=== chriadam|away is now known as chriadam [01:35] hi guys [01:36] does anyone know the status of ubuntu touch in regard of developed apps? [01:36] music and vieeo player [01:36] mail client [01:36] there is a music player yes [01:36] openpgp support and stuff? [01:36] and you can access gmail [01:37] is there an email client with pgp support? [01:37] something like k-9 and APG in android [01:38] not that I know of === _salem is now known as salem_ [01:51] cypherman, I know there is a gmail client [01:51] or web app I should say. [01:53] native email client might be ready on next on april release. === salem_ is now known as _salem === Namidairo`bnc is now known as Namidairo [03:22] Where can I get the images for dlx? === Ursinha is now known as Ursinha-afk [05:49] Hi all [05:49] How can I install ubuntu on Phone === Ursinha-afk is now known as Ursinha [05:54] cwayne: ping [06:08] l624: https://wiki.ubuntu.com/TouchInstallProcess === Larry is now known as Guest50794 [06:33] I am curious if somebody has succesfully ported touch tablet to the new 2013 nexus 7? [06:33] No. [06:35] I'm not certain anyone's actively working on it, either. [06:35] If I would follow the porting guide would I be able to post it to ubuntu? I mean bugs, proformance, logs etc [06:35] maybe even the image [06:35] Probably not, because if it were that simple, it'd be done by now. There are driver issues. [06:35] But yes, that's the correct first step. [06:35] true that! [06:36] Right now *everything's* focusing on phone support, because there are 13 days until that's released. [06:36] After that, there are 6 months to tablet and desktop support. [06:37] so ubuntu touch is not beeing released on the 17th? [06:37] for tablets I mean [06:37] I just said it was. [06:37] No, not for tablets. It has never been planned for tablets until 14.04 LTS. [06:38] Desktop support is planned for 14.10. [06:38] ow ok then I will definetly start porting this weekend. Can't wait to see the result! [06:41] mfisch: ping === chihchun_afk is now known as chihchun [07:03] bit of a noob question but im looking through the core apps on launchpad.net and i cant seem to find system settings anywhere. Where is it? [07:11] GavMickey: ubuntu-system-settings [07:11] https://launchpad.net/ubuntu-system-settings [07:12] thanks just found it [07:12] ogra_, do you know if non-nexus images can use the delta upgrades at this point? [07:14] they cant [07:15] o/ all [07:15] guess I'm a bit early :) [07:16] Hi! Question: How do I restart unity8 without restarting the phone (just running unity8 doesn't seem to work...)? [07:17] pkill unity8 [07:17] it restarts itself [07:18] rickspencer3: I guess. I only just started flashing 80 [07:18] heya popey [07:18] read pauls mail in bed this morning and it inspired me to leap out of bed and flash [07:18] I'm waiting for my next plane at CDG [07:18] my phone [07:18] figure it's early there :) [07:18] popey, does stop unity8 no longer work?IIRC it used to be possible to stop/start it? [07:18] popey: ok, what about if unity8 has crashed? How do I restart it then? [07:19] janimo`: that works? [07:19] popey, no [07:19] hah [07:19] popey, hmmm, looks like Mir by default quite soon? [07:19] popey, I was just looking into how to restart unity8 when I saw this conversation [07:19] rickspencer3: exciting! [07:19] FINALLY [07:19] popey, I thought unity was a service [07:19] i usually just kill unity8, but if there's a better way? [07:19] popey, I want us to have an image "we are happy to ship" by the 10th, so, less than a week :) [07:19] (but I think we will do it) [07:20] i just opened the terminal on my phone and said "restart unity8" and it did [07:20] rickspencer3: yeah, I think so, I think we'll have a big queue for 0-day updates though [07:20] "Big" might be overstating it [07:20] "medium sized" [07:21] popey, as root or as phablet user? Neither works for me [07:21] as phablet [07:21] popey, well, I figure we can keep making new images after the 10th, and if we are happier with one of those, we'll ship it [07:22] but we should have on "in the can" by the 10th [07:22] rickspencer3: I'm also super keen for us to ship core app updates in the store post 10th [07:22] popey, ok, I switched using su - phablet and that made restart unavailable in the path. It works now, thanks [07:22] at least, that's my goal [07:22] popey, also, updates are weird with the image based updates [07:22] oh? [07:22] well, weird in a good way [07:23] I *love* the new update system [07:23] it's like, we have a solid image, and we know that everyone will get that exact image on the exact hardware [07:23] so, it's a lot less hair-raising [07:23] +1 [07:23] BRB Breakfast calls [07:23] once we pair that with phased updates, we'll have a really solid system [07:23] later popey! [07:45] is it possible to flash an old image, like 4,5 or 6 month old? [07:47] tmoenicke: I think you should be able to flash the zip files by adb pushing them and then select them in the recovery menu. Given you can find them somewhere on the servers still [07:47] anyone know if we have an archive of old images? [07:47] i do mzanetti [07:48] http://popey.mooo.com/mirror/ mzanetti tmoenicke [07:48] oh , cool [07:48] http://popey.mooo.com/mirror/cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/ goes back to june === chriadam is now known as chriadam|away [07:51] cool [07:52] popey: mzanetti: thx, i will try it out [07:57] does anyone know the id for the gsm interface on a nexus 4 running the latest build? in ifconfig it shows as a usb interface? [08:00] rmnet_usb0 [08:01] thanks (again) [08:03] xnox, hmm could it be that your dbus change moves the log somewhere else ? seems the tests dont find it anymore [08:10] hi there [08:12] * ogra_ flashes #80 [08:15] hey lool [08:18] xnox: heya [08:18] xnox: so while the dbus fix seemed to be correct, we need to fix it the other way as we're losing important logs now [08:18] Bored.. thought I would see what Mir would do... "terminate called after throwing an instance of 'boost::exception_detail::clone_impl >'what(): display factory cannot create fb display" [08:18] xnox: that is, dbus.log is now empty, and that's where we get a lot of debug information === cul- is now known as cul [08:19] xnox: could we fix it the otherway around? patch dbus to have a flag to NOT fork and revert back to the not forking behavior? [08:19] xnox: or patch it to still output stuff [08:21] \o/ Time setting sticks across reboots with #80 [08:22] popey: :-) [08:22] popey: are you Mir or SF? [08:22] boot time has been degradating lately [08:26] lool: sf [08:35] ogra_: log is in " .cache/upstart/dbus.log" that's where it's meant to be when started under upstart. any other locations means, you were not using dbus from upstart but auto-launched one. [08:35] lool: ^ [08:35] xnox, not here with a freshly flashed image [08:35] root@ubuntu-phablet:/# ls /home/phablet/.cache/upstart/dbus* [08:35] /home/phablet/.cache/upstart/dbus-session === sil2100_ is now known as sil2100 [08:36] this is after the third reboot [08:36] ogra_: wait i'm on my laptop, which wasn't upgraded over night. [08:36] i dont think upstart removes logs :) [08:36] ogra_: ... and you have other logs in ~/.cache/upstart/*.log ?! right. [08:36] so watch the timestamp or move it away [08:36] ogra_: well there is logrotate job for those ;-) [08:36] ah, k [08:36] yeah, its full of the normal session logs [08:37] xnox, lool: Adding "initctl set-env DBUS_DEBUG_OUTPUT=1" to the dbus.conf pre-start seems to allow dbus to produce the usual splurge. [08:38] jodh, yay, thanks ! [08:38] jodh: why set-env, instead of just normal "env DBUS_DEBUG_OUTPUT=1" at the top? [08:38] so all jobs get it ? [08:38] xnox: yup - even better! [08:39] Hi everyone, I have a basic question if that's OK? [08:39] ogra_: they'd all get it if you used 'set-env --global ...'. [08:39] ogra_: " echo env DBUS_DEBUG_OUTPUT=1" > ~/.config/upstart/dbus.override [08:39] ogra_: reboot and see if that works for you =) [08:40] xnox, will do, i just flashed and want to do some other tests before making it writable [08:40] ogra_: above doesn't need to have it writtable. [08:40] ogra_: or does it?! [08:40] oh [08:40] indeed [08:51] Good morning all; happy Friday, and happy Smile Day! :-D [08:52] lool, ping [08:54] xnox, works ! [08:55] ogra_: excellent. [08:55] Saviq: pong [08:55] jodh: i had no idea about the dbug environment variable =) [08:55] ogra_: so I guess another dbus upload.... [08:55] yeah [08:55] lool, ^^^^ i guess thats fine ? [08:56] lool, hey, so can we help with anything? kgunn asked me to give you anything you need to get AP going on Mir [08:56] (to get logging back) [08:56] lool: ogra_: or, do we really want that set on production systems, or should the override be done in testing? [08:56] Saviq: oh yes, help definitely welcome [08:56] Saviq: can we hangout? [08:56] Saviq: finishing an email, will be with you in 4mn [08:56] lool, sure, let me know [08:56] xnox, well, i think until we are sure the quality is good enough i would say we still need the logs [08:56] xnox: it's amazing what you can find reading the source :) [08:57] * ogra_ is afk for a moment === iahmad is now known as iahmad|afk [08:58] jodh: yeah. [08:58] ogra_: ok. [09:04] My messaging indicator turns blue when I receive an SMS, but pulling down the menu still shows "Empty!" Is this known? [09:19] hello [09:19] how do you run unity8? [09:19] Unity8 is the ui you see on ubuntu touch [09:20] I don't see it. That's wy I ask. [09:20] JamesTait: which image? [09:20] What do you see when you boot? [09:20] JamesTait: on 79 mine is blue and on 80 mine isn't blue, but both show messages [09:20] the login prompt [09:22] What device are you running it on and which build? [09:22] # unity8 [09:22] Segmentation fault (core dumped) [09:23] I am running this on cubieboard. Just did apt-get install ubuntu-touch [09:24] well, it's technically not installable but the parts apt complains about are not really important [09:24] popey, build 80. [09:24] calendar-app and likes === iahmad|afk is now known as iahmad [09:25] Try sshing into it and restarting unity8? [09:25] how do you restart it? [09:25] I am sshd in it all the time. it's not helping [09:25] Not sure what else to suggest it's not supported [09:26] To restart unity8 run pkill unity8 [09:26] it's not running [09:26] JamesTait: there are a couple of known issues with the messaging icon's state. I'll fix those today. [09:26] it crashes whenever you run it [09:26] and it's not even run because lightdm fails to start [09:28] Ah that's your problem then. Try disabling lightdm, then ssh into it after a reboot and manually executing unity8 [09:28] it crashes [09:29] and no need to disable lightdm. it does not work anyway. hence the login prompt [09:30] Sorry don't know then [09:30] ok, I guess it's normal that it does not work then [09:30] will look at something else then [09:30] thanks [09:31] apw, strace -o /tmp/strace.out ureadahead --force-trace [09:31] ... [09:31] munmap(0x40149000, 4096) = 0 [09:31] openat(3, "events/fs/do_sys_open/enable", O_RDWR|O_LARGEFILE) = -1 ENOENT (No such file or directory) [09:31] close(3) = 0 [09:31] write(2, "ureadahead: Error while tracing:"..., 59) = 59 [09:32] apw, did we lose the ureadahead patch from the mako kernel ? [09:33] is there anyway to make adb understand that it should not limit itself to 80x25 ? [09:33] sadly not [09:33] ok [09:33] use ssh instead ;-) [09:33] (well, there surely is with a lot of hacking) === b0bben_ is now known as b0bben [09:39] larsu, excellent, thanks. :) Is there a bug I should subscribe to? [09:40] Hi everyone, I've just got a basic question - is it OK to ask it in this channel? [09:40] apw, oh, seems we only added the patch to https://launchpad.net/ubuntu/saucy/+source/linux-maguro/3.0.0-3.9 ... would you mind adding it to mako too (that actually eplainx why maguro boots faster) [09:42] Hey anyone. How would need to get some content like music and videos on my Nexus 7. Few days back when I flashed it think there was some but now with the latest build Music and Videos scopes are just empty. I'm using following flash command: phablet-flash ubuntu-system -d grouper [09:42] KathyReid: on irc, usually one simply asks the quesion one has straight away.... =) so yes, always simply ask / say what you want to say. [09:42] xnox: thanks. [09:43] JamesTait: I haven't triaged them yet (mostly it's been people pinging me about it). There are at least #1234680 and #1220487 [09:43] bug #1234680 bug #1220487 [09:43] bug 1234680 in indicator-messages (Ubuntu) "indicator-messages blue even though there are no messages" [Undecided,New] https://launchpad.net/bugs/1234680 [09:43] I've successfully flashed my LG Nexus 4 with Ubuntu Touch (mako) and it's working pretty well. What do I need to do to help test things / help out further? [09:43] bug 1220487 in indicator-messages (Ubuntu) "Messages indicator does not indicate new messages" [Undecided,Confirmed] https://launchpad.net/bugs/1220487 [09:44] Is there some way to make mako wake me up? I cant find the alarm clock... === vrruiz_ is now known as rvr [09:46] I suspect it would be one of the apps that does not exist [09:47] Go to the clock app and swipe along the top where it says clock [09:48] KathyReid, welcome! It really depends on what you want to do and how knowledgeable you are - there are plenty of opportunities like, continue testing images and look for regressions, or if you find something that's not working, submit a bug report, help fixing it if you know how, etc [09:51] vesar, you can use the MTP protocol to transfer your own music and video to the device. [09:52] vesar, I think the [09:52] vesar, I don't know why the sample movies/music was removed but it was, so you have to put your own content there. [09:53] ok, unity8 still crashes on start with today's updates so that's it I guess [09:53] diwic: thx for the welcome. I'm comfortable in ssh / bash, know a little bit of C, Java etc. Is there a 'here's how to start helping' guide anywhere? [09:55] ogra_, will look into it [09:55] apw, thanks ! [09:55] KathyReid, if you like coding, maybe you would be interested in helping out with one of the core apps: https://wiki.ubuntu.com/Touch/CoreApps [09:55] diwic, thanks. I'll transfer some of my own content there then. Just wondered because I remember seeing some default offline content there in the past. [09:57] Has anyone started working on getting tethering/hotspot working? [10:01] Morning all [10:01] ogra_: you nice refreshed after your day off ;) [10:01] GavMickey, after 13.10 is out [10:01] davmor2, totally .... and completely united now [10:01] diwic: thanks for the tip, the burndown charts are really useful [10:02] Ok thanks [10:03] KathyReid, I encourage you to pick what is in your interest and what you think is fun to do - if so, you're more likely to enjoy it here. :-) [10:04] diwic: thanks again, will do, and thanks too for the welcome - it is very appreciated [10:06] KathyReid, thanks for wanting to help out [10:10] Hello! I own a Nexus 4. Do current Ubuntu Touch builds already provide a desktop environment when the phone is connecterd to an external screen via hdmi cable? [10:10] no [10:10] thats 14.04 or 14.10 material [10:12] Quick question, will running sudo apt-get upgrade from ubuntu touch upgrade to the latest nightly build? === fmasi_afk is now known as fmasi [10:33] ogra_: who do I talk to about mir at this untimely hour for the US [10:33] gema, Saviq perhaps [10:33] Saviq: are you around? [10:34] gema, or #ubuntu-mir :) [10:34] Saviq: ack, will try there [10:36] Adios all will be back soon [10:51] guys, since todays image "reboot" as root of the phone doesn't reboot it [10:51] it seems it gets stuck trying to shutdown [10:51] any idea? [10:52] tsdgeos, i noticed that myself "adb reboot" works as expected [10:52] not sure what changed that made upstart reboots not work anymore [10:53] ogra_: shall i report a bug? against what? [10:55] tsdgeos, ah so you think it's not shutting down even? [10:55] Saviq: not sure [10:56] it is shutting down, you can bnoot it immediately again [10:56] *boot [11:01] stgraber, hmm, i'm trying to get ureadahead to work ... sadly it expects to mount the debugfs under /var/lib/ureadahead/debugfs ... and demands /var/lib/ureadahead to be writable ... adding the latter to writable-paths gets me a writable dir, but the mount in the subdir is gone === fmasi is now known as fmasi_afk === fmasi_afk is now known as fmasi [11:24] mhr3_, you guys were looking for a crash file ? http://people.canonical.com/~ogra/ubuntu-touch/_usr_lib_arm-linux-gnueabihf_hud_hud-service.32011.crash [11:24] ogra_, can you get the stacktrace from it pls? [11:25] i dont think there are any dbg packages installed ... [11:25] install them :) [11:25] but i see the dbus-hang a lot with maguro today [11:25] (image is still readonly only Mir enabled) [11:26] i havent see a single hang with SF [11:26] *seen [11:27] ogra_, mhr3_: let me get the stacktrace, I've a rw image with ddebs enabled [11:27] cool. thanks [11:27] seems to really be realted to Mir [11:27] *related [11:27] and the issue is just happenning on my desktop actually [11:27] SF with working ureadahead boots in ~25sec now [11:27] so can get live data from it [11:28] mhr3_, the hud one? [11:28] Mir takes around a minute [11:28] seb128, yes [11:28] hmm that indeed trashes my Mir theory :) [11:28] seb128, xnox, and DBUS_SESSION_BUS_ADDRESS is not set in its environ [11:28] mhr3_, ok, great [11:29] and paste.ubuntu.com isn't working :/ [11:31] xnox, env of the hud process http://paste.debian.net/50230/ [11:32] ogra_, do you know if the impact of a package on the image disk footprint is ~ the size of the deb? === MacSlow is now known as MacSlow|lunch [11:32] i think we're running multiple dbuses [11:32] ogra_, e.g is it compressed and at the same level? [11:32] mhr3_: what's the output of $ initctl list-env ? [11:32] seb128, the size of the installed deb minus /usr/share/doc/* [11:32] ogra_, thanks, looking at langpacks, Pat said it's ok to get a few on the images, I'm getting the numbers [11:32] xnox, http://paste.ubuntu.com/6191827/ [11:33] seb128, it uses xz compression for the tarball, not sure if it complesses at the same level though [11:33] ogra_, we were shooting for chinese, spanish, german, french [11:33] mhr3_: right, dbus is set. [11:33] seb128, sounds good [11:33] mhr3_: status hud; status dbus; ? [11:33] mhr3_: and does the hud's pid from status, match the pid of hud? [11:33] xnox, no, hud exitted meanwhile, it does that after a while [11:34] but if something starts it again, the env still isn't set [11:34] mhr3_, is it started through upstart, or dbus activated? [11:35] seb128, it's dbus-activated which does "start hud" [11:35] so... let's call it upstart-over-dbus activation? :P [11:35] i mean [11:35] mhr3_: which is wrong, cause it will happen before dbus is started, since you are forcing it to start regardless of what other things are happening or dbus starting. [11:35] broken-upstart-over-dbus activation [11:36] xnox, hm? [11:36] dbus is running my whole session, hud is stopped, if i let dbus do "start hud" the env there isn't set [11:37] i *think* if i do it myself it will be set [11:38] don't want to otherwise i lose that broken state [11:38] also why is ted using bash to exec things /o\ [11:38] and sleeps =/ [11:39] let's not go into that [11:39] the dbus activation script is a red herring, though (no matter whether it's good or bad) [11:39] as mhr3_ says, manually starting and stopping the service reproduces the problem [11:39] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-hud-dbus-hang.png [11:40] i dont get that with SF on maguro [11:40] http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-SF.png [11:41] pete-woods: well, something talks to hud -> dbus-activation-hack.sh bash script is called -> if upstart_session variable is set -> start hud; sleep 10..... else fallback to executing hud-service [11:41] (the latter one actually has the UI on screen under 25sec [11:41] ) [11:41] pete-woods: and the upstart job is to simply start hud. [11:42] xnox, how many dbus-daemon are we supposed to have? [11:42] cause i have 5 of them [11:42] mhr3_: that means, that crap is autoactivating additional onces instead of using the system on. [11:43] mhr3_: we should have at most 2 session dbus (one normal one, one for atk/a11y) [11:43] 690 ? Ss 0:03 dbus-daemon --system --fork [11:43] 1284 ? Ss 0:00 \_ dbus-daemon --fork --session --address=unix:abstract=/tmp/dbus-ic2MIWcWW9 [11:43] 1259 ? Ss 0:00 //bin/dbus-daemon --fork --print-pid 4 --print-address 6 --session [11:43] xnox, which would totally explain this, hud connects to a different dbus [11:43] 102 1176 0.0 0.0 32360 3028 ? Ss Sep30 0:47 dbus-daemon --system --fork [11:43] miso 3125 0.0 0.0 33920 4416 ? Ss Sep30 1:06 dbus-daemon --session --address=unix:abstract=/tmp/dbus-h90A8DWTwU [11:43] miso 3160 0.0 0.0 30748 1916 ? S Sep30 0:01 /bin/dbus-daemon --config-file=/etc/at-spi2/accessibility.conf --nofork --print-address 3 [11:43] root 8077 0.0 0.0 30376 1352 ? Ss Oct01 0:00 //bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session [11:43] miso 32673 0.0 0.0 30508 1264 ? Ss 11:59 0:00 //bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session [11:43] mhr3_: let me explore what's going on, on my grouper after I fix the next one. [11:43] i wonder what the latter one is [11:44] mhr3_: please use pastebin [11:44] mhr3_: upgrade, you are running out of date dbus package?! [11:44] xnox, yes :) [11:44] * ogra_ guesses all these // processes shouldnt be there [11:44] mhr3_: what's the output of: dpkg -l dbus [11:44] xnox, i upgraded already, but running this session for a few days [11:44] mhr3_: well, if you don't apply yesterday's fixes, i'm not sure how else i can help you =) [11:45] mhr3_, you mean you relpaced dbus underneath a running session but didnt restart it ? [11:45] mhr3_: there is no way to restart dbus. [11:45] the session [11:45] xnox, you think your dbus fix is going to fix the multi-dbuses i'm having? [11:46] xnox, isn't the unset DBUS envvar still odd? [11:46] of course it's odd [11:46] xnox, i mean it'll be unset if upstart runs the job *now* [11:46] mhr3_: so unity & hud were starting ahead of dbus-daemon completing initialisation, thus clobbering and auto-spawning it's own. [11:47] xnox, could it be that upstart-dbus-session spawns its own daemon ? [11:47] mhr3_: sure, but not for the jobs that are already running..... upstart doesn't and I guess can't modify the environment of already running processes. [11:47] *bridge [11:48] xnox, but the hud job quits itself after a while [11:48] xnox, why doesn't upstart "update" the env on next start? [11:48] mhr3_: sure, but upon respawn it will be the environment of the dbus-daemon that launches it. and there are multiple ones, how would you know which one? [11:48] mhr3_: if you are sending dbus-send from e.g. unity, you will enherit running unity's environment which will have the wrong dbus in it. [11:49] mhr3_: and dbus-activation hack will not take the upstart path of execution at all... [11:49] mhr3_: please reboot and reproduce the problem, if you can. Such that I can reproduce it locally as well. [11:50] xnox, you know that's not gonna happen :P [11:50] (the reproducing it locally part) [11:50] the problem was that the whole session couldn't connect to the upstart's launched dbus-daemon, and instead gave up and auto-launched it's own dbus and used that everywhere..... [11:50] but at least we know we have too many dbuses [11:50] hmm [11:50] mhr3_: _you have too many dbusses ;-) [11:51] ev, does whoopsie spawn its own dbus daemon ? looking at http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-hud-dbus-hang.png ist suspiciously looks like that [11:51] ogra_: not explicitly [11:51] surely if it were simply the case of having too many dbus's then the DBUS_ADDRESS environment variable would be incorrect, not missing [11:51] ogra_: kind of the crasher, reacts to file events, which happens before dbus-daemon launched. I have merge proposal to fix that. [11:51] ogra_: but it will try to talk to NM over DBus [11:51] xnox, but upstart's DBUS == my session's DBUS env [11:51] as soon as it starts [11:52] ev, and if it cant ? [11:52] it really looks like it fires up one in the botochart [11:52] ogra_: then it will assume the system is always online [11:52] *boot [11:52] ev: ogra_: bug #1234841 [11:52] bug 1234841 in upstart (Ubuntu) "crash reports pop-up ahead of desktop loading, or atleast dbus being available" [High,Confirmed] https://launchpad.net/bugs/1234841 [11:52] aha [11:52] ogra_: I'm not denying that's happening. I'm suggesting it may be happening implicitly. [11:53] ah excellent [11:53] ogra_: file events emitted before dbus is launched, thus whatever non-interactive ui / apport thingy doesn't have dbus and spawns it's own. [11:53] yeah, just read the bug :) [11:53] thanks [11:53] * xnox goes to fix up that merge proposal. [11:53] zsombi, ping [11:54] i doubt it will help the hud though :( [11:54] mhr3_: if you want to, modify dbus-activation-hack and add "env > ~/.cache/dbus-activation-hack-hud-environ.log" at the top [11:55] seb128: is it on purpose that the diagnostic data sending cannot be disabled? [11:55] gema, talk to ev [11:55] I don't know [11:55] ev: is that on purpose? [11:55] seb128: it's on the system settings ui [11:55] I think Laney and ev said that whoopsie needs a plka file [11:56] ev is sorting it [11:56] gema, right, but ev wrote that code and probably knows better the status [11:56] seb128: ahh, gotcha [11:56] it'll be added to pk-d-p [11:56] ev: do you have a bug number for that? [11:56] xnox, http://paste.ubuntu.com/6191903/ [11:56] ofono telephony really needs to be disabled for non 3g/4g tablets [11:56] how can we get a message to the ubuntu-touch devs [11:57] unless there will be a build for phones and a build for tablets [11:57] ogra_, doing "reboot" in an adb shell stopped working it seems, is that a known issue? it seems the tablet power down but doesn't reboot... [11:58] xnox, inside the hud process there's still no DBUS envvar set [11:58] DJJeff, it wont start on modem-less devices [11:58] gema: no, I haven't been using a bug to track it [11:58] seb128, i noticed it too, adb reboot works fine though [11:58] endeavouring to have it in today [11:58] ogra_, oh, I didn't know about that command, danke [11:58] ev: ack, if you could let me know the ETA image, I'll verify [11:58] whenevr you know it === hikiko is now known as hikiko|lunch [11:59] will do, thanks [11:59] ogra_: a couple builds back it caused network-manager 100% cpu [12:00] ogra_: and is also noisy in /var/log/syslog [12:00] DJJeff, well, thats hopefully fixed, if not it should be :) [12:00] did you file a bug ? [12:00] not yet if I see it again I will [12:01] dont wait to long, release is close :) [12:02] ogra_: here is my /var/log/syslog || http://paste.ubuntu.com/6191942/ [12:03] yeah, file a bug [12:03] either cyphermox or awe should take a look at that [12:04] syslog fills up pretty fast with that crap [12:04] so it goes on and on ? [12:04] correct [12:04] (your paste only shows 1min) [12:05] * ogra_ goes afk for a bit [12:05] that paste is from # tail -f /var/log/syslog [12:07] not sure how to file that sort of bug and where :( [12:07] ubuntu-bug network-manager [12:07] via adb [12:07] Laney, seb128: would one of you mind doing a bit of quick code review for me? http://paste.ubuntu.com/6191951/ [12:08] ev: looks correct [12:09] uploading === alan_g is now known as alan_g|lunch [12:13] Saviq: we found a problem that if an app is running on the foreground, a call cannot be answered [12:13] Saviq: bug 1235215 [12:13] bug 1235215 in unity8 (Ubuntu) "cannot answer call when system settings are in the foreground" [High,New] https://launchpad.net/bugs/1235215 [12:15] gema, ok thank [12:15] Saviq: with mir [12:15] MacSlow|lunch, hey, can you have a look ↑ [12:15] MacSlow|lunch, maybe we don't have an InputFilterArea for the notification? [12:16] MacSlow|lunch, it'd explain the other bug where your interaction with a notification pass through to the app [12:16] cjwatson: any objection to me using the proposed-migration hinting to let the above pkla change (policykit-desktop-privileges) through? [12:19] ev: it's not touch-specific, so needs release team signoff, but I'm happy to give that yes [12:19] * ev nods [12:19] it's also not blocked [12:21] Hm? The archive is frozen [12:21] mzanetti, can you try https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1235222 btw? [12:21] Ubuntu bug 1235222 in Ubuntu UI Toolkit "Autopilot suite gets killed after running tests on Mir" [Undecided,New] [12:21] ev: unapproved queue is what stops upload [12:21] ev: not hints [12:21] mzanetti, on SF, that is? [12:21] this was decided last week [12:22] (release team preferred really preventing uploads reaching the archive, unapproved is earlier) [12:22] ah [12:23] jodh, mid to upload https://code.launchpad.net/~jamesodhunt/upstart/bug-1234898 so we can get it in one of todays images ? [12:23] *mind [12:25] ogra_: I've asked for feedback on the bug so if someone aside from me will test it, I'll get it merged upstream then cherry-pick into ubuntu. [12:25] jodh, see -release :) [12:26] xnox: would we also need to make whoopsie start on started dbus, since it will also autolaunch a dbus daemon trying to talk to NM? [12:27] lool, I have invited Saviq to the standup ( Saviq fyi;) [12:27] olli, k [12:27] lool, do we want somebody else in the standup? [12:29] cool [12:29] olli: jfunk? [12:29] oh shit that's in 30mn already [12:30] Saviq: on it [12:32] xnox: actually, I'm not sure start on started dbus in upstart-file-bridge is the right location if the motivation is apport. Apport by itself doesn't talk to dbus - whoopsie does that - and whoopsie is already started and talking to dbus long before any file bridge events [12:32] lool, no, in 90min [12:32] acc to my cal [12:32] oh right, that's my rick 1-1 [12:32] good [12:32] jfunk will be there for the regular standup (at least he used to so far) [12:32] de-panic! [12:32] ;) [12:33] german/panic ? [12:33] unity8 high cpu usage screenshot of htop || http://i.imgur.com/AEkbCc7.png [12:36] Saviq: posted comment [12:36] mzanetti, thanks [12:37] mzanetti, know what's interesting? [12:37] mzanetti, http://reports.qa.ubuntu.com/smokeng/saucy/touch_ro/4541/ [12:37] mzanetti, http://reports.qa.ubuntu.com/smokeng/saucy/touch_ro/4542/ [12:37] mzanetti, only 48 tests there [12:39] Saviq: there are some tests with multiple configurations. not sure if smokeng doesn't run them or if they just don't appear in that log [12:40] Saviq: it's 10 hidden tests, which would make 46, but there are sysetmsettle-before and systemsettle-after on smokeneg. so numbers match again [12:40] mzanetti, ah [12:40] mzanetti, 10 hidden? what's that mean? [12:40] http://paste.ubuntu.com/6192056 [12:41] Saviq: ^ [12:41] ev: session init, not system init. so on the desktop, I get the crash pop-up appear ahead of unity launcher, indicators and gtk initialised. First thing that any gtk dialog does, is start talking to session-dbus. Which if started ahead of dbus, autospawns additional ones. [12:41] ev: so that fix is primarily desktop driven, not touch. [12:41] ah right [12:41] sorry, I appear to be of a single mind today :) [12:41] ev: and file-bridge in the session-init, is actually the only one that _didn't_ start on started dbus already =) [12:42] what were your thoughts on changing whoopsie's job to start on started dbus? [12:42] it does mean that if dbus falls over in a heap, we wont get error reports [12:42] if dbus falls over you might not even get networking (on NM systems) [12:43] Saviq: did terminal-app pass 100% for you on Mir? [12:43] ev: i never proposed to change whoopsie to start on started dbus. [12:43] ogra_: a valid point [12:43] xnox: I was, to solve ogra_ 's problem [12:43] ev: ah, ok =) [12:43] ev: sorry, I appear to be of a single mind today :) [12:43] but maybe I'm misunderstanding what's going on here [12:43] well, not sure if it is a probalme ... its surely a waste to have an extra dbus though [12:44] *problem [12:44] mzanetti, yes [12:44] hmm... ok === alan_g|lunch is now known as alan_g [12:45] ogra_: dbus-autolaunch should maybe do $ start wait-for-state WAITER=$uuid WAIT_FOR=dbus WAIT_STATE=running || normal-dbus-autolaunch [12:45] slangasek: ^ [12:45] cause it seems like it might be out of our control of crap auto-launching it's own session dbuses. [12:46] it really shouldnt :( === MacSlow|lunch is now known as MacSlow [12:48] Saviq, taking a look [12:49] MacSlow, I think it's the same as https://bugs.launchpad.net/unity8/+bug/1233411 [12:49] Ubuntu bug 1233411 in Unity 8 "Clicking on a snap decision action button is causing the click to hit the application underneath the notification area" [Critical,Triaged] [12:49] MacSlow, only SF and Mir behave differently - under SF you get all input, and the IFA blocks the input from getting to the app [12:49] MacSlow, whereas under Mir you don't get input at all in shell unless you have IFA [12:49] ubot5: bug 1233870 [12:49] bug 1233870 in Mir "unity8 cpu spike after blank/unblank (nexus4)" [High,Fix committed] https://launchpad.net/bugs/1233870 [12:50] MacSlow, so it looks like that'd be all we need [12:50] Saviq, SF = SurfaceFlinger, IFA ? [12:50] MacSlow, InputFilterArea [12:50] Saviq, InputFilterArea [12:50] MacSlow, or InputArea like it's called now [12:50] MacSlow, or ShellArea or something - you'll find it [12:51] MacSlow, it'd probably be best if we only created one - covering all the notifications on screen, if possible [12:51] Saviq, no clue what it is, but I'll take a look and try to get an idea [12:51] MacSlow, yeah, it's InputFilterArea still [12:51] MacSlow, yeah, and we have none around Notifications [12:52] Saviq, it's "new" and needed because of mir on the phone?! [12:52] MacSlow, no, it was there from January [12:52] MacSlow, but because of how SF behaved [12:53] only https://bugs.launchpad.net/unity8/+bug/1233411 was visible [12:53] Ubuntu bug 1233411 in Unity 8 "Clicking on a snap decision action button is causing the click to hit the application underneath the notification area" [Critical,Triaged] [12:53] MacSlow, so easy to miss [12:53] Saviq, never knew it's relevant [12:53] sforshee: heya, around? [12:53] mfisch: poke too === hikiko|lunch is now known as hikiko [12:54] sforshee, mfisch: To keep the music-app running when the idle timer kicks in, we are taking an active lock with requestSysState(state=1) [12:54] Saviq, so an InputFilterArea has to "span" the whole notification-bubble?! [12:54] sforshee, mfisch: This seems to keep the screen on though; what's the best way to just stop system from suspending, but let the idle timer shut the screen off? [12:54] om26er: ^ could you hand these guys the bug number when you have it? [12:55] bug 1235241 [12:55] bug 1235241 in QtDeclarative bindings for Powerd "Screen can be brought back up by just tapping on it" [High,New] https://launchpad.net/bugs/1235241 [12:55] MacSlow, even better - all of the notifications at once [12:55] MacSlow, ideally we'd have only one InputFilterArea - you could probably use ListView's contentHeight for that [12:56] Saviq, well I'll see === _salem is now known as salem_ [12:57] Saviq, btw... which flash-target is recommended? [12:58] MacSlow, ubuntu-system, as usual [12:58] Saviq, usually I use "ubuntu-system" [12:58] ah ok [12:58] sforshee, mfisch so that its really handed over :) bug 1235241 [12:58] bug 1235241 in QtDeclarative bindings for Powerd "Screen can be brought back up by just tapping on it" [High,New] https://launchpad.net/bugs/1235241 [12:59] Saviq, the -part2 MRs are both Jenkins-CI approved [13:00] MacSlow, cool, we're in fix-mir-autopilot mdoe [13:00] mode [13:00] MacSlow, so it might need to wait until next week [13:00] Saviq, yeah I know [13:03] mfisch: ping [13:03] mfisch: dpkg: error processing /var/cache/apt/archives/gallery-app_0.0.67+13.10.20131004.1-0ubuntu1_armhf.deb (--unpack): trying to overwrite '/usr/share/icons/ubuntu-mobile/apps/144/gallery-app.png', which is also in package ubuntu-mobile-icons 13.04+13.10.20130925-0ubuntu1 [13:04] mfisch: what's the plan there? ship the icon in ubuntu-mobile-icons or in gallery-app? could you drop it from the right packages and if it's moving add a Replace in the new package against the old ones? [13:06] mfisch: the easiest might be to drop it entirely from gallery-app.deb; it should be kept in there for the click transition though [13:06] sergiusens: ^ FYI [13:07] lool, I'd keep it in the gallery-app so then we can reuse the icon for the package (click) icon [13:08] sergiusens: yes, that's what I suggested as well (drop from .deb but keep in bzr) [13:10] lool, that works for me [13:11] ogra_, if I reboot enought I can lock myslef out of adb, so there is surely a race somewhere [13:11] sergiusens, yeah [13:11] sergiusens, the prob is that we dont handle adbd via the property system i think [13:11] but write to sysfs [13:11] ogra_, I'm changing that fwiw [13:12] sergiusens, to what ? [13:12] we cant use properties that early in the boot [13:12] ogra_, to use the property system [13:12] how ? [13:12] the container isnt there when adbd starts [13:13] and we need it to start that early to be able to debug boot issues [13:13] ogra_, already talked to rsalveti; we might just delay adbd [13:13] bad idea imho [13:13] especially for ports that will be hell [13:13] ogra_, well that's going to be an issue :-) [13:13] right [13:14] Is Ubuntu Touch capable of simulating glass effect? | http://askubuntu.com/q/353621 [13:14] geez [13:14] ! [13:14] glass effects [13:14] ogra_, let me think of something then; quickly though I guess we can enable an early boot mode [13:14] tmoenicke: hi! [13:14] ogra_, I have glasses on; if not the screen looks fuzzy, is that the effect? [13:15] tmoenicke: do you know what's up with the ubuntu-keyboard autopilot tests? I'm running ubuntu-keyboard trunk and it's failing on some AP tests [13:15] sergiusens, dunno, ask on askubuntu ... [13:15] seems its an important feature [13:15] (who needs to make phone calls if he has _galls effects_ ! ) [13:15] *glass even [13:15] sil2100: hi. veebers should know details [13:16] veebers: piiing [13:16] ogra_, saw the question, lets leave that to the design team [13:16] :) === flo__ is now known as w-flo === aaa_ is now known as Guest11034 === flo__ is now known as w-flo [13:23] 1 month until I get a new router and new ISP.. wonder how many reconnects I have to suffer through until then.. sigh [13:26] bfiller: heya, FYI gallery-app has file conflicts with icon-theme; I've pinged mfisch to fix this when he comes up, but feel free to fix earlier [13:26] lool, seems reboot is broken with #80 and i cant really find the cause [13:28] since adb reboot works just fine i suspect a hanging upstart job [13:29] ogra_: I have had this issue indeed [13:29] ogra_: adb reboot is a brutal adbd syscall to kernel to tell it to reboot NOW [13:29] ogra_: reboot is actually init sequence [13:30] must be somethig after the runlevel is switched, since adbd gets stopped [13:30] ogra_: I've pinged slangasek on getting some help from his team on this, but that will be later [13:30] k [13:30] ogra_: yes; stgraber might be able to provide guidance on how to best debug this [13:30] right [13:30] i was hoping for him [13:30] stgraber: ^ latest image doesn't reboot anymore, and we don't know how to track this down [13:30] (also for the ureadahead issue indeed) === kentb is now known as kentb-afk [13:46] What's the dot-file for user "phablet" again to enable mir instead of sf? [13:47] bfiller: FAILED (failures=11) [13:47] bfiller: for gallery-app [13:47] e.g. ERROR: gallery_app.tests.test_album_view.TestAlbumView.test_add_photo [13:47] bfiller: do I need to have empty gallery? [13:47] found it... .display-mir [13:48] lool: I'm awake but not really, I saw the failure last night, should we just remove icon from gallery app? [13:49] ogra_: please talk with jodh and slangasek, since last we discussed it with them the conclusion was not to turn ureadahead on [13:49] mhall119: ping [13:49] stgraber, it saves 5sec on maguro for me ... mako misses the kernel patch [13:50] mfisch: from the .deb, yes [13:50] mfisch: would you think you could look at the AP failures? [13:50] ogra_: I don't know any details about the reasoning, just that slangasek didn't want it on, so I'll let you talk with each other :) [13:51] lool: I'm busy for at least another hour, can look into the reboot thing after that [13:51] lool: I should have some time today pending any other disasters on my team [13:52] mfisch: cool [13:52] stgraber: thanks [13:52] lool: we can skip the meeting today unless you have something [13:54] mfisch: cool; I'm happy to get that time [13:54] lool: ok [13:54] mhr3_: sync cancelled for today [13:55] k [13:55] mzanetti: pong [13:56] lool: do not know [13:56] mhall119: wanted to ask you something about the nexus 7. but found the info on the net [13:56] lool: tests are failing when you are running on device? [13:56] bfiller: https://code.launchpad.net/~mzanetti/python-ubuntu-platform-api/fix-resolutions/+merge/189320 [13:56] ok [13:57] mzanetti: which issue does this fix? [13:57] bfiller: autopilot being broken with Mir [13:58] mzanetti: nice, not sure if this is what lool is talking about, maybe? [13:58] bfiller: I've ran most ap tests on surfaceflinger today. only community apps seem to have failing tests [14:01] xnox: could you prioritize the upstart review? :-) or we can defer to next week [14:01] ogra_: I'm just reflashing to pick up #80 to see if I can work out what is wrong with reboot. [14:01] http://people.canonical.com/~ogra/touch-image-stats/20131003.1.changes or http://people.canonical.com/~ogra/touch-image-stats/20131003.2.changes must have the offending change [14:02] it worked fine on the iage before [14:02] *image [14:02] cjwatson, jdstrand, architecture support has landed in the store, users are required to specify it on upload [14:03] beuno: OK, cool. We have click chroot management support underway; I don't know if it'll be in time for the SDK to support it in 13.10, but the SDK isn't so tightly bound to the release cycle that I'm going to panic about that [14:04] beuno: ack, I updated the review scripts earlier this week for that [14:04] ah, seems it even works on the 03.1 image [14:04] cjwatson, indeed, it's not really a 13.10 goal, so not worried about the exact timings [14:04] http://people.canonical.com/~ogra/touch-image-stats/20131003.2.changes should have the issue then [14:04] (That's the last thing resembling a feature to land in click for 13.10, unless I decide to throw in signing support as well, which I'm not sure I will) [14:05] jdstrand, thanks for checking the apparmor stuff, my /var/lib/dpkg/ content indeed changes when i install bootchart ... so might be a perfectly valid run i see [14:05] seb128: maguro settings says I have 2.1gb of free space. This is a 16 GB version surely I should have more than 2.1GB of free space? [14:06] root@ubuntu-phablet:~# lsusb [14:06] unable to initialize libusb: -99 [14:07] ogra_: ok, glad to hear [14:07] stgraber, btw, regardless of ureadahead, i think system-image needs to learn to remount mountpoints that live in rw dirs [14:07] ogra_, which file shows if I am using a local build of the android parts ? [14:07] sforshee: don't know if it will make it that far east but it's snowing here [14:08] janimo`, define local build [14:08] seb128: http://ubuntuone.com/04QsRW0y1RUNHyhiOF6q12 [14:08] alot of the hardware is not exposed to ubuntu-touch ? [14:08] davmor2, maybe most of the 16gb is used for the "internal SD card" android thingie (and AFAIK you don't repartition the devices yet) [14:08] ogra_, built system.img and boot.img from a locally checked out phablet Android tree [14:08] mfisch: dude. Our highs are supposed to stay above 60, so if it reaches us it should be rain by then. [14:08] ogra_, and then flashed them with fastboot [14:09] janimo`, hmm, no idea, probably rsalveti knows if there are any specific criteria [14:09] sforshee: i pasted you the stuff [14:10] ogra_: what do you mean? [14:10] lool: when you keep the system from suspending like you did there it will still have the timeout on the screen activity [14:10] not much shows up in a lshw hehe | http://paste.ubuntu.com/6192383/ [14:10] stgraber, ureadahead uses a debugfs mount inside a dir that it also needs rw [14:10] stgraber, once the rw mount is there the debugfs mount is hidden [14:11] stgraber, bug 1233281 ... see my last comment [14:11] bug 1233281 in ureadahead (Ubuntu) "crashes on Ubuntu Touch due to /var/lib/ureadahead/ being read-only" [High,New] https://launchpad.net/bugs/1233281 [14:12] davmor2, upgrade your ubuntu-settings-settings, it's fixed in https://launchpad.net/ubuntu/+source/ubuntu-system-settings/0.1+13.10.20131004-0ubuntu1 [14:12] " * [about] In the storage subpanel, check the space for drives Qt [14:12] reports as "Unknown"." [14:12] ogra_: there's an easy fix for that, ureadahead should just use debugfs from the right location ;) [14:12] ogra_: (/sys/kernel/debug) [14:12] seb128: nice one thanks [14:13] stgraber, i know where that lives, but i guess scott had his reasons to use a separate mountpoint ? [14:13] lool, om26er: fyi ChickenCutlass's team owns powerd now, so they should be the first point of contact for bugs [14:13] ogra_: I think the reason was that we didn't mount debugfs by default back then [14:13] ah [14:14] ogra_, So have you be able to recreate the HUD timeout with the new dbus configuration? [14:14] ogra_: how do I get phablet-flash to give me #80? [14:14] tedg, as soon as i switch to Mir [14:15] jodh, --channel saucy-proposed [14:15] ogra_: thanks [14:15] ogra_, ? It happens when you switch to Mir or you'll check once you switch to Mir? [14:15] tedg, it happens as soon as i switch to Mir [14:15] * tedg likes SF more now [14:15] tedg, http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-SF.png is SF [14:16] and the same after touching ~/.display-mir http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-hud-dbus-hang.png [14:17] ogra_, What happens with platform API in the Mir case? When we set up our observers and such? [14:17] (notice how dbus goes into zombie state again) [14:17] tedg, uh, no idea, platform-api -> ricmm/rsalveti [14:18] lool: ok. [14:19] ogra_, lool: your reboot problem is caused by upstart-local-bridge getting stuck in stop/killed state [14:19] sergiusens, ^^^ [14:19] stgraber, do we have a wikipage about nailing down such stuff ? [14:20] ogra_: I'm pretty sure we document "initctl list" in the upstart cookbook [14:20] (debugging the job chain ... ) [14:20] initctl list | grep -v "stop/waiting" | grep -v "start/running" [14:20] k [14:20] that gave me a single entry: [14:21] upstart-local-bridge stop/killed, process 501 [14:21] root 724 0.5 0.0 2316 880 ? S 14:19 0:00 upstart-local-bridge --daemon --event=android-container --path=/dev/socket/upstart-text-bridge [14:21] root@ubuntu-phablet:/# initctl list | grep upstart-local-bridge [14:21] upstart-local-bridge start/running, process 500 [14:21] so the problem is that upstart doesn't track the pid properly for that one [14:22] i guess the exec should live outside of the script block [14:23] and the script should become a pre-start one [14:23] * ogra_ tries that [14:23] nah, that's not the problem [14:23] the fix is even simpler than that [14:23] stgraber, is it a demon/fork one? [14:23] oh [14:23] sergiusens: just drop expect daemon and remove --daemon [14:23] * ogra_ totally missed expect [14:24] no need to use ptrace if you can avoid it === tkamppeter_ is now known as tkamppeter [14:24] I confirmed that http://paste.ubuntu.com/6192429/ works fine here [14:25] rsalveti, do you how can I tell if a device has an officially built or a locally built system.img/boot.img flashed [14:25] rsalveti, I just built the tree locally and fastboot flashed it [14:25] if you really want to use --daemon, then look into using expect fork (not sure that'll work, haven't tried that) [14:25] stgraber, that part of the job I have no idea about [14:26] stgraber: sergiusens: not, expect daemon ?! [14:26] stgraber: "expect daemon" + "--daemon" works fine for that bridge in #78. [14:28] well, i get a proper reboot here with that change [14:29] jodh: it sure doesn't on my phone [14:29] ftr, I'm happy to review any upstart .conf file changes. [14:29] jodh: well, sure doesn't starting with #80 or whatever it's we hae today [14:30] jodh: and what I pasted above clearly shows that upstart failed to track the PID of the process... [14:30] http://launchpadlibrarian.net/152241386/lxc-android-config_0.105_0.106.diff.gz [14:30] thats the change [14:31] stgraber: your comment above about "expect fork" + --daemon confused me somewhat. [14:32] ogra_: so based on that diff, I guess you're right that the mkdir confused upstart... [14:32] ogra_, maybe for that expect daemon thing to work it needs to be out of the script stanza, so then the pre-start thing ogra suggested is better [14:32] ogra_: you know about the apparmor parts in the kernel ? i have merged the apparmor v3 patch, and enabled it in defconfig, but somehow it looks for sure like it doesnt make the module, and sys/module/ is not having a apparmor module [14:32] stgraber, well, i think the job didnt start at all before [14:32] gusch: I'm having issues building gallery-app from trunk due to a missing dep on content-hub 0930.1 [14:32] jodh: so I guess the problem is that the mkdir counts as a first fork, then the first fork of the daemon is considered as the final one and so we track the pid of the intermediate fork? [14:33] stgraber, which is why is didnt hangon shutdown [14:33] at least thats my theory [14:33] kenvandine: where can mfisch get the proper content-hub package for gallery? [14:33] I tried building it here and the tests failed after 25 minutes of watching it build [14:34] ogra_: those 5 lines, did i add to the correct defconfig , without luck : http://paste.ubuntu.com/6192454/ [14:34] anders3408|afk, i dont think that works as module === anders3408|afk is now known as anders3408 [14:34] jodh: though I thought we were clever enough to only track those after the exec somehow [14:34] gusch: there is no 0930.1 build based on the changelog, its either 0930 or 10* [14:34] anders3408, but i have not much clue about apparmor ... try jjohansen [14:34] anyway, for now if dropping --daemon and expect daemon works, I say we go with that, we'll even get working logging as a bonus [14:34] hm.. apparmor config files says something about look for /sys/module/apparmor , still its not a real module like wlan.ko :) [14:34] thanks ogra_ [14:34] jjohansen: you here :) [14:34] (and save ourselves a bit of ptrace and a few forks) [14:35] ? [14:35] * ogra_ wants more spoons [14:35] lool, slangasek, ralsina, ping abeato and I have an interesting finding we would like to share with you guys :) (or not so happy face) [14:35] kentb-afk: see the comment from mfisch about the content-hub version number [14:35] ogra_, are you applying that? [14:35] sergiusens, on it === alan_g is now known as alan_g|tea [14:37] mandel: speak quick or remain quiet for eternity [14:37] ogra_: a good way to start would probably be make bootimage -j1 and confirm if it builds anything or not :) [14:37] mandel: sure, share... how? [14:37] lool, ralsina, QNetworkInfo from qt performs a DOS attack to ofono via DBus, to be more precise to the GetModems call [14:38] ogra_, jodh: Did you guys find interesting bits on the shutdown/reboot issue? [14:38] lool, fix uploaded :) [14:38] ah read stgraber's update [14:38] cool [14:38] lool, ralsina that makes ofono go funny as abeato can explain better [14:38] lool, lxc-android-config 0.108 [14:38] anders3408: I can't really help either, but did you see https://wiki.ubuntu.com/SecurityTeam/AppArmorForPhabletKernels? (linked from https://wiki.ubuntu.com/Touch/Porting#AppArmor)? [14:38] lool, ralsina so using that class makes (that uses NM and ofono) to detect the network state is/might be a problem [14:39] mfisch: so you're saying this should not be happening and it's a powerd bug or a testing bug? [14:39] yes jdstrand i did, not saying much:) [14:39] lool: no, its a real bug, sforshee is looking at it while I fight to get gallery-app building [14:39] ogra_: could you test the upstart in -proposed? [14:39] ogra_: with new lxc-android-config would be cool [14:39] lool, sure [14:39] mandel, lool, ralsina what we see is > 300 calls to interface=org.ofono.Manager; GetModems [14:39] anders3408, IIRC apparmor is built into the kernel for me, not as a module.. maybe you have the same (though that doesn't explain why it fails to work) [14:40] ogra_: remember lxc-android-config is a bit tricky to upgrade [14:40] lool: I've actually known about that for a long time, but never fixed it because unity was supposed to take over the inactivity timer [14:40] but that hasn't happened yet [14:40] when ubuntu-download-manager is executing [14:40] sforshee: so it's a known powerd bug? [14:40] sforshee: is it logged already? [14:40] sforshee: (to merge the two) [14:40] lool: I don't recall if there was another bug filed [14:40] lool, "upstart in proposed" ? [14:40] w-flo: the thing is /out/target/product/find5/obj/KERNEL_OBJ/Security/ doesnt contains anything from apparmor at all :) [14:40] that makes network detection slower [14:41] lool, there is no new upstart [14:41] mandel: Can you give me the punchline summary of what this breaks and how we fix it? :-) [14:41] om26er: where's that bug of yours on the music-app? [14:41] om26er: I checked qtpowerd, and it isn't there [14:41] lool: I already reassigned it to powerd [14:41] sforshee: oh thanks [14:41] anders3408, maybe you're missing some kernel dependencies? have you tried enabling apparmor via "make menuconfig" or similar instead of editing the defconfig directly? [14:41] stgraber, sergiusens: that mkdir should be in a pre-start to avoid disrupting the main process logic as ogra_ suggests. Then use 'expect daemon' + '--daemon' as designed. [14:42] ogra_, ^^ [14:42] sforshee: do you intend to work on it, or should I talk to Chicken to find someone in his tema? [14:42] sforshee: (I doubt we will move it to unity at this point) [14:42] w-flo: make menuconfig says it aint enabled so the lines i added to defconfig was a bit wrong :) [14:42] jodh, heh, i just uploaded the former fix ... will test and change accordingly [14:42] lool, bug 1235241 [14:42] lool: I'm already testing a fix [14:42] bug 1235241 in powerd "Screen can be brought back up by just tapping on it" [High,Confirmed] https://launchpad.net/bugs/1235241 [14:42] ogra_: 16:42 -queuebot:#ubuntu-release- Unapproved: accepted upstart [source] (saucy-proposed) [1.10-0ubuntu4] [14:42] ogra_: https://launchpad.net/ubuntu/+source/upstart/1.10-0ubuntu4 [14:42] ogra_: not yet published yet [14:42] anders3408, hope it works after enabling it in menuconfig :) I also have the apparmor ....HASH option enabled, not sure what it is / if it is required [14:42] sforshee: \o/ [14:43] sforshee: if you like, try music-app; play/pause should take the lock [14:43] lool, hmm, not on -changes ... [14:43] uhm ... w-flo i didnt have that one ? :) [14:43] lool: powerd-cli works fine for reproducing [14:43] ogra_: stgraber is right, but I'd rather we didn't rely on side-effects - if that script gets changed, the job will break again. So best to put all the required setup in pre-start. === Namidairo is now known as Namidairo`bnc [14:43] jodh, ok [14:44] mfisch, gusch: it's in saucy... [14:44] anders3408, it's not in your pasteit [14:44] nope :) [14:44] i didnt see it on mako either [14:44] kenvandine: I see this: content-hub | 0.0+13.10.20130930-0ubuntu1 [14:44] gusch / kenvandine gallery-app needs 0930.1 [14:44] anders3408, so it's not required I guess.. :D [14:44] or later [14:44] jodh: well, my "fix" was not to use expect daemon and simply exec the bridge without --daemon, I don't see how that can possibly break with further changes to the script section [14:44] mfisch, >= [14:45] lool, mainly, we make ofono work like crap if we support network mode detection in u-d-m, possible ways to fix it, patch QNetworkInfo and reduce the number of calls, call NM and ofono dbus apis directly from u-d-m, that second fix is not a good idea because if more than one app uses QNetworkInfo we are multiplying those 300 calls per the number of apps using the class [14:45] abeato, ^^^ [14:45] oh! [14:45] abeato, sounds reasonable? [14:45] mfisch, that never got published [14:45] mfisch, it is in the daily-build ppa [14:45] sergiusens: forgot to mention. i added another to address your MP remarks: https://code.launchpad.net/~doanac/phablet-tools/phablet-config-developer-mode/+merge/188416 [14:46] mandel: how does it affect the image today? [14:46] mandel: does it e.g. prevent receiving phone calls? [14:46] mandel: does it not work? [14:46] mandel: I think we want to fix this properly, but time remaining is super short, so if you can work out the right fix in a minimum time and queue it up for next week, that's good [14:47] mandel: make sure you have an easy test plan [14:47] mandel: (is this just about 3g/wifi detection?) [14:47] lool, AFAIK it works, it makes ofono work slower but is not critical (right abeato) [14:47] lool, yes, just about 3g/wifi/connection detection [14:47] mfisch, however, that build depends could be changed to 0930, dropping the .1 [14:47] kenvandine: yes, that would be helpful, I'll do that [14:47] lool, detecting the connection is important to be able to do correct downloads in not reliable connections [14:48] lool, bug 1233435 [14:48] bug 1233435 in ubuntu-download-manager "u-d-m downloads stall if the network configuration changes" [Critical,Triaged] https://launchpad.net/bugs/1233435 === alan_g|tea is now known as alan_g [14:49] lool, I would fix that bug, go with what we have that is not terrible and do a proper fix [14:49] lool, we can talk about who fixes QNetworkInfo after release [14:49] * mandel hides [14:49] seb128: any idea why system-settings would freeze up on the device? does it log or have any way to enable debug log output to console? [14:50] mfisch, scratch that... [14:50] it does need > 0930 [14:50] mandel, lool right, radio still works [14:50] mfisch, it's tough picking build depends version when you don't know what order things will land in :) [14:50] mandel: please prepare what you want to upload monday/tuesday that is good for release and easy to review/test [14:50] just makes registration slower [14:50] w-flo: make menuconfig , removes 2 lines and add 3 lines : see http://paste.ubuntu.com/6192502/ [14:50] mfisch, so it needs what's in daily-build [14:51] kenvandine: ok, I'l add the ppa [14:51] anders3408, AUDIT... isn't that selinux? maybe apparmor & selinux conflict? just see if it works now :) [14:52] lool, ok, will fix that bug, and I'll open a bug on QNetworkInfo ( abeato can you give me a hand with some data regarding the issue) and we should be good to go (just slow registrations in ofono) [14:52] w-flo: im missing : CONFIG_SECURITY=y :D [14:52] mfisch, i'll get that on the landing asks page [14:52] hahaha guess why it fails :D [14:52] anders3408, oh. haha [14:52] mandel: thanks! [14:52] ogra_: can you ping me once upstart is good? [14:52] ogra_: would like to release it out of proposed [14:53] ah sorry it's not even built [14:53] yeah, was about to say :) [14:54] stgraber: so the upstart issue was the one already in the pipe for upload? [14:54] bah [14:54] lool: the reboot issue and the upstart upload are completely unrelated [14:54] kenvandine: is there a ppa for the daily-build? I cannot seem to find it [14:55] oh sorry, lxc-android-config fix [14:55] https://launchpad.net/~ubuntu-unity/+archive/daily-build [14:55] mfisch, ^^ [14:57] ogra_: any closer to other than nexus devices to make those system images ? [14:57] anders3408, not before release i fear [14:58] ogra_: find 5 need it soooo much :) [14:58] anders3408, unless you know a way to clone stgraber quickly [14:58] (and spread the clones across multiple timezones preferably) [14:58] clone him or his sources ogra_ ? :) [14:59] ogra_: armhf upstart built [14:59] lool, pulling [14:59] ogra_: actually, I've delegated that work to schwuk's team since the OEM guys need the same thing (well, and QA too), but yeah, may still take a few weeks [15:00] ah, so thats semi-cloned then :) [15:00] the good thing is that it's all server side so we can release it whenever it's ready and it'll just work with existing saucy [15:00] mfisch, i added that to the asks page [15:01] anyway, I have it as a nice to do for when I'm done getting back to 100% unit test coverage for the server code and I find some spare time... It shouldn't take me much more than a day to get that stuff implemented with tests but it's hard to find a whole day to focus that close to the release. [15:01] but stgraber those system image , does that in current state working fine on mako devices ? [15:02] gusch, i have a branch that fixes a potential race condition can you review it? [15:02] https://code.launchpad.net/~ken-vandine/content-hub/invoke_on_initiated/+merge/188629 [15:02] anders3408: yep, that's what we recommend using for mako, maguro, grouper and manta and that's what the QA team and everyone else doing testing on Touch uses [15:02] as find5 is pretty much the same as mako :) [15:03] kenvandine: a quick review - yes === dandrader is now known as dandrader|lunch [15:03] it's an easy one, i think [15:03] the reason im asking , is it may be possible to get working on find5 when its working on mako devices :) [15:03] gusch, i just approved your hint branch [15:03] sure, if we had the find5 kernel in the Ubuntu archive and had our android source package spits out a find5 boot/recovery/system partition image, I could add support for it in the production server in a couple of minutes. The tricky part is letting you do that outside of the production server while still having the whole chain signed with GPG, https, ... [15:04] kenvandine: oh - yea that's the spot I had it on the first place [15:04] dobey, no idea about the freeze, you can run it by "adb shell" "su - phablet" "system-settings --desktop_file_hint=/usr/share/applications/ubuntu-system-settings.desktop" [15:06] stgraber: oh yes that would make sense :) so when its ready , it will not be just a source release ? [15:07] anders3408: the source for the server is already public, when it's ready, you'll just have the needed modules in there to import from the central server repack, ... and maybe a wiki documentation on how to setup your server [15:07] w-flo: YAY https://bitbucket.org/anders3408/ubuntu_kernel_oppo_find5/commits/f9c25d803f172998dd9ad2526e7fb759d776f4ef worked :D now it builded the apparmor parts :D [15:07] anders3408, cool :) [15:08] stgraber: let me get this correct , server like a build server , or something phone releated ? :) [15:09] anders3408: the code that generates the structure on https://system-image.ubuntu.com [15:10] anders3408: that thing basically imports all the files needed to produce an image, does the checksumming, generates deltas, manages multiple update channels, ... [15:10] seb128: thanks. it seems to be working now. i'll try that if it freezes again in the future [15:11] stgraber: oh :D well let me work on the remaing issues then :) [15:11] MacSlow: Heya [15:11] MacSlow: when you know what's going on with picking up calls on Mir, I'd love to get an update :-) [15:12] MacSlow: would like to switch the default soon, and that's a blocker [15:12] another one, can i enable Mir without rebooting ? like enable it before first bootup :) [15:13] w/win 118 [15:23] dandrader|lunch: heya, (have a good lunch), I hear you're chasing input issues (volume up/down and webbrowser); would you mind updating me on the fix (ETA, affected package(s))? I'd like to try enabling Mir over the week-end, but this needs to be fixed first [15:23] dandrader|lunch: (mainly the webbrowser input issues need to be fixed) [15:25] YAY :D [15:25] root@ubuntu-phablet:/# aa-status [15:25] apparmor module is loaded. [15:25] 22 profiles are loaded. [15:26] anders3408, congrats ! [15:28] ogra_: :D can dl and install click apps now :D [15:28] awesome ! === gatox is now known as gatox_lunch [15:29] lool: gallery app is still building, 45 mins and counting... hope to have a fix soon [15:30] ogra_: sforshee made a comment on powerd/backlight-settings a week back, can someone review that ? :) i really need it for pwrkey :) and brigtness [15:33] hmm ogra_ if i installs an click package like g+ or xda , i can start it, but even if i pin it to launcher it still not showing up there and when i then close the installed app, the app is gone as if i looks under clicks it says not installed , and if i look under installed , it is missing :) [15:34] mfisch: it's now depending on new content-hub [15:34] mfisch: so will be monday I think [15:34] lool: yep, but I still want to test it today [15:35] cool [15:36] xnox: ..."crap auto-launching its own session dbuses"? like what? [15:37] ogra_: why do you want ureadahead? I believed it was irrelevant on disks of this class. If that's mistaken, we should certainly use it, but it seemed that it was a) failing to run and b) not useful if it did run [15:37] slangasek: well one thing i did catch is apport starting on file bridge event. which is now fixed, but there appears to be more stuff. [15:37] slangasek, it isnt irrelevant [15:38] xnox: but if the file bridge is starting after dbus, we're still ok, aren't we? [15:38] slangasek, i can get maguro to boot 5sec faster when having it set up properly [15:38] ogra_: why not? [15:38] lool: tested and proposed: https://code.launchpad.net/~mfisch/gallery-app/remove-icon-from-package/+merge/189331 [15:38] what is ureadahead doing in this case that is useful? [15:38] slangasek: for that case yes. and mhr3_ is still reporting that dbus launched hud, starts it's own dbus. [15:38] =/ [15:39] hmm [15:39] slangasek: but yeah, we should fixed those cases. i just seems like there is more and more of them. [15:39] slangasek, it loads all files in one big chunk into ram [15:39] * mfisch makes a note never to touch gallery-app again [15:39] slangasek, http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-SF.png [15:39] mfisch: cool, thanks [15:39] xnox, imo something else does, hud just uses it somehow [15:39] slangasek, thats a 25sec boot from vibration to UI on screen [15:40] ogra_ if i installs an click package like g+ or xda , i can start it, but even if i pin it to launcher it still not showing up there and when i then close the installed app, the app is gone as if i looks under clicks it says not installed , and if i look under installed , it is missing [15:40] slangasek: hey, quick question-- libthumbnailer0 is shipping /usr/share/thumbnailer/vs-thumb, but vs-thumb is an ELF binary. I know it should go in /usr/lib, but not sure the best place to recommend [15:40] ogra_: and why should that speed anything up when loading from flash media? There's no seek time involved [15:41] jdstrand: /usr/lib/$arch/thumbnailer/vs-thumb; since it's a shared library package we want it to be multiarch-clean [15:41] seems we have applications honoring multiarch, using /usr/libexec and /usr/lib/ [15:41] slangasek: ok, thanks [15:41] nothing should use /usr/libexec [15:41] if you see any of those, please let me know what they are so I can hit them with a hammer ;) [15:41] slangasek: /usr/libexec/signon-ui/browser-process [15:42] slangasek, dunno, but i know it gained us faster boot on SD cards on the pandas too [15:42] even if only a few secs [15:42] why do i get alot of : request_suspend_state: wakeup (0->0) at 1029588882890 (2013-10-04 15:41:13.105054102 UTC) , init: powerd main process (14557) killed by FPE signal, init: powerd main process ended, respawning [15:42] ? :O [15:42] sforshee: you here ? [15:43] slangasek, in any case, we can non not ship ureadahead (since we depend on minimal) and i prefer to have it functioning if we actualyl ship it [15:43] lool, sure [15:44] ogra_: a few secs is more than enough to be worthwhile, but I would like to understand why it's doing anything useful at all [15:44] anders3408: I recall someone talking about something like that the other day, but I can't remember who. The problem for them was either a bad flash or something on the android side of things. [15:44] ogra_: I don't think anyone is arguing for hacking up the seeds to remove it [15:44] slangasek, all IO happens at one point and then there is nearly no IO throughoiut the process anymore, i guess that buys us something [15:45] slangasek, and i know that Keybuk always encouraged me to use ureadahead even on flash media ... i had that discussion a few times with him [15:45] sforshee: it has done that on every build i have tried so far , sforshee lastest couple of build has somehow not been able to use pwrkey anymore, even i if used your powerd.deb you sent me :( [15:45] (but i sadly forgot his arguments) [15:46] ok [15:46] slangasek: can we worse disable it, such that it doesn't run at all, on touch. and do a boot chart run. Maybe the gain is "ureadahead profiling" vs "ureadahead loading and invalidating cache quickly" [15:47] ogra_: well, that's one appeal to authority I can accept, anyway, since Keybuk certainly did know this code ;) [15:47] the profiling is actually painful [15:47] so yes, we should turn it back on [15:47] and fix any remaining bugs with having it on [15:47] but our first boto in itself is painful [15:47] slangasek, great, will work on that then [15:49] sforshee: apt-get update and then apt-get install powerd fixed it :) [15:50] Ubuntu Touch on Samsung Galaxy Wonder (GT-I8150) | http://askubuntu.com/q/353674 [15:50] ogra_: if we know the device ram-device bank sizes, and format with something like fat, make sure our ureadahead dumps are equalify distributed across all banks (that is ureadhead file physically spread across 8 memery banks (with gaps)) then sequeantial read will actually become a parallel read from all memory banks and gain ~ x8 read improvement. If the firmware on the eMMC device is smart enough and can do that for a large sequantial file even wit [15:50] h ext4 filesystem it can be a win. [15:51] xnox, well, in theory ... [15:51] xnox, practically you usually max out the eMMC bus at 30M/s [15:51] yeah, i'm yet to get that far. [15:52] some might go higher nowadays (i think mako actually goes up to over 40) [15:52] Hi, I would like to know if Ubuntu Touch will support Google contacts and Whatsapp in 13.10 release [15:54] Is it known problem MusicPlayer works fine on Ubuntu phone, but not on desktop? I'm running same code, and it just won't play. [15:54] It's an app I made for the workshop at Deview next week. [15:54] ogra_: if audio stops working , can i just reboot that service ? [15:55] kenshiro: Unless there's a community person working on that, I don't think we'll have whatsapp anytime soon. [15:55] At least to my knowledge. [15:55] anders3408, you cant try to restart pulseaudio from the terminal (no idea what happens)( [15:55] WOAH [15:55] does anyone know of any existing work done porting Ubuntu Touch to the original Samsung Galaxy Tab (7", CDMA/Verizon version, Model SPH-P100). I have googled and checked https://wiki.ubuntu.com/Touch/Devices, but I'm hoping there's something off-radar. === dandrader|lunch is now known as dandrader [15:55] Wed, 17 Feb 2010 12:57:00 +0000 [15:55] karni that will be a problem, most of people is using whatsapp these days :-/ [15:55] thats the last ureadahead upload we had [15:56] karni and Google Contacts support? [15:56] ogra_: the thing is if i tries to start a call all audio breaks :) [15:56] anders3408, most likely a bug in telepathy-ofono [15:56] kenshiro: I'm both Android and Ubuntu user, I don't use whatsapp. I know many people do. They also use twitter or hangouts :( [15:57] ogra_: oh :) hmm... i cant even test mic with video recording as that fails to :) [15:57] kenshiro: honestly, I know nothing about that part, someone else might be able to answer [15:57] karni thank you very much, ok I will wait for other answer [15:57] kenshiro: You are most welcome! [15:58] :-) [15:58] lool, on a first look, I think only mir packages will be affected (regarding that volume keys' fix) [15:58] lool, ETA is hard to tell right now [15:59] kenshiro: btw, quick google gives promissing results [15:59] http://stackoverflow.com/questions/14520776/is-it-legal-to-use-whatsapi [15:59] https://github.com/venomous0x/WhatsAPI [15:59] http://openwhatsapp.org/develop/ [15:59] :) [16:00] slangasek,ogra_: right, ureadahead was the result of Keybuk unifying two different versions of readahead one of which was optimised for rotational disks and the other for SSDs and similar [16:00] yeah [16:00] right [16:01] i know the advantage on MMCs/SD cards is minor, but there is still a gain [16:01] dandrader: are you also looking at webbrowser? [16:01] so what does it actually *do* on non-rotational devices? That's the only part I'm not clear on [16:01] lool, no [16:01] Keybuk did once explain this to me and I'm having trouble paging it back in [16:01] same here [16:01] dandrader: ok, thanks [16:01] and he explained it multiple times to me [16:02] karni yea, I just read the links, thank you again ! :-) [16:02] yw [16:02] but you also clearly see how it does all the I/O business between 8 and 18sec in http://people.canonical.com/~ogra/ubuntu-touch/bootchart-maguro-SF.png [16:02] i guess just the pre-loading in general gains us some improvement [16:05] slangasek,ogra_: I think that the point of sreadahead (then later ureadahead) was essentially that the principal limiting factor on booting the OS was often simply the I/O bandwidth multiplied by the amount of data you had to read, so you win by front-loading as much of that I/O as possible rather than by waiting for programs to initiate it later [16:05] yeah [16:05] slangasek,ogra_: https://groups.google.com/a/chromium.org/forum/#!msg/chromium-os-dev/mUO-hvapcGE/VeQt-kxGDjUJ has an explanation from Scott [16:05] hello [16:05] ogra_, did you see the langpack emails? can you get those added to the image? [16:06] "if the process wins it's just an I/O wait you could have eliminated but didn't" [16:06] seb128, yep, will add them for 82 [16:06] ogra_: the ideal situation with ureadahead is exactly front-loading a load of I/O [16:06] ogra_, thanks a lot! I'm replying to the email to say that and confirm the list [16:06] ogra_: you might like to work out the size of the data it has to read and divide that by the I/O bandwidth to get a lower bound on its time [16:08] ogra_: that is, all of that I/O would (presumptively) still happen, but would instead be spread out over other processes and handled less efficiently [16:09] that's the theory, anyway [16:09] right [16:09] so it's not clear to me that what you're seeing there is actually bad; it could well simply be ureadahead doing its job [16:09] cjwatson, slangasek claims its bad, i'm praising it, it gains me 5sec boot time ;) [16:10] it is weird that the click-system-hooks job apparently starts *before* ureadahead. click-system-hooks is "start on filesystem", while ureadahead is "start on starting mountall" [16:10] shouldn't ureadahead start strictly earlier? [16:10] compared to not using ureadahead and having it spread the I/O across the boot [16:10] it should start on filesystem i think [16:10] at least for us [16:11] slangasek: it basically issues a load of readahead() syscalls for the blocks recorded in the pack [16:11] what intrests me more atm though is, what does mount /var/lib/ureadahead/debugfs [16:11] ogra_: why? [16:11] start on filesystem would be less optimal [16:11] cjwatson,because that doesnt work on ro images [16:11] how so? [16:12] /var/lib/ureadahead needs to be rw [16:12] making it rw means something gets mounted on /var/lib/ureadahead ... which hides the mount in /var/lib/ureadahead/debugfs under it [16:12] i can see /var/lib/ureadahead/debugfs in my mount output .... [16:13] but ls /var/lib/ureadahead/debugfs is empty [16:13] if i unmount /var/lib/ureadahead all is fine [16:13] I guess the mounts are happening in the wrong order [16:13] cjwatson, right [16:13] ureadahead only uses /var/lib/ureadahead/debugfs if /sys/kernel/debug isn't mounted when it starts [16:14] So perhaps making it "start on virtual-filesystems" would be sufficient? [16:14] yeah, well [16:15] i see /var/lib/ureadahead/debugfs in my mount output [16:15] but it isnt in /lib/init/fstab for example [16:16] Sure, but what I'm saying is that this is a *consequence* of ureadahead being started in the wrong place [16:16] ureadahead mounts /var/lib/ureadahead/debugfs only if /sys/kernel/debug/tracing doesn't exist [16:17] oh, the binary mounts it ! [16:17] now i get it [16:18] that was the puzzle piece missing [16:18] ogra_: Furthermore, /var/lib/ureadahead is only needed to be rw when outputting the results of tracing, and that only happens at the end of the boot when we can assume everything's been mounted, so there's no ordering problem there [16:18] ogra_: can i reboot only ubuntu but not the whole phone ? [16:19] ogra_: Now, I think there's still a bit of a problem with getting the sequencing right [16:20] There's a ureadahead instance for each mount; starting the main process on virtual-filesystems does nothing to guarantee that /sys/kernel/debug is mounted before any non-virtual filesystem (it's likely, I suppose, but certainly not guaranteed) [16:20] cjwatson, i enabled /var/lib/ureadahead in writable-paths today already so the profiling should work ... but yeah sequencing is still wrong [16:20] root@ubuntu-phablet:/# ls /var/lib/ureadahead/ [16:20] android.factory.pack debugfs pack [16:20] android.system.pack etc.fstab.pack userdata.pack [16:20] thats what i get with working profiling [16:21] slangasek: Any thoughts on the ordering problem here, if you've been following my train of thought? [16:21] intresting it doesnt profile each and every of stgraber's bind mounts :) [16:22] ogra_: looks like it works per block device [16:22] I suppose we could move the temporary debugfs mount to somewhere that isn't a child of /var/lib/ureadahead [16:22] stgraber, oh, indeed :) [16:22] That way it doesn't matter whether it starts before or after /sys/kernel/debug is mounted and it ought to just work [16:23] well, cant we just always have /sys/kernel/debug mounted ? [16:23] It's mounted by mountall, which runs in parallel with the very early stages of ureadahead [16:23] like ... doing the mount from initrd and move mount it [16:23] I don't think that's a good idea, no. [16:23] We don't need to overcomplicate the standard mounts; we just need to have ureadahead cope [16:24] It would be a much simpler change to move ureadahead's mount to somewhere that won't be shadowed by a RW mount of /var/lib/ureadahead [16:24] That said, I'm not clear on why /var/lib/ureadahead isn't mounted well before ureadahead starts [16:24] Aren't those RW mounts done in the initrd? [16:24] stgraber, so seb128 wants me to add langpacks, i was wondering if you know where my size limits are for the system partition [16:25] ogra_: the size is currently set to 2GB and we're at 57% of that at least on mako [16:25] ok, mako should be fine then [16:26] cjwatson, they are [16:27] * ogra_ wonders if there is a mountall issue here [16:27] So then I'm totally confused, /var/lib/ureadahead/debugfs should just mount on top [16:27] cjwatson: no [16:27] How can I test this locally? [16:27] cjwatson: the initrd generates /tmp/fstab, bind-mounts it to /etc/fstab and then let mountall deal with the ordering [16:28] oh, right, we have an fstab [16:28] Aha [16:28] I guess this is a problem on non-phone systems with separate /var too [16:29] so sounds like if we want ureadahead to work at all, we'd need to have it mounted by the initrd so it's there before mountall runs [16:30] doing that should also make the existing debugfs mount just work as /var/lib/ureadahead will be writable [16:30] (but I hate adding more logic to the initrd, it's already complex enough...) [16:35] stgraber: That's why I don't want to do that. We could just move the mount [16:35] stgraber: How about /lib/ureadahead/debugfs? That's still reasonably namespaced, and guaranteed to be on / [16:35] to /run or something [16:35] that won't actually help since /var/lib/ureadahead won't be mounted by the time ureadahead runs [16:35] so it won't find its pack files [16:36] Oh, I see what you mean [16:36] still, we could use /run/ureadahead/debugfs [16:36] or not ? [16:36] Yeah, that's tricky [16:36] We could [16:37] Though you'd have to add a mkdir [16:37] sure we could, wouldn't buy us anything but we could :) [16:37] Just /run/ureadahead [16:37] stgraber: It would fix bug 523484 [16:37] bug 523484 in ureadahead (Ubuntu) "ureadahead requires /var on root filesystem" [Medium,Triaged] https://launchpad.net/bugs/523484 [16:38] cjwatson: well, it'd stop complaining but it still wouldn't give you any performance improvement [16:38] Actually, no, I suppose it wouldn't [16:38] actually, I'd expect it to make things slower since it'll trace the boot every single time [16:38] as it won't ever be able to read the pack files from the previous one :) [16:39] this is btw my awful quick hack to make it work http://paste.ubuntu.com/6192955/ [16:39] bfiller: hea [16:39] bfiller: would you have a list of apps that were broken by the proposed ubunut-ui-toolkit? [16:39] bfiller: with the hardcoded database path thing === fmasi is now known as fmasi_afk [16:39] Indeed, /run/ureadahead doesn't actually help at all, thinking about it [16:40] its also a virtual fs [16:40] Because /run is a virtual filesystem and will be mounted around the same time as /sys/kernel/debug! [16:40] bfiller: would like to track a landing for the fixed apps for monday as to test this again ASAP [16:40] right [16:40] bfiller: (also, will the fixed app work with old toolkit?) [16:40] In fact, it's listed in /lib/init/fstab after /sys/kernel/debug, so it might well usually be mounted *later* [16:40] cjwatson: hmm, I thought /run was mounted by the initrd [16:41] cjwatson, well, /run/ureadahead would buy us something as replacement for ./var/lib/ureadahead/debugfs [16:41] Mm, could be [16:41] not as replacement for the /sys/kernel/debug one [16:41] ogra_: stgraber's right though, we still have to solve /var/lib/ureadahead or it makes no difference [16:41] ogra_: So there's actually no point [16:41] we need it writable yes [16:41] ogra_: no, you need it readable [16:41] well, both :) [16:41] We need the data in it to be there, which only happens when the persistent mount happens [16:42] ogra_: ureadahead can't possibly give you any performance improvement if it can't read the files stored in there [16:42] cjwatson: ordering problem> you mean bug #523484, I guess :) [16:42] bug 523484 in ureadahead (Ubuntu) "ureadahead requires /var on root filesystem" [Medium,Triaged] https://launchpad.net/bugs/523484 [16:42] stgraber, yeah, understood [16:42] slangasek: Right, as I got there a bit later [16:42] ogra_: so sure we can get it running without /var/lib/ureadahead, but that'll make your boot slower, not faster ;) [16:42] Yep, that patch is bad [16:43] * ogra_ slowly starts to think his hack above might actually be a good option for touch [16:43] I disagree [16:43] it doesnt slow down the boot at least [16:44] It doesn't slow down the boot because you have "mount on filesystem", so /var/lib/ureadahead is available [16:44] right [16:44] But that means that the manual debugfs mount is pointless for exactly the same reason [16:44] cjwatson: so why do we want the debugfs mounted somewhere other than /sys/kernel/debug ? why shouldn't ureadahead just mount it in the normal place? /sys is always handed off to us mounted from the initramfs, so that's not a problem in practice [16:44] slangasek: I've moved on to the belief that the debugfs location is actually a complete side issue [16:45] but getting the packs mounted is obviously still an issue, yes; we would need that mounted before ureadahead starts [16:46] xnox: do you know why upstart-file-bridge is 'start on mounted'? [16:46] Certainly for full correctness ureadahead could test whether /sys/kernel/debug exists and if so mount it there - provided that mountall will still emit the "mounted" event for it when it discovers that it's already mounted [16:46] cjwatson: mountall will [16:46] But I don't (now) think that it does anything much to fix this bug and it shouldn't even be a necessary change [16:46] right, thats in /lib/init/fstab [16:47] slangasek: system-wide one? hm, i guess it needs filesystems to monitor events on?! [16:47] I think fixing 523484 is a prerequisite for this [16:47] bug 523484 [16:47] bug 523484 in ureadahead (Ubuntu) "ureadahead requires /var on root filesystem" [Medium,Triaged] https://launchpad.net/bugs/523484 [16:47] xnox: but that's absolutely not what 'start on mounted' means [16:47] xnox: (It had already been ubotted twice recently so I didn't think it was necessary ...) [16:47] 'start on mounted' in this case means 'start the service as soon as any one of the filesystems has been mounted' [16:48] cjwatson: (i only read up to my latest ping =) so i didn't see it) [16:48] xnox: upstart log suggests that this was just a thinko on jodh's part and he meant 'start on filesystem' - which seems much more reasonable to me [16:48] Scott's suggestion of moving to $MOUNTPOINT/.ureadahead might perhaps have worked then but unfortunately doesn't work here, since it requires the fs itself to be writable [16:48] jdstrand: bug #1235337 for you :) [16:48] bug 1235337 in signon-ui (Ubuntu) "signon-ui ships files in a wrong directory" [High,New] https://launchpad.net/bugs/1235337 [16:49] cjwatson: well, that would just mean somehow pregenerating the pack file along with the ro filesystem :) [16:49] slangasek: thanks! :) [16:49] lool: is there a bug # for the question of apport not picking up /var/crash? [16:49] lool: I think this is an upstart bug, I want to make sure it gets properly in /etc/init/upstart-file-bridge.conf [16:49] slangasek: per each device?! [16:50] slangasek: I think per-device profiling is a better and more maintainable solution [16:50] slangasek: or ship it as addon and make initramfs / inital bootstrap pre-do it? but upon system image updates we'd need to reprofile.... [16:50] xnox: no, that wasn't a serious suggestion at all [16:50] sorry, I apparently need to make my trolls better [16:51] ugh. signon packaging === gatox_lunch is now known as gatox [16:51] slangasek: don't know of a bug [16:54] slangasek, we used to do that for isos [16:54] tedg: https://bugs.launchpad.net/ubuntu/+source/unity-lens-applications/+bug/1229827 [16:54] ages ago [16:54] Ubuntu bug 1229827 in unity-lens-applications (Ubuntu) "Launching click installed apps from the dash runs multiple instances" [Undecided,Confirmed] [16:54] ogra_: if we did, it was terrible and we stopped :) [16:54] well, not exactly ureadahead, it was some squashfs trickery from tollef [16:54] heh, yeah [16:55] right, "we used to do " is usually a good indicator that it was a bad idea ;) === dandrader_ is now known as dandrader [16:55] tedg: this is based on saucy's upstart-app-launch [16:55] like inifinity's "we used to visually notify users to reboot" ?! =) [16:56] slangasek: ^ [16:56] tedg: I wonder: should we be treating x-singleton as defaulting to true? [16:56] tedg: I expect click apps wont have it set [16:56] slangasek: I wonder how terrible implementing Clint's suggestion in 523484 would be [16:57] Moving /var/lib/ureadahead somewhere non-FHS-compliant would perhaps fix that bug for desktop users, but at the cost of more space taken up on /, and it wouldn't fix the touch case [16:58] If we come up with an approach that fixes the touch case with no initramfs code involved, then that should be basically the strictest possible environment, and will therefore definitely fix 523484 [16:59] * cjwatson -> dinner [16:59] * slangasek nods === alan_g is now known as alan_g|EOW [17:09] popey, who can get this in? https://code.launchpad.net/~sergiusens/ubuntu-weather-app/click-in/+merge/189381 [17:10] bfiller: sorry, I was wrong asking for apps; I was looking for tests [17:15] jono, ping..... hi, have you seen my comment here: https://bugs.launchpad.net/click-update-manager/+bug/1235003 or my email? [17:16] Ubuntu bug 1235003 in Click Update Manager "Updates not showing" [Critical,Triaged] [17:19] sergiusens: you removed ubuntu from the name but the desktop file still has it? [17:20] popey, yeah, desktop file name is sort of useless [17:20] k [17:20] popey, when provisioned it's set to com.ubuntu.weather_weather_[version].desktop [17:20] right [17:21] popey, if I change the name, I'll need to fiddle with the debian packaging which I don't want to :-) [17:21] mhall119: can you top approve that? [17:21] Do you guys know if the 'import "components"' is indeed required in an Ubuntu Qt QML app? isn't import Ubuntu.Components 0.1 sufficient? [17:21] popey, oh wait, need to check the autopilot stuff ne sec [17:22] popey, now it's good === jdrab_ is now known as jdrab === psivaa is now known as psivaa-afk [17:36] popey, https://code.launchpad.net/~sergiusens/ubuntu-filemanager-app/clicker/+merge/189389 [17:39] done [17:47] hi is there any info on how too get touch on a galagzy note 3 [17:48] popey, https://code.launchpad.net/~sergiusens/ubuntu-calculator-app/clicker/+merge/189392 [17:55] hi is there any info how to get touch o a note 3 [17:58] !devices | FK_nero [17:58] FK_nero: You can find the full list of devices, official images, community images, and works in progress at https://wiki.ubuntu.com/Touch/Devices [17:58] jdstrand, seems clock added features along the run... http://paste.ubuntu.com/6193275/ [17:58] FK_nero, there should be a page for the note 3 ... and a link to the xda forum etc [17:58] jdstrand, another issue that happens when we delay click :-/ [18:00] popey, https://code.launchpad.net/~sergiusens/ubuntu-clock-app/clicker/+merge/189395 [18:00] k [18:03] popey, so is mhall119 the only top approver? [18:03] sergiusens: no, the community devs can too, mihir did one [18:04] as can I.. but y'know [18:04] popey: i just did :) [18:04] popey, feel free to delegate, I'm reaching to you as the middle man; would be good if the devs ended up approving [18:05] sergiusens: I just approved your MR :) [18:05] so they know what's going on (even though I've been doing this behind their backs for a whiel :-)) [18:05] mihir, thanks [18:05] sergiusens: you are welcome :) [18:05] dpm does to but he's on vacation [18:06] popey, yeah, saw, Qt dev '...' [18:06] Saviq, hey [18:07] Saviq, or mzanetti wrt to the launcher, how deep in the code is the defaults apps in there now? [18:07] if I add something that doesn't exist, it's a nop, right? [18:08] sergiusens, dconf [18:08] Saviq, is it in lp:unity8 ? And for click apps, do I need the full APP_ID or just partial? [18:09] Saviq, since APP_ID includes version (as does desktop file) [18:10] sergiusens: jenkins bot failed could you just check once? [18:11] mihir, I didn't change any logic :-/ [18:12] mihir, from the tests something just went wrong http://91.189.93.70:8080/job/generic-mediumtests/684/testReport/junit/ubuntu_calculator_app.tests.test_simple_page/TestSimplePage/test_hide_calc_keyboard_with_mouse_/ [18:12] sergiusens: yeah I just checked :| i need to ask boiko [18:13] mihir, I wonder how the previous MR got in [18:13] boiko: can you help !! ? === claydoh_ is now known as claydoh [18:14] mihir, oh, I know... SDK [18:15] lool, what's the ETA on the new ui-toolkit? [18:16] sergiusens: I'd say Monday [18:16] sergiusens: its in ubuntu-touch-settings [18:16] lool, bummer [18:16] Laney, nice work on the timezone selection [18:16] mzanetti, thanks [18:18] sergiusens: uhh, the clock is not allowed to use gsettings [18:18] mihir, hmm, come to think of it, it can't be the sdk [18:18] no apps are [18:18] jdstrand, not sure why they have... popey ? [18:18] sergiusens: yeah I am looking into it :) === kentb-afk is now known as kentb [18:18] mihir, if you have a fix to merge first, feel free to do it and reapprove this one after [18:19] plus the e-d-s access is not in a policy group [18:19] network manager I expect and am not worried about [18:19] sergiusens: okay :) [18:19] (we'll silence those denials soon) [18:20] 697 1002 20 0 1192 256 176 R 99.8 0.0 7:27.37 brcm_patchram_p [18:20] whee, so that's back to munching the cpu it seems [18:20] jdstrand, well I'm trying to get these apps in the image instead of the debs today; this should avoid these mishaps [18:20] sergiusens: cool, thanks === dandrader is now known as dandrader|afk [18:22] hey, looking at installing UT on my galaxy tab 2 10.1 but the only guide I can find is from Feb on xda devs. should I just follow from the wiki? [18:24] is there a standard way to file bugs against stuff on ubuntu-touch, to provide reasonable info and such? trying to use ubuntu-bug on the device seems like it will result in some very weird problems [18:29] thanks mihir [18:29] popey: something went wrong with jenkins auto pilot [18:29] mihir: poke fginther [18:29] gently ☻ [18:30] mihir, which build or mp? [18:30] fginther: https://code.launchpad.net/~sergiusens/ubuntu-calculator-app/clicker/+merge/189392 [18:30] lool, We assume click apps are single instance [18:30] lool, Looking at the bug [18:32] mihir, yeah, sergiusens pinged me so I had a quick look. It didn't make any sense. The logged showed that the expected packages were there. Try re-approving to rerun it. I will hopefully have time to take a closer look later [18:33] fginther: okay , because till last MR there wasn't any failuer :| [18:34] sergiusens: happroved [18:34] afk for 10 [18:37] popey: is that fine without Jenkins pass ? [18:40] tedg: hmm indeed [18:40] lool, mzanetti might be able to have a look [18:40] Saviq had to EOD [18:40] I imagine, considering it was 9/24, that it wasn't using u-a-l [18:40] lool, can you point us to the bug? or issue? === dandrader|afk is now known as dandrader [18:42] olli: Saviq thinks it might be https://bugs.launchpad.net/unity8/+bug/1235268 [18:42] Ubuntu bug 1235268 in unity8 (Ubuntu) "Unity8 hangs when a category is expanded and then collapsed in the Apps scope" [High,Triaged] [18:42] I'll try reverting that [18:42] aka https://code.launchpad.net/~aacid/unity8/revert_r388/+merge/189356 [18:43] mzanetti, that's what Saviq mentioned on his way out [18:43] lool, Saviq offered to look tomorrow if still needed [18:44] I'm testing that [18:44] olli: right... the one with SF seems to be this one [18:44] I didn't have to rebuild, will take 1mn [18:44] olli: the one with Mir is the display thing I mentioned [18:45] olli: that branch is approved, which means it should automagically stop once the upstream merger is done [18:45] mzanetti: it's not that :-( [18:45] unity8 is still 50% CPU here [18:46] after reverting the QML files [18:46] oh [18:46] ogra_: I'm stress tesing mtp transferring 3.8GB of music across so I can stress the hell out of media scanner over the weekend I bet it ends in tears :D [18:46] lool, 50% is better than 100% ;) [18:46] olli: it was 50% for me :-) [18:46] when screen is off [18:46] davmor2, heh [18:47] olli, its often enough at 100 [18:47] ogra_, yeah I know [18:47] i cant hold my maguro anymore so hot is it [18:47] but at least it is the same on SF and Mir [18:48] mzanetti, Saviq: So unless you have other suggestions, I am installing the old unity8 + indicator-network [18:48] then will want to upload these [18:48] greyback, ^ [18:48] lool: no, I don't really know more right now [18:50] olli: with respect to what? The CPU usage? I suspect a Mir problem , so kdub is looking. [18:51] greyback, k, just pinged you as mzanetti had mentioned you in a /q [18:51] olli: ah ok [18:51] greyback, CPU usage under SF being a Mir problem is a curious thing though [18:51] ogra_: then it's something I wasn't aware of. Bug> [18:51] ? [18:52] mzanetti, greyback, just chatted with lool, he is going to revert u8 for #81 [18:52] ok [18:52] and we can debug in non-firefightmode [18:52] greyback, no bug open yet i think, but unity is hogging the CPu in the same way for me on maguro no matter if SF or Mir is used [18:52] doing a late Fri night drill won't get us much further [18:53] ogra_: interesting, I'd not noticed that myself yet. I'm getting something like 5% usage when idle, which is bad [18:53] on maguro that is [18:53] greyback, with image 81 ? [18:53] olli: I'll have to revert unity-notifications too [18:54] ogra_: ah 80. [18:54] greyback, i get something like 47% when idle ... and at some point the UI completely dies and top shows above 100% for unity [18:54] and the phone glows [18:54] ogra_: yeah not good [18:54] * greyback flashing to see [18:55] greyback: want a bug? [18:56] (not sure what I'd put in it aside from "unity eating 25% cpu on image 81 on mako" tbh) [18:56] haha getting dpkg errors now [18:56] popey: as in to fix? Cuz I'm busy :) But a bug number would be handy, in case I determine anything in profiling later [18:56] kk [18:57] popey: we've already got this: https://bugs.launchpad.net/ubuntu/+source/mir/+bug/1235190 [18:57] Ubuntu bug 1235190 in Mir "[mako] Unity8 on Mir got slow" [Critical,Confirmed] [18:57] greyback: thats on mir [18:57] I'm not using mir [18:57] oh dammit, I keep forgetting, sorry [18:57] np ☻ [18:57] * popey files bug [18:57] * greyback needs pub [18:57] and doesn't boot now *sigh* [18:59] bug 1235408 fwiw [18:59] bug 1235408 in unity8 (Ubuntu) "Unity 8 on image 81 on mako consumes lots of cpu" [Undecided,New] https://launchpad.net/bugs/1235408 [19:00] * lool reflashes [19:00] greyback: hey, in qtcreator, do you know where the code lives for the ubuntu showcase gallery? [19:01] achiang: it's part of lp:ubuntu-ui-toolkit project. It's not part of qtcreator afaik [19:01] cjwatson: I *fear* we might have an issue with dpkg; on latest touch image, I remounted read-write, downloaded old debs, dpkg -i -O *.deb; I got a bunch of "no file list for package xyz, assuming it's installed" errors; first time I see this; either that, or fs corruption, but I hadnt played that much with this install, and most were fixed recently [19:01] greyback: derp, i think i found it: /usr/lib/ubuntu-ui-toolkit/examples [19:02] lool: I saw something like that a while back. I'm going to go ahead and be very sceptical that it's actually a problem in dpkg itself, though [19:02] yeah [19:02] fair enough [19:02] there was a recent merge [19:02] I'm being paranoid at this time [19:02] I saw it before that [19:02] I'm pretty sure [19:03] achiang: yes you're correct [19:03] lool: Indeed, I definitely saw it before this merge - I haven't actually booted a touch image since then [19:03] I assume that /var/lib/dpkg/info/.list are in fact missing [19:04] I've already reflashed cause it didn't boot anymore [19:04] so fs corruption is indeed likely [19:04] Yeah, that sounds very much like dpkg just being a canary for some other problem [19:04] stgraber: ^ might be another case of not unmounting properly ... after our recent job changes perhaps? [19:07] lool: do you know why 'ubuntu-location-service-examples' is installed on the phone? that's a rather strange name for a package that ships an upstart job [19:08] slangasek: I dont on top of my head [19:08] slangasek: someone trying to pass the clever filters? ;-) [19:09] slangasek, iirc that was added by tvoss or on his request with the rest of the location bits [19:09] greyback: do you know why i might be hitting this error: 'AnimationCircle is not a type' [19:09] greyback: i copy/pasted some qml from the sample into a toy app i have [19:09] but didn't copy any qt code around [19:10] ogra_: and nobody demanded to know why "examples" were being installed on the phone? :) [19:11] achiang: have you "import Ubuntu.Components 0.1" at the top of the file? [19:11] greyback: yes [19:12] mzanetti, Saviq, olli: Confirm that reverting unity8, unity-notifications, indicator-network to their previous versions works; no CPU pegging [19:12] going to upload reverted sources now [19:13] achiang: tbh I can't find that component in the SDK. I'm seeing if it was renamed or something [19:14] achiang: ah it's not part of the SDK, it's part of the examples. There's a file /usr/lib/ubuntu-ui-toolkit/examples/ubuntu-ui-toolkit-gallery/AnimationCircle.qml you'll need in the same directory as your qml [19:14] greyback: ah! thanks [19:19] slangasek, that was at a time where the image was still full of examples [19:22] mhr3_: do you know if it's possible to hide an element of a GenericPreview for example, via a libunity API call? [19:24] ogra_: ok. well the reason I'm asking is because the upstart job is broken and has no 'stop' rule, and the package as a whole looks suspect. Where should I file a bug to get this looked at? [19:24] well, against the package i guess [19:24] ev: so I have the file bridge problem fixed here, but /etc/init/apport-noui has plenty of problems [19:24] ogra_: against the package, to ask for it to be unseeded? [19:24] i dont think we make any use of that stuff anyway [19:25] slangasek, ah ! against ubuntu-touch-meta [19:25] :) [19:25] ok [19:27] ogra_: bug #1235419 [19:27] bug 1235419 in ubuntu-touch-meta (Ubuntu) "seed should not include ubuntu-location-service-examples" [Undecided,New] https://launchpad.net/bugs/1235419 [19:28] triaged [19:28] ev: so, a) there's no /etc/apport/autoreport in the image, which means autoreporting doesn't happen; b) the job has an 'instance', which means it will spawn the job once in parallel for each file it finds, even though the command it runs is 'whoopsie-upload-all'. [19:38] Hey all [19:39] Are the images still read-only filesystems for the most part? [19:42] is there a best way to dualboot mako and utouch ? [19:43] ogra_: I'm upset mtp from saucy to the phone didn't die :'( [19:44] davmor2, we can fix that ! [19:44] is there anyway from the cli to get the rotation mode (portrait/landscape) of the device? [19:44] bryanagee, the official ones are, yes [19:44] Why is that? [19:44] bryanagee, community ports are not [19:45] Thanks ogra_ [19:45] ogra_: \o/ I was just gonna use a bigger hamer but if you can break it :D [19:45] ogra_: Monday when I flash it I'll try the same thing from Rhythmbox that should kill it :D [19:46] if it works at all :) [19:54] ogra_: well that's the fun in trying to break these things right :D [19:54] test I meant test honest [19:59] ev, lool: bug #1235436 [19:59] bug 1235436 in apport (Ubuntu) "/etc/init/apport-noui.conf is non-functional on the phone" [Undecided,New] https://launchpad.net/bugs/1235436 [20:00] lool: do you know the details of the decision to enable auto-reporting on the phone? I don't want to just create this file, in case there was some reason this hasn't been done yet by default [20:06] Ubuntu Touch Dual-Boot Android | http://askubuntu.com/q/353783 [20:08] Got an old Nook Color (encore) off of Fleabay, back in the porting game after I got an older version of touch working a while back and my old hardware fried. [20:08] slangasek: it's best to talk to ev + Cc: pitti on these questions; I don't have the rationale for the current approach [20:08] lool: ok, will wait for ev to be around [20:09] Did I miss anything cool? [20:09] slangasek: and thanks for looking into that, the analysis seems indeed correct; also did not get to the bottom of why it didn't trigger so far, might be due to upstart-file-bridge not picking it up [20:10] lool: isn't the absence of /etc/apport/autoreport "the bottom of why it didn't trigger"? [20:11] slangasek: quite possibly [20:11] lool: well, I'm saying, if you don't have that file then it obviously won't trigger ;) [20:11] but is it enough to trigger? :-) [20:12] hmm I have no crash here to confirm [20:13] slangasek: I touched /etc/apport/autoreport and then /var/crash/foo.crash, but cant find any trace that it has run [20:14] lool: did you get the fixed upstart file bridge? [20:14] that's not in an image yet [20:14] lool: you can simulate by stopping and restarting the file bridge [20:14] this is with upstart 1.10-0ubuntu4 [20:14] yeah, not fixed, I only just uploaded the fix :) [20:14] ok [20:15] slangasek: I'm going to reroll an image pretty soon, is there any risk of regression from the upstart upload? [20:16] lool: ofono will start slightly later in the startup sequence due to the upstart-file-bridge ordering fix; should be safe [20:21] slangasek: ok thanks [20:22] slangasek: BTW got another fs corruption today after installing packages [20:22] with image #81 [20:23] lool: and you think this is a shutdown problem? [20:24] slangasek: getting dpkg errors that some pkg lists are missing [20:24] after dpkg -i some packages and rebooting [20:24] lool: ok. so you set it rw, dpkg -i, didn't set it back ro again, and rebooted? [20:25] does v. 81 have stgraber's fixes for mounting everything ro that the rootfs depends on? [20:25] couldn't set it back to ro [20:25] some running processes open some file read-write after a while in rw mode [20:25] did you see what was holding it open? [20:25] no [20:26] in the past, multiple things did, but I haven't checked recently [20:26] slangasek: I never managed to get to the bottom of this because of this since I can't get any logging information regarding the state of the system right before reboot. [20:26] then on a subsequent reboot the phone didn't boot and I reflashed it [20:26] slangasek: adbd dies pretty early on, we don't have a serial console available and last_kmsg is a joke that never works for me [20:26] stgraber: there's a way to get a serial console on nexus4 BTW! [20:26] with the audio jack [20:26] stgraber: ah. nobody was able to answer the "where did dmesg go" question? [20:27] slangasek: right [20:27] I think ricardo S has one [20:27] the obvious way around is to write that log to the fs, but since fs read-only remount is precisely what we want to debug, it's not that easy without the debugging script changing that part of the shutdown sequence... [20:28] stgraber: you can write to the fs up until the ro remount, and then if the ro remount *fails*, you can write some more ;) [20:29] we probably could push something that does "umount /var/log/ ; mount -o remount,ro / ; mount -o remount,ro /userdata ; sync" and that should fix the problem lool is seeing, but that's not a real fix and I'd prefer to understand exactly what's going on. [20:30] lool: right, so after touching /etc/apport/autoreport, apport-noui runs... and spawns a bunch of gdbs in parallel that crush the system. [20:30] so I guess that means it's "working"? [20:30] slangasek: \o/ [20:30] slangasek: it also means we dont want it right now :-) [20:30] I'm pretty sure our current scripts don't remount everything properly and that we need to add a remount,ro /userdata after / is read-only (since otherwise it'll fail) [20:30] oh, better yet, all three gdbs are processing *the same coredump* [20:30] just writing the .crash files is already killing these poor systems [20:30] lol [20:31] because of that 'instance' bug I mentioned. :P [20:31] right [20:31] I hope apport writes the crash in a tmp file *NOT* named .crash [20:31] and renames it to .crash [20:31] or it will trigger before the write is complte [20:31] yeah, I hope so [20:31] but I don't know offhand... that's a pitti||RTSL question [20:37] lool: and I checked, and no it doesn't do atomic creation of .crash [20:37] straightforward to fix in /usr/share/apport/apport, but needs doing [20:44] slangasek: thanks for checking [20:52] dobey, there aren't too many of those, are there? [20:59] mhr3_: well, we'd like to hide the ubuntu shape in the GenericPreview, as we aren't using that alement, but it remains visible and empty, which makes our errors look ugly (as we're using genericpreview for error display, as there is no errorpreview) [20:59] 1 [21:05] dobey, i was recently asking design for previews dealing with errors and the only reply i got was that it should be like https://docs.google.com/a/canonical.com/document/d/1qjnUirxkswgCvWRT5ervfQoR03zs-7HIIMI2e3Lvuqs/edit# [21:05] Unpacking replacement upstart ... [21:05] dpkg: error processing /var/cache/apt/archives/upstart_1.10-0ubuntu5_armhf.deb (--unpack): [21:05] dobey, but anyway, no there isn't a way to hide an element from the scope api [21:05] slangasek, cjwatson: getting this again; I took care to "sync" after last manual installs I had done, but other writes might have happened later sadly [21:05] dpkg: warning: files list file for package 'python3-gi' missing; assuming package has no files currently installed [21:06] etc. [21:07] indeed the files are missing [21:07] lool: ack. so these are files related to the packages you just installed? [21:07] no [21:07] mhr3_: ok. alecu filed a bug, that gatox will provide a branch for, to at least hide the ubuntushape header in the generic preview, when it is empty. which will at least make our errors less ugly for the time being [21:07] many many failures [21:13] dobey, each and every preview has an image, design will not ack that [21:15] mhr3_: it's the only remaining reasonable course of action to resolve the issue for 13.10, i think. it's way too late to be adding a new preview and new API to libunity, for 13.10 [21:15] dobey, image with an error icon? [21:16] dobey, as i said, you're not the only in need to show errors [21:16] mhr3_: i don't see how that will make it not be ugly. [21:17] dobey, it's not going to make it any less or more ugly than hiding the image [21:18] plus it's going to be consistent [21:19] anyway, i'm not here to argue about it. i just wanted to know if there was another way to resolve the issue for us via libunity. [21:24] Does Ubuntu Touch support dockig? | http://askubuntu.com/q/353812 [21:28] what's the best project/package to file bugs against, for big general interactivity things in ubuntu touch? unity8? [21:28] for example, application/window management [21:31] HELLO I NED HELP WITH INSTAL [21:40] Could anyone help me here please? My Nexus 4 is no longer booting up to ubuntu touch, and is not responding to adb. All I got access to is fastboot and the cwm ubuntu touch gave me on install. ADB sideload is not showing any device om my computer. Any remote chance I can have to get back to a working system? Preferably Android :) [21:46] bitethecoin: boot to recovery mode, and then flash it again [21:48] Dobey: I wish I could. I flashed it from phablet-flash the first time, and have no contact to the storage on the phone :( [21:49] bitethecoin: you should be able to flash at the recovery menu. i've had to do it several times now on my nexus 7 [21:49] and there are instructions on the wiki for how to go back to android, iirc [21:49] Dobey: I have looked. From the PC with phone in recovery mode, I get no adb devices at all. [21:51] don't know then [21:52] I will try some more. And come back here again when/if it fails :) [21:52] lool: ah, so it managed to corrupt the whole directory then; un [21:52] fun [21:53] lool: so if this is with something holding / rw, it would be good to know what that is, and also good to know if the problem is reproducible when / *does* get mounted ro === jhodapp is now known as jhodapp|afk === kentb is now known as kentb-out [22:03] hi, I have tried the phablet-10.1 in my nexus7 but it has even boot [22:03] is this the right branch? [22:07] slangasek: ok === salem_ is now known as _salem [22:21] any idea? [22:24] Hi there [22:25] I have been trying to install the 3rd oct build using the manual method but ubuntu touch does not seem to boot. it just auto shuts down [22:25] please advice === hemant is now known as Guest81376