[00:16] hackersarchangel: SturmFlut SMS goes over a control channel, no need for 2G [00:17] MMS is a different story and a data context is required === _salem is now known as salem_ === ssweeny` is now known as ssweeny [02:38] SturmFlut: Well for some reason SMS is not working. It just sits there spinning like it’s trying but it never sends. [02:39] sergiusens_ thanks for that info. Is there somewhere I can look to check it out? [02:40] I’m going to figure out where the logs are and check them out. === ycheng_ is now known as ycheng === salem_ is now known as _salem === _salem is now known as salem_ === salem_ is now known as _salem === Ursinha is now known as Ursinha-afk [06:33] good morning [07:04] can someone please review https://code.launchpad.net/~pitti/ciborium/use-langpack/+merge/232989 ? [07:04] it's the same simple change as in ~ 10 other branches, but I suppose it needs a formal ack [08:17] pitti: does it matter where X-UseLangPack is defined? [08:17] I'm seeing at the source package level here but at the binary package level in account-polld === sergiusens_ is now known as sergiusens [08:18] * sergiusens checks other packages [08:19] mhall119: rtm is in constant traincon-0 yes [08:41] Good morning all; happy Great Fire of London Day! :-D [08:54] sergiusens: no, it needs to be in the Source: stanza [08:54] sergiusens: thanks for approving [08:54] pitti: yeah, I'll fix that one [08:56] np [08:58] hi, quick question; is there a tip to accelerate the x86 emulator? i see mmcqd taking most of the cpu [08:59] pitti, do yu have an idea why https://launchpad.net/ubuntu-rtm/14.09/+source/ubuntu-touch-meta/1.182 didnt migrate out of proposed ? [09:00] * ogra_ doesnt even know where to look for excuses [09:01] ogra_: http://people.canonical.com/~ubuntu-archive/proposed-migration/ubuntu-rtm/update_excuses.html [09:02] ogra_: yeah, obviously the -desktop metapackages are uninstallable [09:02] ogra_: I figure we should just forcefully copy this into RTM and skip -proposed for meta [09:02] pitti, not sure [09:03] it's no different from the current -meta in RTM, aftera ll [09:03] i suppose the block is valid and we simply miss the new dependency in the rtm archive [09:03] ogra_: the only new dep is the ubuntu-keyboard-serbian one, and that is in RTM (and is not on the excuses page) [09:04] * pitti copies [09:04] hmm, right [09:05] ogra_: thanks for pointing out! [09:10] sergiusens: actually, pkgbinarymangler just does: if grep -q 'X[[:alpha:]]*-Ubuntu-Use-Langpack: yes' debian/control [09:10] sergiusens: so that bit doesn't care; but I'm not entirely sure whether Launchpad looks at that key, too [09:10] * pitti checks [09:11] I'll fix it anyways, this is the package where I have to do the po mangling we discussed yesterday [09:12] sergiusens: ah, then you can just slip that in, yes [09:12] sergiusens: I just see that https://translations.launchpad.net/ubuntu/utopic/+source/account-polld/+imports needs some manual action, doing that now [09:12] sergiusens: but it seems it worked anyway [09:20] hmpf, still no progressbar during system image updates [09:57] ping mvo_ [10:01] hey bzoltan [10:05] mvo_: Good to see you. I hope you had a good time at debconf :) do you mind if I bug you with click chroot related questions? [10:06] alex-abreu: who would be the right person to review https://code.launchpad.net/~mvo/ubuntu-html5-theme/multiarch-fixes/+merge/227751 ? it would be nice to get this into the archive to make creating cross build chroots easier [10:06] bzoltan: I'm back, it was a good conf. I'm a bit jet-lagged, but go ahead and ask questions - I will do my best :) [10:06] bzoltan: I have a pending branch for you as well, should I nag you or zbenjamin about it? [10:07] mvo_: I am working on the autopilot (so functional) tests of the SDK tools. The point would be to offer a test suite what anybody can run on any desktop to verify if the SDK tools are OK. The biggest headache for me is to create test click chroots. [10:07] * zbenjamin hides from mvo [10:07] mvo_: Both are good :) [10:08] bzoltan: is it a headache because it take so long to build? [10:08] mvo_: the ideal solution would be to make click chroots in an alternative directory and the clean it up after the tests. [10:09] mvo_: the headache is that creating click chroot needs a native environment (can not create chroot in a chroot) .. the environment might have chroots [10:09] bzoltan: hm, how about using the click chroot with a transient overlay ? so that it installs only into a temp-write dir ? then the chroot needs to be only created once? [10:10] bzoltan: so you need a native chroot to run the test in? is that correct? [10:10] mvo_: i am not familiar with the transient overlay. [10:11] mvo_: I would prefer not to use chroots at all... [10:11] mvo_: we thought about using overlayfs as well, but the point is that the host system should not be polluted. And that would create config files in /etc [10:11] bzoltan: what would you like to use instead of chroots? [10:12] mvo_: my real Ubuntu desktop [10:12] zbenjamin: right, I'm not suggesting that we use the host and pollute it, thats a no-no :) [10:12] mvo_: agreed ;) [10:12] mvo_: i would love to use the host _AND_ not to pollute it :) [10:12] bzoltan: well, thats certainly quicker than a chroot but the downside is that the developers box might be "unclean" so the tests may fail/work on his box but not on the rreal device [10:13] bzoltan: what would be the reason to use the host instead of a chroot? because its (much) quicker to setup? [10:13] mvo_: in our case, with our product (SDK Tools) the "unclean" environment is the typical target [10:13] (i.e. creating the chroot takes too much time?) [10:14] mvo_: the simple reason is that i wish to verify the SDK in real environments and not in a clean, stock and perfectly untouched environment. [10:14] bzoltan: maybe I don't quite understand then (sorry, jetlag) :) how is the unclean env the target? I thought the tests are there to verify that the app works on the device or am I missing something here? [10:14] mvo_: my aim here is not to create a black box automatic tests running somewhere in a dark lab :) [10:14] bzoltan: ohhhh, you do autopkg tests for the sdk itself? not for apps written with the sdk? [10:15] mvo_: yeps [10:15] mvo_: I tell you in simple words what I wish :) [10:16] * mvo_ is much in favor of simple :) [10:17] mvo_: a new emulatoris about to be released, or a new QtCreator is coming out. The new click package is just about to hit the archive ... etc, etc... I wish to ask those upstream developers to install their release candidate on their environment and run a simple test. [10:18] bzoltan: and the is "install new qtcreator/click with some magic, then run the test (that would build some package I assume?)" [10:18] mvo_: The SDK tests are pretty simple... create an emulator, run an emulator, create an app, open a project, make a click package, build a project, and so on. These use cases are what developers see on a daily base. The problem is that emulator/new device image/click chroots/etc are released without testing the whole app development chain. [10:20] bzoltan: well, if you make sure that all of the new components (qtcreator, plugin, click) can be downgraded with dpkg without issues, then you could simply install on the host, run tests and downgrade to the previous version if the tests fail [10:21] mvo_: the Ubuntu SDK tests are packaged with the qtcreator-plugin-ubuntu already. I just want to enable a simple way of running tests against the installation. [10:21] bzoltan: assuming you don't actually need to modify configuration [10:21] mvo_: my problem is not to install or uninstall the SDK Tools. [10:22] mvo_: Actually it is not the Ubuntu SDK what I mean to test... but the click, emulator, device image ... from the point of app development [10:22] bzoltan: those are debs as well, so you could upgrade/downgrade them too? [10:23] mvo_: like for example you as the developer of the click package release a new version of click. before releasing you could just simple run ubuntu-sdk-tests, lay back, wait few hours and see that with _your_ click release candidate the whole app development chain is still OK [10:24] mvo_: .deb packages are easy ... creating and desztroying click chroots are the problem. Kind of only problem :) [10:25] bzoltan: ok, I'm (slowly) getting there. so as part of the test you need to create a new chroot with the new tools to verify it all still works? [10:25] mvo_: creating/removing emulators is possible, creating/building/removing test apps is easy ... but creating a brand new click chroot for the test process is something i do not know how to do [10:25] mvo_: Yeps, exactly. [10:26] mvo_: like `CLICK_TEST_SANDBOX=/tmp/sdk/ click chroot -a armhf -f ubuntu-sdk-14.10 create` and it would create the chroot in the /tmp/sdk/ ... [10:26] bzoltan: ok, would you mind filing a bugreport? I think we want something like a alternative prefix (or postfix), should be really straightforward to add [10:27] bzoltan: straightforward assming cjwatson_ is ok with the approach of course :) [10:27] assuming even [10:27] *sigh* [10:27] * mvo_ can't type today [10:27] mvo_: I know cjwatson_ is super busy [10:28] bzoltan: yeah, he just needs to nod or shake his head :) the rest I will do [10:29] mvo_: I know many people's first suggestion would be to set up a virtual environment and mess around there. But that would take away the whole point, that our toolchain should work in all kind of setup. [10:30] bzoltan: adding a custom prefix/postfix to the generated click chroots maybe useful in its own right, so I'm fine with that if thats enough to unblock you [10:31] mvo_: the fundamental question if it is possible to create schroot chroots in other places than the standard. As I was looking in the manuals it is not possible. But my understanding is limited. [10:33] sergiusens, gdbus call --session --dest com.ubuntu.Postal --object-path /com/ubuntu/Postal/_ --method com.ubuntu.Postal.Post _ubuntu-system-settings '"{\"message\": \"foobar\", \"notification\":{\"card\": {\"summary\": \"yes\", \"body\": \"hello\", \"popup\": true, \"persist\": true}}}"' [10:33] that doesnt generate a message with the content i define [10:33] but only a predefined thing [10:34] ogra_: well, you would need to modify the push helper too [10:34] ogra_: to bypass [10:35] ogra_ for that Chipaca can loop you in [10:35] I thought the 'generic' helper part was landing too [10:35] k [10:35] sergiusens: no, tedg opposed the generic helper part in the strongest of terms [10:35] bzoltan: I think its not possible, it would be nice though. if you use a unique name it should be ok though. [10:35] ogra_: /usr/lib/ubuntu-push-client/legacy-helpers/ubuntu-system-settings [10:36] ogra_: there you go [10:36] discuss with tedg [10:36] ah, thanks ! [10:36] ogra_: just pass in valid json, it's currently ignored [10:36] ogra_: what're you trying to do btw? [10:36] ogra_: Chipaca it should still be possible to add a taylored part in the helper, right? [10:37] notification for disk usage [10:37] sergiusens: sure, i'm a fan of polynomial expansion [10:37] Chipaca, having a hourly cron job running df on /home and sending a message if we hit 95% [10:37] mvo_: unique name would be a compromise I could live with, I guess ... [10:37] a pretty simply thing :) [10:37] I don't like the cron part though :-P [10:37] ogra_: yeah. There's all kinds of things that should be sent from system settings, hence why i supported making it generic [10:37] sergiusens, you prefer long running daemons ? [10:38] ogra_: well a cron can miss a heavy copy here [10:38] ogra_: heavy/large [10:38] no, we have anacron, that processes missed runs [10:38] the sampling might be too low for when it matters [10:38] oh, you mean between the hours [10:38] yeah [10:39] ogra_: I mean, cron would run every hour, but you can fill up yur disk with a movie in 1 minute [10:39] but we cant constantly monitor the fs either [10:39] ogra_: I think android does that [10:39] ogra_: isn't df cheap enough to run it every minute? [10:39] without producing a lot of overheard [10:39] Chipaca, and wake up the device every minute ? [10:39] happy charging :P [10:39] ogra_: cron would wake the device? [10:39] ogra_: cron won't wake the device up [10:39] yes, once an hour [10:40] ogra_: why? [10:40] ogra_: you need powerd integration and alarms for that [10:40] it does ... well, at least on mako ... which doesnt go 100% into deep sleep [10:40] ogra_: if the device is asleep, nothing is writing to disc [10:40] ogra_: if the device is awake, it's awake already :) [10:40] anyway [10:41] Chipaca, mako never goes fully into deep sleep, it would wake up every hour as long as you have a SIM in the slot [10:41] other devices will likely be differet though [10:42] pitti, how do we monitor on the disk fillup on the desktop ? i assume we also only call df sequentially ? [10:42] ogra_: me, i'd suggest making it more often in general and working around mako's bugs only on mako, but that's just me :) [10:42] Chipaca, well, i'd like it to work on all devices and not eat any of the batteries :) [10:42] ogra_: but, anyway, tell me what json you want to send in so i key off of that to notify users [10:43] ogra_: also tell me what you want in the notification body and text, and action, and stuff. Tell me everything you want :) i'll add it to the helper. [10:44] Chipaca, thanks, will do [10:45] sergiusens, i'm wondering if that couldnt be a ciborium feature actually ... i see it is constantly running anyway [10:46] (and it kind of matches the feature being a disk tool) [10:50] bzoltan, FYI, we are waiting for one final landing on the android side to then switch to the new adbd and have developer mode off by default ... [10:50] ogra_: hmmm [10:50] ogra_: ciborium manages external drives [10:50] bzoltan, just to make sure the SDK side is ready :) [10:50] ogra_: good thing the name is generic enough ;) [10:51] sergiusens, well, it is a disk related tool ... the only one we have ... and it constantly runs so could easily sequentially call "df /home" [10:51] ogra_: but the icon will always be that of an sdcard in the notifications [10:51] sergiusens: ogra_: The branch of the SDK is ready. Please do not land the new adb without that branch. Let's test both your and my changes from the same silo once we are good. [10:52] bzoltan, wont help ... the stuff we will change is deepwer down, we will need one image build with the changes to test (since the process starts at image install time) [10:53] ogra_: OK [10:53] we are changing image defaults with the final step ... thats not testable in a silo [10:53] ogra_: sorry, was in meeting [10:53] ogra_: obviously it will mean that the SDK will not supported older images after that image is released. [10:54] bzoltan: are the sdk changes backwards compatible? [10:54] bzoltan, sergiusens, so we should pinpoint one day this week where we land the switch and are ready to roll back ... and then test the SDK with that image [10:54] bzoltan: you should be able to land regardless [10:54] ogra_: {gnome,ubuntu}-settings-daemon has a plugins/housekeeping/gsd-disk-space.c which calls statvfs() on all "interesting" mounts every minute [10:54] unless we stop caring for anything pre adb change [10:54] pitti, so not different to i.e. calling df sequentially, thanks ... [10:54] ogra_: it uses gvfs to figure out the current mounts, and ignores the virtual file systems [10:55] * ogra_ finds every minute a bit awkward for a phone [10:55] sergiusens: ogra_: i will check if I can make it backward compatible [10:55] ogra_: so in spirit it's calling df (just a bit more efficient) [10:55] right [10:55] ogra_: df still needs some filtering, like /proc, /sys etc.; /dev/shm/ is also interesting, but has a different failure mode (you need to reboot instead of cleaning up files, etc.) [10:56] pitti, well, i would only call "df /home" and split that properly [10:56] ogra_: I can call statvfs if you wat me to do it [10:56] ogra_: so statvfs()ing the /userdata partition is certainly a good first thing; and that's a rather cheap call compared to executing df and grepping its output [10:57] pitti, right ... but i dont think we want to notify the user for / ... he/she cant really do anything if it fills up [10:57] ogra_: where would that live? for anythign which is complied, calling statvfs("/home", &st) and looking at st.f_bavail is certainly easier than callin gdf [10:57] calling df [10:57] ogra_: right [10:57] (for writable space of / a syslog message we can see in bug reports would be more helpful i think) [10:58] (that would be a good "extra" :) ) [10:58] sergiusens, yeah, sounds like the easiest thing [10:59] ogra_: add me the trello task or bug so I dont forget [10:59] probably not every minute though ... 5min might be sufficient [10:59] sergiusens: since some weeks i've noticed that phablet-click-test-setup isn't working anymore because for some reason it tries to check out a branch of the application i want to test that doesn't exist. like in this pastebin: http://pastebin.ubuntu.com/8214478/ Do you know what's wrong with it ? [10:59] I wish this were a bug instead of a feature :-P [10:59] sergiusens, i can make it a bug (or part of one) [10:59] sergiusens: (i'm poking you as your name is in the top of the script file ;)) [10:59] nerochiaro: you are building locally? [11:00] sergiusens, ogra_: I think 1 minute is alright, but we should only do that while the phone is "alive" [11:00] i. e. not locked/sleeping [11:00] we ceratinly don't want to wake up the phone every N minutes for that [11:00] so mtp active or unlocked/display on [11:01] sergiusens: i have built the click locally on the desktop then installed it on the device. then i've been told to run phablet-click-test-setup to prepare the device for running tests there [11:01] sergiusens: maybe i'm out of date on the process [11:01] sergiusens: what i want to do is to run the entire test suite on a device [11:01] nerochiaro: -DBZR_REVNO might be missing, it's defaulting to "latest" which is wrong [11:02] sergiusens: yeah, but i've never had to pass that myself [11:02] sergiusens: why is it checking out the application anyway ? i've never understood that [11:02] nerochiaro: nothing has changed in like a year [11:02] nerochiaro: because it is a ci tool to test on the images [11:02] it's just conincidentially used by devs as well [11:03] sergiusens, bug 1284629 [11:03] bug 1284629 in ciborium (Ubuntu) "filling up $HOME can make the system unusable and does not warn the user" [High,New] https://launchpad.net/bugs/1284629 [11:03] nerochiaro: I would rather have people moving to pitti's work on autopackage :-) [11:04] sergiusens: i wouldn't mind that but i know absolutely nothing about it. i just want to run the AP tests for a change i made to verify i didn't break anything [11:04] nerochiaro: uninstall the click, run the command again, then install the click and shove the tests === MacSlow is now known as MacSlow|lunch [11:05] * ogra_ likes bug recycling :) [11:07] sergiusens: ogra_: so will system settings be sending notifications about disc space? [11:07] mvo_: here you go -> https://bugs.launchpad.net/ubuntu/+source/click/+bug/1364327 [11:07] Ubuntu bug 1364327 in click (Ubuntu) "Title: Custom prefix/postfix to the generated click chroots" [Undecided,New] [11:07] Chipaca, seems we can fit that into ciborium instead (at the cost of not having a freely modifyable icon though) [11:08] cjwatson_: hi -- is it possible to force architecture on a click install? [11:08] sergiusens: the secondary icon might be fixed to be the sender's app icon, but the event icon you can (and should) make to be event-related (so a sd card for an sd, and http://media.tumblr.com/tumblr_m8bpshGEPp1rs4exg.gif if it's home) [11:08] ogra_: ^ [11:09] Chipaca, you sure thats the gif you wanted to post ? [11:09] ogra_: yes :) [11:09] heh, ok [11:14] Chipaca: yeah, just that the app is an sd card icon; si it will always be confusing ;-) [11:42] thanks bzoltan1 [11:59] pitti: hey Martin; can i pass on a shameless plug for http://bazaar.launchpad.net/~oxide-developers/oxide/oxide.trunk/revision/706 from chrisccoulson ? [11:59] "seek hacker extraordinaire to craft quick test suite for new cool SSL API" [11:59] ;) === bregma_ is now known as bregma [12:04] the thought process may be a bit convoluted, but we were looking at http://www.piware.de/2011/01/creating-an-https-server-in-python/ to get started... === MacSlow|lunch is now known as MacSlow [12:07] dbarth: hah, glad to hear that a three year old quick hack is still useful :) [12:07] dbarth, pitti, oh, I wasn't asking for any help on this. I was just pointing out that pitti's blog post was the first hit on google :) [12:07] hi pitti :) [12:07] hey chrisccoulson, how are you? [12:07] pitti, yeah, not too bad thanks [12:08] how are you? I don't think I've spoken to you in ages! :O [12:08] chrisccoulson: quite well indeed, thanks; lots of phone stuff to do recently, I hope I can go back to some QA work again :) [12:11] chrisccoulson: ah ok, so as long as there is a test for the feature, that's the main point [12:11] dbarth, yeah, there's going to be tests for these features :) [12:12] pitti, ah, you're not doing QA for the phone stuff? [12:13] chrisccoulson: most recently a lot of i18n === _salem is now known as salem_ === gnu is now known as Guest28366 === zequence_ is now known as zequence [13:08] sil2100: is there a simple tool to clean the "landed" bits from the speadsheet? they make searching harder than necessary [13:09] sil2100: i. e. searching for "is someone else trying to land project foo already" [13:09] i. e. is "Archive landed requests" that tool? [13:09] pitti: sure, we have a script for that ;) [13:09] YEah [13:09] just dont wipe it accidentially :) [13:09] would it be appropriate to run that now, or do you use the "recently landed" for some weekly report or so? [13:09] pitti: no no, I'll run it now, one moment [13:10] * ogra_ doesnt have the dbus patches he still needs to land except for the linked pastebins in the spreadsheet [13:10] pitti: we didn't do that last week since we were syncing stuff still, so it was good to have the recent landings on the main page [13:11] sil2100: yeah, I don't want to break your workflow [13:11] sil2100: it's not crucial right now, I just wondered [13:11] oh look, Mr. Proper! thanks sil2100 [13:12] :) np! Thanks for poking about that, so busy with things that it really got badly bloated [13:14] sil2100: did I understand your announcement right -- I'd first land stuff in an ubuntu silon N, then you assign an additional RTM PPA, then we can add the "sync:N"? [13:14] err -- s/silon/silo/, s/PPA/silo/ [13:15] pitti: yeah, basically when you get an ubuntu silo N assigned, we automatically (by automatically I mean manually by the LT) give you an ubuntu-rtm silo with an sync:N [13:16] pitti: so you only need to bother with the ubuntu landing, the rtm one we'll fill in for you [13:16] ok, I filled in line 42 [13:19] ah, /me will fold in rvr's branch too [13:20] * pitti watches sil2100 doing the downwards line extension -- good, so I got that right :) [13:20] Hello [13:21] I've installed Ubuntu Touch on my Nexus 5, do you know where I can get some support ? [13:23] kenvandine, bug #1362143, I'm happy to change, maybe check with mpt for "Lock security" None->Swipe [13:23] bug 1362143 in ubuntu-system-settings (Ubuntu) "Change translation for "None" in Lock Security" [High,In progress] https://launchpad.net/bugs/1362143 [13:24] seb128, i think it makes more sense to me [13:25] kenvandine, wfm as well, still need a design ack ;-) [13:25] sil2100: sorry, I think I messed that up: I added vrruiz branch after you assigned the silo already; I pressed build after adding the branch, but I figure it grabs the branches at assignment time already [13:26] sil2100: so I figure that build needs to be canceled and the silo reconfigured? [13:26] seb128, would adding a leading space work as a workaround? :-) (“ None”) [13:27] mpt, I've a fix suggested which adds the context [13:27] mpt, it's just that Ken said on the review that he thinks using "Swipe" would be nicer [13:27] mpt, e.g he challenges your design decision :p [13:27] ichigo-roku: not explicitly. it's not an officially supported device. you can ask questions here of course [13:27] seb128, does french have a word for "Swipe" ? [13:28] * ogra_ guesses translating it to something useful is the hard bit here [13:28] mpt, :-D [13:28] kenvandine, a swipe may be the thing that you “Unlock the phone using”, but it is not a form of “Lock security” — it isn’t security at all [13:28] dobey, OK thanks [13:29] mpt, agreed... but i think it's commonly referred to as "Swipe to unlock" [13:29] ogra_, we would probably use "glisser" (= slide) [13:29] None seems odd to me [13:29] yeah, i wouldnt even know what to use in german [13:29] and i think needing to add context for translators is a clue [13:30] ogra_, what does android say for german? [13:30] hmm, dunno, where do i find that setting :P [13:30] * ogra_ boots an android phone [13:30] not sure in settings [13:30] but on the unlock screen [13:31] kenvandine, the context is not because the option is ambiguous [13:31] it says swipe to unlock [13:31] what does iphone say for "Swipe to unlock" ? [13:31] "finger bewegen" ... [13:31] kenvandine, it's just because "none" can be translated differently in different context, so the different instances can't be grouped under 1 translation [13:31] which would be "move your finger" ... [13:31] and yes, that is as awful in german as it sounds in english [13:31] oh i guess iphone says "slide" [13:32] seb128, i see [13:33] I don’t see how languages translate the word “slide”/“swipe” is related… [13:33] charles_: ping [13:34] kenvandine, there are several other places we could plausibly use “None”. For example, if Notifications settings were more fine-grained, “None” next to an app name might mean “can’t issue notification bubbles or sounds or vibrations”. [13:36] That might need different translations again. [13:37] mpt, it isn't... i just saw the branch and it made me think "Swipe" makes more sense to me [13:39] kenvandine, let's respect the design and land the fix like it is then :-) [13:40] seb128, i'm fine with that :) [13:47] kenvandine, thanks for challenging though. Sometimes I’m wrong. ;-) [13:47] mpt, every once in a while :) === Ursinha-afk is now known as Ursinha [13:55] so - meizu mx4 was announce - but with some shady flyme OS. When is it gonna get Ubuntu touch? [14:01] ask meizu ? [14:02] derek-g: flyme is their android fork [14:06] dobey, android fork. eww, gross, sick [14:07] it's basically just android with some different default apps/ui iirc. [14:08] mterry, on holiday? [14:08] dobey, :) yeah. don't want that. [14:08] Cimi, nope, that was a US thing just for yesterday [14:08] mterry, ok, well I am finishing up here [14:10] derek-g: http://forum.meizu.com/forumdisplay.php?fid=54&page=1 is probably the best place to discuss Ubuntu on Meizu phones at the moment [14:25] sil2100: ok, silo 1 tested and marked so in the spreadsheet; what's the magic to invoke the copying? [14:26] sil2100: just assing a silo and build? [14:26] pitti: ah, so you assigned it yourself :) Yeah, just do an assignment and build [14:26] sil2100: no, I didn't touch the RTM landing [14:26] sil2100: and you assigned the ubuntu landing silo [14:26] or what do you mean? [14:28] Oh, I didn't assign the ubuntu silo [14:28] And since you didn't, then it was Mirv who did ;) [14:28] Anyway, let me just assign the rtm one now [14:28] ah :) [14:29] sil2100: so it's test - assign RTM silo - build (? or just assign?) - wait until it hits that PPA - publish and clean the ubuntu one? [14:29] pitti: ok, silo assigned - once you press 'build' on it it will just sync the packages that are in the ubuntu silo [14:30] sil2100: splendid, thanks! I'll spread the word on the next one who runs into this [14:30] pitti: does this change not require QA sign-off? Just to make sure [14:30] sil2100: I thought trivial changes wouldn't? you tell me [14:31] pitti: depends on how the changes look like, isolated bugfixes don't require that, same for some basically no-op packaging changes [14:31] Let me take a lookie [14:31] sil2100: (btw, no RTM silo assigned yet) [14:32] pitti: huh, 2014-09-02 14:29:18,536 INFO Assigning "ubuntu-rtm/landing-004" as silo. <- the spreadsheet didn't register this yet though [14:32] ah, ok [14:32] sergiusens, rsalveti, hmm, do you remember what we had to do for flo to not make it show up offline in adb all the time ? i think i remember there was some recovery hackery we had to add [14:32] there it goes [14:33] seems krillin hits a similar issue [14:33] when dropping the hacked upstart job [14:33] ogra_: check that the keys aren't set in /data (recovery) [14:33] pitti: I'm still making this silo-sync functionality better, but for now it does everything it needs to - i.e. build the packages from the other silo ;) [14:33] pitti: but it will be more verbose and smarter soon [14:34] sil2100: https://ci-train.ubuntu.com/job/ubuntu-rtm-landing-004-1-build/15/console looks a bit scary (like it's not just copying packages, but actually building things), but I'll let it finish before making a fuss :0 [14:34] pitti: it's not using copy-package [14:34] ogra_: the keys that make it an authorized device that is [14:35] hmm [14:35] pitti: we're not doing a direct source copy because we're actually changing the version number now ;) [14:35] pitti: i.e. appending the ~rtm to the upstream version [14:35] aah [14:35] Hello, all! In latest utopic-proposed I can't type anything in the "search" field of the scopes. Also, even though I set a PIN it asks me for a "passphrase" to unlock. Known issues? [14:37] ralsina, i think mterry is actively working on the latter one [14:37] ogra_: thx [14:37] ralsina, did you set PIN via the wizard? [14:37] ralsina, or via settings? [14:37] mterry: exactly [14:37] mterry: via wizard [14:37] seb128, what did happen to the reboot option in the power off dialog? [14:38] sergiusens, sniff ... there is nothing related to adb :( [14:38] ralsina, ah... this isn't the first report about that... I've got it on my list to look at [14:38] kenvandine, intentionally killed [14:38] ogra_: so user mode adb has this issue? [14:38] mterry: awesome [14:38] ogra_: while in ubuntu or recovery? [14:38] mterry, bummer... [14:38] kenvandine, with the rationale that users shouldn't need to care about rebooting [14:38] mterry: at least the pin works s passphrase :-) [14:38] mterry, but we tell them they need to reboot [14:38] ralsina, yeah... :-/ [14:39] sergiusens, as soon as i drop the bind mount of the adbd upstart job it falls over as "offline" yeah [14:39] kenvandine, during system image update? we do the rebooting for them then [14:39] for a language change, in settings [14:39] kenvandine, oh hrm [14:39] ogra_: maybe special privs are missing [14:39] kenvandine, that was probably overlooked. Maybe file a bug with design about it [14:40] ogra_: afaik, android runs adb as root and fallbacks to user on start, right? [14:40] sergiusens, well, i wonder why the bind mounted hackish file works then [14:40] and if the ro prop is set, adb root is allowed [14:40] sergiusens, no, i only changed the upstart job, i still use the same adbd [14:40] pitti: since yeah, we had some doubts about doing srccopies all the time, since then what you end up with is basically the same version of a package with different insides if the build-deps change [14:40] mterry, i guess it would be even better if the lang change dialog would allow the reboot [14:40] ogra_: udev races? [14:41] ogra_: but now the upstart job is part of the session, right? [14:41] and adbd always runs as root ... in my new adbd setup it drops to the user for the shell but the adb daemon goes on running as root [14:41] pitti: in the very first moments of the ubuntu-rtm branching it was fine as the delta was small between archives [14:41] sil2100: yeah, the binaries would be different; while that's not in principle a problem, having a diffretn version number makes that more explicit indeed [14:41] ogra_: ah, interesting [14:41] sergiusens, no, no udev races ... i'm sure it has to do with init.rc not properly initializing [14:41] kenvandine, true... [14:41] pitti: exactly, that's why we decided to actually do this properly now once we can do it automatically by the infra [14:42] ogra_: can't think of anything, does logcat printout something useful? [14:42] mterry, do you know where I can look for an API to do that? i assume powerd dbus? [14:42] sil2100: that changelog looks weird: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu-rtm/landing-004/+sourcepub/4387111/+listing-archive-extra -- it includes *all* versions ever, while RTM alraedy has 0.2.11+14.10.20140827.4-0ubuntu1 [14:42] sil2100: so it looks like a scary change, but it's just the topmost record [14:42] pitti: don't worry about that, it's an launchpad issue [14:42] sergiusens, nopes, only sensorservice spam [14:43] pitti: if you check the diffs they look ok [14:43] pitti: I write my own changelogs to avoid that issue [14:43] (lots of it though) [14:43] sil2100: right, just looks like a wrong -v with dpkg-buildpackage [14:43] sil2100: to include all previous changelog entries [14:43] kenvandine, yeah or the system-image-dbus service [14:44] mpt, thoughts on adding a "Reboot now" button when changing language? [14:44] sil2100: i. e. not an issue with LP, but with the train (see https://launchpadlibrarian.net/183792525/ciborium_0.2.11%2B14.10.20140902.1~rtm-0ubuntu1_source.changes), but I don't worry anyway :) === dandrader is now known as dandrader|afk [14:44] kenvandine, hot tears of frustration [14:44] haha [14:45] mpt, it's a corner case... i'd imagine [14:45] how often will users change lang after the wizard [14:45] yes [14:45] mpt, but if they do, they need to restart [14:46] nik90_, pong [14:46]