[00:28] I found rotation in terminal can be painfully slow [00:28] +1 [00:31] hi guys [00:31] how is the ubuntu phone system coming along? === chihchun is now known as chihchun_afk === Nitrodex is now known as Guest83359 === chihchun_afk is now known as chihchun === Namidairo`bnc is now known as Namidairo === wilee-nilee is now known as dr_rickrolled === dr_rickrolled is now known as wilee-nilee [05:22] What is being used to currently do Ubuntu Touch image builds? [07:20] dholbach: can I adb remount && adb shell then apt-get update && apt-get dist-upgrade to keep the phone updated? [07:22] good morning [07:23] freeflying, for the time being I think so, but I'm not sure how that's going to work with the image-based updates [07:28] dholbach: thanks :) [07:28] it isnt, apt is disabled on them [07:29] but there is a developer mode that enables apt (which is a permanent change though, image updates will be disabled then) [07:30] ogra_: can I update through system-settings now [07:31] i'm not sure how far that is already [07:31] (and i'm personally not running system images yet) [07:32] uhmm [07:32] you can definitely upgrade them from commandline [07:34] by using the builtin terminal or via adb shell, guess they're the same approach [07:34] yeah [07:35] which command shall I run, btw [07:36] https://www.stgraber.org/2013/07/20/introducing-the-ubuntu-touch-image-based-upgrader/ [07:37] thanks [07:37] though thats not the default yet (as you can see [07:37] ) [07:38] with the default you need to regulary re-flash (all data in the homedir is persistent and kept over re-flashes) [07:38] if you want to be on the safe side ... [07:39] or use apt if you are brave ... but there is always the risk that you mess it up (parts of some libs live in android and are not packaged, changes to them require both sides to be upgraded) [07:39] it'll be fine, I use it for fun, not for daily using [07:40] ogra_: I did use apt over the past couple of days, so far its running ok [07:40] yeah, for daily usage the regular re-flash would be the only safe thing ... if you can bear that it might break use apt :) [07:40] watch out for libhybris and platform-api [07:40] if either of them is in the upgrade you most likely want to flash an update [07:41] trying with system-image-cli now [07:43] ogra_: hi. I wonder if (in absence of didrocks and seb128) you as a core dev could approve tvoss' packaging change so that I could publish it? http://pastebin.ubuntu.com/5987618/ [07:43] Mirv, +1 [07:44] seb128: shouldn't you be having holiday as well? :) thanks [07:44] Mirv, looks fine [07:44] ogra_: thanks [07:44] (i assume this has been testbuilt before ?) [07:45] ogra_: yes, it's via the normal cu2d system, so built and tests run, we just have a deal that a core dev must +1 in case of packaging changes [07:45] ah, k [07:45] now just France and many other places have holiday today [07:46] right, well, just go ahead [07:47] did so, that unblocks well actually everything probably (platform, sdk, unity, apps) that was waiting [07:47] well apps and unity have another run ongoing, but possibly after that [07:49] unity8 has some test failures since yesterday evening, have to look at those / file bugs [07:50] Mirv, yeah, I'm going to be off soon, I was just doing some email/IRC checking in the morning === nik90|Office is now known as nik90|Busy [08:04] hello [08:04] could i ask if i install ubuntu touch preview does it wip internal storage parition? [08:05] *wipe sorry typo [08:05] juliankreuz: it wipes the device [08:05] popey: you cannot reproduce the whoopsie 100% CPU bug now, correct? [08:05] ev: i tend to flash my phone daily, just about to flash it now, will let you know. [08:05] popey: cheers [08:06] ok thank you [08:06] if you can reproduce it, there's a version of whoopsie on its way into the archive (0.2.22) that has an option I'd like you to test. [08:06] just as a heads up :) [08:06] ok [08:07] i just released 14.1 as current btw [08:07] ☻ [08:07] just as I'm flashing --pending ☻ [08:07] (since there was one failure less) [08:07] (which is 14.1) ☻ [08:07] well, yeah :) [08:10] hi, also could i ask if there's any plans on bringing Ubuntu Touch Preview on Flo? [08:11] Flo? is that the new nexus 7? [08:11] yes [08:11] we dont plan to, but anyone in the community could go ahead and port it === schwuk_away is now known as schwuk [08:11] we don't have any devices in the company making it hard ☻ [08:13] ok, thanks for the answer. :) [08:18] So the good news is that I finally decided to try phablet-flash ubuntu-system to my Galaxy Nexus. [08:18] It even installed properly! The bad news is that there's no display now. === dholbach__ is now known as dholbach [08:33] ev: seems okay so far, been running for a while with autopilot tests running, rebooted a couple of times, no whoopsie [08:33] whoop [08:33] sie [08:33] I'm assuming you mean no 100% CPU from whoopsie, rather than whoopsie just not being there [08:33] correct === gusch_ is now known as gusch [08:33] yay [08:33] whoopsie 1544 0.0 0.1 49864 3668 ? Ssl 08:15 0:00 whoopsie [08:34] whoopsie there, but not eating silicon [08:34] Good morning all, happy Lion Day! :-D [08:36] popey: ha, you! you must know who could breath life into a QA machine that seems to have gone down? [08:37] if anyone else is still having trouble with whoopsie (hi plars), do let me know. I've got some things I'd like you to try. [08:38] I'm having hard time thinking of someone who would be awake and not on holiday due to the Assumption of Mary day (which I googled up) a lot of Europe is having [08:54] Mirv: which machine, which datacentre? [08:59] popey: QA lab intel machine, now I've progressed in finding people to the extent that it's just about guesswork on which machine to powercycle :) [09:00] ah ok [09:05] mzanetti: ping [09:12] tmoenicke_: pong [09:21] ok now we'd probably need someone physically around the machine, even powercycling does not seem to help === chriadam is now known as chriadam|away [09:31] Hey guys! [09:32] Who here is amazing with porting Ubuntu Touch? === b0bben_ is now known as b0bben === svip_ is now known as svip [10:06] ogra_: you poop glitter? Now I'm really concerned about your well being :D [10:06] hahaha [10:07] ogra_: you really need to stop eating glitter for a start off, it's not good for you honest :) [10:13] davmor2, no, no ... i *produce* glitter from munching phones indeed :) [10:16] ogra_: that's even less healthy there are toxic substances and all sorts in phones. You just stay exactly where you are till the nice men in white coats with the rear fastening jacket arrive ;) [10:16] is there any good guide to how to install cwm to a nexus 7? i somehow seem to have zapped the recovery image on mine [10:16] when i run adb reboot recovery i end up with a dead android [10:17] hehe [10:17] brendand, you can just manually flash a new recovery in fastboot mode [10:17] adb reboot-bootloader ... [10:18] wget http://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/20130815/saucy-preinstalled-recovery-armel+grouper.img [10:18] * popey stabs ueventd [10:18] sudo fastboot flash recovery /path/to/downloaded/saucy-preinstalled-recovery-armel+grouper.img [10:19] and then boot into recovery [10:31] popey: don't stab the messenger bah-dum-tish [10:33] Don't quit the day job, comedy-dave [10:38] Is there any known issue with the maguro image upgrades? [10:49] popey: Thanks to quote Milton Jones that joke doesn't normally work :) [10:50] popey: it looks to me like people are hating things that end in d more than they ever did with apps that start with p and end in audio :) === MacSlow is now known as MacSlow|lunch === LoLcAt is now known as lolcat [11:26] Can anyone one tell me where i can find a bug list for the Nexus devices? Thanks [11:30] ivanka: beautiful design work on the phone. :) === chihchun is now known as chihchun_afk [11:37] nhaines, there is a whole team working on it: http://design.canonical.com/team/ [11:38] asac, http://paste.ubuntu.com/5988662/ ... [11:38] to compare two dashboard results easily [11:40] ogra_: whats the build id? [11:40] oh got it [11:40] ./get-utah-results.sh 20130814.1 20130815 [11:40] thats how i call it here [11:41] well i needed to use the number from the URL [11:41] 3526 [11:41] and 3536 [11:41] the columns with the id at the top show the error count ... the rest should be self explanatory [11:41] why ? [11:41] the dates didnt do [11:41] it picks them from the dashboard [11:41] just returned saying they are not good [11:42] sh Development/ubuntu/dashboard.sh 20130815 20130814 [11:42] dashboard.sh [11:42] what dates did you give ? [11:42] sh Development/ubuntu/dashboard.sh 3526 3536 [11:42] works [11:42] well not really :) [11:42] err [11:42] new id needs to be last (as the error message says) [11:42] aha :) [11:42] well, i thought comparing wouldnt matter :) [11:43] just a diff i thought [11:43] nah [11:43] its more :) [11:43] hi [11:43] !!11!! 0 0 default [11:43] asac: I am only a bot, please don't think I'm intelligent :) [11:43] it compares the test results line by line [11:43] what does !!11!! mean? isnt good? [11:43] see the elgend at the bottom :) [11:43] *legend [11:43] oh now i see :) [11:43] that means you should take a closer look [11:43] i didnt get that you display for all builds :) [11:44] on ubuntu desktop the current stable 13.04 is code named Raring, what would the current version of Ubuntu Touch be called and what would be it's version number? [11:44] total changed ... cant really automatically compare then [11:44] ogra_: ignore... :) [11:44] big brain farth here [11:44] well, just run it if all tests are done and as long as there is no red we didnt regress [11:45] ogra_: would be good to also see if a test that is in A is not in B? [11:45] e.g. not yet or failed to run [11:45] like marking the line red [11:45] or something else nice [11:45] that would show as !!nn!! again [11:45] since the total was zero last run [11:45] ogra_: nice: http://reports.qa.ubuntu.com/smokeng/saucy/image/3536/default/265277/ [11:45] anyone? [11:45] :) [11:45] systemsettle has landed [11:46] so at that point there is nothing looping at least [11:46] it still needs 3 runs ... intresting [11:46] well we shoot for 99.25 idle average [11:46] thats pretty high bar [11:46] i thought i would have reduce it slightly [11:46] ShapeShifter499, there is only a development version [11:46] (of ubuntu touch) [11:47] ShapeShifter499, Ubuntu Touch will have its first release with 13.10 ... which has the development codename "saucy" [11:47] maguro also needed 3 runs [11:47] http://reports.qa.ubuntu.com/smokeng/saucy/image/3535/default/265363/ [11:47] e.g. 3 minutes [11:47] yeah [11:47] 3 minutes ?!? [11:47] ogra_, that is what I meant, thank you [11:47] wasnt it 10sec per loop ? [11:48] one run is vmstat 6 10 [11:48] 6 repeats with 10 seconds [11:48] ah, k [11:48] then it calcs the average after throwing first sample away [11:48] and sees if that average is better than 99.25 [11:48] * ogra_ knew there were 10sec in the equation somewhere [11:48] 10 runs a 6 secs :) [11:49] and we wait for at most 10 of those 1 minute runes [11:49] ogra_: if we dont succeed at all to settle we will see tops dumbed from after each run [11:49] dumped [11:49] maybe i should just always dump them even if it succeeds [11:49] if we dont settle at all we need to fail and reboot the device [11:50] (and set a reboot counter, so it fully fails after a certain amount of reboots) [11:50] ogra_: atm not... but if it fails i will ignore all the other failures [11:50] because it doesnt really make sense to investigate those [11:50] right, but having a reboot and fail count would help to see if its a race [11:51] so it helps us not hunting down autopilot failures due to noisy system (which was claimed a few times during this week) [11:51] i.e. if it passes after the reboot we found a race [11:51] ogra_: oh ... well, we reboot before every autopilot ... and we will run it on every boot [11:51] so you will get that [11:51] err, i was told we dont reboot [11:51] at all [11:51] it will initially just not be nicely visualized (might even already happen ... check the autopilot runs) [11:51] ogra_: we dont reflash [11:51] but we seem to reboot [11:52] hmm, i thought plars said we dont [11:52] well, i thought we rebooted 2 or three times [11:52] not every test [11:52] e.g. after default,sk,securit, but not after each autopilot [11:52] but i was told its really on every test [11:52] *shrug* [11:52] ogra_: i guess you should be able to find the bootlog in jenkins to confirm [11:53] ah, well [11:53] anyway... agreed is that it gets run before every test [11:53] and i was told we reboot before every test [11:53] so we get what we want :) [11:54] well, in any case looking at my script output it seems like 0815 is publishing worthy [11:54] all tests ran? [11:54] ogra_: successfully tested? [11:55] according to the script output it improved :) [11:55] termial app passes one more [11:55] adn default has a new test (your load thingie) [11:55] ogra_: +1 [11:55] have you tested locally? [11:56] yes, maguro seems fine [11:56] ... and released [11:56] plars: doanac`: so lets really run this before AND AFTER each test... [11:56] if we reboot we won't see left over runaway processes after the test otherwise [11:56] and i am currently seeing unity8 going crazy not after boot, but just after very light use [11:57] so the after will bring us such bad issues [11:58] ogra_: this tool is so nice ... can you also make it produce nice data that i can put in a spreadsheet to make nice drawings over time :)? [11:58] lol [11:59] sure, we can make it produce pie charts with candy and the like :P [11:59] ogra_: can we maybe not use yellow and !! if we grew tests and those are still perfect? :) [11:59] i guess green with a (+) [11:59] (i can indeed improve it further to get long term data for a curve etc) [11:59] or something would better reflect what happened with default [12:00] sure, i can change that [12:00] Hello [12:00] well not important... i just have to learn what it means [12:00] but ... i think everything that is green should be green here as well :) [12:00] it is just that i tink we really should check manually if new tests are added [12:00] and rather put NEW/FIXED/ etc. [12:00] since even if the failure count didnt change it might be that suddenly you have different failures [12:00] to annotate if we went green just now [12:01] so manual check is needed [12:01] or a NEW test was added [12:01] ok [12:01] well.. we should fail if there are less tests i agree [12:01] adding new tests i like :) [12:01] and i want to keep the output as simple as possible so you can easily see the improvements/regressions [12:02] yeah right [12:02] so scratch the thing about the all green [12:02] just think how to best annote what changes [12:02] you could just use NEW etc. and add a column to the left [12:02] if we want more colums and/or more functions i would rather write another tool for that [12:02] where you put those [12:02] hehe [12:02] this one is rather single purpose to make it easier to see at a glance if we got better [12:02] well whatever, what i have now is good :) [12:02] i can probably change it too ;) [12:03] so better work on more useful stuff for now === MacSlow|lunch is now known as MacSlow [12:03] it can surely grow in all directions or get new friends in a bzr tree :) [12:03] maybe later we can think about long term data, but i hope utah is not loosing anything so we can do that anytime (not now) [12:03] yeah [12:04] gema_: whats the retention policy for the data we see on dashboard? [12:04] after all ist justr a screen scraper for the dashboard [12:04] will that get wiped after a month or so? === alan_g is now known as alan_g|lunch [12:05] asac: right now we dont remove stuff === b0bben_ is now known as b0bben [12:05] asac: we are considering removing for old releases, though [12:06] gema_: how much storage do we have/use right now for the jenkins data? [12:06] gema_: also, i assume we dont do backups? [12:07] asac: this is an IS machine, I am pretty sure we do backups [12:07] asac: on both, the dashboard and the external jenkins [12:07] popey: is ueventd really looping right after boot? or does it take a while before you see this? [12:07] asac: I can check for the backup policy of the internal jenkinses [12:07] ogra_, I releasing ubuntu-touch for my device on xda [12:08] to allow open source contribution [12:08] asac: tends to be soon after boot [12:08] popey: what does soon mean? for me thats before you touch the device :) [12:08] root@ubuntu-phablet:/# uptime 12:08:13 up 1:58, 1 user, load average: 2.98, 3.08, 3.03 [12:08] root@ubuntu-phablet:/# ps aux | grep ueventd [12:08] root 473 96.0 0.0 376 196 ? R 10:09 113:51 /sbin/ueventd [12:08] like before unlock or after? [12:08] it doesnt loop [12:08] it hangs [12:08] i dont touch it [12:08] i just flash then adb shell as soon as I can [12:09] and very soon see ueventd [12:09] popey, can you chekc the end of logcat ? [12:09] ogra_: thought it eats CPU cycles? [12:09] does it talk about sensorservice ? [12:09] asac, it does .... but it doesnt loop [12:09] how do i run logcat? [12:09] command not found [12:09] /system/bin/logcat -d | tail [12:09] popey: interesting ... so before the lock screen is even there? [12:09] ta [12:09] see if there is a lot sensorservice [12:09] noise [12:10] http://paste.ubuntu.com/5988757/ [12:10] asac: lemme reboot now and test [12:10] no, looks fine thanks [12:10] ogra_: well, it maybe is in a busylock in the kernel? [12:10] :) [12:11] asac, well, its wrangling with udev about the devices ... can be busylock or some such, yeah [12:11] asac: we believe the external systems are backed up daily, the internal jenkins instances back up every hour [12:11] popey: yeah. well, we only have one dashboard boot on mako, but there it looked like: [12:11] asac: just confirmed with retoaded [12:11] its about time we get that upstart bridge [12:11] popey: http://reports.qa.ubuntu.com/smokeng/saucy/image/3536/default/265277/ [12:11] so was good :) [12:11] lets wait [12:11] for more attempts [12:12] ogra_: steve told me that we can fix our problems without that upstart part if we do what we need to do anyway for the solution :) [12:12] ogra_: from what i know rsalveti has that fix already :/ [12:12] asac, which would be ? [12:12] no, he hasnt [12:13] hmm. lets wait for chicken [12:13] steve wanted to give him th details [12:13] the proper solution would be to have something like "udevadm settle" for ueventd [12:13] but that means a lot of patching [12:13] ogra_: steve said something about stopping ueventd after its done [12:13] on the android side [12:13] yes [12:13] that thats the real fix :) [12:13] the "it is done" part is the hard one [12:13] he said if we do that alone we are already happy [12:14] you need something like "ueventd settle" [12:14] asac: http://showterm.io/0254cd08db04bb375a0a5 <- there you go [12:14] right, but that seems to not involve upstart according to steve [12:14] and rsalveit knows how to fix it [12:14] that waits until it is done and then sends the event to the upstart bridge [12:14] but afaik he didnt add it ... and we wanted to see if the upstart stuff is probably suffiicient [12:15] popey: interesting [12:15] but anyway, even if ueventd would have the patch ... we still dont have that info for udev on the ubuntu side [12:15] we need the bridge in any case [12:15] popey: and that doesnt goaway ... interesting [12:16] yeah, it will stay forever [12:16] popey: can you copy http://paste.ubuntu.com/5988768/ on your device and call it with 99.25 as argument? [12:16] i went for a walk and it was still there 40 mins later [12:16] just to see that it definitely will fail in automation :) [12:16] sure [12:16] popey: well, you can just run vmstat 5 10 and see if your idle ever goes above 90 :) [12:16] oh ok [12:17] in my experience once ueventd hits high CPU on boot, it _never_ comes back down [12:18] I've left my phone for hours and its still running [12:19] yeah, it wont [12:19] popey: right... so vmstat also shows the same idle/busyness as top for you? [12:19] ogra_: kill :) [12:19] yeah, idle is 0 [12:19] that kills the container [12:19] http://paste.ubuntu.com/5988789/ [12:20] ogra_: thats a bad thing? can we restart the container without rebooting the system? [12:20] and it works? [12:20] we can, but that takes all the other bits down with it [12:20] ogra_: could we do that transparently? [12:20] like without user noticing? [12:20] no [12:21] not even if we wanted? [12:21] killing the container will kill the session [12:21] * asac takes a note [12:21] unless you put the check before session statup ... which means a massive boot delay [12:22] oSoMoN: who is doing browser? [12:22] morning [12:22] oSoMoN: bill talked about configs we could set so it behaves nicely in low memory environment [12:22] oSoMoN: guess reducing caches etc. [12:22] oSoMoN: do you know anything about it? [12:22] * ogra_ guesses we always want that [12:23] given how choppy and sloe the browser acts [12:23] *slow [12:23] ogra_: thats intentional atm [12:23] ah, k [12:24] i know we have slowness in javascript ... didnt know that we also should expect choppy scrolling etc [12:24] (or as in case of G= no scrolling at all) [12:24] *G+ [12:26] ogra_: well, does unity8 loop for you? [12:26] no, should it > [12:26] ? [12:26] it does for me all the time [12:26] ogra_: use your indicators once :) [12:26] just pull down and look around [12:26] i think that triggers it [12:27] * ogra_ isnt near his phone atm ... i can test in ~30min [12:27] no need [12:27] davmor2, ^^^ [12:27] just saying, you might have seen that problem [12:27] nope [12:27] when scrolling was super choppy [12:27] asac: what image is that on? [12:28] asac, it is always choppy and slow [12:28] yesterdays for example [12:28] and G+ scrolling doesnt work at all [12:28] (two different bugs) [12:28] ogra_: probably also a not-optimized rendering engine? maybe there is a hwaccell setting somewhere we can turn on [12:28] which would use surfaces for scrolling etc. [12:29] well, ltes wait for Mir and see how it os then [12:29] *is [12:29] *lets [12:29] *sigh* [12:29] Mir fixes everything \o/ [12:29] yeah [12:29] "Shit canonical employees say" tumblr... [12:29]