[00:07] Anyone still awake and know stuff about apparmor? This is blocking mir-mon-by-default [00:07] sergiusens: maybe you know, or know who might know? [00:07] thomi: hi - I can probably help but I don't know what mir-mon-by-default is [00:08] tyhicks: turning mir on by default for the phone images [00:08] so here's the thing... [00:08] When we run with mir, mir is reporting "Permission Denied' when it tries to add the uinput devices autopilot creates for generating user input [00:08] that is, /dev/input/event6 & event7 [00:09] However, the file permissions on those device nodes are identical to all the other device nodes that mir is able to open just fine [00:09] so I'm thinking maybe apparmor is doing something here? [00:10] is there a way I can turn apparmor off, just to eliminate that as a potential issue? [00:10] thomi: check the dmesg/kernlog [00:10] thomi: are there any apparmor denials in the syslog? `grep DENIED /var/log/syslog` should do unless you have auditd running [00:11] thomi: you can disable apparmor with the kernel cmdline apparmor=0 [00:11] https://wiki.ubuntu.com/SecurityTeam/AppArmorForPhabletKernels [00:12] see the end of that wiki for how to update the cmdline on a phablet image [00:13] hmmm, there are no DENIED messages [00:14] ok, so it's not apparmor [00:15] I don't suppose you guys can think of any other reason why mir would be able to open some input devices, but report "permission denied" on others? [00:18] container? [00:18] the device node is the lxc container so its uid is actually different? [00:20] symlink, the node you are looking at is a symlink and actual device nodes has different perms [00:21] maybe the device itself is return EPERM for an unsupported action [00:22] possibly a library failure before the actual syscall? Does strace show its the syscall vs something else? [00:22] unlikely but I am trying to imaging other permissions failures [00:23] thomi: hrmm I don't have any other ideas besides those ones atm [00:24] I was thinking that it was maybe a device specific EPERM [00:24] hmmm [00:24] so, this exact same device node works perfectly under surface flinger [00:24] so I think it's probably something in the mir stack [00:24] thomi: surface fliger is inside the lxc container [00:25] this is sounding like a container bridge issue [00:25] jjohansen: OK, I don't understand that at all, are you able to give me some things to try to figure it out? [00:26] thomi: sadly I am not up to date at all on what we are doing to map devices from within the container for access outside [00:26] stgraber: might know better [00:27] Not sure if it's relevant, but mir can (and does) open all the other /dev/input/event* nodes [00:27] just not the two that get created after mir has started. [00:27] but I guess all the others are "real" devices, whereas the last two are uinput drivers [00:28] the actual device node (type, minor and major) is identical inside and outside the container, though permissions on the files may differ [00:28] how do I check that stgraber? [00:28] thomi: so basically we have a split system. We have an ubuntu system as the base, and an android lxc container. Some devices, and services are run inside the lxc container, but other devices and services are in the regular linux environment [00:29] thomi: ls -l inside and outside the container to check that the node is the same [00:29] stgraber: I was thinking that the uid might be different, but mapping is making it appear the same. Well dependent on how you are looking at the device [00:29] stgraber: how do I get "inside the container"? [00:30] jjohansen: we don't have userns yet, so unless the kernel checks for a specifc uid, I don't see how that'd be possible [00:30] thomi: android-chroot [00:30] ahhhh [00:31] stgraber: oh!, I didn't know we weren't using userns for the container [00:31] need to have a rw FS I see [00:31] jjohansen: we'd love to but we don't even have userns in our current 3.11, so we're very far from having it in those old android kernels :( [00:32] stgraber: ah, right /me not thinking [00:33] the userns backport would be an absolute nightmare [00:33] stgraber: inside the container: http://pastebin.ubuntu.com/6181888/ [00:33] so, permissions look identical to me [00:35] thomi: strace is really your best bet at figuring out what's going on I think [00:36] thomi: just for kicks, how about doing a "getfacl" on the device files that work and the ones that don't [00:37] thomi: getfacl /dev/input/event* [00:37] mdeslaur: won't work on android but I can get the result some other way [00:38] stgraber: no EAs on the fs, or the tool isn't there? [00:38] racarr: stgraber: so... if I make sure the device node is present before mir starts, then mir is able to open the device [00:40] http://paste.ubuntu.com/6181913/ [00:40] any known issues using channel "devel" and actually getting code to build? [00:40] mdeslaur: no tool, so you have to go through /proc//root to access the fs from the Ubuntu side [00:42] stgraber: which process' pid should I use? [00:43] thomi: any process that's inside the container [00:43] thomi: look for sensorservice for example [00:44] stgraber: http://pastebin.ubuntu.com/6181922/ [00:48] thomi: do you have the same for /dev/input/event* [00:48] stgraber: yes [00:49] thomi: ok, you really need to strace that thing :) [00:50] ok, that's next. Any particular strace options I should use? I've only used that a few times before... === claydoh__ is now known as claydoh [00:53] if the failure happens in a thread/subprocess, you'll need -f. In any case use, -o to redirect the output to a file [00:53] so: strace -f -o debug [00:54] then pastebin debug [00:54] ok [00:56] strace incoming [00:57] it's rather large :-/ [00:57] maybe I should have filtered it [00:58] stgraber: http://people.canonical.com/~thomir/ [00:58] err [00:58] stgraber: http://people.canonical.com/~thomir/strace_out.gz [01:00] I see that on startup it opens /dev/input/event5 - event0 OK, and then towards the end it opens event6 & event7, which fail [01:00] you indeed get a EACCESS on /dev/input/event7 [01:00] yeah [01:03] thomi: can you try "> /dev/input/event7"? [01:03] and if /dev/input/event7 is started before mir you can access it? [01:03] jjohansen: that is correct [01:03] stgraber: trying now [01:04] stgraber: that works fine :) [01:05] that's pretty weird [01:05] yeah... [01:05] thomi: what does cat /proc//attr/current show? [01:06] will have to restart mir, one moment [01:07] jjohansen: "unconfined" [01:09] thomi: what do you get for [01:09] ls /proc//fdinfo/ [01:09] and [01:09] cat /proc//limits [01:10] jjohansen: limits: http://paste.ubuntu.com/6181992/ [01:11] jjohansen: fdinfo: http://paste.ubuntu.com/6181994/ [01:11] hrmm, okay no where near bumping up against those [01:12] yeah [01:12] I wondered whether mir was dropping priviledges after it starts, but the mir team tell me that it shouldn't [01:13] and unity8/mir is running as the 'phablet' user [01:13] and that user is in the android_input group [01:14] thomi: what if you change ptrace scope (unlikely but hey worth a shot) [01:14] echo 0 >/proc/sys/kernel/yama/ptrace_scope [01:15] jjohansen: /proc/sys/kernel/yama doesn't exist [01:16] ah right, phablet kernel, well then not even remotely likely :) [01:18] thomi: what device is it? [01:18] jjohansen: it's a uinput generated touch device [01:19] would you like the evdev device spec? [01:19] thomi: actually I was going to poke the kernel code and see if that device had specific cases for EACCES [01:20] ahh well, I'm not sure how this stuff works... from autopilot we open /dev/uinput and ask it to create a device. Then the /dev/input/event* devices appear, which I see is what mir reads [01:26] hmmm. what's interesting is that while mir doesn't seem to be able to open these devices, something else obviously is, because when I generate some touch events, it unblanks the screen [01:28] thomi, stgraber mir implements it's own input handling [01:28] thomi: ah yes, I think its a single open device [01:30] jjohansen: so... only one process can open that device at once? [01:30] thomi: I think so but haven't checked yet [01:30] jjohansen: do you think that is specific to uinput devices, or all evdev devices? [01:30] thomi: you should be able to use lsof to find which process has opened it [01:31] thomi: not sure yet, I am not familiar with the device, but I do know there are single open devices [01:31] and that would explain why Mir can't get it [01:32] jjohansen: having said that, I can do 'cat /dev/input/event7' from a shell and it works just fine [01:32] when I generate events I see garbage scrolling up the screen [01:32] hrmm, right I forgot about that [01:32] hmmm, lsof only shows one process: python (which will be the python terminal I'm using to create the event) === bfiller_afk is now known as bfiller [01:43] thomi: so I don't see anything in the input drivers that should be directly causing an EACCES failure [01:43] :( [01:43] thomi: I have to bail for a few hours, but I will be back later [01:44] ok, thanks for your help jjohansen [01:44] stgraber: I wonder if you have any more ideas? === ejat- is now known as ejat === jono is now known as Guest37503 === Namidairo`bnc is now known as Namidairo === ubot5` is now known as ubot5 [03:03] mhall119, hey, still around? [03:10] how to make it writable ? i want to do syncevolution ... [03:10] for contact [03:15] ejat, you don't need writable image for syncevolution [03:17] [ERROR] /home/phablet/.config/syncevolution: Permission denied [03:18] I had to mount -o remount,rw /dev/root to sync back down from gmail [03:18] ejat, that can only mean you mixed root and phablet once upon a time [03:18] krisabsinthe, you shouldn't need that [03:18] so what is the right way to do sync ? [03:19] to think of it, it was before I actually READ the instructions and sudo su - phablet ;p === jhodapp is now known as jhodapp|afk [03:45] ejat, what exactly are you doing? start from when you type adb shell and don't give me your passwords please [03:45] sergiusens, do you have a good idea who in Europe could help with the mir/ap input bug? [03:46] https://bugs.launchpad.net/mir/+bug/1233944 [03:46] Ubuntu bug 1233944 in Mir "Unity8/Mir is unable to open autopilot uinput devices" [Critical,New] [03:47] olli, not in the mir team nor in Europe, but ricmm_ is a safe bet [03:47] outside of the mir team, thinking of someone more closer to phonedation [03:47] ah [03:47] yeah === jhodapp is now known as jhodapp|afk [03:50] balloons, still around? [03:51] anyhow, mhall119 or balloons can you look at https://code.launchpad.net/~sergiusens/ubuntu-rssreader-app/click_and_tr/+merge/188760 === chihchun_afk is now known as chihchun [03:54] sergiusens: hey, can you remind me how to make ubuntu-system rw please? :-) [03:54] veebers, adb shell touch /userdata/.writable_image; adb reboot [03:54] sergiusens: awesome thanks! === ricmm_ is now known as ricmm [04:02] should I be able to rm /etc/init/ssh.override? [04:03] veebers, maybe, it's a mounted file iirc, so you may be able to empty it instead [04:04] sergiusens: ah good point, cheers === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [05:53] rsalveti: still around by any luck? [05:53] didrocks: yes [05:53] \o/ [05:54] rsalveti: plars told me you were looking at the mediaplayer-app AP tests failing? [05:54] didrocks: yup, checking them locally still [05:54] it seems it crashes when starting [05:55] didrocks: yeah, I was able to reproduce the crash, just trying to understand and will open a bug [05:55] seems maguro is fine [05:56] probably a side effect of the new media stack, as we're now using the hardware decoders [05:57] rsalveti: ok, so we can't promote image 75 I guess [05:57] rsalveti: once the bug is opened, let's see who can work on that [05:57] didrocks: jhodapp|afk probably [05:57] ok [05:58] rsalveti: I would have prefer we blocked the gst upload before getting that fixed TBH :/ [05:59] I didn't plan it before we got all the media stack fixed. Not sure who added it :/ [06:00] didrocks: well, it was planned to land it anyway, we just didn't find the issue with nexus 4 specifically because we don't have CI for the gst stuff [06:01] rsalveti: yeah, but I would have prefered that step by step, we obviously missed the mediaplayer-app crash at first, so let's go back to green before going on [06:01] anyway, I guess too late now, but I'll talk to people modifying the landing plan [06:01] didrocks: yup, that's fine [06:01] keep me posted on the crash please! [06:02] sure [06:02] thx ;) [06:53] didrocks: found the issue and was able to run all the tests locally successfully, just need to get the proper fix in place [06:53] rsalveti: rocking! what is it about? [06:55] gst-hybris keeps the buffers in a list, that's initialized as soon the media hal is up [06:55] this list in theory is static, as only gst-hybris can release the buffers [06:55] and mediaplayer is trying to handle that list? [06:55] but what happens with mako is that the hal is a bit smarter, and ends up releasing a few after the hal is initialized [06:56] breaking the internal list logic [06:56] this is at the gst-hybris/media compat layer [06:56] oh [06:56] indeed, I guess the app doesn't really like that ;) [06:57] so instead of keeping the list internally, I'm asking a copy of it to the hal everytime I need to use it [06:57] invalid index :-) [06:57] indeed ;) [06:57] this only happens with the video format used as sample [06:58] at least when comparing with some other videos I have [06:58] ah? it doesn't know how to handle the other ones and thus, don't go to that magic handling? [06:59] it does, but maybe because of the size x resolution, it just decides not to release anything [06:59] makes sense [07:15] didrocks: bug 1234007 [07:15] bug 1234007 in libhybris (Ubuntu) "[mako] out of index crash when handling media_codec output buffers list" [Critical,In progress] https://launchpad.net/bugs/1234007 [07:15] we should be able to get this fixed as soon jhodapp|afk is up [07:17] rsalveti: excellent! thanks a lot ;) [07:17] you should go to bed I guess :) === chihchun_afk is now known as chihchun [07:26] yup [07:26] * rsalveti gone [07:28] good morning [07:31] Saviq, hey there... [07:31] Saviq, I replied to your comments/questions on https://code.launchpad.net/~macslow/unity8/extended-snap-decisions-part1/+merge/187312 [07:32] Laney, the spacing in "About this phone" doesn't look right. :-P The left edge of the icon should line up with the left edge of the search field and the left edge of the "System Settings" header. [07:48] hey everyone [07:49] does anyone know an autopilot test case for unity or some app which involves "snap decision" popups, like for a phone call? [07:49] they are not generated by dialer-app, so I wonder how to expose that to autopilot [07:51] pitti: does messages app have them for sms notifications? [07:52] pretty sure unity8 ap tests have some faux notification dialogs popup [07:52] popey: ah, but I guess those would run unity8 itself under AP [07:53] ah yes [07:53] popey: nope, no incoming test for messaging app [07:54] ok, something for me to figure out [07:56] I'll start with telling ofono directly to accept the call [08:08] mpt: where are you looking? [08:08] Laney, https://ubuntuone.com/7hFiOAHWEhbC89AFQDl9Yk [08:09] mhm [08:09] I can't tell if that lines up or not [08:10] that's not changed by this code [08:12] indeed, it is indented a little bit === fmasi_afk is now known as fmasi [08:16] Good morning all; happy Farm Animals Day! :-D [08:17] baaa [08:17] moo === fmasi is now known as fmasi_afk [08:22] oSoMoN: i'm looking at https://code.launchpad.net/~osomon/notes-app/fix-pep8-warnings/+merge/188523 and it seems that you do more than fix pep8 warning. you replace a lot of assert with wait_for too === fmasi_afk is now known as fmasi [08:24] oSoMoN: on top of that, jenkins keeps blocking on it, but I guess you know that already due to the large number of build retry i see [08:24] nerochiaro: that’s exactly equivalent (same implementation internally), and it does fix pep8 warnings by reducing the length of those lines to < 80 chars [08:24] nerochiaro: yeah, CI is very frustrating these days… [08:25] oSoMoN: ok, haven't checked the internals.if you say they are identical i'm fine with it [08:34] nerochiaro: see http://bazaar.launchpad.net/~autopilot/autopilot/trunk/view/head:/autopilot/matchers/__init__.py for the internals [08:39] ogra_: we releasing this morning? [08:40] oSoMoN: is that MR on notes+pep8 in the list of the blocked MRs ? [08:41] nerochiaro: let me check [08:41] nerochiaro: no, adding it now [08:42] popey, nope === fmasi is now known as fmasi_afk [08:42] popey, seems gstreamer 1.2 went into the archive last night ... so all media tests we have are moot [08:46] ok [08:46] thats for future images though [08:46] I wondered about the one we tested last night.. 75? [08:47] has mediaplayer regressions [08:47] ah ok [08:47] bummer [08:55] hum, testing 75 ... is that know that if you click, in the apps lens, on an already installed 3rd party software, it's giving you an "install" again like if it was not installed? [08:55] mhr3, ^ [08:56] seb128, not sure, the click scope guys would know [08:57] who are them? ;-) [08:59] ralsina/alecu [09:01] cjwatson, thanks [09:04] ralsina, alecu: ^ known issue? (I don't find it in launchpad) [09:04] Hello? Anyone around? [09:06] Hallo? :) [09:06] ralsina, alecu: cjwatson: I opened https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1234042 [09:06] Ubuntu bug 1234042 in unity-scope-click (Ubuntu) "Unity lens displays "Install" button for already installed clicks" [Undecided,New] [09:15] hmm [09:15] i dont seem to be able to install any click apps on 75 [09:15] tapping install gets me a progress bar but it never moves on from 0% [09:17] i had that too [09:17] * popey confirms seb128's bug === fmasi_afk is now known as fmasi [09:17] I switched off wifi and back on again and it worked === vrruiz_ is now known as rvr [09:19] popey, thanks [09:19] ogra_: do you know what's the plan for android portion going forward? rebase on top of AOSP or on top of cm-10.2? for emulator I want to pull in more stuff from either 4.3 or cm-10.2..... [09:19] ogra_, I had that as well, I put the device on the table and a few minutes later retried and it worked [09:20] seb128, yeah, works here too now [09:20] Perhaps related to download manager flakiness? Should be able to tell from /home/phablet/.cache/unity-scope-click.log [09:20] odd behavior ... i had to try three times [09:20] xnox, not sure what rsalveti planned, i dont think we want to update anything the next ten days [09:21] ogra_: sure =) i understand that. Emulator is a frankenstein already ;-) [09:21] sigh [09:21] the timezone selection is still broken for me on 75 [09:21] yeah, some apps install, some don't [09:21] akari and sleepy time do [09:21] ogra_: (is sigh about emulator or timezone selection?) [09:22] xnox, the latter [09:22] i can select the TZ just fine ... but after a reboot it resets to "Europe/Berlin UTC+0" [09:23] and i cant manage to get it to UTC+2 now [09:23] hmm, yeah, it's flaky, trying again and again, eventually xda downloads [09:24] pitti, ^^^ [09:24] i suspect there is some issue with carrying over or storing the setting on reboot [09:25] ogra, popey: does /etc/writable/timezone get udated alright? does that survive reboot, or does reboot reset them? [09:25] well, i'm not sure yet [09:25] phablet@ubuntu-phablet:~$ cat /etc/writable/timezone [09:25] Europe/London [09:26] looks like it got updated last night when I was testing, yeah [09:26] I can flash my phone later, but I really don't want to spend another half an hour setting up everything that I need for telephony test case writing [09:26] 10:26:03 up 14:06, 1 user, load average: 1.96, 3.94, 8.30 [09:26] not rebooted yet [09:26] root@ubuntu-phablet:/# cat /etc/timezone [09:26] Europe/Berlin [09:26] root@ubuntu-phablet:/# date [09:26] Wed Oct 2 11:26:13 CEST 2013 [09:26] that looks right? [09:26] thats after freshly setting it [09:26] phablet@ubuntu-phablet:~$ date [09:26] Wed Oct 2 10:26:27 BST 2013 [09:26] * ogra_ reboots [09:26] * popey also reboots ☻ [09:27] *twiddle* [09:27] * ogra_ will make some bootcharts today ... reboots start getting annoying [09:27] root@ubuntu-phablet:/# cat /etc/writable/timezone [09:27] Europe/London [09:27] yup, fine [09:27] root@ubuntu-phablet:/# date [09:27] Wed Oct 2 09:27:44 UTC 2013 [09:27] er [09:27] popey, what does system setting show you ? [09:27] not fine [09:27] yeah [09:28] * ogra_ bets for UTC+0 [09:28] popey: cat /etc/timezone ? [09:28] look up ☻ [09:28] thats post-reboot [09:28] popey: and ls -l /etc/timezone [09:28] -rw-r--r-- 1 root root 14 Oct 1 20:00 /etc/writable/timezone [09:28] root@ubuntu-phablet:/# cat /etc/timezone [09:28] Europe/Berlin [09:28] root@ubuntu-phablet:/# date [09:28] Wed Oct 2 09:28:18 UTC 2013 [09:28] thats mine after reboot [09:28] So, just catch me up - are we still looking at switching the core apps to click packages by release? (Specifically calendar, music, rssreader, calculator, clock, filemanager, terminal, and weather) [09:28] popey: not /writable/; /etc/timezone should be a symlink to writable [09:28] ogra_: you mean /etc/writable/timezone ? [09:28] ah [09:28] root@ubuntu-phablet:/# ls -l /etc/timezone [09:28] lrwxrwxrwx 1 root root 17 Oct 2 2013 /etc/timezone -> writable/timezone [09:29] lrwxrwxrwx 1 root root 17 Jan 22 1970 /etc/timezone -> writable/timezone [09:29] ogra_: so, the files are alright now at least *phew* [09:29] Wondering if ubuntu-touch is going to vanish from http://people.canonical.com/~ubuntu-archive/proposed-migration/saucy_uninst.txt by release [09:29] cjwatson, sergiusens works on that if anyone [09:29] so it seems we at least got the initramfs/timedated bits right [09:29] ogra_: echo $TZ ? [09:29] pitti, yeah, that seems alright, but smells like some part is missing [09:29] blank here [09:30] ogra_: if date doesn't obey /etc/timezone, then maybe something sets $TZ or whatever [09:30] root@ubuntu-phablet:/# sudo -u phablet -i [09:30] phablet@ubuntu-phablet:~$ echo $TZ [09:30] phablet@ubuntu-phablet:~$ logout [09:30] ogra_: does the clock in the panel also show a wrong time? [09:30] root@ubuntu-phablet:/# echo $TZ [09:30] nothing [09:30] same [09:30] ogra_: try in the terminal app? [09:30] pitti, the clock in the panel didnt even update when the TZ was right [09:30] err, on my build from yesterday I don't even have a clock [09:30] yeah, known bug [09:30] pitti, that's a known issue [09:31] it sometimes vanishes [09:31] will be back on next reboot [09:32] cjwatson, empty in the terminal-app as well btw [09:32] ogra_: and date still spits out UTC? [09:32] yes [09:32] changing tz works here, and the clock update (on the next minute, the UI seems to refresh only on the minute) [09:32] ogra_: ls -l /etc/zoneinfo [09:32] but my fs is rw, not ro [09:32] and the settings app shows "Europe/Berlin UTC+0" [09:32] I think that's what date actually looks at [09:33] ogra_: ah, so it sounds like timezone is ok, but /etc/localtime is wrong [09:33] ogra_: sorry, /etc/localtime, not zoneinfo [09:33] root@ubuntu-phablet:/# ls -l /etc/zoneinfo [09:33] ls: cannot access /etc/zoneinfo: No such file or directory [09:33] a-ha [09:33] ogra_: localtime :) [09:33] localtime seems fine, its a link to the right place [09:33] (it's a symlink *to* /usr/share/zoneinfo/, I keep mixing up teh names, sorry) [09:34] root@ubuntu-phablet:/# ls -l /etc/localtime [09:34] lrwxrwxrwx 1 root root 18 Oct 2 2013 /etc/localtime -> writable/localtime [09:34] ogra_: i. e. /etc/localtime -> writable/localtime -> /usr/share/zoneinfo/Europe/Berlin? [09:34] * cwayne_ tested timezone yesterday and it worked, though tbh its a bit confusing from the ui [09:34] root@ubuntu-phablet:/# ls -l /etc/writable/localtime [09:34] -rw-r--r-- 1 root root 118 Oct 1 17:55 /etc/writable/localtime [09:34] nope [09:34] not a link [09:34] ditto [09:34] that'd do it [09:34] ah [09:34] root@ubuntu-phablet:/etc/writable# cat localtime [09:34] TZif2UTCTZif2UTC [09:34] UTC0 [09:34] ogra_: sounds like it got restored from the original image on boot then [09:34] mind you, localtime is sometimes a copy, not a link [09:34] pitti: ^- [09:34] and i'm not sure stgraber's code allows links atm [09:35] it's a plain file in the origianl image, once you change it it becomes a link [09:35] and this is sometimes deliberate in order to avoid requiring /usr at boot [09:35] pitti, but not after reboot :) [09:35] 118 bytes is too short for Europe/Berlin though [09:35] ogra_: right, so it seems "synced" overwrites an existing symlink with the original file [09:35] i dont think there is code to carry over links inside synced dirs [09:35] right [09:35] But it's the same size as /usr/share/zoneinfo/UTC (-> Zulu) [09:35] yes, it's Etc/UTC [09:36] so it copies the original file over [09:36] from the ro space [09:37] +# If the target doesn't exist, just copy it over [09:37] pitti, i guess if you could do the copy on your side instead of linking, we'd be fine [09:37] +if [ ! -e "$target/$file" ]; then [09:37] could this be running when /usr/share/zoneinfo/ isn't visible yet? [09:38] * ogra_ checks ... i think / is mounted by then [09:38] cat ${rootmnt}/etc/system-image/writable-paths | while read line; do [09:38] i. e. [ ! -e "$target/$file" -a ! -L "$target/$file" ] [09:38] thats how it starts [09:39] so / should be there [09:39] aah [09:39] ogra_: so the symlink would indeed be broken [09:39] as it's not /usr yet, but /usr [09:39] yeah [09:39] and likewise, it couldn't even resolve /etc/localtime itself (but that shouldn't matter) [09:39] so I guess we need the -a ! -L .. [09:43] uh [09:45] pitti, looks good [09:46] * ogra_ tests the change here [09:46] ogra_: actually, I think it needs to be fixed in two places, not just the ! -e [09:46] if the keyboard wouldnt die all the time that would be easier === chriadam is now known as chriadam|away [09:47] root@ubuntu-phablet:/# date [09:47] Wed Oct 2 11:46:53 CEST 2013 [09:47] after fresh reboot [09:47] with the change in place [09:47] ogra_, pitti: https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1234051 ... if you want to reassign to the right place [09:47] Ubuntu bug 1234051 in ubuntu-system-settings (Ubuntu) "timezone settings are reset on reboot" [Medium,New] [09:47] are we talking bug 1234051 ? [09:47] ah [09:47] seb128: thanks, doing [09:47] pitti, danke [09:48] getting close! [09:49] ogra_: http://paste.ubuntu.com/6183164/ [09:49] at least setting it *ever* works now Laney :) [09:50] ogra_: I guess you already locally applied something like that? [09:50] pitti, looks good [09:50] pitti, the second line, yeah [09:50] ogra_: thanks for cross-checking, uploaded [09:51] pitti, argh " [09:51] !! [09:51] ? [09:51] pitti, we cant just upload ... you need a slot on the landing plan [09:51] you just trashed didrocks' day i guess :P [09:51] ogra_: well, it's in unapproved anyway? [09:51] no [09:51] touch goes directly through [09:52] like universe [09:52] how come I was supposed to directly upload yesterday? [09:52] ah, that was systemd and got held in unapproved [09:52] but it got reviewed pretty much instantly [09:52] right [09:52] pitti: ogra_: it's the systemd changes we discussed yesterday, right? [09:52] well, it was in the right time slot [09:52] didrocks, nope [09:52] didrocks: nope, a fix for the new "synced" mode [09:52] didrocks: in initramfs-tools-ubuntu-touch [09:53] didrocks, bug 1234051 [09:53] hum, what was tested yesterday worked though? [09:53] bug 1234051 in initramfs-tools-ubuntu-touch (Ubuntu) "timezone settings are reset on reboot" [High,Fix committed] https://launchpad.net/bugs/1234051 [09:53] so, it wasn't really tested? [09:53] didrocks, doesnt survive a reboot it seems [09:53] didrocks: within the same session, yes; but a reboot resets the file [09:53] didrocks, http://paste.ubuntu.com/6183164/ fix is minor and works, i just tested it ... but wasnt fast enough to hold the upload [09:53] well, it's uploaded… so too late, but please do additional tests to ensure we don't end up in a bad state [09:54] ok, yeah, the syntax looks good [09:55] didrocks: sorry about the timing then [09:56] pitti: no worry, but please ensure (especially if this is the day afterwards) that we're in a safe mode :) [09:56] I thought approval was given for the change in general, and then people would upload the corresponding fixes [09:56] pitti: some people assumed the same with gstreamer 1.2 for instance :/ [09:56] pitti: yeah, but that was for yesterday, when we thought we can release an image [09:56] which isn't the case [09:56] pitti, https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0Au6idq7TkpUUdGNWb0tTVmJLVzFZd0doV3dVOGpWemc#gid=0 ... it would need a new "landing asks" entry for a new change [09:57] well, it got monotonously better (in the TZ regard anyway), just not fully fixed yet [09:57] theory is that every change that happes to an image build got a full set of test runs first [09:58] which means it slowly moves through the different sheets on the spreadsheet while people test it [09:58] pitti: ITYM monotonically :) [09:58] (unless you mean boringly better) [09:59] cjwatson: monotonomally :) [09:59] cjwatson: (thanks) [10:01] didrocks: at least we should have the /etc/writable/ bits under control mostly, so for the next bug that we'll discover that fails to update a file in /etc/ it should be much easier [10:01] (but still, *ugh*) [10:02] * ogra_ installs bootchart ... [10:02] pitti: excellent, thanks! [10:02] where "much easier" is still "patch the heck out of packages", of course [10:03] Morning all === mzanetti is now known as mzanetti|otp [10:06] didrocks, thanks for those landing emails btw, they're quite helpfu for our team :) [10:07] cwayne_: happy they are helping! :) [10:08] it's nice to have insight as to which images will be published, helps plan which to test for demos === daniel is now known as Guest53096 === mpt_ is now known as mpt === gusch is now known as gusch|lunch [10:26] should I be able to add contacts, in the contacts app, on grouper? [10:27] oh, ignore me [10:27] it's not obvious there is a entry for the name at the top of the UI [10:27] I hadn't put anything in there so the button was unsensitive === mzanetti|otp is now known as mzanetti [10:40] There is no sound sound notification when I receive a call... what changed ? [10:40] davmor2, you see that issue ? [10:40] om26er: yes [10:41] I see the OSD, I can pickup the call and the sound works fine in the call. [10:41] i get no ringing sound [10:41] popey, bug is reported for that? or is that new? I first faced that with yesterday's image [10:41] ogra_, cjwatson we are, only one I have doubts with is the music-app because of all the magic it needs in other places [10:41] nope, go ahead and I'll confirm [10:41] popey, you have mako as well ? [10:41] yes [10:42] om26er: yeap [10:42] davmor2, maguro or mako ? [10:42] maguro [10:44] om26er: I pointed it out to ogra_ last night but had to race off so wasn't able to do anything with it as such [10:44] sergiusens: Wondering if maybe I should special-case proposed-migration to know that music-app is a thing that exists, then [10:45] (But probably no point until the rest of it's done) [10:45] davmor2, right, it first appeared yesterday indeed, I thought maybe since things were broken in that image it was just not worth reporting [10:46] popey, davmor2 bug 1234087 [10:46] bug 1234087 in telephony-service (Ubuntu) "No sound notification for Call/SMS" [Critical,New] https://launchpad.net/bugs/1234087 [10:46] boiko, ping [10:46] cjwatson, it's done, and should start as the infra has been fixed to support testing click yesterday [10:46] I initially reported for telephony-service but I guess if its the wrong package we can move from there. [10:46] cjwatson, I'll keep you posted on this [10:47] Done and in the archive, I mean, but thanks [10:47] jodh, hey, can you help me out with the upstart-local-bridge? I enabled it on android side and it seems not all the events get back to upstart === chihchun is now known as chihchun_afk [10:49] om26er: thanks === chihchun_afk is now known as chihchun [10:57] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png [10:57] hmm [10:58] that dbus zomby process looks owrrying [10:58] *worrying [11:00] ogra_, upstart-local-bridge is also crashing [11:00] yeah [11:00] ogra_, or restarting [11:00] and there is a 10sec sleep in the session startup [11:02] wow [11:02] that unity8 upstart job is full of sleeps [11:04] * ogra_ drops all sleeps there and reboots ... lets see what happens :) [11:04] ogra_: bootchart > cool :-) [11:05] ogra_: what are the steps to get one? I guess writing the initrd is a bit more subtle here [11:05] lool, not at all [11:05] make the image writable and just install it ... [11:05] the upstart job is broken though [11:05] hmm 5secs of apparmor CPU pegging [11:05] on first boot it is 20sec [11:06] ogra_, unity8 sleeps until we get the upstart-local-bridge working properly [11:07] lool, http://paste.ubuntu.com/6183389/ thats my modified bootchart.conf [11:07] sergiusens, unity8 has a hardcoded sleep 12 before running ofono-setup [11:07] ogra_: what did you change there? [11:08] lool, LOGS and i added unity8 as stop gap [11:08] (in --crop-sfter) [11:08] *after [11:08] ogra_, I'm guessing that's for PIN in the greeter and it depends on ofono and that one does take a while [11:09] what is really surprising is that neither CUP not I/O are even remotely busy [11:09] we should be able to do a lot better here [11:09] sergiusens, then it should get its own upstart job and emit an event .... not sleep for 12 secs [11:10] (and unity8 should wait for the event) [11:11] I get slow speeds to ports.ubuntu.com are there any mirrors yet? [11:12] wow [11:12] seems apparmor-parser decides randomly when it wants to run [11:12] this chart has it again with a 20sec run === jhodapp|afk is now known as jhodapp [11:13] DJJeff: No. There's an internal ticket open to create a US mirror. [11:13] 20sec is nothing. just downgrade to some old SOC and enjoy a few minutes of apparmor parsing :) [11:13] I like random === iahmad_ is now known as iahmad|afk [11:14] w-flo, well, the point is that it only should do that once on first boot [11:14] I get 150ms to 200ms ping to ports.ubuntu.com [11:14] and about 120 KB/sec [11:14] I agree.. never noticed it happen more than once [11:15] @ogra & w-wflo What would you guys do if you were me? My touchscreen works great, except for like 1 out of 6 boots... It doesn't swipe.. I put my ts_srv program that drives the touchscreen in android in debugging mode, I swear surface flinger is not always.. accepting my swipes from ts_srv, instead of ts_srv having a bug [11:15] It works great in android.. 100% of the time [11:16] maybe something to do with the elusive bug with services not always starting.. or service manager not thinking surfaceflinger has always started when it has [11:16] that annoted me so much I did this [11:16] OrokuSaki, I'd have no idea what to do. Other than comparing logs :) [11:16] I told service manager.cpp whatever to always say my services are started. [11:16] yeah... [11:16] same here [11:16] mpt: Time and Date says this: "When you select a different item, it should appear as the selected radio item before navigating back to the main “Time zone” screen." - how long should it appear for? [11:16] I did notice that the PID of my process is lower when it works... and higher if it doesn't [11:16] for ts_srv.. there must be a reason why [11:16] OrokuSaki, I think(?) mir will be the default soon, so maybe forget about surfaceflinger [11:16] so I am going to try to find out... oh cool! [11:17] well hell.. nevermind [11:17] thanks w-flo! [11:17] during apt-get upgrade ( Installing new version of config file /etc/pulse/system.pa ... ) [11:17] I hope thats a good thing [11:17] should be unless you modified yours [11:17] Laney, let's say one second [11:17] @DJ [11:17] OrokuSaki, I'm not following this very closely, so maybe plans have changed. :) And last time I tried mir, it didn't work for me, but that was some weeks ago [11:17] mpt: ok [11:17] I was thinking that might be too long [11:18] enough to disconnect you from the action when it navigates away [11:18] so it's jarring [11:18] but let's see [11:18] well.. if were going to be leaving surfaceflinger... evertually.. I can live with 1/6 not working on certain boots.. That is a-okay [11:18] Laney, updated === MacSlow is now known as MacSlow|lunch [11:19] roger [11:19] https://github.com/jshafer817/UbuntuTouch/blob/master/Source%20to%20compile/frameworks/native/libs/binder/IServiceManager.cpp I changed line 140 to say always return a svn name and not this if NULL [11:20] don't know if that was enough.. but now I can get into unity 100% of the time [11:21] I still have to add the sleep 30 seconds to lightdm.conf and ubuntu-touch-session-setup.conf file [11:21] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako-dropped-sleeps.png [11:21] now this one looks a lot better [11:21] and i cant see any functional issues caused by dropping the sleeps [11:22] ogra_: it isn't random. on first after a flash because there is no cache. subsequent reboots iff apparmor policy changes [11:22] jdstrand, well, i seem to have it every other boot [11:22] adding ~20sec boot time [11:22] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako-aaparmor.png [11:22] there is one for example [11:23] ogra_, are you in readonly mode? [11:23] i wouldnt know why the aa policy would have changed between two boots [11:23] sergiusens, nope [11:23] sergiusens, i indeed had to drop ro when installing bootchart ... and kept it that way for now [11:24] ogra_: when did this boot take place? === gusch|lunch is now known as gusch [11:24] ogra_: after a flash? 2nd boot, 3rd boot? [11:24] ogra_, I saw yesterday that it checks on the pkg data changes to run it [11:24] jdstrand, a few minutes ago ... was like the 5th reboot since installing botochart [11:24] ogra_: any other updates? [11:24] nope [11:24] no changes at all [11:24] just reboots === oSoMoN is now known as oSoMoN|lunch [11:25] is there anything in /var/cache/apparmor and /etc/apparmor.d/cache? [11:26] jdstrand, http://paste.ubuntu.com/6183448/ [11:26] ogra_: can you do an 'ls -l' instead? [11:27] jdstrand, oha [11:27] i think i know [11:27] jdstrand, the click hook reinstalls the click packages on every reboot it seems [11:27] or some of them at lest [11:28] http://paste.ubuntu.com/6183451/ [11:28] the timestamps seem to match my reboots [11:29] ogra_, cjwatson: I had another instance of "click install stays at 0%", that's the corresponding ubuntu-download-manager log: http://paste.ubuntu.com/6183456/ [11:30] ogra_: I don't think that is it. aa-clickhook, called from click job, is not launching apparmor_parser [11:31] ogra_: (which it shouldn't) [11:31] jdstrand, well, the timestamps surely match the two runs i have seen in recent reboots [11:31] ogra_: it is apparmor's sysv initscript that is doing it. there is a bug I am waiting on a fix for from tyhicks that could be related [11:32] (teh device was freshly installed several hours ago, initial run has long happened) [11:34] ogra_: what's weird is ai don't see the click-apparmor upstart job [11:34] Saviq, greyback hey, do you know if the path bug has been solved? === iahmad|afk is now known as iahmad|afk|reall [11:35] sergiusens: you'll have to be more specific. The issue causing unity8 to not work with AP I had yesterday maybe? [11:36] greyback, sorry, Exec in desktop file not being in PATH [11:36] jdstrand, i ass aa-clickhook [11:36] greyback, for click packages [11:36] right, but from click [11:36] yes [11:36] sergiusens, bug #1229253 [11:37] bug 1229253 in Unity 8 "cannot find click application when Exec not in PATH" [High,Triaged] https://launchpad.net/bugs/1229253 [11:37] Saviq, that one, thanks [11:37] sergiusens, so no, doesn't look like it [11:37] there is another upstart job that, while very fast in the normal case, should be present on the chart (as a blip anyway) [11:37] click-apparmor [11:37] Saviq, I'll keep adding ./ then, thanks [11:37] seb128: seems highly unlikely that it has anything to do with click itself [11:38] jdstrand, right, and it should be at the very top, given it starts on "filesystem" [11:38] sergiusens: I'm not aware of that bug. Saviq? [11:38] greyback, it's assigned to you :) [11:38] seb128: but, yeah, surely that's "download manager needs to recover from or at least report network errors"? [11:39] or maybe "download manager needs to recover, scope needs to report" [11:39] cjwatson, right, just opened https://bugs.launchpad.net/ubuntu/+source/ubuntu-download-manager/+bug/1234113 [11:39] Ubuntu bug 1234113 in unity-scope-click (Ubuntu) "sometimes download are stuck at 0%" [Undecided,New] [11:39] ogra_, ^ [11:39] greyback, but you can ping it back to me if time is tight [11:39] greyback, looks like a relatively easy one [11:39] seb128, confirming [11:41] ogra_, dake [11:41] danke [11:41] Saviq, larsu, who is responsible for starting the indicators, is that unity or are they starting self contained ? (see http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png ... there is a massive delay before they start) [11:41] dholbach, any chance for 1 more quick review of uwoot? i did it right this time i swear :P [11:42] well, it doesn't matter-- the problem is the initscript [11:42] ogra_, AFAIK they're starting themselves, dednick ↑↑? [11:43] ogra_, what is the big "sleep" in there? [11:43] ogra_: what's the trick with generating that image? install bootchart, then what? [11:43] seb128, see the unity8 upstart job [11:43] seb128, seems to not cause any harm to drop it, but i'll wait for ricmm to elaborate :) [11:44] ogra_, larsu, Saviq: i think they are supposed to be started by a unity8 signal to upstart, but i think some of them are still dbus started. [11:44] there is likelz a reason for having it [11:44] ogra_: they're dbus activated (unless ted's upstart stuff got merged in) [11:44] ogra_, "hackish" is all I'm going to say [11:44] seems thats one of the issues causing the slow boots, do you know if teds stuff is ready for 13.10 ? [11:44] larsu, (no it didn't) [11:44] ogra_, it's not [11:45] :( [11:45] the upstart jobs work only in upstart sessions [11:45] e.g would create issue for ubiquity or unity-greeter [11:45] bah, desktop crap [11:45] lets drop it ! [11:45] :) [11:45] :-P [11:46] ogra_: switching to upstart jobs that are triggered by unity won't make them start earlier... [11:46] root@ubuntu-phablet:/# less /var/log/upstart/click-apparmor.log [11:46] Files /var/lib/dpkg/info/apparmor.md5sums and /var/lib/apparmor/profiles/.apparmor.md5sums differ [11:46] ERROR: Could not parse click manifest. Skipping 'com.ubuntu.developer.adrian-arroyocalle.wikipedia-mobile_wikipedia_1.2.json' [11:46] jdstrand, ^^^ [11:47] larsu, upstart will allow to parallelize the startup, it will surely improve the crawling boot [11:47] seb128: hm? ubiquity? [11:47] sergiusens, btw i piggy-backed a small phablet-demo-setup fix into that MR, is that ok, or would you rather me do a separate MR? [11:48] we are at ~1:20 boot time ... Mir will add another 20sec [11:48] ogra_, Saviq: i think only indicator-network is upstart. at the moment. [11:48] ogra_: they are already started in parallel... the question is _when_ we start them [11:48] larsu, well, the question imho is: do we need to delay unity mapping for them [11:48] cwayne_, well for reviewing yes, it's ok, just make the commit message really clear about it :-) [11:48] ogra_: absolutely not :) [11:48] larsu, having unity up even before they start could buy us 20sec [11:49] oops, it could've been clearer sergiusens, i'll add a comment then to help clear it up :P [11:49] xnox, is the installer an upstart session? do you still use indicators there? [11:49] even though your panel might look jumpy, thats still better than a black screen [11:49] ogra_: no no no no, please don't do that [11:49] seb128: installer is not an upstart session, but I was hoping to make it one. Yes we still use indicators, but since transition to -ng none of them are loading. I think I need to add/request ubiquity profiles on the them. [11:50] larsu, well, mgmt wants us to boot in ~10sec [11:50] larsu, while i think thats utopic, 30 should be doable [11:50] but that wont work if unity stays black for 20sec [11:50] seb128: I will not have time to make ubiquity an upstart session in time for 13.10, unless i hack i together, based on the "upstart session as it is on touch images" [11:50] xnox, well anyway, bottom line "we are not going to migrate indicators from dbus activation to upstart jobs this cycle" [11:51] xnox, no, don't, we would have also to make unity-greeter an upstart session... not this cycle [11:51] ogra_: unity could just hit the indicators up earlier... [11:51] larsu, fine with me [11:51] any improvement will do atm :) [11:51] dednick: ^^ [11:52] ogra_, larsu: I wonder what unity is doing [11:52] the chart has no busy time [11:52] e.g no blue/pink on the unity bar [11:52] seb128, waiting for the indicators [11:52] yet it takes ages before it activates indicators and lenses [11:52] seb128: larsu: i wonder if I can hook into pam_systemd and create upstart session per logind session. [11:52] ogra_, isn't unity what should dbus activate the indicators? [11:52] and it doesnt map unless they are done [11:53] ogra_: wait. How long does it take to start the indicators? [11:53] can I even see that in this chart? [11:53] they shouldn't take more than a couple 100ms each [11:53] if that [11:53] ogra_: yeah-- I have downloaded all the apps and will be filing bugs [11:53] xnox: that sounds .... evil :) [11:53] ogra_: that won't slow things down of course, but noted [11:53] larsu, the "busy time" is the blue on the bar, they seem small enough indeed [11:55] ogra_, seb128: indicators are "mapped" as soon as unity8 starts. Read from file system. They are loaded (shown) only we find a valid indicator at the dbus path given by the indicator service files. [11:55] seb128: ah, thanks. If there's something we can make faster on indicator startup, I'm all for it. Right now it looks like the issue is that they're being started late, not that they take a long time [11:55] dednick, well, take a look at http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png [11:55] is there an svg version of this that I can search in? [11:55] ogra_, seb128: valid being we found a indicator with a menu item saying that it's enabled. [11:55] larsu, each thicker vertical line marks 5sec [11:56] dednick: indicators are started ~30 seconds after unity8 [11:56] it doesn't make sense that nothing happens during 20s and then indicators and lens start [11:56] so from starting unity8 to starting the indicators there is a 30sec gap [11:57] right, that's ridiculous [11:57] ogra_: they are started by dbus. i have no idea why so late. === hikiko is now known as hikiko|lunch [11:57] I wonder if the issue is on the dbus side [11:57] maybe dbus confinement or something? [11:57] and the indicators themselves take another 10 or so [11:57] there is a dbus-activation process that takes 10s [11:57] then blank [11:57] which means the session startup alone adds up to 40sec or so [11:58] ogra_: no... they're long running processes. They're pretty fast in sending the menus and icons to unity when they start [11:58] ogra_, it would be interesting to try with a dbus without the confinment patch, just to see [11:58] ok [11:58] seb128, that zombified dbus-daemon looks very odd i must say [11:58] well, it could be a bug in unity8 as well [11:58] ogra_, yes... [11:59] it looks like dbus activation is taking ages [11:59] or that dbus is in a weird state [11:59] * ogra_ bets on the latter [11:59] ogra_: you can test dbus activation by calling StartServiceByName on org.freedesktop.DBus [11:59] ogra_: if you remove/add a indicator service, it's pretty much immediate to add/remove, so i dont think it's the unity8 side. dbus me thinks [12:00] i mean the icon add/remove from the panel is immediate. [12:00] if unity8 is already loaded. [12:00] dednick: I'm activating stuff all the time when developing, it's very fast. Maybe the confinement patches? [12:00] yes, i dont debate that :) [12:00] but there is definitely something wrong in our session startup [12:00] ya... [12:01] we're just trying to shift the blame :) [12:01] haha [12:01] :) [12:03] do we have bootcharts from pre-confinement (if there was ever a pre-confinement)? === iahmad_ is now known as iahmad| [12:03] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro.png [12:03] thats from june [12:03] right after flipped container became default === iahmad| is now known as iahmad [12:04] dbus looks less odd there [12:04] but overall its not much better [12:04] (and the HUD was broken back then, thats fixed nowadays) [12:06] ogra_: oh, I missed the bit about how the md5sums differed. that makes no sense if you didn't update apparmor in between [12:06] jdstrand, i didnt update anything [12:06] apart from installing bootchart [12:06] (i didnt even apt-get update) [12:06] ogra_: still starting about 10 seconds after hud... [12:06] dednick, right [12:07] ogra_: I'd like to try to reproduce this. if I install bootchart, then where do I see its outpuit? [12:07] ogra_, that old chart is less odd [12:07] its the same === dholbach_ is now known as dholbach [12:07] output even [12:07] jdstrand, flash image, make writable, install bootchart package, copy my upstart job over the original one http://paste.ubuntu.com/6183389/ [12:08] ok, thanks [12:08] jdstrand, then reboot and look in /var/log/bootchart ... grab the tarball from there and process it on your PC with pybootchartgui [12:09] ogra_: what image was this with? [12:09] 74 updated to 75 [12:09] originally installed with 74 [12:09] ogra_: is that in proposed? [12:09] saucy-proposed channel, yep [12:10] Saviq, hey, for the unity scopes not showing up, i understand we're going to the distro-patch route, is that right? [12:10] apparmour_parser... runs for about 8 seconds until about 2 seconds before indicators start [12:10] cwayne_ yes [12:10] ok, I'll try to reproduce on grouper-- hopefully I can cause I don't want to move to proposed on my mako [12:10] Saviq, is there an ETA on when that would land? [12:10] isnt apparmor for confinement? [12:10] dednick, on the old chart you mean ? yeah, that was fixed [12:10] tsdgeos, you didn't get to the Qt fix yet for the missing scopes did you? [12:10] dednick, it is supp[osed ot only do that on first boot [12:11] oh. right. got the wrong chart... [12:11] tsdgeos, I can take care of it, if you'd like [12:11] ogra_: if you wipe /var/crash and do a reboot do you see maliit and hud if you run mir? [12:12] davmor2, currently bootcharting, i dont want to switch to Mir atm [12:13] ogra_: I've heard some excuses in my time ;) [12:13] :P [12:13] i'll test later today after a fresh flash [12:13] ogra_, anything i can help with? i happened to have just done a fresh flash [12:14] davmor2, ^^^ [12:14] cwayne: check /var/crash please === alan_g is now known as alan_g|lunch [12:15] cwayne: then do touch /home/phablet/.display-mir reboot and check /var/crash again ta [12:16] xnox, jodh, i assume using "start on" twice is not really allowed ? http://paste.ubuntu.com/6183598/ [12:16] ogra_, it's hud that's slowing down everything, no? [12:16] ogra_: last one will be used. [12:16] hah [12:17] thats why maliit never reqpwns then [12:17] and you have two stop on. [12:17] ogra_, it blocks unity for ~30seconds [12:17] xnox, erm, thats what i actually menat [12:17] ogra_: "stop on stopping unity9 or session-end" [12:17] yes [12:17] i know [12:17] cwayne_, we'll try and and get it in today / early tomorrow [12:17] mhr3, on the old chart from june, yes [12:17] mhr3, that bit is fixed [12:18] xnox: man you devs are way ahead if you're on unity9 [12:18] davmor2: damn, leaked top-secret TV project. [12:18] Saviq, perfect, thanks! [12:18] ogra_, i'm looking at mako chart [12:18] says 2nd oct :) [12:18] Saviq, as always, feel free to let us know if you need an extra hands testing it [12:19] cwayne_, sure [12:19] davmor2, only foundations ... the UI devs dont know about it yet ... xnox will slowly transmit the data to their brains over the next two cycles [12:19] ogra_: haha [12:19] ogra_, hud started around 26s mark, all the indicators and everything around 55s [12:19] how do I enable Mir on current G4 builds? [12:19] I'd like to reproduce bug 1233944 [12:19] bug 1233944 in Mir "Unity8/Mir is unable to open autopilot uinput devices" [Critical,New] https://launchpad.net/bugs/1233944 [12:20] mhr3, url ? (there are 5 charts [12:20] ) [12:20] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png [12:20] (I'm on image 75, current devel-proposed) [12:20] pitti: touch /home/phablet/.display-mir and reboot [12:20] pitti, adb shell touch /home/phablet/.display-mir [12:20] i see hud-service at the bottom [12:20] and reboot [12:20] davmor2, jibel: merci [12:21] ogra_: apparmor_parser !? [12:21] and apport twice [12:21] pitti, yep [12:22] ogra_: we got rid of that many releases ago in desktop; I wonder how it crawled back in [12:22] mhr3, i must be blind, i only see a 5sec hud-service [12:22] pitti, the mysterious ways of building touch images i'd say :) [12:22] ogra_, i guess that's the thing, it dies, but it shouldn't, dbus is still waiting 30seconds for it [12:23] davmor2, i get a new maliit_server and hud-service crash on mir [12:23] mhr3, oh ! i get your point [12:23] cwayne: thanks [12:23] davmor2, also, do you happen to know what video files will play [12:23] i assume ogv would? [12:23] adb shell [12:23] eek, -EFOCUS, sorry [12:24] Saviq: tsdgeos: just noticed that hud is not invokable when running mir. are you aware of this or have an idea why that is? [12:24] mhr3, then i guess the hud issue from http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro.png isnt fixed, just doesnt produce the crazy IO anymore [12:24] thostr_, yes, mzanetti and racarr are on it [12:25] ogra_, i'd even say that it dies before it has a chance to produce crazy IO :) [12:25] thostr_, input behaviour is different between surfaceflinger and mir [12:25] mhr3, yes, thats how it was "fixed" i suppose :P [12:25] cwayne: I was copying the media file from /usr/share/example-content to the correct places on the phone and trying them I think the video is just and .ogg container [12:25] Saviq: yes it seems it's not getting the swipe event [12:25] Saviq: thostr_: once this is merged I can make use of it in the shell https://bugs.launchpad.net/mir/+bug/1233378 [12:25] Ubuntu bug 1233378 in Mir "Unity requires input injection API from Mir " [Undecided,In progress] [12:26] cwayne: it wasn't working at the time I last tested it I'm assuming it is kinda fixed now though [12:27] mzanetti: thanks. [12:27] so with phone running Mir, is there such a thing as a "Mir server process"? I can't see one [12:28] but /usr/bin/unity8 doesn't link to anything libmir-ish, is that going through libQt? [12:28] and Qt uses the Mir libraries? [12:29] yeah, there is no process to match against [12:29] what oyu can check is if there is no surfaceflinger process [12:29] yes, I checked that [12:30] ogra_: so if unity8 is directly talking to the display server through libraries, how is it able to open input devices? [12:30] they are root:android_input 660 [12:30] right, if there is still something on display you run Mir [12:30] yes :0 [12:30] dunno how input handling in Mir works exactly, could be it still goes through android [12:31] ogra_: oh, through libhybris? [12:31] pitti, ask ricmm [12:31] he knows the stack [12:31] ogra_, pitti, dandrader should know [12:31] or ricmm [12:31] "The code that opens this device is (in lp:mir HEAD) [12:31] 3rd_party/android-input/android/frameworks/base/services/input/EventHub.cpp line 959" [12:31] aah [12:32] pitti, ogra_ mir implemented it's own input based out of android [12:32] but it's not hybris [12:32] dandrader, ricmm: so Mir does not actually access /dev/input/event* by itself, but shovels events through some Android layer? [12:32] sergiusens: how odd.. [12:33] sergiusens: but thanks for the hint; do you know if there's a process for this event hub, or if that's going through binder or something/ [12:33] ? [12:34] pitti, that's as far as my knowledge goes, I can be almost sure that it doesn't go through binder and uses /dev/input directly [12:34] pitti, Mir does access /dev/input/event* directly [12:34] sergiusens: ah, so the phablet user is in the android_input class [12:35] yuck, that's a giant security hole [12:35] pitti, it has a copy of android input inside itself [12:35] s/android input/android input code [12:36] pitti, to it has an inotify for files being created of deleted in /dev/input [12:36] dandrader: may it be that something in libmir drops the android_input group membership after startup? [12:36] dandrader: oh, why not listen to uevents directly? [12:37] pitti, the android blobs need the user to be ablet to access the devices directly [12:37] pitti, but it seems that those uinput files are first created with root:root and later changed to root:android-input. but mir doesn't get notified about that change in permissions and therefore doesn't try to reopen them [12:37] dandrader: yes, there is a slight delay when udev rules are applied to a new device [12:37] dandrader, udev sets the permissions [12:37] hence "listen to uevents" [12:37] yeah [12:38] dandrader: thanks for clearing that up, it all makes a lot more sense now! [12:38] note that our udev starts delayed [12:38] (after the container is up) [12:38] ogra_: that should be fine, autopilot stuff runs much later [12:39] ChickenCutlass, finally got to do some bootcharting today ... http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png [12:39] looks better than it actually *feels* when sitting in front of the black screen [12:39] pitti, yeah, we didnt have probs due to it yet [12:39] ogra_, how did you produce it, with standard bootchart or a customized version? [12:40] ogra_, that's great -- thanks [12:40] ogra_: so I guess nobody tried to attach an USB or bluetooth keyboard :) [12:40] and we need to make udev wait until ueventd is done [12:40] ogra_, still that weird dbus-daemon zombie :/ [12:40] this doesn't seem uinptu specific at all [12:40] ogra_, is this with SF [12:40] pitti, heh, nope, i dont think we even have phone kernels that have OTG host mode [12:40] ChickenCutlass, yes [12:40] ChickenCutlass, i'll do Mir later today [12:40] pitti, just to clarify, is uevents what we listen through using that inotify API? [12:40] awesome [12:41] dandrader: no, not at all [12:41] gatox: ping! [12:41] dandrader: uevents are the kernel's way to tell userland about new, changed, or removed devices [12:41] sil2100, pong [12:41] dandrader: so with libudev you say "I want to get notified about events for the "input" subsystem, and otherwise not be woken up [12:41] gatox: hello! Did you see https://code.launchpad.net/~sil2100/click-update-manager/autopilot_package/+merge/188783 mayhaps? ;) [12:41] ah, an inotify is just for changed in the filesystem [12:41] ChickenCutlass, lots of issues we need to solve in that bootchart [12:41] changes [12:42] sil2100, no, reviewing now [12:42] gatox: actually, I need to push a small fix into it, I see some commented leftovers! [12:42] dandrader: and then your event callback gets called on any event on an input subsystem; you watch out for an added or changed "event" device with the properties you are interested in (like, "is it a keyboard"), and handle it then [12:42] dandrader: uevents are much more efficient and convenient than inotify, and you get the device attributes and properties for free, too [12:42] ogra_, that's good. Means we can make it better [12:42] we do [12:43] gatox: ok, done now [12:43] ChickenCutlass, if everyone (upstream) would have time to focus on this i'd bet for ~30sec [12:43] dandrader: and you also avoid these kinds of race conditions [12:43] ogra_, agree [12:44] gatox: that's generally how we treat autopilot tests, as we move them to a separate package so that those can be installed and ran on test-systems [12:44] gatox: for integration testing [12:44] gatox: so actually we never run them from source, but install the -autopilot package and test without any sources [12:45] ChickenCutlass, saldy i doubt we can get time from upstreams for this ... toomanybugs ... [12:45] ogra_, certainly not for 13.10 [12:45] pitti, didn't know about that stuff. thanks for the info! [12:46] * sil2100 lunch [12:46] ChickenCutlass, for 13.10 some improvement should still be possible ... especially all the artificial sleeps should go ... and the but with the HUD delaying indicator startup by 20-30 sec [12:46] s/but/bug/ === dandrader is now known as dandrader|lunch [12:47] dandrader|lunch: I left a summary in bug 1233944 FYU [12:47] bug 1233944 in Mir "Unity8/Mir is unable to open autopilot uinput devices" [Critical,New] https://launchpad.net/bugs/1233944 [12:47] dandrader|lunch: "FYI" [12:47] sil2100, +1, do you want me to globally approve it too or do you want someone else to review it?? [12:47] dandrader|lunch: I'm out tomorrow and Friday, but I'm happy to tell you more about uevents; pretty much everything else in the FOSS world uses libudev, so there's plenty of examples to look at [12:47] dandrader|lunch: I can toss you some easy ones [12:48] pitti, untiting really hard this year ? [12:48] *uniting [12:48] ogra_: ? [12:49] pitti, celebrating unification for a whole weekend i meant :) [12:49] ogra_: oh, you mean Didier's wedding? :-) [12:49] ogra_: ooh, I get it now [12:49] oh, no, i meant unification day and the long weekend [12:50] :) [12:50] ogra_: yes, German reunion AND didrocks+julie union :) [12:50] that needs a longer one [12:50] whee [12:50] yeah [12:50] thats a lot of unification === CarlosNeyPastor_ is now known as CarlosNeyPastor [12:51] xnox, hey, how much do you know about upstart-local-bridge? === hikiko|lunch is now known as hikiko [12:52] pmcgowan, http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png [12:53] (in case yours didnt work out yesterday) [12:53] ogra_, looks like apparmor still takes a lot of cpu time [12:53] ChickenCutlass, 5sec without holding up anything else in that chart [12:53] thats fine [12:53] ogra_, that's 5 seconds of max cpu [12:54] not good [12:54] slows everything down for 5 seconds [12:54] ChickenCutlass, i have another one where apparmor-parser runs just randomly for a full run (20sec with stalling), jdstrand is already looking into it [12:54] is that every boot? [12:54] ogra_, thanks btw [12:54] thostr_: the backend doesn't work with Mir as far as i know [12:55] seb128: are the stats on welcome screen supposed to be off when you check that option in settings? is the backend there? [12:55] ChickenCutlass, it wont hold up anything else ... the HUD delaying the inciator startup (and thus unity8 startup) which causes a 40sec black screen is the big issue here [12:55] pmcgowan, only after upgrades (if no bug) [12:55] tsdgeos: yes, we figured that by now [12:56] ChickenCutlass, pmcgowan, thats one we urgently have someone look at before release [12:56] ogra_, sure, I agree the HUD is a bigger problem. But anything that takes all cpu slows everything else down. [12:56] +should [12:56] ChickenCutlass, well, the CPU is idle 90% of the time ... same for I/O [12:56] look at the graphs at the top [12:56] ogra_, then thats very bad [12:57] what is the bottleneck? [12:58] pmcgowan, seems hud-service dies and unity8 waits for it for 30sec [12:58] pmcgowan, then it starts the indicators (hud-service is further at the bottom) [12:58] ogra_, well, should be easy to get that time back then, so thats good [12:58] unity8 doesnt map to screen before the indicators are up === alan_g|lunch is now known as alan_g [12:59] thats our long black screen on boot [12:59] * ogra_ thinks even 1min boots are acceptable if you dont have to look iinto a black hole during that minute [13:00] gema, the backend is up for review on https://code.launchpad.net/~mterry/unity8/statsWelcomeScreen/+merge/184153 [13:00] gema, so no, it's not supposed to work until that commit gets in [13:00] seb128: ack [13:00] seb128: what do you think is already there (in devel-proposed) and needs testing? [13:01] gema, I'm not sure anything needs specific testing... [13:01] sergiusens: a bit, what's up? =) [13:02] seb128: ok, so what is there and is supposed to work? === _salem is now known as salem_ [13:03] gema, 90% of the settings: selecting a wifi ap (without password, the password part is not working), selecting a carrier, turning on/off datas, adding a bluetooth headset, selecting a background, selecting a ringtone/message sound/adding online accounts/the device infos in about (including click and disk space) [13:04] gema, the battery charge, turning on/off wifi and bluetooth [13:04] gema, turning on/off online datas for the unity lenses [13:04] gema, updating [13:04] gema, selecting a timezone (though on the current image it doesn't persist after reboot, should be fixed on the next one) [13:05] seb128, the password part works fine for me ? [13:05] seb128: selecting a timezone doesn't persist for me since selection till a step back [13:05] ogra_, not in system settings? [13:05] seb128: it doesn't stick, the selection [13:05] seb128, ah, no, from the indicator indeed [13:06] * ogra_ wasnt aware we duplicated that [13:06] seb128: making a note of all that is supposed to be working, will go through it again [13:06] ogra_, right, the indicator uses the password agent [13:06] xnox, I applied http://people.canonical.com/~sergiusens/upstart-property-watcher/ those two to my system [13:06] xnox, after boot everything works fine [13:06] seb128, yeah, the FBI thing ... it is actually called "secret agent" :P [13:06] ogra_, we don't, system settings uses the indicator backend, it's just not plugged into the secret agent yet [13:06] lol [13:06] seb128: thanks for the "Install" bug. This happens because the dash is not doing the refresh of categories, so the just installed app does not move from "more suggestions" to "installed". I'm marking it as a duplicate of https://bugs.launchpad.net/unity8/+bug/1230352 [13:06] Ubuntu bug 1230352 in Unity 8 "Scope results not updated when scope invalidates them" [High,In progress] [13:07] xnox, but during boot, the upstart-property-watcher reports a bunch of changes that aren't seen by the bridge during boot === oreneeshy_ is now known as oreneeshy [13:07] alecu, thanks [13:07] gema, gema, selecting a timezone (though on the current image it doesn't persist after reboot, should be fixed on the next one) [13:07] gema, you confirm what I was writting there? [13:07] * ogra_ waits for the first press reviews that will dicaover a process called "secret agent" on ubuntu phones :) [13:07] ogra_, https://code.launchpad.net/~indicator-applet-developers/indicator-network/secret-agent/+merge/182898 [13:08] ogra_, I'm not making up the name! [13:08] seb128, i know ! [13:08] sergiusens: correct, there was a lot of discussions as to when and how start local-bridge and the container property-watcher and in what order. [13:08] i think its a very unfortunate choice [13:08] yeah :/ [13:08] ogra_, we should call it prism-data-collector [13:08] ;-) [13:09] seb128: I am on image 75, and yes you are right, I just checked mexico timezone and it stuck [13:09] lol [13:09] ++ [13:09] seb128: the UI was very confusing with the UTC-X [13:09] gema, sorry about that, feel free to open design bugs if the UI is confusing [13:09] gema, it just knows you are lying :) [13:09] sergiusens: does it help if you change /etc/init/upstart-local-bridge.conf to "start on starting lxc-android-config" ? [13:10] seb128: I will try to figure out why I find it confusing [13:10] gema, not that we have a merge request coming that is going to go back to the previous screen when you select the tz [13:10] xnox, will try that [13:10] gema, that should make more obvious that the action worked [13:10] sergiusens: cause it looks like lxc-android-config kicks in very early (when cgroups are available) which can be ahead of "filesystem" event which the local-bridge starts on. [13:10] gema, e.g you click, it displays the change and then after a second go back to the previous panel [13:11] well "previous panel", the one with the settings for datetime [13:11] seb128: I think going back to the previous screen automatically will help [13:13] gema, great, let us know if it feels better with the next update then ;-) [13:13] seb128: will do [13:14] Hi. Do you know whether Ubuntu Touch supports ad-hoc WLAN connections on Nexus 7? [13:15] probably via cmdline and nmcli, there is surely no UI option for it [13:16] seb128: the bluetooth button (on the battery settings) just disappeared on me when I ticked it off [13:16] together with the indicator, I think [13:16] gema, https://bugs.launchpad.net/ubuntu/+source/indicator-bluetooth/+bug/1233626 [13:16] Ubuntu bug 1233626 in indicator-bluetooth (Ubuntu) "shouldn't hide itself when bluetooth is turned off" [Undecided,New] [13:16] ups [13:16] gema, https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1233628 [13:16] Ubuntu bug 1233628 in ubuntu-system-settings (Ubuntu) "disabling bluetooth makes battery's setting switch to hide" [High,Confirmed] [13:16] seb128: ack [13:16] gema, ^ that's your issue, charles is looking at it [13:17] seb128: and the brightness slider not working properly? [13:18] gema, is it working in the indicator? [13:18] I thought larsu had patched the brightness slider? [13:18] seb128: if you encourage it with a few extra clicks here and there it catches up, but the slider stops changing brightness after a couple of ups and downs on the settings [13:18] gema, define "not properly" [13:19] yea, it's working for me both in the settings and indicator [13:19] gema, are you sure it's not https://code.launchpad.net/~larsu/qmenumodel/maintain-action-state-type/+merge/187100 [13:19] xnox, starting it then make the service on android just restart a lot [13:19] gema, e.g just the extremities not working [13:20] seb128: it stays at 100% when I go back to 0% [13:20] gema, right, that's the fix I just pointed [13:20] gema, should be fixed when that code lands [13:20] seb128: ack [13:20] charles, ^ [13:21] you called me? [13:21] sergiusens: hm. [13:21] larsu, only for reference, no action items afaict [13:21] charles: ha, coolio [13:24] seb128: connecting to a new wifi network in the settings wifi area doesn't prompt me for password [13:25] seb128: it shows a wifi indicator with a little lock on it [13:25] gema, known issue, pete-woods needs to land https://code.launchpad.net/~indicator-applet-developers/indicator-network/secret-agent/+merge/182898 === Namidairo is now known as Namidairo`bnc [13:25] seb128: axk [13:25] ack [13:26] xnox, anyways, with filesystem the socket did connect; upstart-property-watcher bails if it can't connect before reporting changed props [13:28] xnox, this is what I get if the bridge starts on filesystem http://paste.ubuntu.com/6183835/ ... the Property changed message can't be reached if the socket isn't avail [13:29] sergiusens: right, then lxc-android-config.conf should "start on started cgroup-lite and started upstart-local-bridge", and revert back upstart-local-bridge to "start on filesystem" [13:30] jodh: ^ [13:31] seb128: I cannot register an account other than ubuntu one [13:31] gema, what happens? [13:32] seb128: I only get a cancel button === fmasi is now known as fmasi_afk [13:33] xnox, if I change that I get restarts again on the android side [13:33] gema, what sort of account do you try to add? [13:33] seb128: twitter, facebook and google , the three of them do the same [13:34] offer me to cancel [13:34] sergiusens: argh. [13:34] :q [13:34] seb128: I am trying to add one of each [13:34] :D [13:34] sergiusens: got your mail - looking... [13:34] jodh: what is suppose to be the correct ordering for local bridge and property watcher? [13:34] gema, check with kenvandine or mardy, that can be a bit slow but it should open a webbrowser on e.g fb [13:34] with the login UI [13:34] xnox, sergiusens: the local bridge needs to start before the property watcher [13:34] jodh: ideally we'd start local-bridge ahead of the android container, such that we do not lose a single event from the container. [13:35] seb128: they are the ones working ont he settings for this? [13:35] xnox: that's how it's supposed to work, yes :) [13:35] gema, yes [13:35] gema, they do online-account-setting [13:35] kenvandine: what's up with twitter, facebook and google settings? [13:35] seb128: ack [13:35] jodh: somehow on android side it keeps on restarting, if we do that. [13:35] hey gema [13:36] should open a page to login, let me try on my device [13:36] kenvandine: I am on image 75 [13:36] jodh, xnox that's for the socket to be available I guess [13:37] jodh, xnox can I move the socket out of /dev/ ? [13:37] jhodapp: hey! how are you? [13:37] gema, do you have network? [13:38] kenvandine: let me open a page, it looks like I do but who knows [13:38] xnox, jodh also seems that if I keep the filesystem event linked to the bridge from whatever race it seems to get connected, which is why I see all the properties being changed in the logs [13:38] kenvandine: positive [13:38] kenvandine: wifi [13:38] kenvandine: and a sim on the device [13:39] gema, and you get the same problem with twitter right? [13:39] kenvandine: yep, let me take some pics and send them to you [13:39] tedg, yo [13:39] kenvandine: I can raise a bug if you want [13:39] gema, don't bother [13:39] a bug sure :) [13:39] gema, kenvandine: wfm with 75, adding a fb account, I get the screen with the cancel button for a second, then the browser with the fb page and the email/password entries and the connect button at the bottom [13:39] works for me too [13:39] kenvandine, gema: that UI really sucks btw, tiny UI elements on that big webpage [13:40] seb128, there is a fix proposed for that [13:40] to signon-ui === oSoMoN|lunch is now known as oSoMoN [13:40] seb128: I don't see anything even after a while [13:40] seb128: the screen goes off with nothing on it [13:40] but, i guess there isn't any error handling there [13:40] ogra_, yo yo yo! [13:41] So... anyone know how I can switch to Mir instead of surfaceflinger... just for kicks? [13:41] tedg: do you know if the abiilty to corecctly handle system settings URLs is supposed to have landed in the URL dispatcher in the current image ? [13:41] tedg, remember that old bootchart, that made you fix the HUD ? http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro.png [13:41] gema, seems like maybe a webbrowser-app issue... [13:41] OrokuSaki, create a file in /home/phablet.. don't remember exactly, something like ".display_mir" IIRC [13:41] nerochiaro, No, not current. [13:41] tedg: ok, thanks [13:42] sweet.. I remember seeing that in ubuntu or lightdm settings.. thnks w-flo! [13:42] upstart conf file === MacSlow|lunch is now known as MacSlow [13:42] tedg, i made some new charts today ... and noticed while the HUD seems to not go wild on IO anymore, it seems that someone forgot to tell unity that it has not to wait for 30sec until it times out waiting for the HUD ... http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png [13:43] seb128: ack, want me to raise it against webbrowser-app? kenvandine email your way with the screenshots [13:43] against ubuntu-system-settings-online-accounts [13:43] kenvandine: ack [13:44] sergiusens: you can move the socket to wherever you want as long as you update both the local bridge config, the property watcher and the lxc config :) [13:44] sergiusens: I need to reflash to pick up your recent changes... [13:45] seb128, actually that fix is in signon-ui trunk now, just not landed yet [13:45] kenvandine, what is it going to change? [13:45] jodh, the changes aren't in yet, it's just in the android-ramdisk, if you tell me what device you have I can make one and give you instructions on how to put it on [13:45] the facebook login will be friendly [13:45] the page [13:45] it's signon-ui [13:46] the qml file was missing an import, so it was failing to set useragent and units.gu [13:46] kenvandine, ok, great [13:46] sergiusens: nexus7 [13:46] jodh, ok, will get it to you soon [13:47] sergiusens: thanks [13:48] ogra_, Do you know if that's the case where HUD "doesn't start" ? [13:48] tedg, well, i can use the HUD in the session [13:48] ogra_, We've been getting bugs on that, but haven't been able to recreate reliably. [13:48] tedg, i just think it is wrong that unity8 waits for 30sec [13:48] Hmm, okay. So probably not that case :-/ [13:48] Yeah [13:49] well, i can see the HUD icon when swiping from the bottom [13:49] usuallz what i get then is empty though [13:49] kenvandine: bug 1234185 [13:49] bug 1234185 in ubuntu-system-settings-online-accounts (Ubuntu) "Login to some online accounts unavailable from settings" [Undecided,New] https://launchpad.net/bugs/1234185 [13:49] but i suppose gettign the icon means it runs ? [13:51] ogra_, No, that's not it. [13:51] ogra_, And it looks like in the chart the hud-service ends. [13:52] tedg, right, but untiy waits for another 20-30sec until it starts to fire up indicators [13:52] tedg, and dbus seems to be in some weird semi zombie state [13:53] ogra_, So that is HUD failing to start. Can you pastebin your ~/.cache/upstart/hud.log for that? [13:53] ogra_, It's the dbus activation waiting for HUD to start, but it failing to. [13:54] tedg, only "__pthread_gettid -2" for each boot [13:55] * tedg has no clue what to do with that [13:55] yeah, not much info at all [13:56] Huh, top hit in Google: https://github.com/libhybris/libhybris/issues/59 [13:56] alecu, is https://bugs.launchpad.net/bugs/1234113 a bug for you as well? [13:56] Ubuntu bug 1234113 in unity-scope-click (Ubuntu) "sometimes download are stuck at 0%" [Undecided,Confirmed] [13:57] seb128, alecu: i just hit that too [13:57] ogra_: that messa is fine [13:57] message [13:57] tedg: [13:57] yeah [13:57] it's a print we have for everything that uses hybris [13:57] i see it a lot without any issues [13:57] tedg, but i just notice i have no HUD atm [13:58] But, why would HUD touch Hybris in any way? [13:59] tedg, http://paste.ubuntu.com/6183931/ [13:59] is that comment wanted ? [13:59] ogra_, "wanted" ? Still valid [13:59] if so, there cant be anything in the log indeed [14:00] since not upstart but dbus starts it [14:00] DBus activation does a "start hud" [14:00] oh, ok [14:00] Could HUD be pulling in Hybris from a dep? What links to it? [14:03] * ogra_ doesnt know how hybris could be involved here ... ricmm or rsalveti might though [14:04] hey didrocks [14:04] ogra_: anything gl/qtubuntu, whatever related [14:04] sensors, not sure [14:04] tedg: unless the hud is using the DBus WindowStack it is using the old platform-api session helpers [14:04] gatox, can bug 1231389 be marked 'fix released'? [14:04] ah, pkatform/api [14:04] bug 1231389 in click-update-manager (Ubuntu) "Click updater app does not run" [Critical,Fix committed] https://launchpad.net/bugs/1231389 [14:04] err [14:04] jhodapp: I think you are currently working on bug #1234007, right? (with rsalveti) [14:04] bug 1234007 in libhybris (Ubuntu) "[mako] out of index crash when handling media_codec output buffers list" [Critical,In progress] https://launchpad.net/bugs/1234007 [14:05] *platform-api indeed [14:05] dholbach, not sure if that is already in the image [14:05] didrocks, yep, rsalveti came up with a fix and I'm verifying it...I'm currently in the process of trying to push it to my repo for submission [14:05] seb128, thanks for the bug. The click webservice seems to be down, so you are hitting a known bug in the scope: #1226021 [14:05] kenvandine: ^ [14:05] jhodapp: can you keep us updated? we are pending an image promotion on it [14:05] seb128: I'm marking your bug as a duplicate [14:06] dholbach, yes [14:06] gatox, I'm on the devel (not 'devel-proposed') image and the updater app does not crash on startup [14:06] jhodapp: hey [14:06] didrocks, yes, np [14:06] hey rsalveti [14:06] dholbach, it's in the image [14:06] gatox, awesome - I'm just going through the list of 'appstore' bugs right now :) [14:06] dholbach, marked as released [14:06] gracias! [14:06] alecu, thanks, sorry for filling duplicates... I checked the bugs list before filing, but didn't spot those for some reasons [14:06] jhodapp: just need to make sure that side effect is covered everywhere where the buffer list is needed [14:07] rsalveti, you mean that it doesn't create any new issues? [14:07] ogra_: are you talking about hud on Mir anyways? [14:07] ricmm, about unity8 taking a minute to start up in either case [14:07] ricmm, trying to get the 30sec sleep of unity8 solved [14:08] ricmm, Hmm, good point. Hadn't thought about platform API there. [14:08] jhodapp: sorry, the side effect of that buffer list being dynamic [14:08] ogra_: right [14:08] mzanetti, bug 1221137 can be marked as 'fix released', right? [14:08] bug 1221137 in Ubuntu UX "[dash] open preview disables scrolling in all scopes" [Critical,Fix committed] https://launchpad.net/bugs/1221137 [14:08] ogra_, You could try building HUD with BAMF support, it wouldn't "work" but it might solve the sleep issue. [14:08] rsalveti, yeah, it's good [14:08] ricmm, http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako.png see how it twiddles thumbs before starting the indicators [14:08] dholbach: yes [14:08] ogra_, It would at least be a good data point. [14:08] rsalveti, https://github.com/jhodapp/libhybris [14:08] mzanetti, thanks [14:09] tedg, well, cant we just tell unity to not wait for it forever ? [14:09] if it is dbus activated on demand anyway, why do we need it at startup at all ? [14:09] fginther: can you twiddle http://s-jenkins:8080/job/unity-scope-click-ci/54/rebuild please? seems it failed to connect to bzr === gustavold1 is now known as gustavold [14:09] hey stgraber, is https://code.launchpad.net/~stgraber/ubuntu/saucy/lxc-android-config/boot-hooks/+merge/188172 planned to land any time soon? [14:10] ogra_, I would agree with you there, but the way Unity8 works today is it allocates a HUD instance and keeps it forever. [14:10] ogra_, So that singleton is created on boot. [14:10] hmpf [14:10] jhodapp: let be build and test that [14:10] tedg: why is it not activated right away when unity8 requires it? [14:10] or does it really take that long to start the hud-service [14:11] rsalveti, k [14:11] dobey, the armhf build hit a merge conflict, you might be better of merging to trunk and pushing a new branch [14:11] ricmm, http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro.png it eats your IO on boot if you enable it right away [14:11] ricmm, we had that before (thats a very old chart) [14:11] jhodapp: did you test with maguro to make sure we don't have any regression? but we shouldn't have I believe [14:11] ricmm, It would be if Unity8 created the HUD query object on need. But, it does take a while with audio init stuff. We've moved it out of the critical path (it delays if not used) but we could/should do more. [14:11] seb128: no problem, thanks for taking the time to fill them. My main list of bugs is the project's one, and that bug was filled against the ubuntu package, maybe that's why it did not show up? [14:11] fginther: weird [14:11] alecu: ^^ seems your confirm-uninstalls branch has a merge conflict [14:12] rsalveti, I did test with maguro, yes [14:12] alecu, right, you should probably subscribe to the ubuntu sources, that's where most of the reports are going to come [14:12] jodh, here it is http://people.canonical.com/~sergiusens/upstart-property-watcher/android-ramdisk.img [14:12] ricmm, we need to talk about the 12 sec sleep in the unity8 upstart job btw ... testing with it dropped doesnt seem to cause any issues for me [14:12] * ogra_ guesses that needs awe input as well [14:13] jodh, just make sure you have a writable image and do adb shell mount -o remount,rw /var/lib/lxc/android/rootfs/system&& adb push android-ramdisk.img /var/lib/lxc/android/rootfs/system/boot/ && adb reboot [14:13] well that wait is *after* unity8 starts, no? for ofonod [14:13] ricmm, it sleeps ... [14:14] sergiusens: thanks [14:14] ricmm, http://people.canonical.com/~ogra/ubuntu-touch/bootchart-mako-dropped-sleeps.png thats what i get with it dropped [14:14] ogra_, can you pin unlock? === CarlosNeyPastor_ is now known as CarlosNeyPastor [14:14] sergiusens, should i ? [14:15] sergiusens, i use a prepaid unlocked SIM for testing atm, would have to switch if thats supposed to work and being tested [14:15] ogra_, right, well the account setup is a one time thing too, could be a separate job itself maybe [14:15] ogra_, wth is happening with dbus on those charts :/ [14:15] seb128, i'd love to know that ! [14:16] its in some weird semi zombie state [14:16] jodh, as it is with the current sync in jobs the race allows it to start (I see it start on manta and rsalveti on mako), while the race needs to be fixed, the socket does indeed connect but not all events get to the bridge === Ursinha_ is now known as Ursinha [14:16] jodh, also adb wait-for-device shell /system/bin/logcat|grep Upstart [14:17] jodh, converted the prints into ALOGs to view them better [14:18] ricmm, oh, and looking at /usr/share/upstart/sessions/maliit-server.conf i think we want to drop one "stop on" so respawning actually works :) [14:19] (looks like a messed up merge or so) [14:19] Saviq, final set of MRs for also supporting wifi-selection ext. snap-decision are up ... I'll turn to the while-list protection now [14:19] ogra_: hack away [14:19] :) [14:19] Saviq, any hints about how to go about this? [14:19] MacSlow, right, I wonder if that's the priority right now... [14:20] Saviq, my gut-feeling tells me so :) [14:20] Gonna assume being unable to pull from the Android repo is a thing for everyone? [14:20] tedg: while on that topic, do you have any time to work on hud with the dbus api? [14:20] dholbach: I believe ogra_ put it on the spreadsheet, so I hope so [14:20] ricmm, heh, and if you dont want to be pinged about these upstart jobs all the time i guess we should use a team address for the author entry :) [14:20] or should we drop the hud from the Mir release [14:21] MacSlow, can you go through the spec first and see what can we easily change so that we're closer to design? like solid background, for example? [14:21] ogra_: I'll happrove the MR [14:21] stgraber, wthats that ? the boot hooks ? [14:21] ;) [14:21] MacSlow, re: whitelisting, we'd need a chat with jdstrand and co. [14:21] stgraber, yep, it's in there, image 77 - thanks [14:22] I just asked because of bug 1215092 and it looked like some of the work was already done [14:22] bug 1215092 in Ubuntu system image "please add support to run postinst-style code on first boot after upgrades" [Critical,In progress] https://launchpad.net/bugs/1215092 [14:22] Saviq, sure... I'd rather have someone with more experience in that domain actually implement it [14:22] dholbach, thats likely late tonight or tomorrow morning [14:22] MacSlow, right, that too [14:22] ricmm, Known issue, we're working on it. [14:22] MacSlow, I don't think we should go for AppArmor protection, as we only want a strict list of core apps to be able to send those up [14:22] great [14:22] ogra_, thanks [14:23] MacSlow, possibly dconf-backed for OEM customization [14:23] Saviq, yeah... AppArmor sounds like overkill [14:23] sergiusens: I think part of the problem is /var/lib/lxc/android/pre-start.sh: it removes /dev/socket before starting the container. But the local bridge must run before lxc starts and it needs /dev/socket. [14:24] MacSlow, on that note, does the spec actually mention who *should* be able to post those? [14:24] tedg, bug 1233381 is for tracking boot speed issues btw, in case you want to add something there for mgmt [14:24] bug 1233381 in ubuntu-touch-meta (Ubuntu) "Touch image boot time is too long" [High,Confirmed] https://launchpad.net/bugs/1233381 [14:25] gema, did you file that bug yet? [14:26] Saviq, no... that kind of detail is not mentioned in the spec [14:26] jodh, xnox rsalveti this is what I get when I sync the jobs, so your assessment makes sense E/UpstartPropertyWatcher( 1626): Failed to connect socket for '/dev/socket/upstart-text-bridge' on fd 10 (2 [No such file or directory]) [14:27] kenvandine: yes, I pinged you with it I think, bug 1234185 [14:27] bug 1234185 in ubuntu-system-settings-online-accounts (Ubuntu) "Login to some online accounts unavailable from settings" [Undecided,New] https://launchpad.net/bugs/1234185 [14:27] gema, thx, missed that [14:27] kenvandine: np [14:28] MacSlow, right so we'll just go and invent something up [14:28] Saviq, :) [14:28] sergiusens, make the bridge job manual and call "start upstart-text-bridge" from /var/lib/lxc/android/pre-start.sh right after the socket dir bind mounting ? [14:29] Saviq, what about the autolanding failure... how often will jenkins try to merge approved branches? [14:29] ogra_, is there a reason to rm them at all though? [14:29] them ? [14:30] ogra_, rm -Rf /dev/socket [14:30] ogra_, in prestart [14:30] lool, I haven't seen any mails in the last week on the "click framework" thread - do we still need to do anything there? [14:30] to make sure you have a clean env [14:31] jhodapp: ok, seems that this change fixes it, autopilot is happy here [14:31] if you would connect to some leftover stuff that android removes or re-adds you are screwed [14:31] rsalveti, excellent news [14:31] MacSlow, as soon as it has the power to [14:31] I'm still getting a crash with mediascanner sometimes, but I'll track that separately [14:31] MacSlow, qmluitests are backed up on the VMs I'm afraid [14:31] Saviq, ok [14:31] rsalveti, I'll link to the github commit in the bug report [14:32] MacSlow, 'cause Mir ci / autolanding takes 2 hrs there... and on 2 out of 3 VMs total currently... [14:32] rsalveti, actually we should probably link from your hybris repo [14:32] dholbach: so xnox fixed the seeds, but I don't think we've actually landed the framework file [14:32] dholbach: that would be good to do now [14:32] jhodapp: well, I'm pushing a new libhybris as we speak [14:32] ogra_, but you are deleting them from the host ? [14:32] lool, awesome! [14:33] ogra_, on the ubuntu side [14:33] rsalveti, yeah, why don't you link to the commit once you push [14:33] dholbach: I guess provided by the ubuntu-sdk-libs package [14:33] jhodapp: yup [14:33] ogra_, what if ubuntu creates a socket there for ubuntu use [14:33] ? [14:33] sergiusens, that code comes from the old liking stuff ... stgraber chnaged it to the bindmount, not sure why we kept the rm there [14:33] *linking [14:33] ogra_, I'm removing the rm [14:33] it might not actually be necessary [14:33] lool, I'll add a task to bug 1219185 then [14:34] bug 1219185 in qtcreator-plugin-ubuntu "Framework "ubuntu-sdk-13.10" not present on system" [Undecided,Confirmed] https://launchpad.net/bugs/1219185 [14:34] please test heavily wghen doing that [14:34] dholbach: so would you think SDK team would upload? [14:34] ogra_, yup, will do, if it boots, bam, DONE [14:34] :-D [14:34] lool, I didn't touch the bug at all - I don't know if qtcreator-plugin-ubuntu needs any changes at all [14:34] helloworld, has anyone tried the following signals and methods form Qt in ubuntu touch => QNetworkAccessManager::networkAccessibleChanged QNetworkInfo::networkStatusChanged and QNetworkConfigurationManager:: onlineStateChange because I'm testing them in my Nexus 7 and on the ubuntu desktop and they do not work as expected (for example, signals are not raised when the network in disabled) [14:34] dholbach: I think it's more of a touch-meta upload [14:34] yep [14:35] I'll reassign then [14:35] dholbach: but I'd rather someone coordinate it, perhaps from SDK [14:35] and tracks removing all the --force-missing-framework things [14:35] lool, any idea who I can get to confirm that those qt methods are/aren't working? [14:35] dholbach: that would be packagekit plugin at least, perhaps other places [14:35] is devel-proposed the extreme end channel ? [14:35] mandel: you mean general qt questions? [14:36] alecu, lool, dholbach beuno wrt to framework, don't we need something in the image so the store filters out apps made with the T, Q, + frameworks? [14:36] mandel: you could turn to either SDK team or upstream I guess? or mail our lists [14:36] sergiusens: that's a good one [14:36] lool, specifically to if the bearer api does work on ubuntu, something that I do not believe and will mean bad news [14:36] gema, do you have an crash files in /var/crash? [14:36] lool, I'll ping sdk, then qt, the list === greyback_ is now known as greyback|shops [14:37] gema, specifically any with signon-ui in the name? [14:38] sergiusens: that's a great point [14:38] gema, i've tried to reproduce that dozens of times this morning, and only managed to once... but that one time did create a crash file [14:38] kenvandine: otp, bbl [14:38] lool, alecu good, so there is some consensus :-) [14:38] sergiusens, yep - that sounds like a separate bug, if we don't do any filtering yet [14:38] gema, i'll add my crash to your bug, but if you have one please add it too === alan_g is now known as alan_g|tea [14:39] dholbach, I don't think we even have the infra [14:39] but as with translations I've been proven wrong... [14:39] dpm, btw, the click apps have their translation files in [14:39] sergiusens, beuno: we are not doing any filtering right now. I can probably add a parameter to the searches that's just ignored for now on the server side [14:41] sergiusens, hum... not sure if I'm missing something, but the framework was specified from since when we first had click packages, beuno told me that we'd have some preliminary automatic inspection of click packages RSN, so we could read it and store it in the index, right? (or is that what you meant?) [14:41] dholbach, the scope needs to filter on it too [14:42] yep, that's right [14:42] davmor2, did you notify awe and diwic about the ringtone issues ? i guess it must be somewhere between ofono and pulse [14:42] sergiusens: the servers need to filter on some parameter sent by the scope [14:42] dholbach, and from what I've read, reading from the clicks is still a bit away [14:42] alecu, yeah, that! [14:43] alecu, not sure why I thought client side :-/ [14:43] ogra_: no I had to disappear, But I need to pick on awe_ mercilessly anyway :D [14:43] alecu, but the scope needs to know what frameworks it supports too, can it today? [14:43] * awe_ ducks [14:43] sergiusens, something like "GET /api/v1/search?q=framework:ubuntu-sdk-13.10,description:awesome HTTP/1.1" should already work (https://wiki.ubuntu.com/AppStore/Interfaces/ClickPackageIndex) [14:43] davmor2, well, it might be pulse [14:43] sergiusens, but yeah, inspecting packages still needs to land [14:43] davmor2, on a more serious note though, I'm actually working through your emails now [14:43] i dont think we had ofono changes in the images recentlyu [14:43] dholbach: good point [14:44] lool, bug 1219185 updated [14:44] bug 1219185 in ubuntu-touch-meta (Ubuntu) "Framework "ubuntu-sdk-13.10" not present on system" [High,Confirmed] https://launchpad.net/bugs/1219185 [14:44] awe_: Ofono crashed dude just as I got used to blaming cyphermox for everything ;) [14:44] hehe [14:44] yea, I read that [14:44] alecu, I just ripped off JamesTait's example :) [14:44] ;( [14:45] dholbach: but, what about two years from now? GET /api/v1/search?q=framework:ubuntu-sdk-13.10,framework:ubuntu-sdk-14.04,framework:ubuntu-sdk-14.10,framework:ubuntu-sdk-15.04,framework:ubuntu-sdk-15.10,description:awesome HTTP/1.1 [14:45] dholbach: *-meta isn't supposed to contain any content itself [14:45] ogra_, or telepathy-ofono, but I'd guess pulseaudio is more likely. Could you update the bug with info on how reproducible it is and if it happens on both mako and maguro? [14:45] If it has content outside dependencies it's not a metapackage ... [14:45] davmor2, ogra_, there haven't been any changes to ofono that would affect ringtones [14:45] dholbach: it will keep growing! [14:45] dholbach: Though I dunno, maybe it's appropriate here [14:45] davmor2, are you getting the notify-osd, but no ringtone? [14:45] davmor2, or are both missing? [14:45] dholbach: OK, ignore my comment, changed my mind :) [14:45] cjwatson, done :-) [14:46] awe_: notify osd popsup just no tone [14:46] awe_: I'm guessing at it being lennarts fault :) [14:46] davmor2, awe_, right, but there were pulseaudio changes [14:46] davmor2, Could you update the bug with info on how reproducible it is and if it happens on both mako and maguro? [14:46] what package/project is the contacts application? [14:46] bah, LP ate my comment, hrm [14:46] ogra_, I don't work on pulse. ;D [14:46] awe_, yeah, guess you know why :) [14:46] diwic: install image 75 have ring the phone [14:47] -have [14:47] diwic, seems to happen on both arches, i saw many reports the last days [14:47] didrocks: it happens on mako and maguro [14:47] (in here) [14:47] diwic: ^ sorry didrocks [14:47] waow, twice in less than 10 minutes diwic/didrocks [14:48] high score today! ;) [14:48] I was thinking of something like "max-framework:ubuntu-sdk-13.10". But that makes it harder to remove support for older versions, so I guess adding all of them makes sense [14:48] sergiusens, awesome! [14:48] davmor2, ok, interesting. I did a pulseaudio upload, but that was just to fix a few crashes on desktop, phone should not have been affected [14:49] alecu, I don't know what a good strategy might be there === alan_g|tea is now known as alan_g [14:49] also, what is responsible for rendering the indicators? [14:49] alecu, maybe it's worth bringing it up on the mailing list? [14:49] Specifically "clear all" in the message indicator didn't work but calling the action manually did [14:49] didn't work → was insensitive [14:49] diwic: music plays fine so I'm assuming it is something not happening in the background [14:50] dholbach: yes, will do [14:50] thanks alecu [14:50] cjwatson, dholbach, alecu I would assume supporting framework+2 means you support framework and framework+1 ... or is that not the design? [14:51] I'm just thinking android and API levels here [14:51] sergiusens: TBH it's kind of underspecified right now [14:51] sergiusens: I *think* the right answer is to leave the framework declaration for a given framework in place until we stop supporting it [14:52] So IOW it's up to ubuntu-touch-meta [14:53] cjwatson, yeah, but then the scope request to the backed would grow with all the frameworks it's supposed to support (unless they start doing ranges) [14:53] barry, ping [14:54] alecu, yeah, adding the parameter would be awesome [14:57] sergiusens: in any case, we will need to start doing ranges in a few years time, and we will not need to change the one on the current image :-) [14:58] alecu, true [14:58] alecu, we could make it easy and break the API in every release :-P [14:58] jodh, I removed the rm from the prestart.sh script and still getting no socket, something else must be happening too [14:58] * dholbach better keeps more useful suggestions to himself. [14:59] awe_: No I meant swipe not swype, I kept opening the camera on the startup screen on android as I was trying to swipe the startup screen to get to the home scope :D edge swiping really doesn't work on android it was most annoying :) [14:59] alecu: I think the general idea was that we would probably try to avoid dropping old APIs more than once per LTS cycle [15:00] cjwatson: great. So the number of supported frameworks will be low (say, less than 8) [15:00] Might need to go for longer though. Android seems to keep its old ABIs around for quite a while. Dunno [15:00] by a given release, that is. [15:00] davmor2, ;D [15:01] I'm guessing that in the beginning it will grow, but after a while it will stabilize a lot [15:01] pmcgowan, do you know if there was talk about preinstalling some langpacks on the touch image? [15:01] seb128, not aware of that no [15:01] I would think not [15:01] pmcgowan, what would be the right place to discuss it? [15:01] sorry otp [15:01] pmcgowan, no hurry [15:02] diwic, what was the pulse change that landed, the fix for the mic mute, or something related to the multimedia code? [15:03] davmor2: where is your bug for the ofono crash? can I take a look? [15:03] awe_, no, completely unrelated stuff. I did remove a dbus interface, but we don't make use of that anyway. [15:03] kenvandine: I only have a unity scope home crash from earlier, this problem didn't generate a crash [15:03] ok, just checking [15:03] awe_, at least not from telepathy-ofono [15:03] davmor2: awe_: I've noticed some cases where ofono goes Attached = false, and never picks data back up [15:03] gema, ok, thx [15:03] awe_, I'm wondering if there can have been some apparmor changes lately that cause permission problems or so? [15:04] cyphermox, you joining the stand-up? [15:04] oh, crap [15:04] yeah [15:05] cyphermox: ofono had completely crashed no bug report bug I can write one up. It was less dropped and more dead in the water :) [15:06] cyphermox, there are definitely issues on the ofono side, and I will be working on them today. I'd like to tweak the default logging to help us monitor the GPRS connection. Also, as I mentioned yesterday, there are several jolla GPRS patches which I will plan on building/testing today [15:06] davmor2: ok, in that case you got syslog at least? [15:07] awe_: if you tell me the patches I can go ahead and test them [15:07] cyphermox: all in the email I can transfer it all to a bug [15:07] cyphermox, I'll let you know once I've done some initial investigation/testing [15:11] awe_: alright [15:12] cyphermox, I should have some stuff for you later today, but I'd like to better understand the problem before just grabbing random patches. That said, I'll build a test package and make it available for you guys to try by end-of-day [15:13] sure :) [15:13] in the meantime I'll look at phonesim quickly, see if I can fix the issue you got, and I need to take a good look at NM 0.9.8.4 [15:14] Graduate Software Developer looking to get involved | http://askubuntu.com/q/352747 [15:15] awe_: I'll retrace davmor2's crash, if you don't mind [15:15] and file a bug for the failure [15:16] ah, scratch that, the follow up gets us a backtrace === greyback|shops is now known as greyback [15:17] cyphermox: I did say I'd try first. I think it might a died at some point but I got as much as I could for you :) [15:18] upstart would have respawned it though, so you should still be able to get online after that [15:18] in theory [15:18] from the output you added, Attached = 1 but the contexts have no IP settings, not sure why that is [15:19] that means no active data call [15:19] cyphermox: nope after I restarted the phone it was fine [15:19] unless because of the crash NM got confused with the state, but it should have brought the connection down [15:19] awe_: yes, but that's likely because when ofono crashed, NM had to clean up state, and perhaps that cleanup went wrong [15:20] cyphermox, sure... [15:20] ah! [15:20] the context isn't active... [15:20] if the context has not Settings, it's not considered active [15:20] s/not/no/ [15:21] well... I think you could have a context that is active but not yet or no longer has IP [15:21] anyway, this is probably a race between NM and ofono because of the crash [15:21] I'll try to reproduce the issue on NM's side now [15:21] ok [15:23] jodh, did the ramdisk work btw? [15:24] cyphermox, awe_ : by the way, in the indicator menu, post crash the only thing it showed was the bars for the phone and where the 3g would be it showed the non connected wifi symbol instead [15:25] wifi available even I couldn't think of the right phrase :) [15:25] davmor2, that makes sense [15:25] awe_: I hoped it would :) [15:25] if 3g is broken, it's going to show one of the Wi-Fi SS icons [15:26] davmor2: awe_: if I manually get ofono to SEGV, it just gets respawned properly by upstart and the connection gets re-established by NM without issues [15:26] cyphermox, I suspect that the crash is a red herring [15:26] yes, but still :) [15:27] and this is more related to network registration / GPRS mechanics; but good to know! [15:27] I would like to know why the connection wasn't established when davmor2 picked up his phone :) [15:27] cyphermox: it might of been because the phone was suspended and had been for a while when ofono crashed maybe? [15:27] maybe... [15:27] everything points (looking quickly) to the fact that it should have been established [15:28] I don't think so [15:28] but yes it's possible [15:28] What are Delta Updates? | http://askubuntu.com/q/352753 [15:28] davmor2, I'm still mod-standup, so again this is my main focus today [15:28] s/mod/mid/ [15:29] davmor2, but that said, I need to spend some time looking at this. As I mentioned earlier, I will produce a test-package for ofono by eod for you & others to test [15:29] awe_: yeap no worries [15:31] awe_: I agree, it was a red herring [15:31] awe_: the phone has been rebooting [15:34] mandel: pong [15:34] mhr3: I installed the latest unity8, should Apps magically move to the top of Home or do I need to change dconf? [15:36] mfisch, you need latest unity-scope-home [15:36] mhr3: okay, and latest unity8 or just the scope? [15:36] barry, small question do you have an xml file with the interface that you expose like the one I provide in the downloadmanager docs, I'm going to give and fix the aa{ss} issue [15:36] mandel: i don't. we don't need no stinkin' xml in python dbus :) [15:37] mfisch, no need for unity8, the reordering still isn't implemented [15:38] barry, ok, I'll deal with it in another way [15:39] mhr3: hooray it works! thanks [15:39] changing isps [15:39] I think I got throttled, brb [15:40] mfisch, so it's just an update to the scope? [15:40] cwayne: yeah [15:40] cwayne: and it updated 1 dep [15:41] mfisch, i still think we should have explicitely sorted it based on the inode number :P [15:41] I love that idea [15:41] pmcgowan, any news about the print dialog? [15:42] mhr3: I'd also be okay if you sorted them based on md5 sum of the file [15:43] mfisch, you like challenge, don't you? :) [15:44] tedg: hey ted, i'm trying to use the url-dispatcher tool in a script, but i can't connect to the bus name as the phablet user [15:44] tedg: what am i missing? [15:45] dbarth_, URL dispatcher? :-) [15:45] tkamppeter, product managers dont see it as a need for the phone, but possibly for the tablet later on [15:45] ah, obvisouly [15:45] tkamppeter, I think we won't do anythign short term [15:45] dbarth_, Is it running? "status url-dispatcher" [15:45] * dbarth_ palmfaces [15:45] dpm: hey [15:46] dpm: about to roll an image, but music-app changes dind't make it in [15:46] dpm: a) can we kick this or b) can we do it by hand? [15:46] tedg: hmm, says it's running though [15:46] dpm: https://code.launchpad.net/~lool/music-app/qtpowerd-rename/+merge/188728 is the one that didn't make it [15:46] dbarth_, Hmm, okay. What are you using to call it? [15:46] tedg: any apparmor specifics [15:47] tedg: i'm just calling it with no argument right now, maybe that's the main issue [15:47] balloons: could you help merge https://code.launchpad.net/~lool/music-app/qtpowerd-rename/+merge/188728 + upload it? [15:47] dbarth_, Something like this? gdbus call --session --dest com.canonical.URLDispatcher --object-path /com/canonical/URLDispatcher --method com.canonical.URLDispatcher.DispatchURL foo://bar [15:47] dbarth_, Or are you using the utility in tools? [15:48] tedg: the utility [15:48] lool, sure, doing it now [15:48] dbarth_, With no URL doesn't it just return "Usage:" ? [15:49] balloons: thanks [15:49] dbarth_, how do you become the phablet user btw ? (should be "sudo -u phablet -i" to have the right env (including the dbus addresses etc)) [15:49] win 1 [15:49] tedg: i'll continue debugging; i wanted to see if there was something obvious [15:49] ogra_: yup, i'm running in a sane environment [15:49] dbarth_, I don't think so, but the gdbus command above might give more info. [15:50] it does, yes; i'll ping back when i see the light [15:50] pmcgowan, OK, I was thinking about after version 1, too. I only want to know whether it is worth to work on the dialog or not. [15:52] lool, jenkins should merge and land it now [15:52] balloons: cool [16:01] balloons, popey: http://91.189.93.70:8080/job/generic-mediumtests/606/console failing [16:01] balloons, popey: can we ignore this right now and just merge + trigger a build? [16:01] it seems the autopilot tests are not installed or something [16:02] Mirv: you still around? [16:03] lool, it should merge. let's see [16:04] tedg: ok, got it to almost work now [16:04] tedg: but that url makes it unhappy [16:04] url-dispatcher application:///com.ubuntu.developer.webapps.webapp-twitter_webapp-twitter_1.0.2 [16:04] still it's the full app_id as i understand it [16:04] lool: I'd like to see that pass [16:05] given I couldn't test locally for the same reason [16:05] dbarth_, Yeah, you need the ".desktop" at the end. [16:05] really? [16:05] ah [16:06] lool, ok merge job is running, let's see [16:07] mfisch: today yes, normally no :) [16:07] sergiusens, happrove pls? https://code.launchpad.net/~cwayne18/phablet-tools/phablet-config-timezone-fix/+merge/188727 [16:07] popey: did you have -autopilot installed? [16:07] yes [16:08] Mirv: okay, I sent you the question in email, looking for a deb for the Qt fix [16:08] Hi!, is there any way to restart unity on nexus4 other then "sudo reboot" ? It hangs/freeze quite a lot lately... [16:09] jibel: apport-cli /var/crash/_report.crash fails to upload as it doesn't open the webbrowser by the look of it. I'm might resort to installing w3m, how are you getting around that? [16:11] lool, you are merged and it's in queue to build: https://launchpad.net/~ubuntu-touch-coreapps-drivers/+archive/daily/+build/5069757 [16:13] mfisch: answered, ie upgrade from ppa:canonical-qt5-edgers/qt5-daily once the armhf has built in about 2 hours [16:14] Mirv, we'll be sure to test it and get you some feedback :) getting this bug fixed would be a huge win for us [16:14] balloons: thanks [16:14] cyphermox, I think I see the possible cause of the ofonod crash [16:14] rsalveti, ^^ [16:15] cool [16:15] has to do with the way we register the various atoms, using a delay [16:15] popey: will dive in the AP issue later; could this be due to AP versions? [16:15] hm [16:15] balloons: that's built now [16:15] rsalveti, I had a suspicion we might get bit by this technique ( hence the comments in the code all over the place ) === boiko_ is now known as boiko [16:15] lool: I'd bow to balloons superior domain knowledge there [16:15] awafaa: right [16:15] awafaa: sorry :-) [16:15] awe_: right [16:16] balloons: ^ :-) [16:17] lool, popey I'm afraid I've missed the question [16:17] cjwatson, great, ty [16:17] Mirv: thanks [16:17] balloons: we dont understand why the music-app MP failed to run the AP tests; popey has the same issue [16:17] Mirv: it would take more > 2 hours to build Qt so thats perfect [16:18] balloons: this might be due to new AP version, it got published to PPA for a couple of hours earlier today, but it was gone when that test started [16:18] fginther: ^ [16:19] lool, ahh, well that's simple enough :-) ubuntu-ui-toolkit-autopilot is still in a frenzy [16:19] cwayne_ mfisch: with our new builders the qt5-daily might be ready even quicker qtdeclarative is not as big as qtbase [16:20] lool, it's re: everything we spoke about this morning [16:20] ah [16:20] Mirv: what packages do I need to upgrade once the fix builds? [16:20] balloons: would that explain the odd "can't import music_app" though? === CarlosNeyPastor_ is now known as CarlosNeyPastor [16:21] balloons: the import music_app fialing is odd [16:22] cjwatson: hey, did you say click packages with Click-Version < 0.4 are not a problem? [16:23] lool, as far as why it failed in jenkins yes. I'm not sure what other errors you are seeing. [16:23] balloons: popey: ok, the issue was likely missing qtpowerd in the PPA and PPA not being included [16:23] fginther is including that now [16:24] balloons: ping [16:25] mhall119, pong [16:25] * balloons assumes music will play now after the screen is off :-) [16:25] balloons: hey, I need somebody to help me make an autopilot test case for https://code.launchpad.net/~mhall119/ubuntu-calendar-app/fix-newevent-month-1233442/+merge/188488 [16:25] is image 76 still on track for today?? [16:26] mhall119, ahh, I confirmed that bug ;-) I remember it [16:26] I took a look as the autopilot tests that are there (few) and honestly I was pretty lost [16:26] mfisch: hmm, right, you don't necessarily want to dist-upgrade from archives but use the flashed image otherwise as is? the PPA is clean enough in itself. [16:27] let me look at your diff mhall119 .. you apparently didn't break any of the tests [16:27] ohh.. tiny :-) [16:27] Mirv: we'd like to update as few packages as possible [16:28] so mhall119 you could modify the new_event test to check and ensure the date is correct [16:29] mfisch: apt-get install libqt5qml5 libqt5quick5 qmlscene qtdeclarative5-folderlistmodel-plugin qtdeclarative5-localstorage-plugin qtdeclarative5-qtquick2-plugin qtdeclarative5-test-plugin qtdeclarative5-window-plugin qtdeclarative5-xmllistmodel-plugin should do it. check that the version number being fetched is the 5.0.2-6ubuntu3~saucy1~test1 from the PPA [16:29] Mirv: thanks, we'll have that in our backup plan in case this doesn't land in the image [16:29] balloons: should I modify new_event or create a new test case for this specific bug? [16:30] mfisch: yep, plan B:s and C:s are always useful. [16:30] I would modify new_event in this case, because it should be checking that after creating the event. [16:31] mhall119, ahh ok, yes looking at the test. It verfies the event shows up, but not the date [16:31] balloons: so for example, the new_event test has: [16:31] self.assertThat(self.main_view.get_new_event, [16:31] Eventually(Not(Is(None)))) [16:31] balloons, popey: So the root cause is: autopilot tests are run on amd64, qtpowerd not built on amd64 yet, and PPA not being enabled for dependencies of AP tests before (but now is); we should retry later when amd64 bits are available [16:31] fginther: Thanks a lot for fixing [16:32] what is self.main_view.get_new_event calling? [16:32] ah, that makes more sense [16:32] ok, :-) I didn't see that error just the ui-toolkit one [16:32] lool: I can score more things up on request. [16:32] (hint) [16:33] sergiusens: coming [16:34] mhall119, that assert is related to timing. The toolbar is opened and the new event button is clicked. That assert is waiting for the new event box to appear, and ensure that it does [16:34] cjwatson: :-) would be nice then: https://launchpad.net/~ubuntu-touch-coreapps-drivers/+archive/daily/+build/5069700 [16:34] cjwatson: amd64 [16:35] the code looks like it could be a bit more readable, but yea, it's all there. So for this after everything is entered there is an assert to make sure the event exists [16:35] lool: oh, well, it was already "start in 3 minutes", but scored up anyway [16:35] mhall119, you see the #verify that the event has been created in timeline assert? [16:35] balloons: yeah [16:35] mhall119, I would add an additional assert right after to make sure the dates are correct [16:36] that should be all that is needed to make sure this doesn't regress [16:38] mhall119, I walk you through how to do that if you are curious. If not, well, let's make sure a bug exists to get it added [16:39] balloons: I'll give it a try later today, and come back to you if/when I need help [16:39] mhall119, kk.. :-) [16:39] mhall119, autopilot launch and autopilot vis are your friends. [16:40] balloons: is "autopilot launch" what I need to run the tests? [16:40] (Not sure it's the right place to ask.) How can I disable the fact that Page { } is flickable? It's stealing my touch events and I dislike that. Is there a way to disable the vertical movement? [16:41] mzanetti: I know you're busy, I thought I could use your brains real quick if you're around :> ↑ [16:42] what's the channel to use to get daily builds for ubuntu-system? [16:42] Oh, it's not even the property of page. It still has vertical movement, when Page { } is not the parent (I commented it out) === gatox is now known as gatox_lunch [16:44] mhall119, autopilot launch -i Qt qmlscene path/to/qmlfile.qml [16:45] mhall119, autopilot vis, select dbus connection [16:45] mhall119, that let's you look at the object tree so you know the object and properties you want for the date assertion [16:46] ok [16:46] karni: afaik Page is not flickable [16:47] mzanetti: yeah, it's the MainView that somehow allows vertical movement in its "body" (header sticks) [16:47] karni: that's new to me :D [16:48] mzanetti: I suspect I might not even be able to turn that off ;) Hahah I see :) Ok then, thanks for chiming in! [16:48] karni: you really sure you haven't got a Flickable in there yourself? [16:48] mzanetti: NACK :) Just a MainView with Page with a GridView [16:48] I commented out Page, and it still has vertical "flex"-like movement. [16:48] karni: ok, the GridView is flickable [16:49] mzanetti: right, but It's smaller than the MainView content [16:49] karni: can you paste the file? /me curious [16:49] mzanetti: so it's not grid view "scrolling", the whole grid view is "flex"ing when I move it up and down. sure thing! [16:50] mzanetti: http://paste.ubuntu.com/6184654/ It will be a sound board. [16:51] mzanetti: But when you touch one of grid items, and move your finger, the onCanceled should be called on the MultiPointTouchArea. It's not, so instead of struggling with that, I was hoping to disable that "lost focus" at root, which is that MainView body vertical flex-like movement. [16:51] I hope it makes sense haha. [16:51] karni: put "interactive: false" in the GridView [16:51] its the GridView that's moving [16:51] \\\\o/// [16:51] * karni tries [16:52] * karni hands tasty cake to mzanetti !! [16:52] mzanetti: <3 thank you [16:52] :) [16:52] mzanetti: Even though it fits the screen. I wouldn't know it was in fact the grid view. Lot's to learn! :) [16:52] mzanetti: Thanks again. === kentb-out is now known as kentb [17:00] issues with some patch | http://askubuntu.com/q/352780 [17:07] I can't find any information about Ubuntu Touch on the Sprint LG Optimus G [17:07] does anyone know anything? [17:10] how can i file a reasonable report of "system-settings keeps locking up" exactly? [17:13] didrocks: will 76 have the new thumbnailer for music/videos? It's in the plan but not your email [17:14] lallenlowe: is it listed at https://wiki.ubuntu.com/Touch/Devices ? [17:14] mfisch: it's part of image 76 [17:14] didrocks: meric [17:14] mfisch: just nothing is using it :) [17:14] didrocks: merci even! [17:14] mfisch: de rien, watch the archive, it's already in ;) [17:14] didrocks: so we dont have functional thumbnails in 76? [17:14] yeah, no functional thumbnails yet [17:15] just the service for it [17:15] cwayne_: ^^^ [17:15] didrocks: thanks === alan_g is now known as alan_g|EOD [17:15] yw [17:19] mfisch, yep, was watching it earlier, its in archive [17:19] cwayne_: yeah but it doesnt do anything for it [17:19] cwayne_: for us [17:21] mfisch, it needs sdk + unity8 branches which didn't land yet [17:22] mhr3: I think we can live without it [17:27] popey: no [17:27] popey: I was wondering if there were any more than those [17:34] popey: with Francis' fixes, we're back to the usual set of failures: http://91.189.93.70:8080/job/generic-mediumtests/613/ [17:36] anders3408: you got your package? [17:37] yes mhall119 i did :) and thanks its nice :) [17:37] great! happy hacking :) [17:39] yeah, make Mir work ! [17:39] :D [17:40] Mir on the OPPO Find 5 would be pretty sweet [17:40] I assume it could use whatever Android graphics drivers there are for that device [17:40] ogra_: thats also the next thing i should get working :) [17:41] mhall119, i think the oppo isnt to far from the mako === gatox_lunch is now known as gatox [17:41] ogra_: correct :) [17:42] note that the Mir guys needed some kernel patch fro fixing a flickering issue [17:43] (foor mako that is) [17:43] ehm... do we talk Mir aka microphone ? :) [17:43] rsalveti: got time to review a patch for NM? http://paste.ubuntu.com/6184847/ [17:43] anders3408, nope, Mir aka graphics stack === lolz is now known as Guest54911 [17:44] oh., that was the thingy that needed some changes to get it to use it ? ogra_ [17:44] touch /home/phablet/.display-mir ... [17:44] reboot [17:45] see if you still have a device or if its a brick now :) [17:45] ogra_: well, thats not going to happens :) [17:45] (adbd will still work) [17:45] reboot fails you know :) [17:45] and removing the touched file gets you back to surfaceflinger [17:45] having a race between lxc-container and ubuntu , which parts is the first to get mounted the /system partition :) [17:46] first bootup lxc is fastest, 2nd bootup ubuntu is fastest, which results in a failed boot as then lxc will not mount correct [17:46] oh, you still fight with that one ? [17:46] so far i havent found anywhere i can add a delay that will get it going, so yes i am :( [17:47] ogra_: somewhere there must be something that triggers the device to know its a "first bootup" will it be impossible to trick it to always thinks that ? [17:47] cyphermox: sure [17:48] anders3408, there will surely be ways out of that [17:48] im not knowing enough about ubuntu touch , to fiugurer that stupid issue out :( [17:50] hi [17:51] can someone tell me how to remove apps like facebook, amazon... doesnt work through gui [17:56] ogra_: going afk for a bit, but get notifications so ping me if you want any images testing [17:56] popey, 76 is just finishing, but likely not releasable due to ringtone still being broken [17:56] so take your time :) [17:56] \o/ [17:56] * popey makes dinner [17:57] Guest54911, long press the icon, then uninstall [17:59] pmcgowan: doesnt work... child process exited with code 7 [18:00] works here [18:00] but doesnt really update the application lens ... [18:02] Saviq: ping [18:02] bfiller, pong [18:03] cyphermox: seems fine, does it fix a specific bug? (not sure we have one opened for it already) [18:03] Saviq: any chance of getting a fix for this one soon? https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1234323 [18:03] Ubuntu bug 1234323 in unity8 (Ubuntu) "hide OSK after searching and clicking on result" [Undecided,New] [18:03] cyphermox: would be nice if awe_ could test that as well (http://paste.ubuntu.com/6184847/) [18:03] rsalveti: yeah I'd just need to get the right bug number; one was filed by davmor2 already IIRC [18:04] bfiller, duplicate of https://bugs.launchpad.net/unity8/+bug/1213034 - in progress [18:04] Ubuntu bug 1213034 in Unity 8 "Can't dismiss keyboard by tapping outside of search entry" [High,In progress] [18:04] Saviq: thanks [18:04] rsalveti: awe_ has tested but we've hit some ofono issues as well while testing [18:04] cyphermox: for what sorry? [18:04] davmor2: auto reconnection [18:04] cyphermox: so this is a possible fix when we get a working ofono :-) [18:05] but looks low risk [18:05] cyphermox: let me check for you [18:05] when we get a working ofono?! :) [18:06] cyphermox, thats surely 14.10 material :P [18:06] hehe [18:06] be nice, awe_ is working hard :) [18:06] cyphermox: https://bugs.launchpad.net/ubuntu/+source/ofono/+bug/1230218 [18:06] Ubuntu bug 1230218 in ofono (Ubuntu) "Maguro: Ofono/network-manager issue when a provider covers multiple networks" [Undecided,New] [18:06] cyphermox: is that the one I think it is [18:06] wooo [18:07] seems right [18:07] davmor2: is that what you've tested with my fix though? [18:09] cyphermox: Yeap I haven't done the same kinda distance as I did that weekend and likely won't till this weekend. But I hit drops and e → 3g and it handled all of that, I would need to move between towers to have the network name changes [18:09] alright, awesome [18:09] just making sure [18:10] cyphermox: but I'm assuming that as long as there is a change that nm should pick up on it and be happy right? [18:11] correct [18:13] cyphermox: so What I'll do tonight is check and see if the ip address changes on me at all like it did with android [18:17] davmor2, I'm not so sure the issue is the multiple networks, I think it's a more fundamental issue with the my GPRS implementation that just happens to act really bad on your provider's networks [18:17] rsalveti, cyphermox, not a big fan of reviewing pastebins. ;)- [18:18] awe_: me neither [18:18] but that said, does this correspond to 0ubuntu22 cyphermox? [18:18] ;D [18:18] yeah [18:18] precisely [18:18] coolio [18:18] if you're so enclined, check rev 767 on lp:~network-manager/network-manager/ubuntu [18:18] rsalveti, I have tested, and it does make things better, however as cyphermox mentioned, I'm looking at a couple of gprs patches now [18:18] or you know, the latest rev [18:19] cyphermox: +1 then [18:19] cyphermox, right now I'm more focused on the ofono side, but will test everything together... [18:19] yeah [18:19] let me know when you got ofono, I'll test as well [18:19] ogra_, lool, slangasek i have fix committed all 4 critical s-i bugs and i could upload s-i 1.8 at any time. if you need me to wait, i can also try to sneak in a few more bug fixes. please give me a landing slot [18:20] cyphermox, awe: let me know to and I'll break^Wtest it to :D [18:27] does anybody know if bugs milestoned 13.06 will be fixed? like https://bugs.launchpad.net/touch-preview-images/+bug/1131794 for example [18:27] Ubuntu bug 1131794 in touch-preview-images "No incoming calls when hiding callerID" [High,Confirmed] [18:28] stgraber: was there some change in how the json files for system images get updated again? something caused the automated tests to get triggered twice with the build that just came out [18:28] lenios, milestone 13.06 is an old milestone... I'll add this to my list to take a look at, although no guarantees [18:29] so cwayne_ tells me that album art sometimes work, although it's never worked for me, is there a trick or hack to make it work? [18:30] lenios, that said, have you tried this on a recent image? This bug was reported a *long* time ago ( Feb ) [18:30] is system-settings freezing up a lot on the device, for anyone else? [18:30] mfisch: have a network conection and music with album art? [18:30] it'd be nice to have confirmation that this is still a problem... [18:30] i remember a bug saying that there was no check when a sms was sent, so it if failed to deliver, there was basically no sign something went wrong [18:30] yes, that's true too [18:31] davmor2: I did not have a network connection on first boot, but I do now and yes, I have album art on these. [18:31] cwayne_: you're using the standard ones from sevilerow rigt? [18:31] i really wish this kind of bug can be fixed by the final release next month... [18:31] next month? [18:31] try this month [18:31] ;D [18:31] oh wait [18:31] plars: can't think of anything for the last image, though yesterday I've been testing set-phased-percentage which does change the json and may tricker your stuff [18:31] it's already october [18:32] mfisch: I just import tracks from my .ubuntuone folder and they all work fine [18:32] lenios, we're trying to land as many bugs fixes as we can, without breaking things... [18:32] lenios, *exactly* [18:32] yeah i know [18:32] keeping an eye on the mailing list [18:32] plars, my updater also only offers me 76 [18:33] davmor2: do you know what tool is supposed to load the art? just unity8? [18:33] plars: if you have some spare cycles, it probably would be worth finding a way not to rely on the json timestamp, because I really can't guarantee it'll only change when we add something... the rewrite made this be the case 99% of the case but that's still not guaranteed and won't ever be [18:33] mfisch, yes [18:34] yes re: using the standard SR ones [18:34] ogra_: isn't 76 the latest image? [18:34] mfisch: it's grillo and media scanner off the top of my head [18:34] stgraber, it is ... all fine imho [18:34] ok [18:34] * ogra_ taps "install and restart" [18:35] davmor2: thanks, i'm looking at the mtp logs [18:35] stgraber: well, it's not a disaster if we accidentally run the tests twice, just not necessary and hard to tell until we get there if it's really two images that triggered it, or just some random update [18:35] barry, i can add an entry to the landing asks page ... it wont be considered before tomorrows landing team meeting though [18:36] stgraber: from our side, there are a few other options we could use for triggering, such as a script that runs a tool that checks the json file, but that relies on the output not changing (which we've seen happen before) [18:36] stgraber: one option that I was hoping to use, is that we could trigger off of a change to a field in the json - but I don't think there's a global latest_image key, or at least nothing like that last I checked [18:37] stgraber: but if there were, you could update the json 100 times a day for all we care, as long as that only gets bumped when there's really a new image [18:37] barry, though i understand that lool already worked with you on getting the last s-i bit in so if he doesnt mind i'll leave it to him [18:37] plars: nope, though sorted({image['version'] for image in index['images']})[-1] will give you that and is pretty reliable (since it depends on the one field we can't possibly change without breaking everything) [18:38] cwayne_: can you share your mtp log with me? [18:39] I see that the DB is being built okay and knows where to find the icon even [18:39] where is it [18:39] stgraber: right, that's not a static key we can tell the jenkins plugin to look at though [18:40] plars: right, you'd need to do the two stage trigger we discussed a while back. Have change to .json trigger a jenkins job which runs a python script parsing the index and then if that version is different, trigger the real jenkins job [18:40] stgraber: yeah, it just triggered again after I canceled the dups, we're definitely seeing this again [18:42] cwayne_: in application-legacy-music-app-.log I have lines like this, note how Cover is empty as is Genre, does it work for you? http://paste.ubuntu.com/6185110/ [18:42] plars: yeah, I think there's a bug somewhere, it appears to get updated at every import-image run (so every 5 minutes) [18:43] mfisch, where's that? not var/log [18:43] plars: actually depends what you monitor, are you looking at devel-proposed or saucy-proposed? [18:43] stgraber: that seems consistent with what I'm seeing [18:43] stgraber: devel-proposed [18:43] cwayne_: phablet upstart logs bro [18:43] stgraber: I thought they were equivalent though? [18:44] plars: it's an alias, not a symlink, so the index is actually slightly different and one of the difference is a bit of code I added yesterday which may be causing what we're seeing (related to phased updates), looking into that now [18:47] mfisch, i dont have that log.. [18:47] cwayne_: maybe b/c I opened the music app [18:48] mfisch, i even rm'd Cover.jpg and it still shows up [18:49] why does this phone hate me [18:50] ssweeny: after you merge I'll kick jenkins [18:50] does anyone ever run any of these touch apps from a console? [18:50] I sometimes do [18:51] do you file bugs about all the qml errors that get printed? [18:52] I will if it's something i know is wrong like they typo'd the icon path or similar, other stuff I'm not sure if it's known already or harmless given my limited Qt experience [18:53] mfisch, mediascanner-service[1944]: ESC[1;34mINFOESC[0m info/metadata: Using "grl-lastfm-albumart" from "/usr/lib/arm-linux-gnueabihf/grilo-0.2/libgrllastfm-albumart.so" to resolve metadata. [18:53] fginther: any problem with 10.97.2.10? it seems to hang here... [18:53] mfisch, it's using lastfm to get the albumart [18:53] cwayne_: hmm, I'm on the interwebs [18:54] i wonder if grilo needs it in a certain place [18:54] I just told it to scrobble the cover whatever that does [18:54] stgraber, yes, something is going on, someone is investigating [18:54] ogra_: sounds good, thanks. if i have until at least tomorrow, i have a few more things i can fix up. [18:54] cwayne_: did you setup u1 on there? [18:55] mfisch, no [18:55] davmor2, who would know more about grilo? [18:56] mfisch, merged [18:57] mfisch, mv Cover.jpg to Josh\ Woodward\ \-\ The\ Simple\ Life.jpg [18:57] mfisch, i think that should get picked up by the grilo local-metadata plugin [18:58] cwayne_: thomas stych I think and I've probably spelt his name wrong so I apologise for that [18:58] -> quick break [18:59] fginther: please let me know when it's fixed, I had to disable the system-image cronjob for the time being [19:00] stgraber, ack [19:11] Anyone around? Just quick few questions regarding build 76 [19:11] Process: 668 root 20 0 268 4 0 R 101.1 0.0 2:09.93 upstart-propert [19:12] is hogging 100% CPU [19:12] reboot does not help [19:12] timppa, bug already found and being fixed [19:12] plars: anyway, I found a bug in the current system-image code and fixed it, we'll know if that's enough to fix the json file updating every 5min once the Jenkins server is back to life and I can turn the cronjob back on [19:12] ChickenCutlass: ok, will it be available today? [19:13] timppa, should be [19:13] stgraber: which jenkins server does this run off of? [19:13] ChickenCutlass: great! [19:13] plars: none, my code runs on the cdimage builder, but it imports bits from 10.97.2.10 (OEM customization tarballs) and that server is hanging at the moment which in turn makes my importer hang [19:13] s/cdimage builder/cdimage master/ [19:14] And the another thing is that the clock is still randomly missing, should that be fixed already? === jhodapp is now known as jhodapp|afk [19:24] cwayne_: I'll try that rename now [19:38] awe_: I don't know if you wanted a bug to work against but I've given you one https://bugs.launchpad.net/ubuntu/+source/ofono/+bug/1234364 [19:38] Ubuntu bug 1234364 in ofono (Ubuntu) "ofono will randomly drop the connection" [Undecided,New] [19:39] Hello, since yesterday I am trying out ubuntu touch on a nexus 4. Today the suggested apps for install don't show up anymore. looks like they're disabled. But I can't reenable them. Click Package scope is enabled [19:40] tl;dr: no more suggested apps, how to re-enable [19:41] phiphi: connect to wifi, reboot the phone (it's a little temperamental) [19:43] is there a reboot functionality in the gui? [19:44] phiphi: nope open the terminal and sudo reboot or press and hold the power button till it shuts down the npower it back on again [19:45] Saviq: can this be top-approved? https://code.launchpad.net/~saviq/unity-notifications/fix-themed-icons/+merge/188483 [19:46] bfiller, how should I know :D [19:46] bfiller, someone should review it ;) [19:46] bfiller, but yeah, probably yes [19:46] Saviq: it was reviewed and approved [19:46] Saviq: by macslow but not top-approved [19:47] bfiller, apparently Mirco didn't want to make the call [19:47] Saviq: ok, I will :) [19:47] after testing it [19:47] bfiller, thanks :) [19:52] ogra_: your boot slowdown where apparmor_parser started-- was it only after rebooting as writable? [19:56] davmor2, how is this different than the previous bug? [19:56] ( ie. the 3g is flaky on EE networks )? [19:57] davmor2, also can you add build #, how-to-reproduce, ...? [19:58] ie. was Wi-Fi every on, or is the bug, boot the phone, check 3g, wait several hours and it's the connection is gone, never to be re-established? [19:59] davmor2, that said, I believe there's definitely a problem, and we're working on it [19:59] I checked the nemomobile GPRS patches, and I'm not sure they'll help... [19:59] that said, I'm building a test package as we speak... [20:00] and will also supply you with a debug test package that logs more events by default [20:00] sergiusens: ping. Do you know what's the right call to get the arch for the machine? like "armhf" [20:01] I'm trying "uname -i", but get "armv7l" on my nexus [20:02] alecu, only way I know how to is with dpkg-architecture [20:02] alecu, I bet jdstrand would know of other ways [20:03] * jdstrand was going to suggest dpkg-architecture [20:04] cjwatson: perhaps from the scope I should be asking click for this too? [20:05] fginther: looks like jenkins is working again? [20:05] stgraber, it just came back up [20:05] mfisch, \o/ [20:06] i think i got it [20:06] cwayne_: nice [20:06] cwayne_: something I can try? [20:06] mfisch, maybe, im not 100% sure [20:06] mfisch, yeah, 1 sec [20:07] mfisch, have that album-art saved where i told you before, and do this: gsettings set com.canonical.mediascanner metadata-sources " [('grl-tmdb', 'grl-tmdb', {'api-key': 'TDB:REPLACE-DUMMY-KEY'}), ('grl-lastfm-albumart', 'grl-lastfm-albumart', {}), ('grl-local-metadata', 'grl-local-metadata', {})]" [20:08] plars: so I probably just triggered another of your Jenkins runs by landing my fix, but I then ran the import script 2-3 times and didn't see a bump in the timestamp for the json file, so I "think" I solved the issue [20:08] plars: let me know if you still see changes and I'll look into it some more [20:11] ogra_: do you know if the mediaplayer fix is confirmed fixed in 76? [20:11] barry: ok, so we are actually waiting for a fix to go in before we can promote an image [20:12] cwayne_: i need to reinstall first, I've messed with it too much [20:12] barry: I doubt it will done today, but perhaps if we're quick [20:12] barry: would you upload a system-image ~ppa1 snapshot to ubuntu-unity/daily-build PPA in the mean time? [20:12] barry: I'll ready you a slot for tomorrow [20:13] barry: are you ok that I upload it tomorrow morning? I would take the snapshot .dsc and upload it with a new changelog entry [20:13] mfisch, alright, im gonna go afk for a bit, ping me when you try it out [20:13] cwayne_: will do [20:13] lool: tell you what. i'm going keep fixing bugs until near my eod, then i'll tag and upload to that ppa (although i think i had some permission problems with that last time?) [20:14] barry: PPA >> ok, then to another PPA, it doens't really matter where; it's the one we use for staging builds, but I'd just upload by hand anyway [20:14] lool: running 76 here, what is the test case for mediaplayer? [20:14] popey: that's what I was asking ogra; I think it's when playing long movies [20:14] ah [20:14] popey: FYI there's a regression in 76 anyway [20:15] there is [20:15] phone doesn't ring or beep on calls/sms [20:15] lool: this is the packaging branch. i'll leave this in a state where you can easily `bzr bd` the package at my eod, and it's fine to upload it whenever you get online in the morning. https://code.launchpad.net/~ubuntu-system-image/ubuntu-system-image/client.pkg [20:15] popey: there's an upstart-property-watcher issue and the missing ringtone [20:15] barry: ok cool [20:15] ok [20:15] lool +1 [20:15] barry: I might not be able to commit there either, but that's ok [20:16] lool: hopefully you won't need to, but i can try to fiddle the perms if you want [20:16] barry: so looking at the list, this seems really good; just missing manifests? [20:17] phased updates and alias tracking are there [20:17] download progress [20:17] and auto download again [20:17] mfisch, btw just updated qt5 on my phone from that ppa [20:17] lool: manifests? [20:17] lool, what is the sound issue [20:18] cwayne_: and? [20:19] mfisch, first boot, scopes were there [20:19] mfisch, gonna try 10 reboots again before i go afk [20:20] ChickenCutlass, any idea why mediascanner doesnt use grl-local-metadata? [20:20] ChickenCutlass: (Ricardo is on it) [20:21] cwayne_, because you didn't write the code [20:21] popey: apparently mediaplayer AP tests were enough to trigger [20:21] k [20:21] ChickenCutlass, it's a gsettings key brochacho [20:21] codes already written [20:21] popey: Ricardo S said these pass fine in 76 [20:21] cwayne_: okay I'm freezing sevile then [20:21] mfisch, wait and see if this fixes albumart for you [20:21] cwayne_, I really don't know -- bro, not sure who owns that [20:21] cwayne_: yep [20:22] if it does, we can override the gsettings key and have albumart work [20:23] stgraber: looks good now, thanks :) [20:24] sergiusens, jdstrand: dpkg-architecture is a perl script that does a lot of stuff, and ends up calling "dpkg --print-architecture" [20:24] I think I'll be using just that [20:25] cwayne_: testing now [20:25] cwayne_, sounds like a bug [20:25] pmcgowan, yeah, it would make sense to parse the filesystem for available metadata as well as online sources [20:26] at least makes sense to me [20:26] cwayne_: I wonder if we can just re-use the bug I filed? [20:26] cwayne_, for any sample data provided makes sense [20:26] although it's 2 separate issues [20:28] alecu, sounds good, as you don't need the triplets [20:30] cwayne_: okay so I moved the file like you said and ran gsettings and still no cover art [20:30] alecu: that sounds very reasonable IIUC that you are filtering by arch [20:30] mfisch, does grep local ~/.cache/upstart/mediascanner.log show anything? [20:31] cwayne_: I see it doing a scan, anything I should see? [20:32] cwayne_: and the rename was to that crazy path or to cover.jpg [20:32] ? [20:32] the crazy path [20:32] album-blah-blah.jpeg [20:32] mfisch, btw with the new qt5, 10 reboots, scopes showed up 10/10 [20:32] 10/10 would scope again [20:32] cwayne_: ncie [20:32] alecu: Hm, maybe this should be wrapped using click; but for now, you should use "dpkg --print-architecture" (not dpkg-architecture, which is a developer tool) [20:33] sergiusens,jdstrand: ^- for reference [20:33] (ah, and now I see that you got there in scrollback) [20:34] cjwatson, always good to get re assurance :-) [20:35] cwayne_: any ideas? [20:36] mfisch, are you sure you set the key right and have the right filename? [20:36] cwayne_: I pasted your crazy path: /home/phablet/.cache/media-art/album-4cbb87d0b1f826c892c7a95518c54aea-d0bb19d7e29f213abb959593165765db.jpeg [20:37] mfisch, and what about the gsettings key [20:37] com.canonical.mediascanner metadata-sources [('grl-tmdb', 'grl-tmdb', {'api-key': 'TDB:REPLACE-DUMMY-KEY'}), ('grl-lastfm-albumart', 'grl-lastfm-albumart', {}), ('grl-local-metadata', 'grl-local-metadata', {})] [20:38] cjwatson: great. Yup, found that too, thanks! [20:39] mfisch, huh, goddamnit [20:40] Do we downscale art resources on the fly in UTouch? Are they cached? [20:41] * karni expects a question @Deview why are we not conserving CPU if we're downscaling all resources on the fly, if that's the case [20:42] karni: not sure [20:42] is there a Right Way to do environment variable-based debugging logs in Qt, a la g_debug()? [20:42] qDebug() seems to only be togglable at compile time [20:42] It is either install time (which I doubt, but I wish!) or runtime. [20:43] mfisch, it seems we don't set XDG_CACHE_HOME [20:43] veebers: I just added a comment to your MP: https://code.launchpad.net/~veebers/autopilot/fixing_backend_being_none/+merge/188762 [20:43] mfisch, which grl-local-metadata uses [20:43] doanac: ack, will check [20:44] let me know if you need help re-creating [20:44] cwayne_: we set most everything else in an upstart job [20:45] mfisch, yeah but we dont set that [20:45] okay I'll try it [20:45] mfisch, i just added it to our custom-env.conf, lets see if it does anything for me [20:45] i wish i knew how to get logs from grilo [20:46] cwayne_: there's a job called xdg-dirs already [20:46] mfisch, and it's not set in there [20:48] ogra_: I just can't reproduce your issue. 25 boots, 25 no apparmor_parser. reviewed scripts, reviewed bootcharts. this was on grouper. I am going to try with mako now, but have to do it on stable instead of proposed for now [20:50] ogra_: your output from the click-apparmor upstart job though-- that shows that apparmor was updated, or that maybe there was a problem with accessing the md5sums [20:50] reboot [20:50] meh [20:51] mfisch, it worked for me, i even made sure to disable the lastfm one [20:51] unless it's just cached.. let me remove the music and re-push it maybe [20:51] cwayne_: nope [20:51] cwayne_: still fails here [20:51] take out your sim card [20:52] no [20:52] this is my only phone [20:54] mfisch, just tried recopying it over and moved the albumart out of the way, making sure it fails first [20:54] cwayne_: nobody's gonna call you! [20:55] mfisch, MY MOM MIGHT [20:55] mfisch, ok, failed as expected, moving the file back and rebooting [20:57] mfisch, \o/ [20:57] works for me [20:57] App Showdown winners announced! http://developer.ubuntu.com/2013/10/winners-of-the-2013-ubuntu-app-showdown/ [20:58] What is the proper way to ask for device orientation? Is there something better than: property bool isHorizontal = width >= height; ? [20:58] http://www.reddit.com/r/Ubuntu/comments/1nlx69/winners_of_the_2013_ubuntu_app_showdown_announced/ too [20:59] mfisch, im gonna push up a branch, will you try it and see if it works for you? [20:59] cwayne_: sure thing bro [20:59] mfisch, i wonder if my disabling lastfm somehow made it work [20:59] cwayne_: I've got a meeting but I can do it during [20:59] like maybe it tries lastfm, it doesnt work, then says fuck it and bails [20:59] mhall119, nice [21:00] yay mhall119 [21:00] cwayne_: and what is the errno for "fuck it"? [21:00] mfisch: -69 maybe [21:01] Anyone knows the _proper_ way to query current device orientation? [21:02] http://design.ubuntu.com/apps/get-started/responding-to-orientation sure does say "implement orientation", but is limited to "set this flag to true to enable it" [21:02] mfisch, is an upstart job in .config/upstart/ run the same way was in /usr/share/upstart/sessions/ ? [21:02] It might be helpful when desining flexible layouts to know the orientation, and I was wondering if a simple dimention comparison was the way to do it. [21:05] cwayne_: I think so [21:06] mfisch, lp:~cwayne18/sevilerow/local-art [21:06] mfisch, now im actually gonna go afk, like i said i would an hour ago :) [21:06] ill do a fresh flash in the meantime, and see if this branh works when i get back [21:07] hello... :) [21:08] after adding the patch for apparmor v3 to my kernel , ubuntu touch fails to boot , lxc container always fails to mount .. even with apparmor=0 is added [21:12] karni: I think using Qt's Screen is the way to check: http://91.189.93.79/api/qml/sdk-1.0/QtQuick.Window.Screen/ [21:12] anders3408, not sure if this helps: check /var/log/kern.log for apparmor DENIED messages [21:13] w-flo: aa-status says module not loaded, and no apparmor related lines in kern.log [21:14] then that's obviously not your problem... *shrugs* [21:15] mhall119: Thank you [21:16] mhall119: That's some sort of staging server? Why the ip? [21:17] karni: yeah, it's staging for what will be on developer.ubuntu.com son [21:17] soon [21:17] mhall119: thank you, I'll note it down [21:18] w-flo: correct , thats what i dont get :) another question for you :) : first bootup normally is fine, if i reboot it fails due to ubuntu mounts /system before lxc-container mounts system partitions, so in the lxc container it fails to mount it. do you have any idea on what i can do to that ? [21:18] hm.. why does it fail in the container? I believe it simply mounts twice for me [21:19] w-flo i dont know , and i dont know [21:19] ups :) [21:23] mhall119: https://plus.google.com/107310930968828806488/posts/CFQn91RRJ1W [21:23] anders3408, tbh I have no idea what causes /system in the container to be populated... I'm just glad it works for me [21:23] w-flo: check link from above :) [21:24] so is this after reboot and it works now? [21:24] nope :) its just a post with my 2nd find5 i got from mhall119 today :) [21:25] oh, cool :) [21:26] anyway, I feel like firing up euro truck sim 2 now.. so.. bye bye :) [21:26] anders3408: looking good :) [21:26] good luck with fixing the mount.. took me more than a month to fix mounts because I was using an ext3 partition instead of ext4 without noticing it :P === salem_ is now known as _salem [21:34] mfisch, any luck? [21:35] cwayne_: not yet, still on call [21:40] cwayne_: will try in 10 [21:41] mfisch, k [21:47] cwayne_: should this just work? [21:50] mfisch, in theory but it doesn't, it's not setting the dconf key [21:50] and i cant figure out why [21:52] Not sure if this is the right place, but I want to know where to grab the Ubuntu App Showdown winners to run on my PC. [21:52] fighting with msising scopes [21:52] why on your pc? [21:53] cwayne_: yes, still broken [21:53] mfisch, i just told you it would be [21:53] i cant figure out why the dconf key isn't updating [21:53] even if i add a lock [21:55] veebers: did my feedback on your MP make sense? [21:55] cwayne_: let me look [21:56] i wonder if i typo'd somewhere or something [21:56] doanac: hmm, let me check my logs as I seem to recall issues with the indicators_client tests too [21:56] doanac: how are you running that tests? [21:56] veebers: i've got it down to just running: phablet-test-run -n -v unity8.indicators_client.tests.test_battery.TestDisplayMenus.test_auto_bright_switch [21:57] doanac: right, that's the failing test I have in my log too for both my branch and saucy version of autopilot [21:57] doanac: I think it might be an issue with the test itself (or something related) let me investigate further [21:58] mhr3: Hey [21:58] cwayne_: mediascanner may be changing it? [21:58] veebers: hmm. its working in daily image testing, and I think i confirmed it worked earlier [21:58] mhr3: do you know stuff about icon search path for application scope? [21:58] mfisch, but i even added a lock on mine [21:59] mhr3: any idea why https://code.launchpad.net/~dobey/click-update-manager/install-icon/+merge/188447 wouldn't work? [21:59] doanac: hmm, maybe the difference is the system running (i.e. I used: phablet-flash ubuntu-system --channel devel-proposed --no-backup) for my run [21:59] doanac: would that be different to what your running? [21:59] veebers: hmm. we use --bootstrap [22:01] lool, this one should work: https://code.launchpad.net/~mfisch/click-update-manager/fix-icon/+merge/188496 [22:02] lool: it does work but there were some unresolved arguments [22:04] doanac: odd. Well his is the failure I see (using either autopilot): http://pastebin.ubuntu.com/6185809/ [22:04] mfisch, cwayne_: Right, was trying to find out from unity folks why pixmaps wasn't used [22:04] just to get to the bottom of the argument [22:04] which, to me, wouldn't indicate a regression in that autopilot branch [22:04] veebers: you probably need to install the unity8-autopilot package [22:05] mfisch, cwayne_: Actually I see it's just an abstain [22:05] mfisch, weird, dconf read /com/canonical/mediascanner/metadata-sources returns nothing [22:05] Hello. [22:05] I have a question about the Ubuntu Touch. [22:05] ralsina: Hey [22:06] Will it be available to American users with Sprint PCS soon? [22:06] lool: hi [22:06] ralsina: Please would you top approve https://code.launchpad.net/~mfisch/click-update-manager/fix-icon/+merge/188496 again (see previous upstream merger error); also would you please grant me right to top approve other branches in future lp:click-update-manager? [22:06] lool: that has a needsfixing from CI? [22:07] oh, right, it's the autolanding failure [22:07] hello [22:07] veebers: its odd to me it works when you run the whole suite. its only a problem when you try to run that testcase [22:07] lool: I'll add you to ubuntuone-hackers I think that should do it [22:07] Hello? [22:08] doanac, veebers --bootstrap and --no-backup are aliases [22:08] Netscape128: is sprint pcs cdma? [22:08] for 3g, yes [22:08] doanac: that is odd :-\ [22:08] Netscape128: then you should be able to use it on one of the supported phones [22:08] sergiusens: thanks for the clarification [22:09] lool: added you to ubuntu one client engineering which should indirectly let you top-approve [22:10] thanks [22:10] so if i switch to android and want ubuntu touch i can just download it? [22:12] oh, it's not out yet... [22:13] mfisch: because I don't have Ubuntu installed on my phone at the moment and probably won't for another week, but I want to test the Reddit client anyway. [22:13] mfisch, this is very frustrating [22:13] or is it? [22:13] it's not out yet for the average user? [22:14] Netscape128: it's still a developer preview for now [22:14] Netscape128: 1.0 is this month [22:14] oh darn [22:14] oh wll [22:14] Netscape128: what device do you have? [22:14] well* [22:14] October 17th, in fact. [22:14] an iphone at the moment but i'm gonna switch to complete linux [22:15] Netscape128: keep this page in mind. https://wiki.ubuntu.com/Touch/Devices [22:15] meaning, some distro on my computer, some distro on my phone, etc. [22:16] cwayne_: I had to re-flash then I'll catch up and help [22:16] mfisch, k [22:17] I'll certainly keep that in mind. [22:17] Wait, it will work on the first generation Kindle Fire? [22:17] Cool! I have one, but the pins are bent on it. [22:18] oh my god the keyboard is so much better now [22:18] Might be worth playing with, but it looks like it's not well supported: https://wiki.ubuntu.com/Touch/Devices/otter [22:21] lool, i do remember that /usr/share/pixmaps is not searched by qt [22:21] mhr3: Ok; there is actually a solution in the pipe that I thought didn't make consensus, but it was good enough after all [22:25] lool, fwiw icons are usually installed in /usr/share/icons/hicolor/ that branch makes the icon theme-specific [22:26] sergiusens, happroval? https://code.launchpad.net/~cwayne18/phablet-tools/phablet-config-timezone-fix/+merge/188727 [22:26] mhr3: right, the actual branch is https://code.launchpad.net/~mfisch/click-update-manager/fix-icon/+merge/188496 [22:26] mfisch, i don't get how i got this to work last time [22:27] lool, yea, that's the one i meant by "that branch" [22:27] cwayne_, in my queue [22:27] sergiusens, ta [22:39] mhr3: so we should be installing to hicolor/apps/144 instead? [22:40] mfisch: Would you be tempted to fix the path as suggested by mhr3 above and to fix the dh_install failure (need to update .install I guess) [22:40] dh_install: usr/share/icons/ubuntu-mobile/apps/144/update_manager_icon.png exists in debian/tmp but is not installed to anywhere [22:41] cwayne_, in line 56 there's an extra space after false [22:46] lool: hmm, I built that package and tested it [22:47] lool: let me redo it [22:47] lool: can qml packages be arch all instead of any? [22:48] mfisch: yes [22:48] mfisch: see e.g. music-app [22:48] okay I'd like to fix that too, it makes it super annoying to build [22:49] mhr3: okay the icons need to go into hicolor? [22:50] they were pretty set on not adding a .install file and fixing this in CMakeLists [22:51] mfisch, should be good, yea, but pls check on the phone, i'm not sure it's common there as well [22:52] Suppose my QML app has a dependency not in ubuntu-sdk-13.10. (whatever dependency, really) How would I go about including in in the package? Can I do that? [22:53] mhr3: almost every app has icons in ubuntu-mobile [22:53] mfisch, yea, cause that's the theme phone is using currently [22:54] and which size of the multitude of sizes in hicolor shall I choose? [22:56] mfisch, ideally a scalable and a bunch of rendered pngs with specific size [22:57] mfisch, but if you have just 144, just put it in ubuntu-mobile as everyone else, it's not time to start doing things properly few days before release :P [22:59] okay well I can do the 3 that gallery does [22:59] 64, 128, and 256 [22:59] but I can't do it right this second, it will be in a few hours [22:59] mhr3: will anyone be awake who can top approve it then? [23:00] europeans won't :) [23:00] yeah I know [23:08] Would you guys say ubuntu-html5-theme is a framework? It's a bit more than styling, but wouldn't be a far fetched selection of words? [23:08] alex-abreu: ↑ You might know :) [23:09] sergiusens, fixed [23:09] cwayne_, happroving [23:09] sergiusens, thank you sir, and as always, sorry for pestering :) [23:10] cwayne_: can you test this package of icons? [23:11] cwayne_: its in ~mfisch/tmp on people.cc [23:11] lool: I've got all the icons packaged up [23:13] mfisch: cool [23:13] I'm off to bed though [23:14] lool: okay, I'll do one more test and then push it [23:16] lool: why do I need to specify files in .install when cmakelists adds them for me? [23:17] // We disable the geolocation API for now, as it's still under development [23:17] // for Ubuntu Touch [23:17] Does that still hold? [23:17] This is from the QR code scanner example. [23:17] mfisch, testing [23:19] mfisch, +1, works [23:19] cwayne_: thanks [23:20] achiang: Sent you an e-mail. I think that'll be it for today. Do you think you could go through it and possibly reply before EOD? [23:20] achiang: I can start making stuff pretty tomorrow. [23:21] Stealing pictures from our design sites, etc. [23:25] Hey everyone [23:25] I have a question about software compatibility with Ubuntu Touch on a Nexus device [23:26] Is the software I have available on my x86 desktop PC also available on a Nexus device with Ubuntu touch? [23:28] Depends on the layer. Ubuntu Touch doesn't run X, so desktop-level software typically doesn't run unported. [23:28] Things that are command-line-only are generally also available for the armhf architecture and would work, although you'd need to remount read-write (and sacrifice system-image updates) to install them. [23:29] Okay, because I wonder if I could just use it as a notebook/tablet combo with a docking station [23:29] Would Java desktop applications work? JetBrains IDEA to be more specific [23:30] No. [23:30] We do plan to make that kind of converged use case work, but it's in the nearish-future plans rather than something that will work today or in 13.10. [23:37] Is it possible to install X and have Ubuntu Touch run Gnome 3? [23:38] It wouldn't be Ubuntu Touch. Er, you could try to do something like the old Nexus images before the Ubuntu Touch project really got started. It would take you quite a long time. [23:38] (That was just Nexus 7) [23:39] Yea... it seems like Ubuntu Touch is not really what I want. It [23:39] It's possible Xmir would be feasible once the touch images switch to Mir [23:39] I'm not an expert on that [23:39] *It's cool to have a command line and now artifical boundaries [23:40] but actually I just want a Linux Desktop on a tablet [23:40] That kind of thing is in the 14.04/14.10 set of goals [23:40] (I forget exactly where it is right now) [23:41] oh, really? That sounds interesting [23:41] Under the general heading of "convergence" [23:42] cjwatson: XMir-on-phone requires a driver we haven't (yet) written. [23:42] ^- expert [23:43] Also, generally no OpenGL support on phone or tablet hardware, so desktop GL apps won't work, and I don't expect we'll try the huge task of implementing desktop GL on GLES :) [23:46] Okay, I never used any GL software on my current notebook I know of right now [23:56] Okay, well thank you a lot cjwatson and RAOF for your time you were a big help!