in call log when call is unknow, in the toolbar at the bottom, add another 'Add' button so user can add a number to contacts. === chihchun is now known as chihchun_afk === chriadam|away is now known as chriadam === fmasi_ is now known as fmasi === fmasi is now known as Guest96304 === chihchun_afk is now known as chihchun [03:38] Ubuntu touch on Nexus 4 | http://askubuntu.com/q/473737 [03:39] pmcgowan: mhall119: it is a known limitation of the old application deployment model. In the new system we create .click package when you run the app, install it on the device, run with the appliction launcher, wait for the app stops and remove the package. It is more error safe solution, less sensitive for the changing shell rules, but comes with a performance cost. [03:43] pmcgowan: mhall119: when we started with the remote deployment I have realized that the (1) package name, (2) project name (4) app ID, (5) .desktop file and (6) main qml file was safer to have the same name. That is how the templates are in the SDK. I know it sounds silly, but that was the only way to ensure that the apps run, build and package up for most of the apps. [03:49] mhall119: your project is a cmake project I assume. because for qml project this issue is gone since malta. The cmake support of the new Run modell is WIP [04:10] mhall119: The new run modell was released a week ago in Utopic, I made a fresh backport to the SDK PPA for Trusty users. Could you please verify that the problem is gone or not? [06:14] cwayne1: ponf :) [06:47] Hi, when I try to flash my device I get the message, "Can't flash recovery image"... A google search was no help. Is anyone familiar with this error? [06:55] mardy, ping === ValicekB_ is now known as ValicekB === Ursinha-afk is now known as Ursinha === Guest96304 is now known as fmasi === chriadam is now known as chriadam|away [07:42] mterry_, howdy [07:42] seb128, hello [07:43] mterry_, is the phone greeter a different source/binary? how do I tell lightdm to use that instead of unity-greeter? [07:43] seb128, it's in the unity8-greeter package [07:43] seb128, once installed, it will tell lightdm to use itself [07:43] thanks === brendand_ is now known as brendand [08:15] sil2100, hey, can I get silo 19 published? [08:16] oSoMoN: sure! Wanted to do that in a moment, in some meetings right now [08:16] Sorry for the lag [08:16] sil2100, thanks! [08:19] Good morning all; happy Learn About Composting Day! :-D [08:19] oSoMoN: done! [08:19] thanks! [08:24] Morning all === vrruiz_ is now known as rvr [09:08] mterry_, hey, what's happening in silo 2? can I help? [09:09] robru, three things: cgmanager needs to get promoted to utopic (waiting on lxc/kernel fixes first), I need to fix a bug found yesterday with the launcher, and QA needs to review the silo [09:10] mterry_, any word on those kernel fixes? can it be today? [09:10] version 50 of ubuntu touch works smoother on touch and speed on my nexus7 [09:11] robru, kernel landed today. But cgmanager still has to wind its way through. [09:11] and update of shorts app also fixxed the crash [09:12] mdeslaur, nice job with tipcalc! I noticed the following warning with it when listing the available click packages: http://pastebin.ubuntu.com/7542637/ [09:13] mterry_, is cgmanager in proposed already? [09:13] robru, yes [09:14] Hi guys, I have a bit of a problem trying to install ubuntu touch on grouper (I know it's deprecated, but the stable channel should still work...) When running ubuntu-device-flash --channel=stable --bootstrap --device=grouper I get the following error after waiting for ages: "Can't flash recovery image" [09:14] Does anyone know what may cause that error? [09:15] mterry_, hmmm it looks like cgmanager is blocked :-( http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html [09:15] zeorin: on wich device? [09:16] robru, right, it's been blocked since monday, waiting on the lxc/kernel stuff [09:16] robru, I'm re-running the lxc autopkg tests now [09:16] mterry_, oh ok, great. I guess we have to poke release team to rerun the test in britney? [09:16] If I don't specifiy device=grouper then it doesn't recognize that my device is connected. Device is grouper (Nexus 7 2012, wifi only, 8gb). If I run fastboot devices it's listed successfully. [09:16] zeorin: well i have channel=devel installed, did you try that one? [09:17] actually I didn't, as I was told that it's no longer supported and that there's no devel image for it (although I did find some trusty images on a server for grouper, so I'm confused about it's support status) [09:18] zeorin: what does that grouper do exactly do? [09:18] robru, hrm, it failed again [09:18] http://d-jenkins.ubuntu-ci:8080/view/Utopic/view/AutoPkgTest/job/utopic-adt-lxc/ [09:18] robru, do you know why that might be? Log doesn't give much [09:18] stgraber, ^ [09:18] zeorin: the way i have been told its best to install channel=devel for you to receive further updates [09:18] mterry_, we're close ... kernel is in ... now lxc and cgmanager need to move ... [09:18] ogra_, lxc autopkgtest fails again, but for mysterious reasons ^ [09:18] sigh [09:19] lotuspsychje, Well, if I try to flash without specifying the device flag, then I get the message: "Expecting the device to be in the bootloader... waiting" [09:19] so it re-ran already ? [09:19] mterry_, bah I don't have the VPN setup on this laptop! [09:19] mterry_, can you pastebin that log for me>? [09:19] zeorin: did you adb reboot before? [09:20] lotuspsychje, if I specify the device flag (--device=grouper) then after trying to flash it eventually (usually after an hour or so), tells me "Can't flash recovery image". [09:20] zeorin, so did you boot it to the bootloader? [09:20] lotuspsychje, my device is in fastboot mode when I try this [09:20] ok [09:20] zeorin: can you just try channel=devel [09:21] robru, http://paste.ubuntu.com/7542689/ [09:21] lotuspsychje, I'm pretty experienced when flashing things, I have many many devices with Cyanogenmod on it, a dualboot of Firefox OS and CM on my Galaxy S2 [09:21] mterry_, thx [09:21] dpm: hrm, interesting...thanks, I'll take a look [09:22] zeorin: this is the guide i followed, but channel=devel instead of trusty channel: http://itsfoss.com/install-ubuntu-touch-nexus-7-2013/ [09:22] lotuspsychje, If try the devel channel and don't specify the --device flag, I have the same error: "Expecting the device to be in the bootloader... waiting" and below: "Exit status 1" [09:22] zeorin: grouper is no longer supported [09:22] the stable channel shouldn't work, we haven't supported nexus 7 2012 since long before then [09:23] oh right it only works for nexus7 2013 right? [09:23] yes [09:24] dpm: that's weird...what's the contents of your /opt/click.ubuntu.com/com.ubuntu.developer.mdeslaur.tipcalc/0.2/.click/info/com.ubuntu.developer.mdeslaur.tipcalc.manifest file? [09:24] I am aware of the fact that it's 'no longer supported', however I was told that stable ought to still work. In any case the problem is that ubuntu-device-flash doesn't even recognize the device, it doesn't even get to the part where it actually tries to flash an image. Also, if it's no longer supported, what are the grouper images on this server for: http://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/current/ ? [09:24] zeorin: how about sell the 2012 version and buy cheap 2013 somewhere, new touch improved alot! [09:24] dpm, camera-app and gallery-app landed with latest translation updates [09:25] the old stab;e might work, as in 13.10 - but the 14.04 one wont [09:25] Kaleo, hi :) [09:25] The docs say that ubuntu-device-flash will automatically pick the latest image in the selected channel for the device [09:26] mdeslaur, it's empty [09:26] did you guys fixxed better touching support to version 50? [09:26] yes, and that can be overridden with --revision=NN [09:26] oSoMoN, awesome, thanks! [09:26] after updating things going real smooth here on my n7 === rickspencer3_ is now known as rickspencer3 [09:26] dpm, you’re welcome [09:26] Indeed, if I specify the --device=grouper flag, it does download an image === vying is now known as Guest54473 [09:26] zeorin: which image? it tells you at the start of the output [09:26] mterry_, sorry, did you paste the wrong link? your paste shows 0 failures 0 errors... [09:27] robru, yup! [09:27] dpm: wow, not sure how you managed to get a empty file there....0.3 is now in the store, try upgrading and see if you still have that error? [09:27] But I think that specifying the --device=grouper flag skips the device detection completely, as it will try to install even if there's no device connected when I use that flah [09:27] robru, I'm guessing the failure is because of the signal 15 at end? [09:27] Flashing version 10 from ubuntu-touch/stable channel and server https://system-image.ubuntu.com to device grouper [09:28] mdeslaur, no idea. I can't see 0.3 under System Settings > Updates yet, but I'll try later on [09:28] mterry_, bah, what? i guess that's infrastructural or something? why would it get killed with 15? [09:28] robru, I dunno! [09:28] mterry_, who can we poke about this? release team? [09:28] dpm: that's weird too...it's been there since yesterday [09:28] dpm: and others have updated it [09:28] robru, who maintains the autopkgtest infrastructure? === tedg is now known as ted [09:29] mterry_, dunno who is the official maintainer but I sometimes see pitti working on it... [09:30] mdeslaur, hm, I didn't have it installed, it seems! Which is even more strange, as I don't remember having uninstalled it [09:30] mdeslaur, after installation click list is happy again [09:30] dpm: not sure how that happened [09:31] me neither, but nevermind [09:34] every start of ubuntu touch on my nexus7 brightness resets to halfway dim, any plans to fix this? [09:38] popey, lotuspsychje, I think that it may be related to my usb cable... The results I was getting previously are now inconsistent... [09:39] mterry_: it's known [09:39] mterry_: even though the fix is in the archive, the adt VMs aren't running the new kernel yet [09:39] mterry_: (they upgrade to it but never reboot so the tests still run against the broken kernel) [09:39] stgraber, ah... huh. But the tests pass [09:43] mterry_, jibel is fixing the testbed [09:43] apparently not trivial, so it might take a bit === ted is now known as tedg === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === fmasi is now known as Guest68919 [10:17] davmor2: can you try something for me for mms? [10:18] sergiusens1: I can in a bit looking at something else right now, but if you leave instructions I can :) [10:19] davmor2: wget http://people.canonical.com/~sergiusens/mms/bug_1324182/nuntium; adb push nuntium /usr/bin/nuntium; reboot or restart nuntium and resend message [10:19] davmor2: I have successfully decoded your messages [10:19] but want to make sure it works for you :-) [10:19] davmor2: or we can wait for the silo and I can ask you to test the silo if it pleases you better [10:20] sergiusens1: no I'll test it in a minute for you [10:20] thanks === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === Ursinha is now known as Ursinha-afk [11:04] sergiusens: I'm not having much joy here with that [11:05] davmor2: can you phablet-shell, stop nuntium and just run it manually from a the cli? [11:09] sergiusens: d'oh that'll be why nuntium wasn't executable [11:09] lol [11:10] that explains a lot :-P [11:10] I'll try again now then :) === chihchun_afk is now known as chihchun [11:13] sergiusens: \o/ [11:13] yay [11:14] sergiusens: although the text is wrong [11:14] sergiusens: let me screenshot it [11:14] davmor2: just says text_0.txt or something like that? [11:15] davmor2: if it is, it's just a missing feature in the UI [11:15] sergiusens: http://davmor2.co.uk/~davmor2/screenshots-phone/device-2014-05-29-121443.png [11:15] davmor2: yeah, you can log a bug specific to the messaging app for that one [11:15] sergiusens: will do [11:16] davmor2: there's a media bug as well if you send a video [11:16] davmor2: in case you want hints for new bugs ;-) [11:16] sergiusens: like I need help :D [11:18] :-) [11:18] * sergiusens moves === Ursinha-afk is now known as Ursinha [11:23] sergiusens: feel free to point Jim Hodapp at the media issues :) [11:32] cyphermox: have you moved from room to room with a device does your wifi get dropped? [11:53] Does ofono need to conflict with modemmanager? [11:54] alternatively, does powerd need to recommend ofono? alternatively, should we remove powerd on desktop-next? [11:57] Laney: ofono and modemmanager would fight for control of the same modems [11:57] mardy, ping [11:57] so it's powerd recommends ofono and network-manager recommends modemmanager [11:57] and they conflict [11:57] so then one, or both, to not worl [11:57] so what is the right thing to do there? [11:57] not sure [11:57] have some alternates? [11:58] do we need powerd on desktop right now? [11:58] don't know, but it seems broken anyway [11:58] I could drop the conflicts [11:58] and have NM alternate recommands ofono and modemmanager [11:58] well, or really [11:58] yeah [11:59] with MM first I guess [11:59] yup [11:59] if that works for you it works for me [11:59] that should work sufficiently, NM should be able to handle iether or [11:59] this makes -next builds fail atm [11:59] so if you could do it soon ;-) [11:59] Laney: sure [11:59] no guaranteeing some other problem doesn't come up after this [12:01] ogra_: ok, so what is pulling dictionaries-common onto the phone in the first place (ref: emacsen-common)? It's not in the ubuntu-touch task, and it's not shown at all in the ubuntu-touch germinate output [12:03] slangasek, my guess would be the onscreen keyboard [12:03] though its weird that germinate wouldnt show it [12:03] we are also seeding gnome langapcks [12:03] *langpacks [12:07] ogra_: why would the osk not be part of the task? [12:10] ssweeny: ping [12:11] slangasek, it should ... [12:11] and it seems to be [12:11] and tasks recurse [12:11] slangasek, ubuntu-keyboard is the package (and there are plenty lang specific ones) [12:11] so it's not the keyboard [12:11] then it must be the langpacks [12:13] i think via hunspell or so [12:13] ogra_: no, the langpacks also have the ubuntu-touch task [12:14] weird [12:16] slangasek, it is aspell via sdk-libs [12:16] http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu-touch.utopic/sdk-libs [12:17] now how did that end up there [12:18] sigh ... any why does the germinate output show recommends again ... [12:19] cjwatson, didnt we fix that twice now ? [12:19] oh [12:19] ogra_: the sdk-libs seed should be unrelated to what's on the phone, surely [12:19] ah, no, touch does inherit from sdk-libs [12:19] i guess sdk-libs needs "* Feature: no-follow-recommends" at the top too [12:20] STRUCTURE:touch: minimal sdk-libs [12:20] its a dep of touch [12:22] ogra_, for some reason you seem like the person to ask, let me know if I am wrong ... [12:23] ask away :) [12:23] ogra_: yeah; I see that the STRUCTURE file also lists 'feature no-follow-recommends', dunno why that doesn't already do what's expected (given that it's not a syntax error) [12:23] I seem to have some weird hang in the emulator as I am adding a U1 account [12:23] which is the best way to log a bug? [12:24] slangasek, we needed to add that extra entry in the touch file directly too ... i think colin wanted to look into that though ... seems sdk-libs needs it as well ... its not an issue apart from the germinate output itself ... the install actually forces recommends to off in apt [12:24] rickspencer3, calling ubuntu-bug i guess ... but i think we have something open for that ... davmor2 or popey would know [12:25] ogra_: then that still doesn't answer the question of how it got on the phone in the first place [12:25] yeah [12:25] rickspencer3: pass, not used the emulator much [12:26] popey, no, but i think i remember we had some bug for U1 accounts open [12:27] independent from the emulator [12:28] ogra_, popey I'll try to find the package myself [12:29] rickspencer3, oh, you just look for the package name ? [12:29] ogra_, yeah [12:29] rickspencer3, account-plugin-ubuntuone [12:29] thanks ogra_ [12:30] dpm: btw, that bug i mentioned yesterday: https://bugs.launchpad.net/savilerow/+bug/1324296 [12:30] Ubuntu bug 1324296 in ubuntu-keyboard "Password prompt is shown repeatedly after properly entering the password using pinyin input method" [Undecided,New] [12:30] slangasek, added no-follow-recommends to sdk-libs now, next germinate run should look better [12:31] (probably giving a better overview) [12:34] rickspencer3: apport-bug but if it is the black screen on clicking back in the accounts app there is a bug for that already [12:36] slangasek, i'm pretty sure its pulled in by the localized keyboard packages through hunspell-$lang [12:37] slangasek, ubuntu-keyboard-german -> hunspell-de-de -> dictionaries-common [12:37] there you go [12:38] sergiusens: care to confirm https://bugs.launchpad.net/messaging-app/+bug/1324498 [12:38] Ubuntu bug 1324498 in messaging-app "MMS: I now recieve an image with a recent fix but the text is not recieved" [Undecided,New] [12:38] ogra_: where do you see that? [12:38] on the device [12:38] cwayne1, ah, sorry for not coming back to you earlier. I tried to reproduce it this morning, but I could not figure for the life of me how to switch the IM in the keyboard [12:39] ubuntu-keyboard-$lang are seeded in touch [12:39] dpm: press and hold the globe on the bottom of keyboard [12:39] dpm: btw you can now get the maedemo image by flashing the latest devel-customized-demo [12:39] cwayne1, oh, man I tapped and tapped, didn't think of hold! [12:40] yeah, not exactly intuitive :) [12:40] oSoMoN: hi! [12:40] artmello: hello! [12:40] cwayne1, we have worse unintuitive bits :) [12:40] oSoMoN, artmello: are you guys around? I need some gallery-app devs to assess something for me ;) [12:41] * ogra_ just learned this week that press and hold gets rid of the X when closing apps [12:41] sil2100: hi there [12:41] oSoMoN, artmello: so the story is - the new UITK broke gallery-app tests, we're getting a solid 19 failure rate in gallery-app... so bzoltan knows about it and thinks gallery-app might need to adjust to the new version somehow [12:42] oSoMoN, artmello: so what I would want is to have someone look at the failing tests and assess if changing this in gallery-app would take a lot of time and effort or not [12:43] sil2100: sure, no problem [12:43] oSoMoN, artmello: since we might just revert UITK quickly, but first I want to know if it's needed - since if it's like a 15 minute change in gallery then I'm ready to risk it [12:43] bzoltan: what is the problem? [12:43] artmello: thanks :) You can find the failures here on the dashboard, but also can be easily reproduced locally: [12:43] http://ci.ubuntu.com/smokeng/utopic/touch/mako/53:20140529:20140523/8277/gallery_app/ === gnu is now known as Guest43510 [12:44] artmello: just running gallery-app tests on image #53 results on the failures happening [12:44] artmello: http://pastebin.ubuntu.com/7543771/ [12:47] bzoltan: ok. What changed on UITK? [12:47] ogra_, rsalveti which package has the UI for the indicators (power in particular) ? [12:47] slangasek, what is the problem with dictionaries-common btw ? (i didnt really get why we are researching it) [12:47] jdstrand, ping [12:48] janimo, indicator-power ... :) [12:48] artmello: loads of things [12:48] balloons: hi! [12:48] ogra_, that was my guess after typing in the question :) [12:48] heh [12:48] jdstrand, hello.. I'm back from the dead (or so it feels) [12:48] ogra_, but since I am not sure how much touch UI uses in common with desktop it's worth asking [12:48] hehe [12:48] ahayzen_, let's chat with jdstrand in here [12:48] balloons, o/ [12:48] jdstrand, so autopilot has landed, but I heard something about the apparmor changes still need work [12:49] so what's up? [12:49] balloons, this was part of my conversation with jdstrand http://pastebin.ubuntu.com/7543796/ [12:49] ogra_: any idea when ubuntu-touch-session will land? :P [12:50] cwayne1, hmm, did i forget about it ? [12:50] sorry sorry sorry ... will try to get that in this week before flying home [12:51] balloons: it shouldn't be a difficult change, just have to create some dirs so that the apps can write their app-specific dirs under them [12:52] jdstrand, but why would the test have permission to write them? Because the phablet user should since it's under HOME? [12:52] mkdir -p $tmpdir/.cache $tmpdir/.config $tmpdir/.local/share $tmpdir/confined [12:52] yes [12:52] balloons, this was my diff to our AP tests i didn't know whether i did it right? http://pastebin.ubuntu.com/7543804/ (to this branch lp:~music-app-dev/music-app/use-mediascanner2.0) [12:52] you are already creating the tmpdir [12:52] jdstrand, yes indeed.. Sounds fine then [12:52] just sprinkle a few things under it [12:53] cool [12:53] ahayzen_, so let's tweak your branch [12:53] ahayzen_, do you have an mp we can look at? [12:53] ogra_: maybe, still not seeing my change on the image [12:53] balloons, well we have https://code.launchpad.net/~music-app-dev/music-app/use-mediascanner2.0/+merge/214140 [12:54] balloons, but i didn't push my changes yet [12:54] janimo: yeah, indicator-power, and probably some logic in unity8 as well [12:54] balloons, as i couldn't get it to work locally [12:54] cwayne1, yeah i think it got lost in my travel preparation ... [12:54] balloons, i could push and then u can edit (as the branch is owned by all of us) [12:55] rsalveti, who would know where in unity? I see indicator power has code for getting the average charge of two batteries [12:55] rsalveti, I am looking into ignoring one of the two detecte batteries as it has no useful info at all [12:56] hi...how to create a final app in ubuntu-sdk? [12:57] ogra_,slangasek: heh, yeah, last time I looked I discovered that "feature no-follow-recommends" wasn't actually implemented in germinate and I wanted to get round to doing that; so you need to do it in the individual seeds [12:57] yup, already done [12:58] right, just following up that it's a germinate bug IMO [12:59] janimo: we had that issue with some other phone as well, I think nexus 5 [12:59] janimo: upower was able to see 2 batteries [12:59] janimo: usually the kernel should only export one, or you can ignore it in upower, or you can change indicator to just show one [13:00] Wellark: ping === tedg is now known as ted [13:23] artmello: how's the gallery/UITK sitution looking? You think reverting UITK is best? [13:23] Or is it something easy fixable? [13:27] sil2100: gallery does not seem to be broken, since we are able to change tabs. it looks like that we have some problem with the AP tests, we are looking for a element (header) that does exist but we are not able to find it [13:27] sil2100: I am talking with QA guys to see how we can fix the AP tests, should be simple [13:27] artmello: do you think it can be simply fixed in the test? Like, removing this check somehow? [13:27] Awesome [13:27] sil2100: if we can wait a little bit more I guess we can fix it, no need to revert UITK [13:28] sil2100: we cannot disable since it is the AP code to change tabs [13:28] We'll wait for some moments still then [13:44] sil2100: robotfuel was helping taking a look on the gallery/UITK issue. it seems that is not an issue on gallery itself, and he is taking a look on how to fix the problem [13:44] artmello, robotfuel: thanks :) [14:02] artmello: sil2100: elopio broke the header and he is fixing it. he can't expect apps to have a main_view object. the gallery app has a main_screen [14:03] artmello: sorry for breaking your tests. [14:04] I'll propose a fix. [14:04] urg. [14:04] elopio: robotfuel: anyway good to hear it can be fixed, I prefer that then reverting [14:05] so i got my nexus7-2013. and i followed these instructions to root it, but now all i gets is the bouncy balls and i can't even get to the bootloader menu. [14:05] sil2100: can we wait for a new release of the toolkit, or should I duplicate the fix on the gallery app? [14:05] http://itsfoss.com/root-nexus-7-2013-ubuntu-linux/ [14:05] sad [14:05] elopio: so, the actual fix is needed in the toolkit? [14:05] sil2100: yes it is [14:05] i've rooted a bunch of devices before, an di've always been able to get back to the bootloader, but not now. any ideas? [14:06] the problem is on the test, that assumes too much. [14:06] elopio: just to understand, it can be worked-around in gallery, right? [14:06] sil2100: it can, yes. [14:07] I can override the default behavior of the test. [14:07] elopio: ok, so... I would propose this: could you work it around in gallery temporary and fill ina fix in UITK as well? [14:08] Since releasing UITK might take longer (as we need to run like all APs) [14:08] And it would be nice if the issue wouldn't cloud real possible problems in the meantime [14:09] Since if I understand correctly, it's not a problem in the actual behvior, but in something what autopilot does, right? [14:09] sil2100: yes. https://bugs.launchpad.net/gallery-app/+bug/1324556 [14:09] Ubuntu bug 1324556 in Ubuntu UI Toolkit "The header autopilot helper will fail if the parent is not called MainView" [Undecided,In progress] [14:09] I'll propose branches to both. [14:10] elopio: you're our hero [14:10] Thanks [14:10] elopio: np :) [14:10] ;) [14:10] robotfuel: thx [14:10] sil2100: well, I was the one who broke it. So I'm a lame hero. [14:11] elopio, the UITK is an amazing accomplishment.. [14:11] keep your hero status :-) [14:15] elopio: if you have the fix pushed into gallery, could you poke sergiusens and popey to push it out to the store? [14:15] We could kick a new image then [14:15] sil2100: working on it. [14:17] * ogra_ hugs elopio [14:17] awesome ! [14:17] Hi all, just want to mention that despite many times being told many times here it's not possible, I have Ubuntu Touch installed on a grouper (Nexus 7 2012, wifi only), which is deprecated. [14:17] zeorin_, who said it is not possible ? [14:17] ubuntu-device-flash --channel=stable worked [14:17] ubuntu-device-flash --channel=devel has no images available [14:17] it is not supported and we will not care for bugs from it ... [14:18] but it is surely possible with the old images [14:18] but the manual install method works with the images available from these instuctions https://wiki.ubuntu.com/Touch/Install#Manual_Download_.26_Installation [14:18] I currently have Utopic (14.10) installed. [14:18] jdstrand, I tweaked music to do what was requested, but I still see some apparmor errors :-( http://paste.ubuntu.com/7544237/ [14:19] calendar too; http://paste.ubuntu.com/7544305/ [14:19] ogra_: would be nice if we could get gallery released and an image built before mterry's greeter... [14:20] well [14:20] jdstrand, the click.rules file seems updated properly [14:21] sil2100, i'm just scared that then something else breaks again before the greeter landed [14:21] :) [14:21] balloons: ok, let me look [14:22] ogra_: we won't wait for the test results, I just want to like have an image building before the greeter just to make sure, just hope gallery will be released in like the nearest 20-30 minutes... [14:22] balloons: ok, so, seems like .config wasn't created correctly [14:22] Since mterry wants to QA sign-off greeter anyway [14:22] So we might have some time still [14:22] sil2100, ok, thats fine then [14:23] balloons: the others are really mocking up what the music app should access since mediascanner never ran [14:23] that one is a harder problem, but the music app should also be the outlier [14:23] jdstrand, yea, I moved to calendar to simplify things to see what was wrong.. no mediascanner there [14:23] right [14:24] were there denials with the calendar? [14:24] sil2100, though i'm just seeing cjwatson talk about procps issues in another channel ... we should probably wait with an image build until we are sure the archive is ok again [14:24] jdstrand, yes, http://paste.ubuntu.com/7544305/ [14:24] I imagine probably with .config. do you have a branch? [14:24] ogra_: ok, right... damn, a bad day for image builds it seems [14:24] well, for the archive rather ... [14:25] * ogra_ guesses desktop users that use debs suffered more today :) [14:25] jdstrand, lp:~nskaggs/ubuntu-calendar-app/fix-ap-env-setup [14:26] sil2100, oh, i see in -devel that it was catched by -proposed ... so we should be good [14:27] balloons: ah, tempdir is ~/autopilot/fakeenv. you wanted it as ~/autopilot/fakeenv// [14:27] balloons: basically, do the mkdirs after running TempDir [14:28] balloons: and use what TempDir creates as the parent of all these paths [14:28] example rule: [14:29] owner @{HOME}/autopilot/fakeenv/*/.cache/@{APP_PKGNAME}/ rw, [14:30] jdstrand, I figured it was something simple.. critical thinking is a challenge today :-) Tweaking now [14:30] hehe [14:34] ted: Wellark: ping, any chance someone could take a look at this bug? https://bugs.launchpad.net/savilerow/+bug/1324296 it's quite important to be fixed for MAE [14:34] Ubuntu bug 1324296 in ubuntu-keyboard "Password prompt is shown repeatedly after properly entering the password using pinyin input method" [Undecided,New] [14:40] jdstrand, ok made some tweaks, but still see apparmor errors http://paste.ubuntu.com/7544447/ [14:40] everything is pushed [14:41] jdstrand, it seems we still need the APP_PKGNAME in there, so I guess that's on me [14:42] balloons: yeah, seems like just need to keep recreated the test environment so the apps can start creating their own directories rather than their parent dirs [14:43] I'm a bit surprised by media-art coming in, but seems to be needed [14:43] cwayne1: pong [14:44] Wellark: any possibility of looking at this bug? https://bugs.launchpad.net/savilerow/+bug/1324296 [14:44] Ubuntu bug 1324296 in ubuntu-keyboard "Password prompt is shown repeatedly after properly entering the password using pinyin input method" [Undecided,New] [14:44] balloons: (essentially a click app is never expected to be launched within a totally empty environment-- all these things get created long before a click app starts, and we don't want to allow these accesses because it weakens security) [14:45] balloons: another thought that might work better in terms of maintenance is instead of creating all this hidden dirs in the TempDir, to copy all of the existing '.' dirs into the TempDir [14:45] s/this/these/ [14:46] jdstrand, well, the issue with cloning the environment is end up with old application data in there [14:47] balloons: you just remove TempDir after the test run [14:47] tests* [14:48] you want to do that anyway I would think [14:48] jdstrand, I meant data from the old home.. Or I am missing what you are saying. Are you saying to copy /home into the temp home or something else? [14:49] I was saying something like 'cp -a $HOME/.[a-zA-Z0-9]* $HOME/autopilot/fakeenv/$TempDir/' [14:49] jdstrand, right.. so won't that also copy the configs for the click apps.. thus we don't have a pristine environment [14:50] that way you use what is already created and you don't have to worry about all the new stuff being added (ie, you don't have to maintain what you put in to TempDir as often) [14:50] cwayne1: the secret agent passes the text from the prompt unchecked and unmodified to network-manager and if the prompt is poppping up again it means that NM didn't like what was entered to the prompt [14:50] cwayne1: do you have this running somewhere? maybe I could pop by and take a look [14:50] balloons: true, you'd want to remove those if you went that route [14:51] maybe it wouldn't save a lot [14:51] Wellark: i wasn't invited to malta unfortunately :) [14:54] jdstrand, right, I mean I could specify what to remove, but that's not much better than were we where with moving files around (backup/restore) [14:55] elopio: how's it progressing? [14:55] it might if you could make some assumptions (rm ~/.config/com.*) [14:55] cwayne1: ;( [14:55] mterry, so sil2100 says that he usually uses om26er for QA signoffs in this TZ but he's not around. we'll have to tackle jfunk and get him to allocate somebody [14:55] Wellark: so i wonder if maybe it's an ubuntu-keyboard issue then [14:55] it wouldn't be as robust as I initially was thinking, but might be better [14:55] Elleo: hey, would this seem like a keyboard issue maybe? https://bugs.launchpad.net/savilerow/+bug/1324296 [14:56] Ubuntu bug 1324296 in ubuntu-keyboard "Password prompt is shown repeatedly after properly entering the password using pinyin input method" [Undecided,New] [14:57] jdstrand, is APP_PKGNAME not an env var? [14:59] cwayne1: could you check out lp:unity-notifications and run examples/sd-example-password-entry.py [14:59] sil2100: almost there [14:59] Wellark: on the phone? [14:59] cwayne1: not sure, it's possible that the text isn't getting committed since the pinyin keyboard always has suggestions enabled; I have a feeling there was a bug relating to text not getting commited when changing field [15:00] although I thought that had been fixed [15:00] Elleo: any logs i could get or anything i could do to verify whats the problem? [15:02] cwayne1: well you can run maliit-server with MALIIT_DEBUG=true for some debug info, not sure if there'd be any output that'd help here though [15:02] mterry, davmor2: so, we usually ask for QA sign-off only on traincon0, but there's nothing bad with a lander asking for that in other occassions if he's not sure if he can assure for nothing-broken [15:03] mterry, davmor2: but that's up to the lander if he wants to wait or not [15:03] balloons: no, but APP_ID is. APP_PKGNAME=split(APP_ID, '_')[0] [15:04] cwayne1: aha, just did a quick text in the contacts app and it seems preedit stuff does get commited if you change field but it doesn't if you click a button [15:04] jdstrand, the sad part is I don't know the APP_ID until I launch the app, which by then is too late [15:04] cwayne1: yes, on the phone. [15:04] cwayne1: with that input method [15:05] Elleo: hm, so is that fixable? [15:05] cwayne1: you can branch the repo on your desktop and adb push that one file to the phone [15:05] balloons: why do you need APP_PKGNAME? to clean out a cloned dir in TempDir? [15:06] cwayne1: yeah, just seeing what signals the keyboard gets under those circumstances to see if we can commit [15:07] jdstrand, I guess I'm getting lost on how we fix the issue. apparmor is still complaining the app is making dirs under /home/phablet/autopilot/fakeenv/tmpcky05tii/*. However, the app should be allowed to write to things under this directory, so long as it's under the app pkgname [15:07] balloons: right so, the app is allowed to create: [15:08] /home/phablet/autopilot/fakeenv/tmpcky05tii/.config/$APP_PKGNAME [15:08] so you don't have to [15:08] the app is not allowed to create: [15:08] /home/phablet/autopilot/fakeenv/tmpcky05tii/.config/ [15:08] so you have to [15:09] if you look at the denial, you can see which directories you need to create. you shouldn't have to worry at all about APP_PKGNAME for creating dirs, cause the app itself will [15:11] jdstrand, ahh.. right.. so all these sub dirs the app is trying to create I need to provide. However, that can get huge with subdirs [15:11] balloons: also note that directories end with '/' in apparmor logs [15:11] hence our conversation.. whew I'm slow today. So fine, is there a master list somewhere of what the /home env looks like for an app? [15:11] balloons: right, which is why I thought it might be easier to clone (and prune) [15:13] balloons: we know that APP_PKGNAME starts with com.ubuntu. and com.canonical. which could be useful for pruning a cloned dir [15:14] jdstrand, yes if you go the pruning route, you have to know everything that could cause you to mess up.. it's just not as clean [15:15] for instance music uses mediascanner which lives in it's own location.. still perhaps it's the way to go [15:15] I'll play a bit [15:15] boiko: hi. have you taken a look at the call hint request for the telephony-service? [15:18] awe_, https://wiki.ubuntu.com/SecurityAndPrivacySettings#SIM_PIN [15:23] awe_, https://wiki.ubuntu.com/SecurityAndPrivacySettings?action=diff&rev2=40&rev1=39 [15:25] thanks mpt! [15:26] jdstrand, ok so as proof of concept, it seems to be working now. I don't see errors anymore [15:26] I just kept creating the dirs.. now we'll have to decide how to maintain it [15:32] cyphermox, just curious, do you have a reference for ad-hoc networks actually delivering malware? [15:33] sure, hold on [15:33] (as opposed to just causing Windows XP to set up a network with the same name) [15:34] balloons: nice! :) [15:36] mpt: I checked, seems like there is only free public wifi in nm-applet; but: http://www.npr.org/2010/10/09/130451369/the-zombie-network-beware-free-public-wifi [15:36] this kinds of networks can just be abused anyway [15:37] cyphermox, “Unintentionally creating or connecting to the ad hoc network isn't inherently harmful, despite its virus-like spread. It does, however, provide an access point for hackers to come in and check out the user's files.” By which they mean the Windows XP user that created the network. [15:37] indeed [15:37] however, it's easy enough for someone to make use of such a network for malicious purposes, depending on what is exposed on it [15:38] (on the device that connects to it) [15:38] How is that different from any other ad-hoc network though? [15:42] elopio: hi! Sorry to poke again, but I was disconnected - is there any progress? [15:44] sil2100_: yes. waiting for jenkins to confirm that it works [15:44] https://code.launchpad.net/~elopio/gallery-app/workaround1324556-get_header/+merge/221415 [15:45] sil2100_: but the workaround is extremely ugly. Autopilot messes with the inheritance making it too hard to override. [15:45] so I would recommend to wait for the proper fix. [15:46] jdstrand, so, migrating to music, I see apparmor denial trying to open the mediascanner db.. it's not under the appname dir ofc [15:47] here's the 3 errors; it's the third; http://paste.ubuntu.com/7544803/ [15:52] elopio: I would say it's good enough, it's well documented as a workaround, and since you also will push a merge for the right fix we might even try incorporating it to the already-prepared UITK landing [15:52] elopio: so that this could land tomorrow [15:52] popey: are you still around? [15:53] rsalveti: hi, did you have time to look at those two commits for hammerhead's kernel? [15:53] jdstrand, anyways, I'm curious why the app can't even do an open operation on the db.. it will certainly need to be able to read it [15:54] sil2100_: ya === chihchun is now known as chihchun_afk === oSoMoN_ is now known as oSoMoN [16:33] mterry: is the split greeter stuff going to land this week? [16:34] cwayne1, yeah... maybe tonight (malta time) [16:34] okay, so then that would enable getting system language from accountsservice right? [16:36] cwayne1, yeah for the greeter session yeah [16:38] mterry: i thought it was going to take the system locale from there as well? [16:38] cwayne1, uh isn't the user session already grabbing language from AS? [16:38] cwayne1, we can't change system language via AS [16:39] cwayne1, but greeter pays attention to user's language in AS [16:39] last time i asked for how to set language, i was told to set it in ~/.pam_environment, not AS [16:42] cwayne1, yes... I think AS is a copy of pam_environment's info? I'm actually not sure where system-settings stuffs the language such that it gets used next time === ted is now known as tedg [16:57] hello [16:57] anyone here [16:59] does anyone know how to develop for ubuntu [16:59] while running mac os [16:59] i don't want to run a VM === tedg is now known as ted === ted is now known as tedg [17:24] ping tyhicks === Ursinha is now known as Ursinha-afk === tedg is now known as ted === ted is now known as tedg [19:21] balloons: so the mediascanner needs adjustments to the click-rules file [19:21] s/mediascanner/mediascanner access/ [19:21] hey jdstrand [19:22] jdstrand, I'm concerned because I'm also seeing apparmor errors for 'open' and 'mknod' operations [19:22] this is where having a second user would have made things easier. basically, there are additional accesses in the policy group to things in ~/.config that are now being redirected into fakeenv [19:24] jdstrand, well, hmm.. technically if things get autopkg'd properly, a second user might not be near the issue it was [19:24] since the test would be properly self-contained and easily provisioned to a new user [19:25] so those rules also have to be added to fakeenv, *but* if the app is supposed to be tested under confinement, we shouldn't just automatically add them to the policy unless they were in the app's policy when not directed [19:25] right [19:25] there might be an alternative [19:26] right.. we can't add policy just for testing :-) [19:27] no, the alternative won't work. I was thinking that we could use an alias rule, but that has to happen in the preamble [19:27] what would work is a symlink [19:28] ln -s /home/phablet/.cache/mediascanner-2.0 /home/phablet/autopilot/fakeenv/tmp_7dfprxt/.cache/mediascanner-2.0 [19:29] the symlink gets resolved to /home/phablet/.cache/mediascanner-2.0 so the policy doesn't have to change. but, it is using the actual data (but, in this case it is ro) [19:29] jdstrand, I also noticed something very odd.. Calculator which doesn't use any of this fakery was blowing up with apparmor errors trying to play around in tmp dirs it didn't create [19:30] I'd need to see the errors === forcev is now known as FunkyPenguin [19:30] err, denials [19:30] I don't want to sidetrack the discussion..heh but http://paste.ubuntu.com/7546195/ is an example [19:31] I just removed the entire fakeenv folder and the tests seem happier now. [19:31] that is using fakeenv... and it is the same issue [19:32] jdstrand, but it's not.. this is weird [19:32] the xdg dirs had to be set or left over from a previous run or something [19:32] jdstrand, ahh.. that's it [19:32] I'll bet it's leftover indeed.. the cleanup didn't happen when it crashes [19:33] * jdstrand was referring to the environment (which is likely also true) [19:33] ok, anyways back to it.. so you are thinking of using a symlink [19:35] the symlink will work, absolutely and doesn't have the issues with the additional policy, but, it does mean using the actual files. that is perhaps ok- before it was all user files, fakeenv is totally clean, symlinks with fakeenv is in between [19:36] you can just : ln -sf $HOME/.config// $HOME/autopilot/fakeenv/tmp.../.config// [19:37] for all the dirs that you encounter issues with that aren't the app specific directories [19:38] well of course for mediascanner, we are actually attempting to supply our own db.. we can't reuse the existing.. it's the point of isolating :-) [19:39] right, but in this case, maybe it is ok to use the existing-- it is ro access. it depends on what you test, or if the db can become corrupted in test runs [19:40] that likely isn't ideal. you could move the original aside, mock the new one, run the test and then put the other one back [19:41] that is doable, but an error could leave the db out of place (you would also want to be sure not to overwrite the backup if it exists) [19:45] jdstrand, for music at least, the primary purpose of the isolation besides providing a clean env, is to allow us to create a known music library. We do this by supplying the music files and the db entries [19:46] jdstrand, your suggestion of moving, mocking, and replacing is exactly how we originally did things. This is fragile and indeed has caused us issues [19:48] this feels a bit like it's gotten more complex [19:48] balloons: a rule could be added for the music library. while it would be correct for music-app (for now, soon it will use dbus with the new mediascanner apparmor integration), it would not be correct for say, the calendar [19:48] yeah :/ [19:48] it could be made to work, but there are some limitations [19:49] jdstrand, well ideally what we do can be applied universally [19:49] yep [19:50] with the route we've taken, we need to mirror the apparmor rules for unredirected access to redirected, but only the ones that are specified in the profile [19:51] well, what is this mediascanner apparmor integration you speak of? [19:52] could music stay unconfined until this was all in place? [19:52] rather than accessing the db directly, it accesses it over dbus (I think) [19:53] ahh.. interesting [19:53] it could, but we will hit this issue with something else [19:53] filemanager :-) [19:53] it isn't confined yet, (but should be) [19:53] maybe serguisens worked to confine it. I know we talked about it [19:54] jdstrand, right.. but the same issues will happen there. The mocking works for now because it's unconfined [19:54] yeah [19:55] jdstrand, hey, since you are here, does system-settings app need to have any apparmor permissions like to access ofono or urfkill type stuff, or since its unconfined it doesnt need them [19:56] balloons, i think the new new mediascanner2 uses dbus...but i was hitting apparmor issues when attempting to test the silo :/ [19:56] balloons, there were a few typos that broke things [19:57] pmcgowan: right now system-settings is unconfined and considered trusted. however, ofono is going to be light confinement very soon that only allows certain apps to talk to it. system-settings could be added to that list if needed [19:57] s/to be/to have/ [19:57] jdstrand, ok it will need that [19:57] * jdstrand adds task to bug [19:57] pmcgowan: is the source package 'system-settings'? [19:57] ubuntu-system-settings [19:58] ah, someone already added it [19:59] mterry_, Saviq: hi guys, any news on the greeter split? [19:59] sil2100, last minute problem in testing a fix I wrote today, I am rewriting it. You had another big silo you wanted to land? Maybe land that now, and I'll have split tomorrow [20:00] ahayzen__, jdstrand I'm thinking of pushing the an updated version of music, with the current mocking and unconfined.. The ultimate solution is going to be a bit more though, but we need an update to hit the store this week [20:00] mterry_: ACK, the other big silo was from ricmm - but it's not set as ready [20:00] And I'm not sure if ricmm is around to give me a +1 on that [20:01] balloons, will this be much easier when we move to mediascanner2 ? [20:01] balloons, and that is using dbus? [20:02] balloons: that's fine. starting to think the old method (or a combination of it and some things we learned) is likely the path of least resistance until new user is in place [20:02] ahayzen__, not at all.. but in theory it might become possible. As it stands, apparmor configs are getting complicated. It would be best to have everything in place as it should be.. I don't want to pursue interim apparmor solutions [20:03] jdstrand, right.. I'm not sure how painful it will be for everyone to have a test user.. [20:03] balloons, ah ok ... i guess whtever is the easiest solution until the proper one can be found...as you said we need to push to the store ASAP due to the demos next week [20:03] ahayzen__, so I'll pull trunk and give it a whirl.. That way something will be up no matter [20:04] balloons, cool, will we be able to pull ur changes into our mediascanner2 incase so it will be ready to land as well? [20:05] ahayzen__, I'm basically just releasing trunk as-is, unless something is broken [20:05] balloons: eg, 1. if ~/bak doesn't exist, mkdir ~/bak && mv ~/* ~/bak 2. create/m[D[D[D[D[D[D[D[D[D[D[D[D[1~[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[B[B[A[Bock all the dirs/etc you need 3. run test 4. restore from back up [20:05] balloons, ah ok but unconfined? [20:05] erf [20:05] ahayzen__, yes, with that one change :-) [20:05] balloons, cool [20:06] balloons: 1. if ~/bak doesn't exist, mkdir ~/bak && mv ~/* ~/bak else rm !~/bak 2. create/mock all the dirs/etc you need 3. run test 4. restore from back up [20:06] ahayzen__, so it'll be a simple mp.. I'll probably ask for you to approve [20:06] balloons, :) [20:07] jdstrand, yep we have a set of try;except;else blocks we did that with. [20:07] if you do that, it is fairly robust. the test harness could be sure to restore from ~/bak [20:07] calculator still uses it [20:07] well, I guess it is the same process. if you traceback, you are left with ~/bak, but the above rights itself next time [20:07] I see [20:08] jdstrand, yea, we just left it inside the app dir itself however, not ~/bak [20:08] like ~/.local/share/com.ubuntu.calculator/backup [20:08] the idea here is to do the fakeenv idea without actually changing XDG* [20:08] I think we evolved to blowing away everything.. not sure [20:09] right. we could do quite complicated things in the test harness (the thing that runs aa-clickhook --include.../click.rules) [20:11] such as cp click.rules $tmpdir/click.rules, then sed the profile in /var/lib/apparmor/profiles/click_com.ubuntu.music_music_0.1.234 > $tmpdir/profile [20:11] that would accomodate actually using fakeenv as you have it now [20:12] basically you create a new apparmor profile directly based on the existing one [20:12] doing sed on anything that was redirected [20:13] then load that profile in to the kernel [20:13] then at cleanup, just load the profile in /var/lib/apparmor/profiles/... [20:13] (harness cleanup) [20:14] click.rules could go back to what it was. the scripting to get that right would not be trivial, but should be nearly maintenance free [20:14] you get to keep the phablet user, fakeenv, etc [20:17] if done right, it should be a reasonable representation of the apparmor profile [20:18] food for thought [20:18] * jdstrand steps away for a few [20:19] is the HUD still going to be a thing in Ubuntu touch? Just the new bottom edge+HUD doesn't look like it would work well together but I'd miss the HUD if it goes completely [20:40] balloons: hey - did you still need something from me or did jd strand help you with everything? [20:52] ChloeWolfieGirl: we don't know what's happening to the HUD. But right now the focus is on the bottom panel. I guess time will tell what happens to the HUD [20:53] Nik [20:56] nik90_, fair enough, Some of the stuff I've seen look really cool, but I'd still like the HUD for alot of the advantage's it gives you, I think Openness, scopes, gestures and HUD are the 4 biggest selling points of Ubuntu Touch! [20:57] ChloeWolfieGirl: agreed [21:09] has anybody successfully triple-booted an HP touchpad with cyanogenmod, ubuntu, and webos? [21:20] that sounds like work [23:45] Hello [23:46] I'm on utopic #50 on my nexus 4 [23:46] and it generally is running like molasses in antarctica [23:46] what can I do to speed things up a bit? [23:54] What's the best channel to flash? [23:54] trusty? [23:57] Has anybody had experience with Ubuntu Touch and the HP Touchpad? [23:58] Hatten: No, but they both have Touch in the name if that helps [23:58] * alfonsojon is being useless [23:58] haha, good one. [23:58] What OS does the Touchpad run? [23:59] I have it dual booting WebOS and CyanogenMod Android [23:59] If it can boot CyanogenMod [23:59] It can probably boot Ubuntu Touch (provided the ROM exists) [23:59] if it does not exist, it can probably be built. [23:59] (lots of probably) [23:59] It exists, I'm just wondering if anybody's tried it before.