[00:40] Hey all, I don't suppose there's any update to the porting process? [00:40] I have a question what should be the output of the flipped image build? should there be a zip file on the out/target/samsung/epicmtd or w/e? === racarr is now known as racarr|themanbur === racarr|themanbur is now known as racarr [00:45] dejello: i havent heard anything except probably by the end of this week a updated guide. been waiting the whole week to finish this port thought i am still on the building phase. === chriadam|away is now known as chriadam [00:53] Thanks, I've been waiting as well and trying to see if I could get a build working :) [00:58] dejello: what device you porting too? [01:06] Nook Tablet (acclaim) and Nook HD+ (ovation) and after I get ovation again will be the HD (hummingbird). I've had acclaim for a bit, just had ovation a day before they added the ubuntu-boot.img stuff. Admittedly I haven't worked too terribly much as I have been exhausted from work and such... [01:07] I am learning all this as I go as well, new to building and porting so it's an experience [01:07] reading, reading and more reading [01:11] rsalveti: you think .2 is good? :) === dejello is now known as dejello|ghost [01:11] doanac: maybe do some love to the QA jobs in case they are stalled :) [01:11] mako/maguro .2 [01:11] i guess [01:13] asac: needs testing still [01:15] rsalveti: yeah. lets hope :) [01:17] rsalveti: is it ok if live-build hooks are in python? [01:18] sergiusens: don't know, something to try, only used it with shell scripts [01:19] rsalveti: ok, I'm guessing yes because of the shebang, either that or I need to install curl or similar, hmmm, or cat > < EOF a python script :-P [01:20] sergiusens: why do you need to run python in there? [01:20] rsalveti: because I get urllib out of the box [01:21] rsalveti: to download click packages [01:21] oh, right =\ [01:22] rsalveti: right now I just hacked in curl to the install list, seems I have a different problem to solve first :-P http://pastebin.ubuntu.com/5882703/ [01:25] oh, weird === LarrySteeze|Away is now known as LarrySteeze [01:37] rsalveti: solved, was a permission issue [01:38] sergiusens: great [01:39] rsalveti: so if we have curl added to the ubuntu-touch meta I can get this script in to see if it can work from the builders (++ chances it would) [01:39] or use a python script, going to give that a go right now [01:39] to avoid adding unnecesary deps [01:40] sergiusens: right, as a hack you could just install curl before running your script [01:40] but yeah, if python is a go it should probably be a better way to go [01:40] rsalveti: yeah, and uninstall post, feel dirtier :-P [01:40] sure, just a temporary hack [01:57] rsalveti: or anyone is this the correct output of a working build? http://pastebin.com/mdHB7s0C or should there be a zip file somewhere in there? [02:00] annerajb: most files are correct, but in the end you should also get a zip file called cm-10.1--UNOFFICIAL-.zip [02:00] I'd probably guess that the build failed at some point [02:00] http://paste.ubuntu.com/5882777/ [02:01] this is how it looks here after successfully finishing up a build [02:02] rsalvetti: yeah it failed but every time i ran make again it worked and got me to this point so i thought running make again was fixing it. apparently is making it skip the failed target. [02:02] rsalveti: ill run again a make clean and log the places that it's failling [02:02] right, might be easy to track, just need the failed output [02:03] usually a broken dependency somewhere === chihchun_afk is now known as chihchun === salem_ is now known as _salem === Kurlon_ is now known as Kurlon === _salem is now known as salem_ === salem_ is now known as _salem [03:08] rsalveti: can you check https://code.launchpad.net/~sergiusens/livecd-rootfs/click/+merge/175178 ? [03:11] sergiusens: sounds fine, just "scares" me that we're pulling stuff from your people.canonical (in the official images), but other then that it looks fine [03:11] sergiusens: you might want to ping cjwatson to review it [03:12] rsalveti: ack, although he was the one with the original idea to copy them to my people/~ [03:12] yeah, just to get this reviewed by someone that knows more about the image policies than me [03:12] rsalveti: would be good to get a final ack from him [03:13] rsalveti: or ogra_ when he wakes up [03:13] :-) [03:13] yeah :-) [03:13] hellow [03:14] rsalveti: well if you can add a comment so the MR doesn't feel so lonely that would be nice :-) [03:15] anybody can tell me, how can i install apps and where to find it?, and any google hangout alternatives? [03:17] sergiusens: sure [03:18] rsalveti: http://paste.ubuntu.com/5882904/ It didnt require me to do make multiple times. i just typed make clean and make and it compiled on the first pass. or at least it didnt stopped building. [03:20] * rsalveti trying to open paste.u.c [03:20] sergiusens: just added a comment there [03:21] annerajb: interesting, might be because the image uses the yaffs2 format [03:22] not sure if you'd get a zip file because of that [03:22] which phone is this? [03:22] epicmtd [03:24] annerajb: what command did you run? [03:24] annerajb: mka bacon? [03:25] sergiusens: . build/envsetup.sh;repo sync; brunch epicmtd; make; [03:26] annerajb: brunch or breakfast? as brunch would do the build and not need to run make after [03:26] annerajb: do a 'make bacon' now [03:26] hm, but cyanogenmod produces a zip, something might be missing in our env [03:26] sergiusens: there's the entire build log in http://paste.ubuntu.com/5882904/ [03:26] i ran mka showcommands and it didnt show the commands, but let me run make bacon now. [03:27] rsalveti: yeah I'm seeing it [03:27] make otapackage might be useful as well [03:28] rsalveti: make has droid as the default target [03:28] sergiusens: this is the last lines of the mka bacon [03:28] i thought that was fixed on a commit done friday [03:28] rsalveti: you need bacon for the zip [03:28] sergiusens: right, bacon calls otapackage first [03:28] bacon is also the target when doing brunch [03:29] sergiusens: also found this after the last lines make: *** Waiting for unfinished jobs....No command 'arget' found, did you mean: [03:29] arget? [03:29] this is a fresh install so there may be missing dependencys dpkgs [03:30] yeah, you should get 'Running custom releasetool' or 'Running releasetool' printed at least [03:30] let me save the output of mka bacon [03:30] build/core/Makefile, check the INTERNAL_OTA_PACKAGE_TARGET rules [03:31] sergiusens: this is the output http://pastebin.com/wAi6EpQ7 [03:31] rsalveti: ack [03:31] ah, if that's disabled it may be why no zip [03:32] annerajb: isn't that trimmed? [03:32] should have more [03:32] maybe it just stopped [03:32] while it's downloading the initrd [03:32] sergiusens: doubt it (thought it's not a make clean); i ran mka bacon > output.log [03:32] rsalveti: i see what you mean. [03:34] rsalveti: sergiusens i confirmed if i run mka bacon it ends at http://pastebin.com/Ky5SA5h9 [03:35] now need to debug why the ota rules are not being used [03:35] might be disabled by your device config files [03:36] rsalveti: where is that file? [03:37] rsalveti: this is the output of mka otapackgage http://pastebin.com/36pb3W41 [03:37] device/$vendor/$device [03:38] check if it's setting TARGET_RELEASETOOL_OTA_FROM_TARGET_SCRIPT for example [03:38] but that would still echo something [03:38] would need some not so fun makefile debugging [03:39] rsalveti: grep -RiI 'TARGET_RELEASETOOL_OTA_FROM_TARGET_SCRIPT' BoardConfig.mk:TARGET_RELEASETOOL_OTA_FROM_TARGET_SCRIPT := ./device/samsung/epicmtd/releasetools/epicmtd_ota_from_target_files [03:39] mka showcommands otapackage [03:41] that's new to me [03:41] rsalveti: it unhides all the hides [03:41] cool [03:42] quite verbose indeed [03:42] fginther: https://code.launchpad.net/~sergiusens/cupstream2distro-config/messaging_and_dialer/+merge/175184 [03:42] that should help [03:42] sergiusens: http://pastebin.com/MS2ExEkq [03:42] fginther: some will fail to build, but we can't land in PPAs anymore [03:43] ake: *** No rule to make target `/home/annerabj/android/out/target/product/epicmtd/ubuntu-boot.img', needed by [03:43] cm_epicmtd-target_files-eng.annerabj.zip [03:43] annerajb: rsalveti seems a zip is possible [03:43] yeah, this is the issue [03:43] I say repo sync if you haven't in the past 6 days [03:45] and then run mka showcommands bacon [03:45] that will show if it was able to download the package successfully [03:48] sergiusens: i repo sync quite a few times in the past few days since i installed this vm yesterday. but ill do it again [03:48] oh, vm [03:48] annerajb: make sure it has networking and that you can reach launchpad [03:49] sergiusens: it has network because repo sync donwnloaded stuff which launchpad url should it reach the http one? [03:53] annerajb: if when you do a . build/envsetup.sh you get a hybris download, you are ok [03:53] * sergiusens is off to bed [03:53] way past bed time [03:53] same here thanks for the help [03:55] annerajb: after repo sync, paste the output of 'mka showcommands bacon' [03:55] that should help finding the issue as well [03:57] rsalveti: this is the output of mka showcommands bacon after the repo sync http://pastebin.com/TS9Ebi5a [03:58] * annerajb be back in 15 minutes [03:59] yeah, it's trying to use ubuntu-boot.img before downloading the file [04:00] rsalveti: this is the latest message i got from you since i DC [23:59] yeah, it's trying to use ubuntu-boot.img before downloading the file [04:00] BRB [04:00] which was in theory fixed with http://phablet.ubuntu.com/gitweb?p=CyanogenMod/android_build.git;a=commitdiff;h=93a7af46f77e0f05187b9aabefa6c72b37a49511;hp=38f399e8bd0292024cdae7cabbdab5aa04fa141b === chihchun is now known as chihchun_afk [04:17] rsalveti: back === annerajb__ is now known as annerajb [04:18] annerajb: which was in theory fixed with http://phablet.ubuntu.com/gitweb?p=CyanogenMod/android_build.git;a=commitdiff;h=93a7af46f77e0f05187b9aabefa6c72b37a49511;hp=38f399e8bd0292024cdae7cabbdab5aa04fa141b [04:18] can you check if you have such commit as part of your 'build' repository? [04:20] i saw that on the mailing list and i think it was there when i checked two hours ago [04:21] rsalveti: your change is on my working copy [04:23] right, your board declares BOARD_CUSTOM_BOOTIMG_MK [04:23] probably because it's creating the fs as yaffs2 [04:24] i assume that should be removed for ubuntu so it can create a zip right? [04:24] no, guess that's just a corner case I'm not yet handling properly, let me investigate that more [04:25] rsalveti: all right [04:26] annerajb: just to check if this is indeed the issue, remove line 1217 from build/core/Makefile [04:27] annerajb: http://paste.ubuntu.com/5883021/ === Namidairo`bnc is now known as Namidairo [04:27] rsalveti: and run what?mka bacon? or mka showcommands bacon [04:27] could be the showcommands [04:28] rsalveti: that fixed it [04:28] Package complete: /home/annerabj/android/out/target/product/epicmtd/cm-10.1-20130717-UNOFFICIAL-epicmtd.zip md5: 6010ed46fff7db28d07b7830297b00af [04:29] annerajb: that's why: http://paste.ubuntu.com/5883028/ [04:29] your device defines BOARD_CUSTOM_BOOTIMG_MK, which is basically then using a custom tool to build the image [04:29] that's from device/samsung/epicmtd/shbootimg.mk [04:30] hmm i see. so is that zipe file and flashing recovery all i need? [04:30] annerajb: are you trying with flipped or unflipped? [04:31] rsalveti: no idea dont know how to know which one i am using [04:32] you should probably try unflipped first, which would mean you now need to install the produced zip via your recovery [04:32] and then install the ubuntu zip, which you can get from http://cdimage.ubuntu.com/ubuntu-touch-preview/raring/monthly-06/raring-preinstalled-phablet-armhf.zip [04:33] your case is a bit special, as the file system is yaffs2 (mtd) [04:33] all right so to try the unflipped image i use regular clockwork recovery (not the recovery on the out folder) [04:34] after that install the zip i just built [04:34] and then raring-preinstalled-phablet-armhf.zip [04:34] yeah, then the ubuntu zip [04:34] let me also reply the mailing list === chihchun_afk is now known as chihchun === chriadam is now known as chriadam|away === chihchun_afk is now known as chihchun [07:18] When I try to follow the instructions on number 1 here: http://developer.ubuntu.com/get-started/ I get an error Depends: Dropping letters, anyone else seen this? === diplo_ is now known as diplo [07:50] Depends dropping-letters error when trying to get the ubuntu-touch sdk preview | http://askubuntu.com/q/320989 === alan_g|EOD is now known as alan_g === hikiko is now known as hik|afk [08:32] Good morning all, happy International Criminal Justice Day! :-D === hik|afk is now known as hikiko [08:37] Today I'm not seeing any /etc/device-services file in a nexus4 device. Is this because of a change in the image? [08:38] ogra_: Maybe you know this? ^ [08:43] jcollado, we switched to upstart user sessions (actually on monday iirc) [08:44] ogra_: I'm not sure of what do you mean, but phablet-test-run still relies on those files to enable/disable the shell.I guess I'm missing something. === jcollado is now known as jcollado_afk [08:47] jcollado, do you know if you use phablet-flash -b for provisioning the devices for the dashboard tests ? (i.e. if home gets wiped) [08:47] * ogra_ was kind of assuming that, but the test failures for the last image indicate otherwise [08:50] asac, so it looks like $HOME isnt cleared for the dashboard tests , which means there is an old .bashrc sourcing dbus info from the wrong place, i just uploaded a fix for that case (but it might not make the new build since that runs already it seems, i'll trigger a new one immediately after) [08:51] jcollado_afk, right, to disable the shelll you need to use upstart commands now [08:54] yay, my new S4 arrived [08:57] I guess "When working on flipped images, detection does not work and would require -d" is an outdated error message at this point? [08:57] got it just now because the phone preferred to stay charging and phablet-flash doesn't like it [08:57] yes, it should work now [08:58] ogra_: do we have a new image from last night? [08:58] a more appropriate message might be "If the phone is charging, shut it down and try again" [08:59] fwiw "adb reboot" is what I do to resolve the situation [08:59] popey, yes, but it has dbus issues still, and i wasnt fast enough with the fix to get it into the running build either i fear [09:00] when are builds created? from when (what time+timezone) is 20130716 ? [09:02] ogra_: ok [09:02] timp, the default cronjob starts at 08:32 UTC and runs ~50min-1h [09:03] ogra_: ok, thanks. so 20130716 is more than a day old [09:03] timp, note that there is job queuing though, which means that if a former build runs longer we start delayed [09:03] yes [09:04] there were 16.1 and 16.2 builds [09:04] ok, thanks. then probably the image that I need is building right now [09:04] jcollado_afk: ogra_: so can we do the cleaning in the test suite? [09:04] or in some infra script [09:04] sounds easy enough i fthats the case [09:04] but... did you reproduce? [09:04] ogra_: that is odd, phablet-flash downloads 20130716 for me, not .2 [09:05] https://bugs.launchpad.net/phablet-tools/+bug/1202126 [09:05] Launchpad bug 1202126 in Phablet Tools "Outdated flipped image error when flashing while charging" [Undecided,New] [09:06] timp, because phablet-flash by default downloads the blessed image from /current ... 16.2 wasnt tested, so it wasnt blessed ... running phablet-flash --pending would give you .2 [09:06] ogra_: rsalveti: so guys can you take a device, learn how to reproduce what is runon infra [09:07] and do that? [09:07] ogra_: ok. I need something that went into ui toolkit yesterday late, so I'll wait for 17 : [09:07] and in the meantime stopp the presses [09:07] nothing more going int [09:07] all stopped [09:07] asac, well, whatever autopilot does atm to stop the shell (why is that needed and not being fixed btw) needs to be switchedd to initctl/upstart [09:07] but I guess I'll need --pending for that [09:07] ogra_: "blessed" gives me shiver flashbacks of trying to boot apple's crippled EFI, with partitions required to be "blessed" [09:07] ogra_: well, stop the presses [09:07] didrocks: ^^ [09:07] asac, ? [09:07] we have to stop daily-release or something i guess [09:07] hum [09:07] what? [09:07] what ??? [09:07] ogra_: nothing more goes in until we have fixed the dashboard [09:08] we seem to still not know whats going on [09:08] more stuff flows in [09:08] we haven't backed out [09:08] asac, without the upload i just did the dashboard wont be fixable [09:08] asac: sorry, can you give context? [09:08] so please let me go on building [09:08] asac, we dont need to release the builds [09:08] didrocks: dashboard is read. autopolit doesnt run. folks broke it with a dbus related upload [09:08] they have been digging around [09:08] no success [09:08] nonsense [09:08] we had no green on dashboard since last week [09:09] asac: here, some tests are passing [09:09] didrocks: that doesnt matter [09:09] asac: well, I think ogra_ just don't rebuild any image [09:09] or publish them [09:09] sometghing might run somewhere.... what matters is that what we run doesnt work [09:09] asac, the dbus issues are clear, what wasnt clear was that autopilot doesnt run on freshly provisioned devices [09:09] didrocks: no [09:09] no more packages [09:09] only packages related to fix this problem [09:09] asac: well, you are telling we have "false red" [09:09] not "false green" [09:10] didrocks: depends [09:10] asac, so there is old config cruft in ~/ which the new upload works around [09:10] good [09:10] but [09:10] so lets wait till that stuff is in [09:10] and we hit green [09:10] when i uploaded the new daily build was already running [09:10] (and we cant stop live builders) [09:10] in the meantime we cannot accept other stuff. [09:10] so the next iage is needed [09:11] *image [09:11] * didrocks is finding the discussion puzzling [09:11] or confusing at best [09:11] asac, just calm down, its all known and being handled, just takes a bit [09:11] ogra_: well, every day all was known and handled, but then something new happened before we hit green [09:12] give the fixes a chance :) [09:12] so if now mr. Y uploads something else [09:12] and it busts you [09:12] the story continues [09:12] how are we supposed to debug if not with having the dashboard tests run [09:12] i dont upload something else [09:13] i react to a flaw in the test setup [09:13] ogra_: you are not doing that [09:13] but others [09:13] are uploading still [09:13] http://10.97.0.1:8080/job/autopilot-saucy-daily_release/469/ [09:13] so right now didrocks might just have uploaded the next bomb [09:13] or someone else [09:13] thats fine as long as they dont add additional breakage [09:13] the issues are in one specific area nothing else touches [09:13] if everything in the sdk is passing tests and that's wrong, we should really question our tests [09:13] how can we get this uynder control after 4 days of constant breakage if not by saying that folks have to stop something? [09:13] and none of the other uploads will touch that area [09:14] just looking for answers [09:14] asac: do you mind sending an email to everyone telling that we are not releasing anything anymore? [09:14] ogra_: but they might juyst shoot yoru stuff [09:14] asac, there was definitely a communication issue [09:14] asac: I don't want to be help responsible for that [09:14] as last time I asked for a freeze, everyone got crazy [09:14] well, if folks would be willing to back stuff out [09:14] I'm fine putting everyone on manual publication meanwhile [09:15] everything* [09:15] then we could continue uploading and backout everything one by one [09:15] asac: just please warn them so that I don't get a ping every 30s about "why dailies didn't happen for us?" [09:15] didrocks: sure... if we do it we warn them.. [09:15] asac, i dont know if anybody warned QA about the switch to upstart user sessions, thats the firs issue here .... the other is that a hack was dropped that exports the dbus info into non local logins since there was no knowledge that the tests depend on it [09:15] ogra_: so lets wait for this [09:15] Hey any one who is working on Calculator core app there ? [09:16] asac: so, ack on putting everything on manual publication? [09:16] if this doesnt fix it, we do it and start backing out. [09:16] asac, what jcollado_afk mentioned above might have additional impact though ... [09:16] which backing out wont solve [09:16] ok, I'll wait for asac's red alert ;) [09:16] jcollado_afk: can you coome to keyboar? [09:17] autopilot needs to be ported to use upstart session mgmt to kill the shell .... [09:17] * didrocks puts his red phone next to his laptop [09:17] seems we cant run it while the shell runs [09:17] lol [09:17] and atm they use a hack that rips unity8 out of the session startup [09:17] Mihir__: I think most app-development talk goes on in #ubuntu-app-devel [09:17] gema: is there anyone else who knows about autopilot and utah who is available right now? [09:17] as i understand [09:18] we need to teach ogra and folks how how to use utah and exactly reproduce what we see there [09:18] asac, well, i can read logs [09:18] ogra_: do you have access to the lab? [09:18] no need to run it for that :) [09:18] gema, i have vpn access [09:18] cool [09:19] no magners account though ... in case thats needed [09:19] ogra_: so whilst jcollado comes back I will show you how to find the right job [09:19] and utah command [09:19] ok [09:19] ogra_: then he can tell you how to set up your devices and stuff if it is not clear from the job [09:19] gusch: ping! Hi! [09:20] sil2100: pong [09:20] ogra_: give me an autopilot job as an example? [09:20] whichever you are trying to figure out [09:20] gusch: since today, one AP test for gallery-app seems to fail [09:20] gema, for what issue exactly ? [09:20] gusch: gallery_app.tests.test_photo_viewer.TestPhotoViewer.test_nav_bar_share_button [09:20] gusch: http://10.97.0.1:8080/job/autopilot-saucy-daily_release/470/testReport/ [09:20] ogra_: I don't know, I guess you are trying to diagnose something [09:20] sil2100: and I have no idea why [09:21] gema, not really, diagnosis should be done already :) [09:21] gusch: nothing changed in the UITK by any chance? [09:21] ogra_: or else why does asac wants it right now [09:21] what is the rush ? [09:21] ogra_: the problem is that you still guess when looking at logs. doanac explained how to run the stuff exactly like in utah yesteday [09:21] sil2100: and it's not the only one, there have been some other (random) failing ones as well [09:21] 'NoneType' object has no attribute 'visible' [09:21] gema: ogra thinks the current seen problem is now fixed [09:21] gema, one was a dbus issue https://jenkins.qa.ubuntu.com/job/saucy-touch-maguro-smoke-gallery-app-autopilot/8/console.... for which fixes are in but not in an image yet ... the other is a code issue that running wont solve [09:22] but we of course dont know what is next without runnign exactly what utah runs a few times [09:22] s/running/running utah/ [09:22] and reproducig the issue [09:22] sil2100: well - the sharing component - no idea if that got an update, or the SDK [09:22] sil2100: works fine locally for me [09:22] asac, we know what is next but i need jcollado_afk to fconfirm, then we can fix utah [09:22] gema: so that :) [09:23] asac: understood [09:23] we need someone to land stuff in utah for us. folks have been digging around for days [09:23] and when we landed a fix another breakage came etc. [09:23] asac: who is digging? [09:23] asac, in multiple different issues [09:23] everyone here [09:23] asac, dont mix them up :) [09:23] sure, just saying that if we cannot get it under control [09:23] asac: ok, what do you mean land stuff in utah? [09:23] we have to set all development to manual [09:23] mode [09:23] e.e.g nothing goes in until we fix this issue etc. [09:23] gusch: did you try with the most recent SDK and stuff? [09:23] asac, we can and we will [09:24] asac: ok, so you want people to be able to reproduce with utah locally [09:24] gema: they say they cannot see the issue [09:24] asac: not necessarily manually, but definitely on their devices [09:24] sil2100: at least with what's in saucy [09:24] i forwarded a mail to them telling how to run utah-autopilot not sure if they tried it [09:24] ok, I can work with andy in creating a video training [09:24] for that [09:24] asac: us teaching every developer one by one doesn't escale === vrruiz_ is now known as rvr [09:24] gusch: it might be a recent change, so maybe try adding ppa:ubuntu-unity/daily-build and use the SDK components from there [09:25] gusch: since I didn't publish SDK yet because I want to make sure all Apps are ok [09:25] sil2100: "file:///usr/share/gallery-app/rc/qml/MediaViewer/MediaViewer.qml:459: ReferenceError: caller is not defined" that is weired - and weired it's only for the sharing [09:26] gema: i dont know how you do that. [09:26] its just that we need to know whats going on right now [09:26] asac: we create a hangout and make it public, andy explains exactly how to run those things locally on their devices [09:26] ogra_ and rsalveti upload without having ever reproduced this [09:26] maybe we have someone like ogra or rsalveti asking questions [09:26] so its completely digging in the dark [09:26] and then we make it available to everyone for watching [09:26] i am not swure. i have sent instructions [09:26] asac: ok, then let's see if those have any effect [09:27] first [09:27] gema, asac, the current issue is that the exiting dbus fix assumes the tests run on a freshly installed device that has no pre-existing stuff in $HOME [09:27] sil2100: "ppa:ubuntu-unity/daily-build" will kill my machine - I played that game way too often - sorry [09:27] which doesnt seem to be the case, so it still fails [09:27] ogra_: you got myu mail from last night? [09:27] asac, yes [09:27] it has the instructions how to run autopilot in utah style [09:27] again i dont need to run utah to know whats wrong [09:27] ogra_: the tests right now are not reprovisioning [09:27] ogra_: but you can check now if after your upload everything is good [09:28] ogra_: and that is done on purpose to have a more realistic environment [09:28] this would calm me down [09:28] gema, yeah, there is a fix that will handle that on our side [09:28] and give me believe :) [09:28] gema, just waiting for it to hit an image [09:28] ogra_: ok, sounds good [09:28] if folks look at log, upload the symptom they see, they might hit another symptom [09:28] asac: all is good, ogra_ and I have an understanding [09:28] ok super [09:28] that things are going to be better [09:28] and I will track that [09:28] gema, but there is another issue that jcollado_afk pointed out [09:28] so will we test before upload if our upload fixes stuff? [09:29] ogra_: yes, I know, the apps crashing sometimes [09:29] gema, no [09:29] ogra_: not handling options properly [09:29] * asac goes off for a bit [09:29] ogra_: ok, another one? [09:29] to not make everyone completely zero productive :) [09:29] there is some weird bug in unity8 that forces you to stop the shell to test apps [09:29] asac: thanks ;) [09:29] what i need are new dashboard results [09:29] thats all i need [09:29] freseh, well run dashboard results :-P [09:29] due to that bug you guys have a hack in autopilot that stops the shell before running tests [09:29] ogra_: yes [09:29] * asac looks away to not see how that is done [09:30] the code for stoppin the shelll needs to be adjusted to work with upstart sessions [09:30] ogra_: do you have a bug for that? [09:30] asac, well, the underlying bug needs to be fixed [09:30] I can nag thomi and veebers to get this fixed overnight [09:30] gema, no, i didnt have a chance to file one since asac hammers me on IRC since i heard about the issue :P [09:31] ogra_: no worries he is going to be quiet for a while, make sure you file a bug that is clear and has all the info in it (or the NZ timezone will not allow this to happen quick) and I will make sure it does [09:32] right, i need to look at the code and can probably even come up with a patch [09:32] ogra_: that'd be even better [09:32] first i want to see if the fiormer fix makes the dbus issues pass though ... it is uploaded and needs a new image (missed the currently building one, so it will take another 90min or so) [09:33] ack [09:34] ogra_: I understand that you and doanac are in touch regarding starting to run tests on --pending? [09:34] yep [09:34] cool [09:34] gusch: well, if you only enable it for a moment and install UITK from it and then disable it, there's no risk [09:35] gusch: since the UITK and SDK elements in daily-build are now ready for publishing [09:35] So today you'll have the same thing in saucy anyway [09:37] sil2100: well then - there is no difference in waiting a bit ;) [09:38] sil2100: I'm in no hurry :p [09:39] gusch: but I don't want to publish the SDK if I have no 100% surness that it doesn't break apps! ;p [09:39] sil2100: argh - we really need to run app-tests when integrating SDK ... [09:40] gusch: we run, but not all of them [09:40] gusch: the ones that we run did not fail, but hm [09:40] sil2100: as part of jenkins? really? nice [09:40] Actually I wonder what could be the problem with that test [09:40] gusch: we run them daily, before daily release [09:41] would be even better to run apps tests before approving a merge for SDK [09:41] https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1192944 [09:41] Launchpad bug 1192944 in Ubuntu UI Toolkit "Run application tests when merging a branch into ubuntu-ui-toolkit trunk" [High,Confirmed] [09:43] gusch: I also poked om26er to help out, so maybe he'll know what could be up [09:46] hmm, whats up with the milaing list ? [09:46] *mailing list [09:46] * ogra_ just got a bunch of really old mails [09:46] sil2100: there is no newer SDK in the daily build === MrDHat|afk is now known as MrDHat [09:52] ogra_: somebody cleared unapproved / non-member posts? [09:52] yeah, probably [09:52] do you see it too ? [09:53] else i might have to blame my mailserver [09:53] gusch: there is: https://launchpad.net/~ubuntu-unity/+archive/daily-build/+packages?field.name_filter=ubuntu-ui-toolkit&field.status_filter=published&field.series_filter= [09:54] ogra_: [09:54] ogra_: yeah... [09:54] ah, good [09:55] ogra_: well ubuntu-phone mailing list is the one i guess you are referring to. [09:55] * ogra_ fires off a 20130717.1 build [09:55] xnox, yep, thats the one [09:55] Morning all [09:55] moin moin [09:55] how are the images this morning :) [09:56] gusch: you did an update first? Since I remember I was fixing things due to new dependencies in the UI toolkit [09:56] davmor2, ignore 20130717 and wait for 20130717.1 (currently building) [09:56] om26er: could you coordinate with gusch about the AP issues of gallery-app? [09:57] ogra_: good to know thanks :) [09:58] sil2100, I am working on fixing that test and for future I think you can ping me for test failures for Apps [09:58] sil2100 didrocks weired, now it's there ... installing [10:00] mardy, are you around and can you join us on #ubuntu-app-devel for a question on Online Accounts? [10:01] sil2100: even with this latest SDK, it works fine locally [10:01] om26er: what's the issue with that test? [10:02] gusch, the mouse moved to the share button on the toolbar but the click somehow did not register hence the menu did not appear [10:02] mostly due to slowness in jenkins [10:03] gusch, I think the workaround (fix?) would need to happen in click_item() increase in the delay before clicking will likely fix [10:04] "hey go to the share button but wait Xseconds before clicking because our VM is slow" === jcollado_afk is now known as jcollado [10:06] om26er: we should start using "self.pointing_device.click_object()" instead of the click_item() [10:06] gusch, I created that method to workaround this issue in jenkins before [10:07] om26er: no idea if that would help, but it would be the proper function to use [10:07] om26er: ah - I thought it's a part of autopilot [10:07] gusch, click_object() would result in more failures [10:11] gema, ogra_: Hi, my ears were burning :-) Are you talking about the code in phablet-test-run that stops the shell etc.? [10:11] veebers: I think ogra_ is , yes [10:12] sil2100, hey, so all these tests run in a single VM, one after the other? how much ram does the VM have? [10:12] gema, ogra_: I believe sergiusens or om26er would be the ones to talk to re: that. I think they're familiar w/ that code [10:12] om26er: if you are running tests on a VM on an overloaded jenkins server and the tests do not pass because it is too slow [10:12] veebers, gema I am supposed to fix that, once service unity8 start/stop works [10:13] om26er: the solution is not to increase the timeout [10:13] om26er: we need to talk about that [10:13] om26er: ah cool [10:13] om26er: cool [10:13] om26er, should work with the next image [10:13] 17.1 [10:14] ogra_, would it work with dist-upgrade or something android side will be needed? aka reflash [10:15] gema, I think its fine for the cases where there is only one app' tests are run. but for the cases where we run autopilot tests for the whole image, we may need to increase the RAM of the VM [10:15] dist upgrade to the latest ubuntu-touch-session (0.59) whould do it [10:15] *should [10:15] om26er: no, if VMs are too slow for this kind of testing then we test on bare metal [10:15] ogra_, I reported a bug yesterday where 'service restart ubuntu-touch-session' would break dbus somehow i.e. qdbus says connection refused [10:16] om26er: but we don't mess with reasonable timeouts that will then make failed tests pass in hw [10:17] om26er: it doesn't matter how much memory you give to the VM if the server that runs them is overloaded, that is an environmental issue that needs to be resolved outside the tests [10:17] om26er: unless you think the timeouts are unreasonable even for HW [10:17] gema, timeouts are not needed for our desktops as all tests pass. so its indeed an environment issue [10:18] om26er: then don't change them, let's figure out where to run those VM tests [10:18] gema, i mean if i run tests locally they always pass, but they do have a tendency to fail in jenkins (due to slowness) [10:18] gema, ok, great === alan_g is now known as alan_g|afk [10:18] om26er: that is a confusing statement, since jenkins is just a glorified cron [10:18] om26er: tests fail on the virtualised environment , rather [10:18] right? [10:18] om26er: do the tests pass when you run them on a VM at home? [10:18] gema, yep, I meant that [10:19] om26er: let's try to be precise with these things or else we all end up blaming jenkins or utah or the lab and get confused [10:19] gema, never tried in a VM here, always run natively [10:19] om26er: try on a VM [10:20] gema, will do [10:20] om26er: thanks [10:21] np :) [10:21] asac, ogra_: I'm back. [10:22] jcollado: I think we don't need you [10:22] jcollado: ogra_ has everything under control [10:22] gema: Ok. Thanks. [10:22] jcollado, thanks for pointing it out though [10:23] * ogra_ wasnt aware you hacked /etc/device-services, i thought you just killled unity8 === alan_g|afk is now known as alan_g [10:33] asac: gema ogra_ it's still failing because the change that was mentioned yesterday didn't happen here: http://bazaar.launchpad.net/~utah/utah/dev/view/head:/examples/run_utah_phablet.py#L142 [10:36] sergiusens, ah, yeah, the dbus-session needs to be sourced from the new location ... and it needs to use initctl to stop unity8 [10:48] I spy 20130717.1 [10:48] ogra_, hey! I have the latest ubuntu-touch-session now but: 'service unity8 restart' still says "unity8: unrecognized service" [10:49] om26er, are you the phablet user ? [10:49] you need to be in the users session [10:49] (root cant see user sessions) [10:49] tried both way. as phablet user with ssh, and over adb as root [10:49] and you used "initctl stop unity8" ? [10:50] ogra_, THAT, I was using 'service unity8 stop' now it works [10:50] (initctl list should also list all apps in the session) [10:50] yay [10:50] om26er: they are not running on VM but on real hardware, beefy machines with 4 to 8G of RAM… [10:52] hmm [10:52] phablet@ubuntu-phablet:~$ initctl list|grep unity [10:52] unity7 stop/waiting [10:52] unity-panel-service stop/waiting [10:52] unity8 start/running, process 632 [10:52] now thats really intresting [10:52] I just noted that [10:52] om26er: most of the thing which induced slowliness is autopilot itself when run with rmd [10:52] where do unity7 and unity-panel-service come from [10:52] its not a biggie since they dont run, but something seems to register them in upstart [10:53] didrocks, I am thinking what's going to be the solution for that. should we just delay the click ? [10:53] om26er: not sure, you can ssh/use kvm on the machine, you should have videos as well [10:53] on the jenkins job [10:53] didrocks, I saw the video [10:53] ok, maybe this delay is induced by unity7 [10:54] as we are seeing a lot of stuff running slowly on it [10:54] om26er: does the video seems slow? [10:54] didrocks, the video seems to also show that the environment is loaded quite heavily given the video is slow [10:54] didrocks, yeah [10:54] om26er: the video is 7fps [10:54] http://10.97.0.1:8080/job/autopilot-saucy-daily_release/label=autopilot-intel/470/artifact/results/autopilot/videos/gallery_app.tests.test_photo_viewer.TestPhotoViewer.test_nav_bar_share_button%20(with%20mouse).ogv [10:54] om26er: otherwise rmd was getting crazy [10:55] didrocks, i am sure if that suite is run again the same test will pass. it wen't haywire that time only due to environmental issue [10:55] om26er: hum, weird though that one both config, it failed, isn't it? [10:56] didrocks, that's the head scratching part :/ [10:56] sergiusens: ok... don't give up getting that in :) [10:56] thanks! [10:56] didrocks, i do think increase in the delay will fix though [10:57] om26er: so I don't trust a new run will fix those, autopilot doesn't introspect the app to know that something is loading? [10:57] sergiusens: maybe you can tell gema exactly what was discussed? [10:57] unkless you already did [10:57] * asac out again doing other stuff [10:58] didrocks, we do make sure (in the test) that before moving the mouse make sure the toolbar is fully visible. What we may need to also assert is that the toolbar is not animating before moving the mouse [10:58] om26er: yeah, not sure how feasable it is though [10:59] It could be that the "spread" state of the toolbar becomes true even when the minor part of the toolbar is visible on screen. [10:59] om26er: I would do either way: have a WI to have this done and implemented. If the media apps team needs to get their stuff quickly, timeout as workaround, but noted as such and removed timely :) [10:59] om26er: right, that would make sense [11:00] sil2100: so at leat, it confirms we can publish the sdk stack ^ [11:01] didrocks, sil2100 can you point me to the sdk change? I want to run this specific test with the latest sdk you guys talking about [11:01] om26er: oh, you didn't try with it? I thought you did [11:01] om26er: ubuntu-unity/daily-build [11:01] grab latest sdk packages [11:01] sil2100: wait then! ^ [11:04] Hi everyone, just trying to sort out google contacts sync with syncevolution. Not having muuch luck. The directions here spew errors at me: http://sergiusens.github.io/posts/google-contacts-on-ubuntu-touch.html Anyo [11:04] Anyone sort this on a recent build? [11:06] didrocks, that's actually a regression in the sdk :p [11:06] oh the irony [11:06] om26er: ah, so sil2100 was right to refrain it :p [11:06] file:///usr/share/gallery-app/rc/qml/MediaViewer/MediaViewer.qml:459: ReferenceError: caller is not defined [11:06] om26er: stop blaming jenkins/the machines running tests :p [11:06] Kaleo: timp ^ [11:07] sil2100: so, no forced release of the sdk please :) [11:07] didrocks, right ;) it has happened for us in ps-qa jenkins jobs in the past though [11:07] om26er: I wonder why gusch can't reproduce though [11:07] om26er: trust daily-release infra ;) [11:09] * greyback hates that nasty click when his nexus seems to get stuck in a reboot loop [11:09] * om26er didn't knew there was a UI toolkit update in the context as well. [11:13] * ogra_ notes popey answers to mails from feb. [11:14] oh [11:14] didnt see the date [11:14] has someone let a load of moderated mail through? [11:15] ogra_, hey [11:16] ogra_, can we pass arguments while starting a service? I would like to start unity8 with -testability for example [11:16] didrocks: I see that "caller" error only on jenkins (and I do have meanwhile the new SDK) [11:16] om26er, hmm, i think so, i guess through a variable parameter that you need to export when callin initctl... xnox do you happen to know that ? [11:16] gusch, I have the same issue here as well. [11:17] popey, looks like, there are really ancient mails among them [11:17] i blame callum [11:18] om26er: start unity8 MY_VAR=foobar [11:18] om26er: will make $MY_VAR available in the environment. [11:19] xnox, what would MY_VAR be for a command option [11:19] i.e. if you want unity8 --foople [11:19] om26er: your job file, then should use a variable. [11:19] om26er: alternatively you can use an override file "echo 'exec unity8 --foople' > unity8.override" [11:20] xnox, ah, thanks, i guess we should change the unity8 job then to accept something like $OPTIONS and use it if set [11:20] om26er: you can place override files in any of XDG_CONFIG_DIRS/upstart or XDG_CONFIG_HOME/upstart locations: ~/.config/upstart, /etc/xdg/xdg-ubuntu/upstart, /etc/xdg [11:21] xnox, cool, thanks. Will try and let you know [11:21] i guess ~/.config/upstart will do [11:23] * popey updates https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0ArLs7UPtu-hJdDZDNWliMmV1YUJ3Zk1pQlpDdGp4VFE#gid=0 === popey changed the topic of #ubuntu-touch to: Ubuntu Touch Support & Discussion | Installation https://wiki.ubuntu.com/Touch | Devices https://wiki.ubuntu.com/Touch/Devices | Release notes https://wiki.ubuntu.com/Touch/ReleaseNotes | Porting guide (advanced) https://wiki.ubuntu.com/Touch/Porting | File bugs at https://bugs.launchpad.net/touch-preview-images/+filebug | Under active development! | Nexus Status: http://bit.ly/18kIrhM === MacSlow is now known as MacSlow|lunch [11:26] didrocks: :D [11:26] * sil2100 sometimes learns from experience [11:27] sil2100: nice call ;) [11:28] sergiusens: [11:28] Preparing to replace phablet-tools 0.15+13.10.20130712-0ubuntu1 (using .../phablet-tools_0.15+13.10.20130716-0ubuntu1_all.deb) ... [11:28] Unpacking replacement phablet-tools ... [11:28] dpkg: error processing /var/cache/apt/archives/phablet-tools_0.15+13.10.20130716-0ubuntu1_all.deb (--unpack): trying to overwrite '/usr/lib/python2.7/dist-packages/tests/__init__.py', which is also in package hud-tools 13.10.1+13.10.20130717-0ubuntu1 [11:32] Laney: bug in both packages, as top level "tests" module should never be shipped. === alan_g is now known as alan_g|lunch [11:32] * xnox was pondering to write lintian check for it, as too much automatic setup-tools does that a lot. [11:32] yes [11:36] gema, https://code.launchpad.net/~ogra/utah/fix-dbus-location/+merge/175254 ... (not sure who can approve it, but that should solve all dbus issues) [11:37] (it wouold be a lot better if it would just rely on the users environment that explicitly exporting though) [11:38] (but before asac wants to stop everything again, i rather leave the hack in since it is known to work :P ) [11:39] jcollado: ^^ [11:39] jcollado: can you review and approve? [11:39] or reject, whichever you feel is better :P [11:40] gema: Yes [11:40] jcollado: thanks [11:40] the get_phone_shell_cmds() function needs to go or at least be rewritten from the ground up [11:40] but thats something jcollado knows already :) [11:42] (in shell i would just do something like: "pgrep unity8 >/dev/null 2>&1 && initctl stop unity8" [11:42] ) [11:43] sergiusens, do we still want to keep compatibility for non-flipped images in phablet-test-run or should I just remove it ? [11:44] just drop it [11:44] ogra_: 20130717.1 seems good here [11:44] non flipped will be completely dead by end of the week [11:44] (it is currently in a weird zombie state) [11:45] popey, yeah, but i fear we will still see red tests on utah ... [11:45] xnox, ogra_ thanks, that worked fine. [11:45] there are teo fixes that need to go in .... (one is the MP above, the other the unity8 initctl stuff above) [11:46] k [11:46] s/teo/two/ [11:46] ogra_: Ok, so no more updates to /etc/device-services (which doesn't exist anymore) and just "initctl start/stop unity8" ? [11:46] ogra_: is dbus-session hack specific to ubuntu touch? [11:46] ogra_: cause i don't have ~/.cache/upstart/dbus-session here =) but wish I had. [11:47] ogra_: it's for autopilot ubiquity, in my case. [11:47] jcollado, right, you should be able to test with initctl status unity8 (or pgrep as shown above) if it runs [11:47] xnox, dbus creates it [11:47] ogra_: but that location seems odd, shouldn't it be, $XDG_RUNTIME_DIR/dbus-session ? [11:47] make sure to have the latest [11:47] ogra_: ah, I didn't log out in a while. [11:48] xnox, feek free to improve, but note that utah currently hardcodes the location [11:48] *feek/feel/ [11:48] not sure if XDG_RUNTIME_DIR always reliably points to ~/.cache/upstart/ [11:49] ogra_: "pgrep unity8" returns a pid, but any initctl command returns "initctl: Unknown job: unity8" [11:49] ogra_: no, it points to a logind assigned location in /run/, e.g. /run/user/1000/ [11:50] jcollado, how do you call it ? [11:50] ogra_: $ initctl status unity8 [11:50] initctl: Unknown job: unity8 [11:50] hablet@ubuntu-phablet:~$ initctl status unity8 [11:50] unity8 start/running, process 1459 [11:50] ogra_: as we don't want to keep that file persistent acroos multiple logins. [11:50] jcollado, are you the phablet user ? [11:50] it only works in the users env [11:50] jcollado: you are probably not running as the phablet user. You need to be on the same UPSTART_SESSION. [11:50] not as root [11:50] ogra_: Yes [11:51] phablet@ubuntu-phablet:~$ env |grep UPSTART [11:51] UPSTART_SESSION=unix:abstract=/com/ubuntu/upstart-session/32011/532 [11:51] you shoudl see something similar [11:51] jcollado: initctl list-sessions [11:52] jcollado: and then export UPSTART_SESSION=$whatever is above, e.g. "unix:abstract=/com/ubuntu/...." [11:52] $ env | grep UPSTART [11:52] UPSTART_INSTANCE= [11:52] UPSTART_EVENTS=filesystem [11:52] UPSTART_JOB=android-tools-adbd [11:52] phablet@ubuntu-phablet:~$ initctl list-sessions [11:52] 547 unix:abstract=/com/ubuntu/upstart-session/32011/547 [11:52] xnox: Let me try that export [11:52] jcollado, make sure to have ubuntu-touch-session 0.59 or bigger installed [11:52] from there on the var should be in the users env [11:53] xnox: After the export, seems to work fine. [11:54] hmm, ubuntu-touch-session definitely should do that export [11:54] in all cases [11:54] ogra_: Version 0.58 [11:54] ah, phew [11:54] :) [11:54] upgrade it then and all should be fine [11:55] ogra_: From what I see in the channel, my understanding was that I needed to "phablet-flash -b", but the flash failed (not sure why). Do I need to re-flash or just upgrade the packages? [11:56] just apt-get update && apt-get install ubuntu-touch-session [11:57] that should give you even 0.60 [11:59] is there a way to login as user (and not root) with adb on flipped images [11:59] nope [11:59] ogra_: Upgraded to 0.60, rebooted to make sure and still I get the "Unknown job: unity8" [11:59] only sudp -u phablet -i [11:59] *sudo [11:59] jcollado, do you see UPSTART_SESSION in your env ? [12:00] ogra_: Nope. Same output as above: [12:00] $ env | grep UPSTART [12:00] UPSTART_INSTANCE= [12:00] UPSTART_EVENTS=filesystem [12:00] UPSTART_JOB=android-tools-adbd [12:01] ls /etc/profile.d/ [12:01] is there an upstart-phablet.sh file ? [12:01] http://popey.com/~alan/device-2013-07-17-125732.png made me smile today ☻ [12:01] I did indeed take 1 photo! [12:02] ogra_, cool, this did what I wanted: adb shell '$@' sudo -u phablet -i initctl start unity8 [12:02] yay [12:02] popey, well, does it change if you take another one ? [12:04] ogra_: i haven't dared take another photo :D [12:04] heh [12:04] (yes) [12:04] love this [12:04] ogra_: Yes [12:05] $ ls /etc/profile.d/ [12:05] dbus-source.sh upstart-phablet.sh [12:05] hmm, then i dont get why it isnt executed for you [12:05] jcollado, it clearly works for om26er [12:05] (and for me) [12:06] jcollado, is that via adb (and then sudp -u phablet -i) or via ssh [12:06] *sudo [12:06] ogra_: adb [12:06] flipped ? [12:06] hopefully [12:06] unflipped on nexus is dead and done since weeks [12:06] :) [12:07] phablet@ubuntu-phablet:~$ env|grep XDG [12:07] XDG_RUNTIME_DIR=/run/user/32011 [12:07] do you have XDG_RUNTIME_DIR in your env ? [12:08] ogra_: I tried using ssh and that works [12:08] well, adb should too [12:08] Is Mirv on holidays? [12:08] jcollado, did you log out and in after installing the new package ? it indeed needs a new login for the changes to take effect [12:09] ogra_: I rebooted [12:09] since we source from /etc/profile.d [12:09] hmpf [12:09] thats really strange [12:09] $ env | grep XDG [12:09] XDG_SESSION_ID=c3 [12:09] XDG_RUNTIME_DIR=/run/user/32011 [12:09] (with adb) === jdstrand_ is now known as jdstrand [12:09] ok, thats good [12:10] i dont get why UPSTART_SESSION isnt there though, it is exported from the exact same place that sets XDG_RUNTIME_DIR [12:12] jcollado, /sbin/initctl list-sessions | awk '{ print $NF; quit }' [12:12] does that retunr anything [12:13] unix:abstract=/com/ubuntu/upstart-session/32011/532 is what i get here [12:14] (and thats also what the prifile.d snippet should put into UPSTART_SESSION [12:14] ) [12:14] $ /sbin/initctl list-sessions | awk '{ print $NF; quit } [12:14] unix:abstract=/com/ubuntu/upstart-session/32011/547 [12:14] *profile [12:14] wow, so weird [12:14] why is it not exported then [12:14] like for everyone else [12:16] ogra_: jcollado: we only setup upstart-user session for graphical login, adb / ssh / sudo will not get you an upstart-session. [12:16] xnox, wrong :) [12:16] thus initctl list-sessions hack is needed to fetch the currently running session. [12:16] ogra_: what have you done ogra and where?! =))))) [12:17] root@ubuntu-phablet:/# cat /etc/profile.d/upstart-phablet.sh [12:17] # Join the phablet Upstart session [12:17] uid=$(getent passwd "$USER"|cut -d: -f3) [12:17] export XDG_RUNTIME_DIR="/run/user/$uid" [12:17] export UPSTART_SESSION=$(/sbin/initctl list-sessions | awk '{ print $NF; quit }') [12:17] the best of upstart cookbook =) [12:18] xnox, every login in touch should automatically join the session that way [12:18] ogra_: but that assumes that a session is present. if the session leader is topped (e.g. unity8) the session dies. [12:18] the session leader isnt unity8 :) [12:18] its ubuntu-touch-session [12:19] unity8 is just one app in it [12:19] s/app/job/ [12:19] so you can kill the shell and still have the rest of the session up [12:19] (and can restart the shell etc) [12:19] * xnox clearly needs to upgrade to latest image and poke around it. [12:20] this all sounds way too awesome =) [12:20] yeah, we started actually making use of upstart :) [12:20] though that still doesnt tell what jcollado's issue is [12:20] i dont get that [12:21] jcollado, initctl list [12:21] doe that return a list of jobs ? [12:21] (if not, whats the error) [12:21] ogra_: initctl --user list [12:22] is better, as that restricts to talking to user-session init only. [12:22] not if i actually am the user and UPSTART_SESSION exists [12:22] the export should properly restrict it already [12:23] ogra_: http://pastebin.ubuntu.com/5884013/ [12:23] ogra_: sure the default is to look at environment vars, but if we are troubleshooting a problem with environment..... without vars $ initctl list will default to system init. [12:23] jcollado, that looks like you are not properly sudoed to phablet [12:23] jcollado: yeah, that's system wide init. [12:23] jcollado, how did you become phablet ? [12:23] su phablet [12:24] ah no [12:24] /o\ [12:24] sudo -u phablet -i [12:24] jcollado, hey [12:24] or su phablet - (but that will likely mess up your path) [12:25] jcollado: don't use su =) ever. it doesn't setup / enherit proper environment. Unless that's your goal to get a clean environemnt, which doesn't correspond to user environment at all. [12:25] jcollado, have you recently run autopilot tests on touch ? [12:25] ogra_, xnox: That was indeed the problem. Thanks a lot. [12:25] YAY !!! [12:25] * ogra_ dances [12:25] \o/ win [12:26] I see this http://paste.ubuntu.com/5884025/ [12:27] om26er: does DISPLAY=:0 help at all? [12:27] * xnox checks [12:27] om26er, oh, wow, why would autopilot fire up dbus at all ? it shoudl just attach to the runnign session bus [12:27] om26er: I'm kind of desperate with that. phablet-test-run doesn't work for me right now. At first I though it was because of phablet-test-run not being updated to use the upstart commands and ignore the /etc/device-services file. However, even running autopilot from a terminal app in the phone itself return dbus problems. [12:28] where is the autopilot source [12:28] ogra_: it does everything over dbus, and it can / should connect to running one. but it depends on how the wrapper is written. [12:28] ogra_, lp:autopilot [12:28] om26er: i'd like to know where is phablet-test-run is at? [12:28] xnox, yeah, i suspect another hardcoded dbus session address case [12:29] xnox, lp:phablet-tools [12:29] xnox, phablet-tools [12:29] om26er: cause you are launching that from host system, instead of the phablet@ user. [12:29] om26er: where is your MR? [12:29] sergiusens, didn't propose yet, its without the MR [12:30] xnox, it should sudo -u phablet during run ... [12:30] om26er: branch? [12:30] om26er: jcollado: do not use phablet-test-run, it's for unflipped / legacy only.... [12:30] xnox: no, it should work on both [12:31] xnox, sergiusens: Yes, there's some code to detect if the image is flipped or not. [12:31] sergiusens, http://bazaar.launchpad.net/~om26er/phablet-tools/adapt_to_latest_changes_in_touch_image/revision/131 [12:31] xnox: there's a chroot check, it will fail wrt to the new dbus/upstart change [12:33] tests run fine if logged into the phone, they only fail for phablet-test-run === MacSlow|lunch is now known as MacSlow [12:33] this error: DBusException: org.freedesktop.DBus.Error.NotSupported: Unable to autolaunch a dbus-daemon without a $DISPLAY for X11 [12:34] that error appears when autopilot tests are tried to run as root [12:35] om26er, make phablet-test-run switch to the phablet user first [12:35] at least for all tests that need dbus [12:35] om26er: that should be the default [12:35] ogra_: ^^ [12:35] let me see if anyone changed that [12:35] what should be the default ? [12:35] sergiusens: but exec_with_adb is not guarded, and does chrootcmd. [12:35] running as root ? [12:35] ogra_: all autopilot tests need dbus [12:35] ogra_: ssh phablet@localhost "autopilot ..." [12:36] ah [12:36] well it's not "just" that. [12:36] ssh -o NoHostAuthenticationForLocalhost=yes -t $USER@$TARGET_IP -p $TARGET_SSH_PORT "bash -ic \"$@\"" [12:36] yeah, that shoould work unless you then firsse up a subshell with cleared env [12:36] ogra_: ???????? no permissions [12:36] xnox: autopilot depends on dbus, the object tree is exported there [12:36] *fire [12:36] what does that usually indicate? [12:36] popey, no permissions i would guess :) [12:36] * xnox ponders if "bash -ic" wrapping is playing up. [12:36] popey, fastboot ? [12:36] popey: device keys need to be accepted [12:36] FUNNY MAN [12:37] thats from doing adb devices [12:37] on a samsung galaxy [12:37] popey: against pure android ? [12:37] no, its one i flashed yesterday [12:37] remember ☻ [12:37] popey, it indicates that the udev rule is not catching ut [12:37] er, monday [12:37] *it [12:37] raring machine with adb tools from ppa [12:37] also that [12:37] unplug/replug maybe [12:37] popey, sudo adb kill-server;sudo adb start-server (used to fix for me) [12:37] yeah ta [12:37] restart adb as root [12:38] right, what om26er said [12:38] kk thanks chaps [12:38] (you could figure out the right runes and send a patch for adb's udev rule if you feel like [12:39] unfortunately the device is remote from me, in london [12:39] if I get hold of a device at oscon next week I may [12:39] ah, well, then restarting adb should just do === _salem is now known as salem_ [12:41] ogra_: is 20130717.1 safe for me to recommend to design people to flash? [12:41] popey, yes, it will only fail autopilot until we fixed that [12:41] asac: ogra_ jcollado gema heads up that once the new image upgrader is in, all this may break again [12:41] great, will still need --pending yes? [12:41] popey, meaning by asac definition it wouldnt be safe to release :) [12:42] sergiusens, yeah [12:42] pfffft, he worries too much [12:42] dont tell me :) [12:43] well, he wants all gree on the dashboard (which is what we are working on atm) [12:43] *green [12:43] but a broken testsuite doesnt necessarily mean bad image :) [12:44] sergiusens: who lands that? [12:44] didrocks: what? [12:44] asac, foundations [12:44] sergiusens: whoever lands that should prep with QA so that doesnt happen [12:44] and coordinate that [12:44] asac, there will likely bee a massive breakage once that happens [12:44] slangasek: cjwatson: stgraber: ^ [12:45] due to the fact that so many parts use hardcoded assumptions all over the place [12:46] (and due to the fact that likely not every single bit is prepared to have the rootfs completely readonly) [12:46] asac: I believe that's on stgraber [12:47] stgraber: can you comment on that? [12:47] sergiusens, so once we have that autopilot stuff back running, i'd like to merge your click package installer ... note that it would be nicer to not use the hooks for new cod but properly integrate it into live-build/auto/config|build ... but we can do that later [12:48] stgraber: lets chat a bit later... have to run to something for 1.5h [12:48] ogra_: sure, I will fix on cjwatson comments I'm just reading now [12:49] asac, stgraber will likely not be able to do much, the tests will fail and need adjustment etc ... i guess that should be a cross team effort across all teams [12:49] jcollado, sorry missed you message, so you have your device running flipped image right ? [12:49] sergiusens, oh, i didnt notice colin had commented [12:49] ogra_: yes, but not in cowboy/firedrill mode inside the archive; prep outside, accept some pain, keep everyone happy [12:49] good deal :) [12:50] ricmm: can you send an email to the lists wrt upstart user session changes? [12:50] asac, too many pieces to do it outside of the archive ... we should instead prepare for a day or two with broken image to speed it up ... after we have one rocksolid build (imho) [12:50] asac, but your call in the end :) [12:51] no ... we have been through that discussion :) [12:51] yes, it will cost us a lot more time to land it then [12:51] a firedrill is more expensive than anything else [12:51] people bleeding out, working all night etc. [12:51] having less energy afterwards [12:52] so that must never be the plan. [12:52] if it is planned in advance i dont see an issue [12:52] the unplanned breakage harms us [12:52] not the planned one [12:52] we cnanot have images broken without a firedrill mode [12:52] thje costs for that are engineering wide [12:53] doing it properly is always more efficient long term [12:53] properly meaning without breaking the main integration baseline [12:53] we dont release broken images ... there will just not be an update released for one or two days (i belibve we can even do it in one with poroper planning and reviews in advance) [12:54] thats not acceptable. we want to get confirm about issues fixed from the images in dashboard each and every day. every day we dont see that we have to get folks do stuff manuallyt etc. eg.g. costs all over the place [12:54] anyway have to run [12:54] bbiab === Gh0sT is now known as mahabal [12:57] mzanetti, ping [12:59] om26er, he's away on Qt Contributors' Summit [12:59] Saviq, ok, thanks [12:59] om26er, he does pop in sometimes, but not often [12:59] mzanetti, can we help maybe? [13:00] om26er, btw, the unity8 job for upstart is a user-session one, so it's there in /usr/share/upstart/sessions [13:00] om26er, but unfortunately it's not working due to some ways the session is set up [13:01] Saviq, that fixed now. I have it working [13:01] Saviq: should be working today [13:01] om26er, ah cool [13:01] sergiusens, cheers [13:01] Saviq, there is a feature request for the unity8 upstart job though. we should be able to give command line parameter [13:03] sergiusens, om26er shouldn't the unity8 upstart job live with unity8 btw? [13:03] !!! infographics are real !!!! [13:03] pete-woods, you ROCK [13:03] "1 photos captured today" [13:03] hehe [13:03] Saviq: please confirm bug 1202165 ☻ [13:03] bug 1202165 in touch-preview-images "Welcome screen says "1 photos" - should be "1 photo"" [Undecided,New] https://launchpad.net/bugs/1202165 [13:04] Does anyone know if Ubuntu phone plans to have xmir in the future? [13:04] Azelphur, the phone will run mir [13:04] om26er: doesn't it already run mir? [13:04] Azelphur, not yet, I hope a few weeks (if not days) [13:05] o.O, what is it running on now? [13:05] Saviq: that's an open question, I may think so [13:05] Azelphur, SurfaceFlinger currently [13:06] om26er: I'm going to reconstruct your MR and drop unflipped [13:06] I see :) [13:06] sergiusens, ok [13:06] Azelphur, and i dont think there are plans to installl XMir by default (only Mir) ... but you will be able to do that yourself in developer mode [13:06] are the apps in the additional downloads section supposed to work? I clicked on most of them and nothing happens [13:07] 14.04 might have it [13:07] ogra_: sweet [13:07] nah, they are just pretty pictures [13:07] So, tethering support on the roadmap at all? [13:07] * marcoceppi uses this feature extensively [13:07] click package integration is on its way, then there will be installlable apps in the shell too [13:07] sergiusens, also fix exec_with_ssh() :) [13:07] om26er: I'm on that [13:07] popey, indeed [13:08] ta [13:08] * cjwatson is sweating through revising the click hooks design/implementation at the moment [13:08] ogra_: sweet, it's looking cool, might dual boot my phone so I can watch the development [13:08] om26er: but it's easier without unflipped [13:08] Azelphur, I think of them as teasers. every cuple days more teasers turn into nice things :-D [13:08] (literally; heatwave) [13:08] heh, here too [13:08] kalikiana: yea haha, I noticed how the stuff that is done is really nice, and the stuff that isn't is non-existent :P [13:08] not that hot though ... we're still below 30 [13:09] Azelphur, the settings app is like that. I love the looks but many pieces don't work just yet [13:09] sergiusens, seriously, drop all that unflipped code [13:09] we wont go back :) [13:09] kalikiana: yea, I went around pressing all the buttons :P [13:09] only barely below here; it's 29 [13:09] phew [13:09] ogra_: I know, I'm dropping :-) [13:09] Is there a page with documentation for Ubuntu phone app developers? [13:09] :) [13:09] Saviq: :D [13:10] Azelphur: developer.ubuntu.com [13:10] Azelphur, this one maybe http://developer.ubuntu.com/get-started/ [13:10] thats the startpoint [13:10] cool, ty [13:10] Saviq: now we just need to get lots more data sources added! [13:10] Azelphur: and feel free to ask in here or #ubuntu-app-devel [13:11] * Saviq needs to get the freakin' microSIM adapter to start dogfooding... :/ [13:11] cool :) [13:12] i would borrow you mine, but my arm is to short to hand it to you [13:12] * ogra_ streches .... see ... [13:16] Kaleo: I think sil2100 already told you, but the sdk stack regressed some media autopilot tests [13:17] didrocks: not yet, but now you did ;) I think I need to add this one test to the list of tests ran on SDK [13:17] Since it slipped through our fingers [13:17] sil2100: yep [13:18] didrocks: so, what's the bug report [13:18] ? [13:18] om26er: ^ ? [13:19] Kaleo: just the messenger here when people don't transmit the message :p [13:19] Kaleo, in the gallery-app clicking on the share-button does not show the dialog [13:20] Kaleo, terminal says: file:///usr/share/gallery-app/rc/qml/MediaViewer/MediaViewer.qml:459: ReferenceError: caller is not defined [13:20] didrocks: https://code.launchpad.net/~sil2100/cupstream2distro-config/media_stack_fix/+merge/175278 <- fix packages list and that one additional test [13:21] sil2100: great! approving, feel free to deploy once merged :) [13:21] kenvandine: FYI, sil2100 refreshed the list ^ [13:22] sil2100, thanks! [13:22] kenvandine: ah, you're online! Np ;) [13:26] xnox: ogra_ what am I doing wrong? http://pastebin.ubuntu.com/5884222/ [13:27] sh -c probably ? [13:28] what is the upstart event on touch to know that wlan is ready to use? [13:28] sergiusens: $DBUS_SESSION_BUS_ADDRESS is interpreted by bash/sh before you executes sudo [13:28] execute* [13:29] (so it's $DBUS_SESSION_BUS_ADDRESS from root's env) [13:29] didrocks: makes sense! [13:29] thanks [13:29] yw ;) [13:31] ogra_: didrocks what about now? http://pastebin.ubuntu.com/5884232/ [13:31] reason autopilot is failing is because the env isn't being exported [13:31] ah, this is more interesting :) [13:31] quoting ... [13:32] yep, what ogra says [13:32] root@ubuntu-phablet:/# sudo -u phablet -i sh -c 'env |grep DBUS' [13:32] DBUS_SESSION_BUS_ADDRESS=unix:abstract=/tmp/dbus-nRITxeJvsr === ckpringle_ is now known as ckpringle [13:34] didrocks: the people being sil2100 ? [13:34] sil2100: where is the bug report? [13:34] or om26er ? [13:35] ogra_: a copy/paste of that returns SUDO_COMMAND=/bin/bash -c sh -c env\ |grep\ DBUS [13:35] Kaleo: well, I don't know the exact problem, but om26er would know more [13:35] let me see what's going on with my env [13:35] sil2100: that's not my question :) [13:35] Kaleo: I just found that newer SDK caused an AP failure [13:35] sil2100: didrocks: om26er: when there is a failure, first thing to do is bug report [13:35] sergiusens, well, worst case just read the file direclty and source it [13:35] Kaleo: right ;) om26er should do that, as he probably knows most [13:36] sil2100: good [13:36] om26er: you there? :) [13:36] reporting [13:36] Kaleo: I would just submit a bug report like 'AP test fails' [13:36] ogra_: yeah, but I was trying to avoid that [13:36] om26er: thanks [13:36] sergiusens, from /home/phablet/.cache/upstart/dbus-session [13:36] yeah, thats definitely better indeed [13:36] ogra_: yeah, read it in .bashrc [13:36] sil2100: om26er: so I assume then that nobody has been working on the issue then? [13:36] Kaleo: moooost probably [13:36] sil2100: om26er: when did we notice? [13:37] Kaleo, an hour ago [13:37] om26er: ok :) [13:37] fresh [13:38] Kaleo: hum, we did notice 4 hours ago rather, gush tried first to reproduce it but couldn't [13:38] Kaleo: then, om26er reproduced it 2h30 ago [13:38] didrocks: so there is something wrong in the process [13:38] from the logs [13:38] didrocks: first, bug report should be made [13:38] ogra_: if I add what's in .bashrc to .profile it works fine [13:39] didrocks: then I think since it seems that SDK is marked as culprit, a person from the SDK should work on it right away [13:39] didrocks: how can we make sure that happens? [13:39] didrocks, Kaleo previously we have been thinking it a general test failure, the dialog not appearing was discovered ~1h ago or a bit longer [13:39] Kaleo: you and timp were pinged for that on that channel, not sure if any other communication has been done [13:40] 13:06:09 om26er | didrocks, that's actually a regression in the sdk :p [13:40] om26er: it's 15:40 now here [13:40] didrocks, wow time flied or something [13:41] Kaleo: I don't know, I hope that sdk will have better regression tests soon to not rely on autopilot tests from apps [13:41] sergiusens, ?? === alan_g|lunch is now known as alan_g [13:42] sergiusens, we ship whats in bashrc in /etc/profile.d/ snippets now [13:42] didrocks: that's unrelated [13:42] ogra_, asac, sergiusens: What's QA doing that'll fail with read-only? Read a good chunk of the backlog but I'm still confused :) [13:42] didrocks: if there is an issue, it needs to be dealt with properly [13:42] sergiusens, are you on the last ubuntu-touch-session package ? [13:42] Kaleo: it clearly is seeing the regressions we had lately [13:42] stgraber: installation of additional things without enabling developer mode [13:42] stgraber, we dont know, thats the point :) [13:42] Kaleo: and the workaround we had to put in place in the tests with that [13:42] didrocks: it's not [13:43] stgraber, we switched to upstart sessions and the whole testing fell over ... now we try to remove all the hardcoded paths etc [13:43] didrocks: so timp did not see the ping because his internet & irc died [13:43] bug 1202201 [13:43] didrocks: I did not because I was sleeping [13:43] bug 1202201 in gallery-app "Share dialog does not appear with latest SDK" [Undecided,New] https://launchpad.net/bugs/1202201 [13:43] stgraber, i expect that switching to image upgrades will cause similar issues [13:43] sergiusens: ah yeah, that'll definitely fail unless they can get their stuff running from a standalone directory on the data partition (so not installing extra stuff) or run in developer mode (which would be bad as test results will likely differ due to / being rw) [13:44] stgraber, due to the readonly nature, due to different FS structure, due to no apt available etc [13:44] Kaleo, is it fine for gallery-app or should I add ubuntu-ui-toolkit as well ? [13:44] stgraber, the point is that we need to check in advance [13:44] stgraber: what you should do is really check with qa how the run the image and which tools [13:44] ogra_: fs structure "should" be fine, we've got a bunch of extra stuff but in theory I have at least the same stuff as you currently do, but yeah, they'll have to cope with readonly, I don't think switching to devel mode is the right thing for them to do [13:44] and give it a try and see [13:44] if there are issues work with them [13:45] in that way you can protect from the most issues. there might be more if it hits real infra [13:45] stgraber, right, which means we will likely need changes in the test suites [13:45] tmoenicke, ping [13:45] stgraber, and we need to identify them [13:45] but :)... we caan minimize the risk by prepping [13:45] asac: who's writing those tools/tests? [13:45] QA [13:46] MacSlow: pong [13:46] ogra_: QA is a rather big team ;) [13:46] stgraber: gema can connect you to someone to assist you... right now we work with doanac and jcollado, but there might be someone better. [13:46] gema: stgraber would like to check with someone to see how he can see what impacts his big landing has [13:47] stgraber, well, there are plety f areas we are currently fixing just for the upstart switch [13:47] stgraber, (they all use the same hardcoded paths etc) === dandrader is now known as dandrader|afk [13:47] stgraber, so it really spans across a multitured of code trees [13:48] *multitude [13:49] asac: ok, I'll see if I can find some time to test that QA stuff, though I only have two days of work left before I leave for London and then I'll be on vacation and conferences for 3 weeks after that (well, still working 1-2days a week) and unfortunately I have higher priority bugs and features to get to by then, so I can definitely help someone in QA setup a readonly device and give pointers as to how to fix the tests but I doubt I'll be [13:50] oh, I have a feeling that one will get cut... let me know where and I'll paste the rest... [13:50] stgraber: well, we definitely don't want to land stuff that will break our QA infrastructure [13:51] Cimi, Saviq: overlay-scrollbar bug fix commited, jfyi [13:51] seb128, cheers [13:53] stgraber, ... give pointers as to how to fix the tests but I doubt I'll be ... [13:53] thats where it was cut [13:53] ogra_: thanks [13:53] "I'll be able to do much more than that" [13:55] stgraber, did you ever try to run the tests that are shipped with phablet-tools ? [13:55] i think getting that to work would be a good start [13:55] if there are changes required they should be easy to port over to utah and other bits [13:55] asac: sure and I'm certainly not for rushing the feature in. On paper we're only supposed to have the thing working end-to-end on all devices by end of August (deadline for landing the UI), until then, I don't really mind having both systems in parallel and switch once the blockers have been resolved (proper testing on all devices, regression free and click packages working by default at least) [13:56] (at least thats the case for the dbus and upsatart changes we fix atm) [13:56] hello [13:56] stgraber, oh, please not parallel images again [13:56] stgraber: i would suggest to check whats going on now [13:56] and then we can assess looking on the details [13:57] stgraber, also landing it in august will be way to late to shake out bugs [13:57] ogra_: well, that's what I'm doing at the moment ;) but yeah, I'd like to avoid having half the users on each [13:57] this is a massive architectural change of the whole image foundation [13:57] what we need to check though if that decision then has good/complete enough data, e.g. did we check what the impact on our QA infrastructure is? are there other areas with impact that we haven't thought about [13:57] there will surely be lots of bugs to shake out we will only get through user testing [13:58] so landing as early as possible is a req. imho [13:59] it really depends on the impact. hence i would like to know the impact first :) [13:59] asac: yeah, I plan on testing the phablet tests as soon as I can, which likely means late this afternoon (today is my meeting day so I'm stuck until 2:30pm) [13:59] stgraber: yeah. if you have questions just ping me... happy to talk [14:00] asac, the impact of switching in august only and thus having bad bugs by release is surely worse than switching asap and having time to shake them out [14:00] I don't have enough details to make such assessment. but note that you also imply that the only other solution from landing today will be to land it late in august :) [14:00] or something' [14:00] thats also not decided [14:00] known [14:00] etc. [14:01] asac, well, we change the complete underlying architecture of the image to something we never did before [14:01] in the past such a change would have happened across two cycles === chihchun is now known as chihchun_afk [14:02] just leaving us 6 week to fix the issues (in case we switch end of august) is extremely risky [14:02] not to mention that this is only the preparation for another heavyweight change that will come after this [14:03] (repartitioning the device and working with partitions instead of images) [14:03] we would only do the second part mid sept. then [14:03] jhodapp: hi! Do you know if I can build/compile qtubuntu-sensors on amd64? Since it's missing some headers it seems [14:03] which well, might give us less then 4 weeks to fix issues [14:03] so as you say, the risk is very risky, because we cannot even assess the risk yet :) [14:03] e.g. hence, we need to prep better [14:04] check exactly what will be broken [14:04] asac, well, then we should not release by 13.10 [14:04] ensure we have a plan and resourcing to really address that. then make a decision where to do that === dandrader|afk is now known as dandrader [14:04] these two changes are so massive that they would have deserved a full release cycle for testing them imho [14:05] it looks all good on paper but we have no real world testing yet [14:06] they are as bad as the container flip was ... which we are far from having shaken out completely yet [14:06] ogra_, asac: just took a look at phablet-test-run, it indeed won't work at all on readonly images as it relies on being able to do in-place changes to files and install packages with apt and dpkg. I expect this would work perfectly if in developer mode but then we'd be in a pretty different mode from our standard devices so I'm not sure how much value the test results would have [14:06] yeah, that wont test what the enduser gets at all [14:07] stgraber: right. we need overlayfs or massive rewrite of infrastructure? [14:07] ugh [14:07] we basically need to be able to produce custom images outside and then flash those [14:07] overlayfs is a no no [14:07] an-overlay-fs ... thats about approach [14:07] it could be changed to do a first prepartion pass where it installs everything in devel mode, then reboot in user mode (readonly) and run the tests which should work for most of the cases, however that means we wouldn't be able to test the upgrader (as the delta images won't work on a device that has been in devel mode) [14:08] it will kill your battery and performance [14:08] how to ensrure we can reasonably well test our read-only iamges [14:08] asac, by completely reworking the test suite [14:08] asac: overlayfs would effectively mean writable / (even if not persistent) which would still hide most of the bugs we hope in catching through QA [14:08] stgraber: yeah. better would really move the isntall and prep stuff out so we can run them up front [14:08] we need to test the actual setup the enduser gets [14:08] before flashing the image [14:08] else the tests are pointless [14:08] asac: so effectively the same thing as running in devel mode except that you also get the overlayfs bugs on top :) [14:09] yeah. [14:09] and you need to have hacked up kernels [14:09] and someone needs to maintain overlayfs for 3.0 kernels ... [14:09] i think we need an officially vetted way to oproduce customized image derivates [14:09] that add tests [14:09] and also daily-release stuff [14:09] for 14.04 that sounds like a good plan [14:09] so that the image we test is as close as it can get [14:10] or we move all tests to click packages [14:10] asac: ideally, QA should rework their tools to be runnable from a standalone directory and not require package installation, that way we can run them on a perfectly clean device and at the end just flush the directory without any impact on the device [14:10] the image we test needs to be the image the enduser gets [14:10] asac, the testsuiote needs to be adjusted [14:10] yeah. so they would want a testing pack [14:10] that we unpack [14:10] ogra_: can you add curl to the touch meta? [14:10] instead of the images [14:10] that would include autopilot stuff and libs [14:10] and stuff [14:10] sergiusens, uh, shudderm, cant we use wget ? [14:10] and the tests [14:10] that thing can then be unpacked to a directory [14:10] stgraber: ? or what would you imagine? [14:10] ogra_: oh yeah, sorry, let's use wget [14:10] sergiusens, only half the deps and so much saner [14:11] :) [14:11] asac, so we should have a test click package [14:11] asac: yeah, that'd work fine. We have writable persistent directories they can use (/userdata/...), so they'd need to rework their setup to be self contained and ship it as a tarball or something that's unpacked there, test runs and then we flush it [14:11] asac: with the advantage that anyone can run that on their device without leaving any trace of the testing tools afterwards [14:11] stgraber: ok. however [14:11] instead of fiddlign with the image architecture to have something completely non std which makes testing moot [14:11] thats nowhere shorterm [14:11] also we cannot afford to turn off dashboard for amonth :) [14:12] so compromise MUST be that we start testing with developer mode [14:12] ++ [14:12] so gyuess yoiu just need to focus on checking whether dev mode can be enabled well [14:12] but need to move the suite asap [14:12] and that it works with that [14:12] and we will pickup the next level after getting this in [14:12] dev mode will be the same as you have today [14:12] asac: right, we can run in developer mode for now, it's just a "touch /userdata/.developer_mode && reboot" away, so that's easy, however this will hide bugs for sure (any bug triggered by the device being in read-only mode) [14:12] yeah i know i know :) [14:13] (except that yoou have loop mounted imgs instead of the weird setup we use now) [14:13] stgraber: ok then just check in your testing discussed above that your landing doesnt cause autopilot regressions [14:13] rsalveti, hey i got the bootstrap to install fine this time. [14:13] otherwise talk to QA folks to address whatever that is [14:14] doanac: awake? [14:14] :) [14:14] asac: if we have regressions it'll likely be bugs on my side, my goal from the beginning was to be identical to flipped images but read-only, so if they run in read-write, the result should be identical (or we indeed have something to fix) [14:14] asac: anyway, have to go to my meetings now but can help run tests later [14:15] stgraber: yeah, so lets try to check before landing so we know the impact and then decide later today or tomorrow whats next [14:15] yeah [14:15] lets check later [14:15] ogra_: do you know if the infrastructure stuff for the dbus dashboard thing landed? [14:15] rsalveti, unfortunately the ubuntu raring link you told me to install goes on a reboot loop and ends the third time on the recovery [14:16] asac, i submitted an MP ... one sec didnt check if jcollado has landed that [14:16] ogra_, asac: Not yet. [14:16] jcollado: whats up? [14:16] asac, not yet then :) [14:16] stuck? [14:17] asac: No, I just want to check with doanac since he knows that code better than me. [14:17] doanac: !! :) [14:17] sergiusens, seeded and updating meta now [14:17] ogra_: thanks [14:18] asac: We have our daily meeting in 12 minutes, so I'll talk to him then. [14:18] so you waited all these hours? [14:19] anyway [14:20] check with him quickly and also kick off all those jobs again once it landed [14:20] thx [14:20] sergiusens :how can i debug a reboot loop since this phones have no serial? [14:21] well, at least these 15 tests that ran were all successful :) [14:22] I: Checking Release signature [14:22] gpgv: Signature made Mi 17 Jul 2013 16:06:44 CEST using DSA key ID 437D05B5 [14:22] gpgv: BAD signature from "Ubuntu Archive Automatic Signing Key " [14:22] E: Invalid Release signature (key id 40976EAF437D05B5) [14:22] network issue [14:22] * ogra_ glares at his scree [14:22] n [14:22] most likely [14:22] or not? [14:22] well, you can still see me [14:22] or not ? [14:22] lol [14:22] same machine :) [14:22] thought you saw that on the phones in lab [14:22] Device detected as /bin/bash: getprop: command not found [14:22] * ogra_ tries again [14:22] just run update one more time [14:23] oh [14:23] yeah, it is the new PPA [14:23] we had that during image build too last night [14:25] silly coreapps PPA ... has to go .... we need click ! [14:25] it's very common with ppas [14:25] :) [14:25] ogra_: comming [14:25] any idea where I can find the "getprop" command? [14:25] sergiusens, i lierally saw that the first time last night ... in 8 years that i work with our infrastructure [14:28] okay, using -d [14:28] ogra_: 20% perhaps of the jenkins build failures come from that [14:29] wow [14:29] i'm lucky we never used PPAs in images before then i guess [14:29] bah, i clashed with Laney [14:30] * ogra_ fixes his liecd-rootfs branch [14:33] anyone else encountered this? http://askubuntu.com/questions/320989/depends-dropping-letters-error-when-trying-to-get-the-ubuntu-touch-sdk-preview [14:33] yes [14:33] working to fix it [14:34] * popey leaves same as a comment ☻ [14:36] popey, ok cool, thanks, no need to be sorry :) [14:36] popey: what's the issue with it? [14:37] mhall119: read your email ☻ [14:37] morning btw [14:39] popey: that's just precise though right? [14:39] Silly question: does PolicyKit work at all on ubuntu-touch? [14:39] Not as in can you actually get prompted, but does it know that the phablet user is active? [14:40] yes mhall119 and quantal [14:40] ah,I didn't notice that tha AU question specified 12.04 [14:47] jcollado: did doanac show up? [14:47] can we land and try if that helps? [14:48] no need to "try" we know it does :P [14:48] didrocks: gusch: om26er: this bug (https://bugs.launchpad.net/gallery-app/+bug/1202201) I have seen for weeks on the device; nothing new there [14:48] Launchpad bug 1202201 in gallery-app "Share dialog does not appear with latest SDK" [Undecided,New] [14:48] ogra_: well, then it should just land [14:49] cjwatson: it doesn't work [14:49] and someone should still ensure that the jobs finish successfully [14:49] how important is this?? on the ubuntu bootloader i get E: Cant' open /cache/recovery/ubuntu_command sergiusens ? [14:49] gusch: om26er: is that the issue you are working on for a few hours now? [14:49] annerajb: not important [14:49] Kaleo, it was working for me today, till I updated the sdk [14:49] asac, my MP is only half of it, i think jcollado was working on the other bit (stopping the shell to be able to run the tests) [14:49] om26er: it was not on my phone for weeks really [14:49] Kaleo: yeah, om26er did before/after comparison [14:49] * annerajb wonder why it dosnt boot raring then... [14:50] om26er: you testing on the nexus S? [14:50] asac, i dont know the status of that second part (though i know jcollado had the right fix) [14:50] ogra_: I updated the MR/MP btw for click [14:50] annerajb: you need to use saucy [14:50] sergiusens, will merge after meeting [14:50] sergiusens, oh that's what rsalveti gave me yesterday [14:50] ogra_: ok. well i would at least like a status update [14:50] annerajb: my guess is missmatch of hybris and platform api [14:50] Kaleo, desktop [14:50] didrocks Kaleo om26er working on that one (Action.onTriggered does not have the caller) [14:50] asac, you need a secretary ... :) [14:50] om26er: not good :) [14:51] sergiusens, wait i am using raring [14:51] om26er: I actually have yesterday's image on my phone [14:51] om26er: and it's broken too [14:51] Kaleo, I agree its been broken last 5 images [14:51] raring-preinstalled-phablet-armhf.zip [14:51] fwiw [14:51] pmcgowan: more I think :) [14:51] Kaleo, so what you are seeing on the phone is a dialog that depends on having a sharing service setup [14:51] gusch: so is it an issue in the gallery or the toolkit according to you? do you need help? [14:51] Kaleo, that's why you see a transparent black thingy on the phone [14:51] Kaleo: btw. ToolbarButton can't be disabled === alan_g is now known as alan_g|tea [14:52] gusch: it can be I guess: enabled: false [14:52] Kaleo: well - no Idea if Actions should pass the caller or not [14:52] gusch: (do you mean it does not *look* disabled?) [14:52] annerajb: I'm 99% sure it's because of a missmatch in libs or missing hw drivers for graphics [14:52] Kaleo: it grays out, but onTriggered is still fired [14:52] gusch: ok, bug report [14:52] sergiusens, how can i debug this since this is not a desktop i can attach a serial too :( [14:52] asac: i'm on. in a meeting, but am going to fix things today [14:52] Kaleo: where are your bugs? [14:53] annerajb: adb logcat? [14:53] om26er: so that means that the whole feature has been broken for a long time [14:53] gusch: https://bugs.launchpad.net/ubuntu-ui-toolkit [14:53] doanac: today might be not before my end of day [14:53] Kaleo: thx (just discovered 5 minutes ago) [14:53] well, the fixes are well known [14:53] gusch: so, do you need help? [14:53] shouldnt take long to merge them [14:53] i want to see it go green and get the results [14:53] Kaleo, well the share menu on the desktop is just a dummy icon. on the device however its W.I.P. [14:53] right. hence we should merge them [14:53] :) [14:53] Kaleo: I can sort it out [14:53] gusch: ok [14:54] gusch: I'll leave you to it then [14:54] sergiusens, i am at the bootloader right now and it says exec line 1 logcat not found.and when the phone is restarting i cannot connect to it thru adb since it's restarting [14:54] Kaleo: but is the Action{ onTriggered()} supposed to pass the caller, or not? [14:54] Kaleo: the ToolBarButton{ onTriggered()} does, but for Actions it might not be the case - no idea about the design there [14:55] Kaleo: but for consistency it should be the same I'd say [14:55] gusch: I don't understand what you mean [14:55] sergiusens, every time i reboot i endup on the ubuntu logo cwm based recovery menu [14:55] gusch: you are taking about http://developer.ubuntu.com/api/devel/ubuntu-13.10/qml/ui-toolkit/qml-ubuntu-components0-actionitem.html#triggered-signal ? [14:55] gusch: it gives you the caller when fired, yes [14:55] Kaleo: ah - ok - so it should [14:55] annerajb: flipped or unflipped? [14:56] gusch: same for http://developer.ubuntu.com/api/devel/ubuntu-13.10/qml/ui-toolkit/qml-ubuntu-components0-action.html#triggered-signal [14:56] Kaleo: but no - does not anymore since yesterdays SDK [14:56] sergiusens, rsalveti told me yesterday to try unflipped first if i recall correctly how can i be sure which one i have? [14:56] doanac: ok thats good... once the stuff land, please repoke the jobs of todays image etc. [14:56] and get everything "greenish" :) [14:56] Kaleo: and that's why this bug showed up: https://bugs.launchpad.net/gallery-app/+bug/1202201 [14:56] annerajb, do you end up an an android root with adb shell ? [14:56] Launchpad bug 1202201 in gallery-app "Share dialog does not appear with latest SDK" [Undecided,New] [14:56] annerajb, that would be unflipped [14:57] gusch: I think I understand [14:57] ogra_, the 4 devices which of those did we have problems with the kernel command ling just not being changable ... when we wanted to turn off apparmour we just couldn't do it from the kernel command line cause it had no effect one of the devices [14:57] timp: still around? :) [14:57] ogra_, when i type adb shell i endup at ~ # [14:57] Kaleo: so according to the documentation, that's another bug [14:57] apw, manta iirc [14:57] gusch: so solving this would solve the gallery issue right? [14:57] gusch: so it's definitely a toolkit bug; can you add that info to the bug report? [14:57] Kaleo: yep - from my POV it's a regression [14:58] Kaleo: yep [14:58] * ogra_ relocates for meeting [14:58] ogra_, ok thanks ... and for any of the others are we seeing the bootload always slamming on any other parameters ? [14:58] Kaleo: I could move that bug to the SDK [14:59] sergiusens, when i type adb shell i endup at ~ # is this a flipped image?? [14:59] apw, not sure, i think rsalveti saw them having no effect with mako recently [15:00] these bootloaders SUCK === alan_g|tea is now known as alan_g [15:01] apw, ++ [15:04] annerajb: root@ubuntu-phablet:/# [15:04] thats what a flipped image shows for me [15:05] then i guess i have a unflipped image. since i dont see that. thought i am not at the OS since the OS dosnt boot i think i am at the bootloader/recovery adb shell [15:08] Kaleo: but anyway I will add my workaround, as I knowh who's calling (and I like it more to use the id directly) [15:09] rsalveti, ogra_, sergiusens: so I have some work planned to add a image bootstrap option to phablet-flash so that people can use that to retrieve the latest recovery partition image, flash it, push the latest full image and get a working image based system. Is phablet-flash the right place to add that or would you prefer a standalone tool? [15:10] stgraber: everyone will want to use that, although there is a lot of legacy stuff in there [15:10] stgraber: if you outline the process I can get it going if you want [15:12] sergiusens: that'd be great, I can give you a bit of python code that does what I want when passed the model name, then if you can add that as a --bootstrap-image option (or similar), that'd save me quite a bit of time [15:12] stgraber: sounds good to me [15:12] gusch: thanks for all the work, we'll take it from there [15:12] stgraber, any reason for that ? [15:13] stgraber, i thought we will have the new recovery in all builds anyway [15:13] ogra_: we do, that's why I want the outline ;-) [15:13] so a normal phablet-flash -b would get you what you need [15:13] sergiusens, oh sorry, didnt read all of it [15:14] Kaleo: thank sil2100 didrocks and om26er [15:14] gusch: no worry, thanks you! [15:15] didrocks: I would have kept blaming jenkins ;) [15:15] ogra_: yeah, I expect we already have code to grab the recovery image, the bits I'll need to add is basic parsing of system-image.ubuntu.com and initial flash of those files. It's going to be a bootstrap-only option as once that's done, you should update from the device with system-image-cli instead of phablet-flash on your machine [15:16] stgraber, yeah [15:16] ogra_: I'll also make that option flush any existing data on the device and put it behind a pretty scary warning to avoid most people using it until we make it the default [15:16] gusch: ahah, trust it ;) [15:16] stgraber, uuh, evil [15:16] stgraber, that means people that want to keep their data wont test it [15:16] (I mostly want it there so I can get QA and other very advanced testers to use it instead of having to manually push stuff to their device with fastboot and adb) [15:16] ogra_: that's kind of the point at the moment ;) [15:17] also note that QA actually seems to require to keep stuff in the homedir [15:17] so you would break that [15:17] ogra_: right, initial flash will flush their data, but upgrades done after that (from the device) won't [15:18] ogra_: so I guess that'll be fine for QA. For the other testers, once we're confident enough that the image setup works, we can drop the "format data" call to the upgrader and add some data transition code [15:18] how do i know if my bootloader is installed correctly if the recovery menu has the ubuntu logo at the background?? [15:18] sil2100, I don't know about qtubuntu-sensors on amd64, although I'm not sure what the point is in building it for amd64 [15:18] stgraber, right, but as i understand the data needs to be on the phone before flashing [15:19] stgraber, we wouldnt have had so many issues today if we culd just have wiped /home/phablet/ [15:22] ogra_: hmm, ok, I guess I'll see what I can do once I understand how their stuff works... I'd have expected they run the --bootstrap-image which wipes the device clean, then boot it, push stuff to /home/phablet, install all their tools and then from there on, it's going to be persistant (unless they call --boostrap-image again instead of using system-image-cli for the updates) [15:22] jhodapp: I built it already, but there seem to be some problems with headers - but that's a different story [15:22] Thanks! [15:23] ogra_: I mean, it's simple enough not to wipe /data clean, but then we have to figure out exactly what to move where which is much trickier, so as a first implementation, having bootstrap wipe everything is much easier ;) [15:23] mhall119: Hey ! :) So I tried the Argument api by the SDK team. It work nice ! But I will have to ask them an I/O Api to manipulate files :) So I can't really implement the arguments api for now [15:23] stgraber, no, i was told they want "real world tests" which means they only flash but dont bootstrap (which preserves all of /home plus the network connections) [15:23] sil2100, np [15:24] ogra_: ok, then on the image based devices, they'll just call "adb shell system-image-cli" which will update the device but won't flush the data [15:24] stgraber, i guess its fine if home is preserved over all subsequent upgrades [15:24] so they have to bootstrap once [15:24] ogra_: with the image based setup they should really only use phablet-flash for the very first flash, not after that [15:24] just need to make sure thats the case indeed [15:24] right, good, so we agree === rachelliu_ is now known as rachelliu [15:36] http://pastebin.com/yhwdivmp this is theoutput i get when i type adb shell on what I think is the boot loader does it look correct? [15:37] annerajb, thats clearly an android environment [15:39] bfiller: hi! [15:39] sil2100: hey [15:40] bfiller: since Ugo seems to be away, could I poke you about something? Since there is the ubuntu-ui-extras project, which lp:ubuntu-ui-extras is an empty branch, but there are some sub-branches in it [15:40] bfiller: I would like to know what is the status, and if any of the sub-branches should be pushed and made as lp:ubuntu-ui-extras ? What about daily-releasing this? [15:40] https://code.launchpad.net/ubuntu-ui-extras [15:41] sil2100: going to have to look into this after I finish the call I'm on. will get back to you soon [15:41] bfiller: ok, thanks [15:42] sergiusens, wget is seeded and meta uploaded, so only your MP is missing (and an image rebuild) [15:43] ogra_: and colin's latest click upload (if it's not in yet) [15:43] there was a tiny bug [15:45] ah, k [15:46] ogra_, that means that my image is unflipped right? [15:46] annerajb, do you have a google logo on screen while adb'ed in ? [15:46] or is it just black [15:47] ogra_, on the backgroudn of the cellphone i have the ubuntu circle logo and it says at the top cwm-based recovery v6.0.2.8 [15:48] annerajb, well, thats recovery, what do you get in a normal boot [15:50] ogra_, i dont think i get a normal boot. this is the only thing it boots too no matter how many restarst i do. it always ends up here [15:51] annerajb, what devie was that ? [15:51] epicmtd [15:51] *device [15:51] and you are porting it yourself ? [15:51] ogra_, yeah [15:52] so did you produ.ce a zip for the device already ? [15:54] ogra_, yes yesterday with the help of sergiusens and rsalveti i got the makefiles to produce cm-10.1-20130717-UNOFFICIAL-epicmtd.zip [15:54] ok [15:54] and did you flash it already too ? [15:55] gusch: hello! [15:55] gusch: can I poke you about one more thing? This time related to qtubuntu-cameraplugin-fake [15:55] sil2100: shoot [15:56] gusch: would you mind if I got rid of the cameraplugin-fake-tests package? Since it only installs unit tests, and I don't think we need a package that installs unit tests [15:56] hello. [15:56] Those should be only ran during build, right? [15:57] Does anyone know what is needed for providing multitouch support under Debian 6.0? [15:57] Currently it's only supporting single touch operation. [15:57] Using kernel 3.10.1 [15:57] ogra_, yes yesterday and after i flashed it it changed the backgroudn logo from the cyanogen mod one to the ubuntu one. [15:57] WebVisitor-0: kinda offtopic here ☻ [15:57] WebVisitor-0, linkely nobody in thei channel ... [15:57] sil2100: get rid of them - truth is - they are only there, becauso I didn't manage to get rid of them ;) [15:57] *this [15:58] gusch: ACK ;) [15:58] WebVisitor-0: maybe ask in #debian on oftc ? [15:58] annerajb, right, but recovery is different from the normal boot ... so if you boot, do you see thr google logo for a while, then it reboots and goes into recovery ? [15:58] Already asked in Debian foruns. But no response in there. [15:59] WebVisitor-0, well, this channel is rather for the ubuntu phone OS [15:59] (and tablet OS) .... [15:59] ogra_, i see the samsung logo and after that the galaxy s and cyanogen mod animation and then it ends up on recovery. [16:00] I got this link from the bottom page of https://wiki.ubuntu.com/Multitouch [16:00] Thats why I asked here :-( [16:00] ahh, we stole the channel from that project [16:00] annerajb, ok, try the following from your pc: adb shell cat /proc/last_kmsg >logfile.txt [16:01] annerajb, then dump the content of that file to a pastebin [16:01] http://popey.com/~alan/device-2013-07-17-170108.png <- note "Carrier" is "Aubergine". My carrier is Orange in the UK. Has someone done a global s/Orange/Aubergine/g or something? [16:02] popey do you know where can I get more info on MultiTouch then? [16:02] ogra_, been asking for that the whole morning let me go ahead and get the paste bin last line says attempting watch dog reset. [16:02] popey, that would be awe_ then :) [16:02] Is there some dedicated channel for that? [16:02] WebVisitor-0: no [16:02] annerajb, yeah, the last line doesnt help much :) [16:03] ogra_, also a bunch of messages about permission denied http://pastebin.com/XqxGp3Bv [16:05] OK guys [16:05] I will keep looking [16:05] thanks [16:06] annerajb, so by the stuff it tries to start this looks like you are actually not building the userspace for your device (looks like it tries to start some services a maguro device would use) [16:07] init: cannot execve('/system/vendor/bin/pvrsrvinit'): Permission denied [16:07] this is the first error i see [16:08] awe_: a possibly amusing bug for you https://bugs.launchpad.net/touch-preview-images/+bug/1202279 [16:08] Launchpad bug 1202279 in touch-preview-images "Carrier is "Aubergine" should be "Orange"" [Undecided,New] [16:09] annerajb, pvrsrvinit is a tool used by OMAP chips , you rarely find them in samsung phones ... [16:09] annerajb, i think while you already build the right kernel, your userspace does not use the right repo yet [16:09] hmm thats weird because i only did breakfast epicmtd so i should have anything thats from another device unless my device makefiles are wrong [16:09] ogra_, is the user space the raring-preinstalled zip? [16:10] no [16:10] thats the ubuntu userspace [16:10] what i'm talking about is the minimal android rootfs you need [16:10] which is in your cm-10.1-20130717-UNOFFICIAL-epicmtd.zip [16:10] which is the one produced by brunch epicmtd [16:10] right [16:10] something is wrong there [16:11] ogra_, so i should start looking into why my userspace is using omap tools like pvrsrvinit [16:11] popey, looks like a bug for seb128 [16:11] popey, no i18n translations in the ofono layer [16:11] ;) [16:11] you seem to use the maguro build in userspace while using the epicmtd kernel [16:12] awe_, oh, so a french aubergine is an english orange ... now that finally makes sense :) [16:12] awe_, popey: the backend of that screen is simply not implemented [16:12] haha [16:12] ah okay [16:12] awe_, how do we can the carrier name btw? ;-) [16:12] ogra_, which makefile specifies that maguro is what it should build? [16:12] popey, nothing on that screen actually do anything (maybe we should try to communicate that) [16:13] seb128, pretty sure it comes thru as a latin-1 string from the operator [16:13] annerajb, i think you simply miss some CM branch for your device in the manifest.xml [16:13] i guess the build just falls back to maguro [16:13] awe_, what I mean is "does ofono provide a dbus api I can call to get the info" ;-) [16:13] because that bit is missing [16:13] ogra_, i used the automatic tools i never touched the manifest.xml myself let me open it up [16:14] annerajb, you didnt follow the porting guide ? [16:14] seb128, yes... you should. Looks like that's already being used though... [16:14] seb128, I think the issue is that settings gets the operator name from ofono, and it's "Orange". and it [16:14] #machinelearning === dandrader is now known as dandrader|lunch [16:14] it gets translated to Aubergine? [16:14] ogra_, i did so if the porting guide tool (i think it was brunch) didnt got the right file itś the tools fault :P [16:15] awe_, no, "Aubergine" is a placeholder label, we just did the UI in the panel [16:15] lol [16:15] WOW ! [16:15] awe_, if you tell me what I can call I can replace that by actual datas [16:15] awe_, which is why I'm asking if you have a dbus method or something I can call :p [16:15] doanac: jcollado: not so long ago ^^^^^ very far in the morning here we were discussing how to make autopilot tests work and properly source dbus-session var from the new location and properly stop unity8 inside the user session init. [16:15] =)))) [16:15] * ogra_ just accidentially hit enter when typing porting into chromiums url bar (instead of navigating to the right line) [16:15] seb128, one sec... [16:16] guess what the second hit on google for "porting" is [16:16] xnox: i'm try to test people's patches right now [16:16] :D [16:16] ogra_, phablet-dev-bootstrap i think was the command that i ran [16:16] right after wikipedias description of the word [16:16] rsalveti: https://code.launchpad.net/~mathieu-tl/phablet-tools/exclude-tests/+merge/175328 [16:16] ogasawara, after that i typed breakfast epicmtd [16:16] annerajb, well, you need to tell the tree that it should build for yur device somehow [16:16] seb, you need to query org.ofono.NetworkRegistration properties for the current operator details [16:16] ^ I ran into a file conflict with hud-tools as I upgraded. [16:16] sergiusens: ^^^ [16:17] ogra_, :i passed my device name to both of those commands [16:17] seb128, if you haven't, install 'ofono-scripts', and run ./list-modems on a phone with a valid SIM [16:17] annerajb, well, cross check against https://wiki.ubuntu.com/Touch/Porting#Manual_.28deprecated.29 [16:17] or 'get-operators' [16:17] ogra_, already on it [16:18] annerajb, if the automatic way doesnt work, follow the manual one to check what the automatic one did wrong :) [16:18] seb128, I'll assign the bug to you. You can close it out if we have a work-item for the task already [16:18] awe_, ok, thanks for the info [16:18] awe_, yeah, assign the bug to me [16:20] seb128, done [16:20] * awe_ really goes for lunch now [16:20] awe_, thanks [16:20] awe_, enjoy ;-) === brion_ is now known as brion [16:32] ogra_, this is how my .repo looked before doing it by hand http://paste.ubuntu.com/5884766/ [16:36] tvoss_: piiing! [16:36] tvoss_: hello ;) [16:37] tvoss_: we're having some problems with using the platform-api headers [16:37] doanac: jcollado: any status update? i saw a few results coming in, but not sure if that was luck :) [16:37] will drop off soon (so dont worry:)) [16:38] ricmm: ping [16:39] ricmm: maybe you would be able to help, as libplatform-api-headers on amd64 installs the headers to /usr/include/ubuntu-1, but all headers inside use ubuntu/blabla.h [16:39] asac: i'm trying to get my phone set up to test [16:39] its running now [16:40] awesome [16:40] popey, hey [16:40] popey, are you QAing the phone images? [16:40] popey, hide ! [16:40] lol [16:40] * jono puts hand in front of ogra_ :-) [16:40] sil2100, how come? about to jump onto a hangout. Mind pinging ricmm? [16:41] jono, popey and i currently do manual tests until the automated ones fullly work [16:41] ogra_, right [16:41] ogra_, so phablet-flash only installs the blessed daily? [16:42] does --pending bring in the auto-built daily? [16:42] right, --pending gives you the most recent build [16:42] default installs what is in /current on cdimage instead [16:42] tvoss_, ricmm: libplatform-api-headers installs, for instance, /usr/include/ubuntu-1/status.h [16:42] (which is the blessed one) [16:42] tvoss_, ricmm: while insides have things like: #include [16:42] sil2100, ack, what's the issue? [16:43] Sorry [16:43] #include [16:43] sil2100, yeah, there is a symlink created on package installation [16:43] (since I modified by hand to point to ubuntu-1 [16:43] ) [16:43] Oh [16:43] ogra_: did 17.1 build in the end and does it work? [16:43] tvoss_: could it be that during upgrade the old directory was not removed and now the symlink cannot be installed? [16:43] ogra, asac: what image do i need? I'm running 20130717.1 and it still uses /home/phablet/.dbus-session [16:44] sil2100, not sure ... didrocks ^, can you help here? [16:44] apw: I only had issues (cmdline) with manta, there others worked fine when changing the kernel cmdline in the abootimg config file [16:44] tvoss_: siiince, I have /usr/include/ubuntu, but it's a directory [16:44] apw: so yeah, we can't necessarily trust we'll be able to dynamically change the boot arguments [16:44] didrocks: ^ [16:44] ogra_, asac: n/m [16:44] davmor2: see last link in the topic [16:44] sil2100, weird, then your theory should be correct [16:44] rsalveti, thanks thats what i wanted to know for what i am being asked to do anyhow, thanks [16:44] doanac, it just doesnt delete it if it exists :) [16:44] tvoss_: sil2100: do we need to have ubuntu-1, do we expect installing/supporting multiple version of the -dev API? [16:45] annerajb: so does it work after just flashing the cm zip? [16:45] I mean, you should at least get a black screen [16:45] if so, please paste your logcat output [16:45] didrocks, you told me that having a versioned include-dir is good style :) [16:45] once we know everything in there is working as expected, we can test it with the ubuntu image [16:45] rsalveti, ogra_ has been helping me out and apparently my userspace is different than the kernel [16:45] tvoss_: are you sure? I bet you mix people :p [16:45] rsalveti, http://pastebin.com/XqxGp3Bv thatś the log cat [16:45] ogra_: annerajb's device might be a pain to port to flipped, as it's the 'mtd' type (using yaffs2 instead of ext4) [16:46] might have some other specifics as well, we need to check that in detail later on [16:46] didrocks, you said on the mp your eyes are bleeding less :) [16:46] tvoss_: didrocks: i'm not sure there is any point in having versioned include-dirs. If sonames are different, runtime is co-installable. We can even name libfoo1-dev if we want, and some packages do that. But don't use versioned header paths, in the source tree. [16:46] rsalveti, argh [16:47] xnox, didrocks otp now, will come back to you later [16:47] annerajb: oh, a bunch of permission denied, weird [16:47] tvoss_: didrocks: instead, configury / cmake /pkgconfig can point to -I /usr/include/ubuntu-sdk/1/ [16:47] tvoss_: interesting, maybe I didn't express myself good enough, but I think we shouldn't [16:47] rsalveti, well, his log shows a lot of maguro stupp (pvrsrvmgr) [16:47] rsalveti, or are there other samsung omap devices ? [16:47] xnox: we need to have a stragegy for those headers, we'll discuss it [16:47] tvoss_: didrocks: but developers in their source code should still simply write #include [16:48] ogra_: should have, let me find what is the soc used there [16:48] might even be omap3 [16:48] xnox, +1 [16:48] rsalveti, ouch ... [16:48] tvoss_: didrocks: throught magic of gcc -isystem they can be anywhere, even in a local dir / repo /checkout / anywhere on filesystem. [16:48] annerajb, so ignore what i said then, seems your usespace might be fine [16:49] rsalveti, Platform: Samsung Exynos 3 < that? [16:49] tvoss_: didrocks: e.g. it's used a lot around android. From gcc point of view all paths are relative, and flags determine the locations of what is "considered systemwide", don't think too much in terms of Debian Policy and typical packaging properties. but instead look at other sdk's. [16:49] ogra_: 1GHz Samsung-Intrinsity S5PC110, with PowerVR SGX540 [16:49] e.g. all of apple's andorid's window's SDKs, simply ship a tarball with "usr/include" in it. [16:49] ah [16:49] ok [16:50] Hummingbird [16:50] annerajb: yeah [16:50] that explains the pvrsrvmgr stuff then [16:50] and then you just point your compiler: gcc -isystem path/to/my/local/unpakced/../current/(symlink)/usr/include [16:50] ogra_, but we still have the permission issue [16:50] yes [16:50] thus a package in it self should just installer headers into $prefix/include without any versioned dependenices, those are added later. [16:50] tvoss_, ricmm: can you guys change that in lp:platform-api then? I think that sounds safer than ln -s, as it just prooved to be buggy ;p [16:51] annerajb: so first goal is to have a useful android image, which can start surface flinger and such [16:51] sil2100, yup [16:52] sil2100, mind filing a bug and assigning it to me? [16:52] you can even use the surface flinger tests we have [16:52] then once we know it all works, we can flash the ubuntu one [16:52] tvoss_: working on it! [16:52] rsalveti, all right link to the surface fligner test image? [16:52] sil2100: hardlinks should be acceptable, soft links not so much. [16:52] rsalveti, he cant even boot into android userspace yet [16:52] ogra_: yeah, that's the first step :-) [16:52] :) [16:53] annerajb: mmm ubuntu/hybris/compat/surface_flinger/ [16:53] sil2100: e.g. andoird's 14 abi/api levels are generated by "adding changes into lowest possible level , and hardlinking into upper ones" [16:53] that should give you /system/bin/direct_sf_test [16:53] xnox: maybe you should just do the MP to fix this? It seems it will be the quickest :p [16:53] didrocks: where is the MP in question? me just jumped into a conversation? =)))) [16:54] xnox: there is none, it's something that needs to be fixed [16:54] in lp:platform-api [16:54] so that sil2100 can land https://code.launchpad.net/~sil2100/qtubuntu-sensors/lib_private/+merge/175331 [16:54] popey: the last link is wrong I tell you ;) [16:55] rsalveti, my /system/bin/ is empty [16:55] what? something is very wrong in there [16:55] maybe if failed when flashing your zip? [16:56] rsalveti, yay foudn something critical process service manager exited 4 times in 4 minutes rebooting into recovery [16:56] annerajb: check if out/target/product/epicmtd/system has everything in there [16:56] you can also extract the zip file to compare [16:56] and then understand why it might have failed the flashing via recovery [16:57] davmor2: fixed! :D [16:57] rsalveti, you mean the zip i build or the one i downloaded from ubuntu.com? [16:58] annerajb: the one you built [16:59] you don't need the ubuntu one yet [16:59] didrocks: sil2100: right i need to poke that in-depth. I'm EOD and need to run out. Can I just claim bug #1190698 as I don't think rpath is the right answer here. [16:59] bug 1190698 in qtubuntu-sensors "should libqtubuntu_sensors.so.1.0 be a private lib?" [Undecided,In progress] https://launchpad.net/bugs/1190698 [16:59] popey: Nexus Status: Pop :D [16:59] rsalveti, ok so there must have been a issue flashing because the zip/system/bin contains a bunch of binaries [17:00] right [17:00] xnox: sure, add the platform-api bug task then to fix the ubuntu-1 [17:00] tvoss_: FYI ^ [17:00] didrocks: ubuntu-1 is a well typo, and has been removed from proposed fix already. [17:01] davmor2: /topic [17:01] bah [17:01] davmor2: you don't see "Nexus Status: http://bit.ly/18kIrhM" [17:02] renato_: can you please clear this up?? what piece of the stack for the network indicator actually talks to NM to list access points and stuff? [17:03] popey: I see pop http://ubuntuone.com/2BYe6VAgirY0lV2nOwztAz [17:03] rsalveti, what should i clear before trying to flash the recovery? [17:03] rsalveti, wipe cache? data/factory? [17:03] davmor2: file a bug in your irc client ☻ [17:03] annerajb: both [17:03] annerajb: which recovery are you using as well? [17:04] tvoss_: https://bugs.launchpad.net/platform-api/+bug/1202309 [17:04] cwm? [17:04] Launchpad bug 1202309 in platform-api "libplatform-api-headers needs to make its headers accessible through #include " [High,New] [17:04] popey: my just be a refresh thing [17:04] rsalveti, he has an ubuntu logo, so that part worked [17:04] rsalveti, yes 6.0.2.8 but it has a ubuntu logo on the background [17:04] oh, so it's our own recovery [17:04] sil2100, thx [17:04] popey: hmmm no .17 builds for my phone then and will this auto update each day? [17:05] rsalveti, yes should i be using it? itś the one that was part of the build i did yesterday [17:05] annerajb: in theory, yes :-) [17:05] annerajb: I know recovery dumps the log in /cache/recovery [17:06] you can get inside that when booted in recovery, adb shell [17:06] maybe you can find more there about what might be happening [17:06] rsalveti, it failed saying something about symlinks failed but the first time it ran fine the second time is when i usually get that message so i think something may not be clearing up correctly [17:07] rsalveti, got it OOM [17:10] rsalveti, http://pastebin.com/8cxgJ9Ba this device has a smaller /system partition so it may be running out of memory because i either failed to apply a patch correctly or i have extra stuff that i didnt delete when i made the bootstrap [17:10] hey.. [17:11] what is other channel for development ? [17:11] discussion ? [17:11] ubuntu-app-devel === alan_g is now known as alan_g|EOD [17:14] annerajb: weird, in theory this zip should be smaller than the original one from CM [17:14] annerajb: can you try flashing that with another recovery? [17:14] rsalveti, in that case i may have inputted the wrong numbers on the partition space change i did. [17:15] right, could be as well [17:18] bfiller: hmm, would you mind if I remove phone-app from the daily-release config/ [17:18] bfiller: since I guess it's still valid that this package gets removed and is deprecated, yes? [17:20] sil2100: well we still need it the build until we have a replacement, fine with it staying in the ppa until we have a replacement as it won't go in the archive [17:20] rsalveti, i verified and i have the patch applied that was applied to get quantal to build i am going to try another recovery. [17:21] bfiller: but no new merges will go in related to it, right? [17:21] bfiller: since it's already in the archive right now, so at least it won't disappear [17:22] Wait, no [17:22] I'm wrong [17:22] Scratch that... [17:22] sil2100: there might be some merges that still need to happen as bug fixes until we land the newer stuff [17:23] bfiller: right, ok, so let me leave it for now [17:24] bfiller: any news related to ubuntu-ui-extras ? [17:25] sil2100: not yet, been in meetings for last hours, will look as soon as off phone [17:25] bfiller: ok, if anything, send me an e-mail about it if you can, since I need to EOD pretty soon [17:29] rsalveti, BTW i have to do a factory stock rom because clock work recovery cannot install because it needs the /system partition also the version i have of clock work recovery is 5.0.2.7 [17:30] rsalveti, if i try to install with that version of CWR it fails because it can´t read the device property from the /system partition [17:30] oh, quite old [17:30] maybe twrp [17:30] all right let me update to a later CWR [17:31] thought this one installed fine the quantal build [17:32] cyphermox_: rsalveti happroved [17:35] rsalveti, i looked at the clock work website and the latest version they have for my device is the one i have 5.0.2.7 maybe the latest recovery 6.0.2.7 uses too much memory making it go oom while installing? [17:36] yeah, not sure, you could also try flashing the original cm image (10.1.2) with the latest cwm you have [17:36] and see if the error happens there [17:36] otherwise it might be something specific to your port [17:36] alright, time for some lunch now [17:36] rsalveti, original cm image from where? [17:37] rsalveti, bona petit [17:37] annerajb: http://www.get.cm/?device=epicmtd [17:37] as your port is probably based on that code anyway [17:37] and the image there might be even bigger [17:38] thanks :-) === Namidairo is now known as Namidairo`bnc [18:02] om26er: https://code.launchpad.net/~sergiusens/phablet-tools/test_update/+merge/175355 [18:03] ogra_: habemus build? [18:05] ogra_: also https://code.launchpad.net/~sergiusens/ubuntu-seeds/locationless/+merge/175320 [18:08] boiko: hey, where are we at? [18:09] sergiusens: telephony-service packaging MR passed and was approved by tiago [18:10] sergiusens: for the other two (dialer-app and messaging-app) CI passed, but tiago needs to fix some things in the MR itself [18:11] Saviq, the gsettings-qt api break has landed, the unity8 branch for background settings should be safe to land now [18:14] kenvandine, cheers [18:14] Saviq, the latest CI run failed so i triggered a rebuild [18:14] i'll top approve it when that passes [18:15] sergiusens, ogra_ : can one of you review this to add address-book-app to the seed: https://code.launchpad.net/~bfiller/ubuntu-seeds/ubuntu-touch.saucy-address-book/+merge/175359 [18:15] kenvandine, there's no review from our people yet [18:15] kenvandine, so please let us do our stuff :) [18:15] oh [18:15] sorry... i thought it was :) [18:16] it's not waiting on anyone else, so likely not in anyone's queue [18:16] want me to add a reviewer? [18:17] bfiller, merged [18:17] ogra_: thanks === gatox_ultra is now known as gatox [18:18] boiko: ack, thanks [18:18] sergiusens: thanks for looking into that [18:18] i'll wait with the meta upload since i know that there is another seed merge from ricmm waiting [18:18] ogra_: yeah, this one: https://code.launchpad.net/~sergiusens/ubuntu-seeds/locationless/+merge/175320 [18:19] sergiusens, merged too [18:25] sergiusens, why would we have a new build btw ? [18:25] ogra_, no 17 today? [18:25] ogra_: I thought we wanted to try the click packages [18:25] sergiusens, as i understood asac he wants to see the dashboard green before i can start a new one [18:26] gema, asac: i've got utah tested and working on touch images again. the unity8 fix is in, the dbus session is in, the --pending is in, and now i'm waiting for the PPA to build [18:26] 0.14.2 will be released shortly and I'll be able to kick off smoke testing again [18:26] (that is, the tests need to work first) [18:26] pmcgowan, not released, you could install it using --pending [18:26] ogra_, anything wrong with it? [18:27] doanac: heads up on the image updates stuff that is comming [18:27] (17 failed dues to a PPA error, 17.1 is fine) [18:27] pmcgowan, nothing except the new rule is that we are not allowed to release them if the dashboard isnt gree [18:27] n [18:27] ogra_, ok ty [18:27] pmcgowan, http://reports.qa.ubuntu.com/smokeng/saucy/ [18:28] pmcgowan, it works absolutely fine though [18:28] (teh test suite is broken) [18:28] gotcha [18:28] yar [18:29] heh [18:29] now thats intresting [18:29] http://reports.qa.ubuntu.com/smokeng/saucy/image/3038/ [18:29] Summary of smoke testing for saucy touch image 20130717 ? on maguro [18:29] ogra_: so first step is to get a new build with tests and then add the click hook? [18:29] i wonder why it adds the question mark [18:30] sergiusens, dunno, i thought doanac would be able to re-run them once he finished his loacl tests [18:30] sergiusens, but i can also just roll a new image anyway i guess ... let me do that livecd-rootfs merge :) [18:31] sergiusens, did cjwatson's fix land already ? [18:31] ogra_: let me check, oh, I have a merge conflict I'm just seeing now :-/ ... changelogs have been change! Go figure :-P [18:31] change[d] [18:32] yeah, i would have handled that :) [18:32] * ogra_ saw it already [18:36] doanac: sounds good === dandrader_ is now known as dandrader [18:36] doanac: let me know when it lands in the lab so that we know which images are tested with the --pending change [18:37] gema, seems the system picks them up anyway and several people want a fresh build [18:37] gema sure [18:38] gema, so i'll just trigger a build later and utah should just do its job :) [18:38] (that will then be 20130717.2) [18:39] ogra_: note: the 0.14.2 utah ppa still hasn't finished building. so i still need time before I can deploy to our server [18:39] ogra_: awesome [18:39] doanac, we still need a merge of sergiusens stuff before i can start a build, that will still take a while, no worries [18:40] i guess you will be done before we are ready [18:40] ogra_: https://launchpad.net/ubuntu/saucy/+source/click/0.1.5 [18:41] sergiusens, ah, still in proposed [18:41] (and waiting for ppc it seems) [18:41] Fix infinite recursion in ClickUser.click_pw. [18:42] yeah :) [18:42] https://launchpad.net/ubuntu/+source/click/0.1.5/+build/4802230 [18:43] start in 5h [18:43] that doesnt look good [18:43] ogra_: I wonders why powerpc builders build private jobs :-P [18:43] ogra_: https://launchpad.net/builders/ [18:44] security stuff [18:44] i guess [18:53] rsalveti, i ran df -h on a loop while insatlling the bootstrap i build yesterday and the system partition ran out of disk space. ill have to look into trimming that image. [18:55] sergiusens, merged, but i'll refrain from releasing livecd-rootfs until click is in [18:55] ogra_: that's good enough [18:56] i dont really want to risk an endless click loop at buildtime :) [18:58] Hi there [18:58] rsalveti, mfisch: another powerd MR: https://code.launchpad.net/~sforshee/powerd/clean-up-cli-output/+merge/175376 [18:58] sforshee: ah, I forgot to to the last one too [18:59] sforshee: I'll do both now [19:00] mfisch: thanks === racarr is now known as racarr|lunch [19:05] sforshee: commented on the upower on [19:05] sforshee: can we delete asac's reviews? [19:06] yes [19:07] sforshee: I'm +1 on removing the "We get signal" call, but when MMS works, we'll need to retest it all, it might be a different signal [19:09] sforshee: approved the CLI review [19:18] sergiusens: who is working on/owns the browser app? [19:19] mfisch: oSoMon [19:19] sergiusens: thanks [19:20] mfisch: we don't need the "we get signal" print. If you enable debug the same information will be printed for any unknown signal. [19:20] mfisch: responded to your comments on the upower MR [19:26] sergiusens: here's the bootstrap script for the readonly image based setup: http://paste.ubuntu.com/5885313/ [19:27] sergiusens: we plan on sending a call for early testers (for people who have a spare device and don't mind loosing data) on Friday, so it'd be great if we could have that code as a --bootstrap-image option in phablet-flash by then (otherwise I'll just link to that script directly) [19:28] stgraber: will we have a way for those folks to install apps? [19:30] mhall119: not until click packages are fully working (I'll make it very clear in the blog post, that it's read-only with no way of running apt) [19:30] ok [19:30] rsalveti: the apps I added to the ubuntu-seeds branch, those are still being installed from .debs not .click packages right? [19:30] what I really want to get is feedback on all 4 supported models and making sure that basic functions (telephony, network, ...) work as well as they do on flipped [19:30] then hopefully soon after that, we'll get click working and we can go with a bigger pool of testers [19:30] and soon after that, just make it the default and be done with it :) [19:32] stgraber: I'll let you know if camera and audio don't work on the N7 :) [19:32] stgraber: ack [19:32] stgraber: it's going to happen today :-) [19:32] stgraber: I'll be testing on two devices and also check what's needed for testing since it's a hot topic ;-) [19:34] sergiusens: cool, thanks! [19:34] sergiusens: I tested on mako and grouper. grouper had a problem (cache partitioning not mounting) which I now fixed, so it'll be good in the next image. [19:35] sergiusens: if you test with grouper, it'll work, however further updates with system-image-cli won't, so you'll need to re-bootstrap with the next daily to get the fix [19:35] ok, no worries, these are not final systems :-) [19:35] maguro and manta may be affected by the same bug, not sure, in any case, I expect a bootstrap to work and updates to be broken, so as long as you re-bootstrap tomorrow, you'll be fine [19:37] I've got to run for very late lunch and some errands but should be back in an hour or so [19:38] sergiusens, where can i find direct_sf_test in the raring-preinstalled-phablet-armfh*.zip ?? === dandrader is now known as dandrader|afk [19:39] sforshee: approved the upower mp [19:39] mfisch: thanks [19:41] mzanetti: hey, are you around? [19:41] ogra_, now when i boot it seems to be on android since it shows root@android:~# [19:42] ogra_, and it has the ubuntu partitions loaded [19:42] great [19:43] ogra_, thought i get a black screen where is /var/log/messages?? [19:43] black screen is good that means surfaceflinger works :) [19:43] (something else doesnt though) [19:43] use logcat [19:44] i am watching the logcat output dont see anything meaningful let me figure out how to get this into a file [19:47] ogra_, http://pastebin.com/Zb8Yy04x i dont see anything apart fromt he time stuff that's error level messages [19:48] annerajb, hmm, i dont see it starting any of the ubuntu-android bits ... like uchroot or ubuntuappmanager [19:48] rsalveti, ^^^ [19:49] ogra_, where should those binarys be let me take a look if they are there === Guest99716 is now known as balloons_ [19:49] ogra_, i found uchroot in /system/bin [19:50] yeah, android in the unflipped images should execute it [19:51] i tried running ubuntu_chroot shell and it said cannot determine your username. [19:51] mhall119: yes, sergiusens is working on converting them to click packages [19:51] and if i run uchroot it hangs and dosnt appear to do anything ogra_ [19:52] uchroot will only work if you also installed the ubuntu zip [19:52] but first let's make sure the android one is working as expected [19:52] so it seems the device is now up and running correctly [19:52] SF runs [19:52] annerajb: can you run ps to see if you have ubuntuappmanager running? [19:53] rsalveti, it's running [19:53] cool [19:53] then just flash the ubuntu zip now === dandrader|afk is now known as dandrader [19:53] hmm i did lol rsalveti this is running the ubuntu zip [19:54] rsalveti, a whole ps http://pastebin.com/ZsWanKkF [19:54] which zip did you download ? [19:56] annerajb: I mean, did you also flashed the raring zip I pasted you yesterday? [19:56] raring-preinstalled-phablet-armhf.zip [19:56] rsalveti, yes i did [19:56] yeah, that's the one [19:56] then check if you have the ubuntu rootfs under /data/ubuntu/ [19:57] i do rsalveti [19:58] annerajb: then just need to check why 'ubuntu_chroot shell' might not be working for you [19:58] annerajb: did you enable all the required kernel configs? [19:58] otherwise it'll fail when enabling the container === racarr|lunch is now known as racarr [20:01] rsalveti, i applied some kernel changes but werent from the wiki page was from the last guy that got the image working for quantal i am missing some namespace kconfig variables that have to be enabled [20:02] so ill do all the kernel changes and recompile and see if i can get another build tonight [20:05] annerajb: cool [20:06] thanks for the help rsalveti and others [20:06] getting close :-) [20:08] rsalveti, how should i build now make clean; make? or make clean; brunch epicmtd? [20:08] rsalveti: by using the raring image, won't platform-api break since the refactor was after? [20:08] sergiusens: don't think we had any api/abi changes in there [20:09] the refactoring happened mostly in the ubuntu side [20:17] rsalveti, if i would like to have this stuff on git. i assume i shoudnt merge request that into cyanogen mod since it will break the android version so where should i store the android changes make a fork on github and commit the changes there? [20:18] annerajb: pushing stuff in your personal github account would be a good start, we can then merge it in phablet.u.c later [20:18] sergiusens: Yeah, sorry, we're temporarily one builder short on powerpc - that'll be sorted out later this week [20:19] all right. rsalveti also didnt tell you how i fixed the running out of disk issue.. [20:19] we hope to have gerrit in place soon, then once merged you'll be able to propose merge requests there [20:19] annerajb: what was the problem? [20:19] sergiusens: I've scored click/0.1.5/powerpc up insofar as it'll make much difference [20:19] rsalveti, had to delete all the apks in /app and /media/audio /media/bootanimations.zip after that when installing the bootstrap i build yesterday it was left with 300k of disk. [20:20] annerajb: right, cool [20:20] cjwatson: thanks for the update [20:21] rsalveti, are those apks really necesary when installing the bootstrap? i assume there are other phones where there may not be a lot of space ont he system partition. [20:21] annerajb: probably not [20:24] rsalveti, is the network indicator broken today? [20:24] pmcgowan: it shouldn't, using which image? [20:24] 17.1 [20:24] I think we didn't yet publish today's build [20:24] I am connected but it says not [20:24] hm, let me flash that [20:25] but I know ogra_ tested it already [20:25] bfiller, I dont see thumbnails in history yet [20:27] pmcgowan: checking [20:28] rsalveti, i didnt test todays build since it was clear that asac didnt want it released, popey tested though [20:28] right, flashing it now to see [20:28] ogra_: next one should be all good right? [20:28] seems doanac fixed all the issues already [20:28] rsalveti, hopefully [20:29] :-) [20:29] rsalveti, i would trigger a build but then we cant have sergiusens click work in [20:29] needs this fix https://launchpad.net/ubuntu/saucy/+source/click/0.1.5.... which is stuck waiting for a free powerpc builder [20:30] oh, wget will now be installed by default [20:30] awesome [20:30] hahaha [20:30] that was so annoying [20:31] ogra_, cant we disable powerpc builds for things [20:31] :) [20:31] how long will that take? [20:31] pmcgowan, thats the archive, no we cant [20:31] click n the ppc link ... it should tell you the ETA [20:31] yeah, don't trust that [20:31] last time i checked it was 5h [20:31] 2h now [20:31] 18 jobs (2 hours 40 minutes) [20:32] ah, so it moved forwatd 3h within 1h [20:32] queue, so should be done soon indeed [20:32] I scored it up [20:32] But I'll just force it into the release pocket [20:32] cjwatson, thanks [20:32] great [20:32] so i can upload livecd-rootf then [20:32] +s [20:32] i held that back to not hit an endless loop :) [20:33] pmcgowan: working for me, check your versions of these: http://pastebin.ubuntu.com/5885486/ [20:34] bfiller, let me make sure it flashed, probably no space [20:34] bfiller, bah nm [20:34] rsalveti, how can i setup a proxy for downloading the ubuntu-ramdisk.img??? i am behind a proxy and it's not downloading it [20:34] gets me every time [20:35] pmcgowan: although I am seeing some issues I wasn't seeing during testing, some thumbnails are empty [20:35] livecd-rootfs uploaded [20:35] annerajb: hm, need to check that [20:36] * ogra_ sets an alarm to be notified once it hits the achive and goes afk [20:36] it downloads the debs with urllib2 === rmeyerri1cks is now known as datachomper [20:40] rsalveti, i setup the http and apt proxys let's see if that fixes it [20:41] rsalveti: should I also move the upstart job to maliit-framework or should that be a new upload? [20:41] so we can sync it better [20:41] sergiusens: let's take that as another upload [20:42] annerajb: in theory urllib2 will use the http_proxy env variable [20:44] just need to check if it's indeed using http there, instead of https [20:44] There, you'll have click 0.1.5 in saucy after the next publisher run [20:45] \o/ [20:45] Which will start in ~3mins [20:46] annerajb: yeah, using https [20:48] rsalveti, all right just setup all the proxys in upper case and lower case and it's been building for a while in retrospective i shold have run make bacon [20:53] xnox: you said you'd be replacing the pull-lp-bin tool as well, any progress on that? [20:54] that might handle proxy better as well [20:54] rsalveti: he's EOD iirc [20:54] hahha, he's always on-line [21:14] WebbyIT: thanks for the review [21:15] I’ll take the liberty of top-approving myself, seeing that it’s trivial enough and has been approved by one person + jenkins already [21:18] oSoMoN: yeah, I think that you can do it, certainly does not break anything [21:34] is ummmmm usb tethering possible? http://i.imgur.com/FszXfdO.png === salem_ is now known as _salem [22:25] ho [22:25] ttyt [22:27] rsalveti, sergiusens, build running [22:34] ogra_: to finish tonight? [22:34] 50min [22:34] popey, dont worry about testing [22:34] no? [22:34] we need to wait for the dashboard anyway [22:35] well, you can if you feel like [22:35] balloons_, do you have a second for some hand holding with autopilot (trying to get playing a song in music-app to work)? [22:37] sergiusens: so does that look easy to integrate? [22:37] stgraber: I'm on it, I was just using python2 though, never migrated since I need to work on precise and never check [22:38] sergiusens: should be easy enough, just import urlretriever from urllib instead of urllib2.request [22:38] I believe that's the only py3 specific bit in that code [22:39] vthompson, shoot [22:41] balloons_, excellent. So I have an UbuntuShape that I want to test that the image.source changes to a different icon when clicked. However, I can't seem to get access t the UbuntuShape's image attribute. [22:42] Is there any docs for dual booting Ubuntu touch? [22:42] balloons_, I get the following: AttributeError: Class 'UbuntuShape' has no attribute 'image'. [22:42] what's the qml file show? [22:43] Azelphur: i have seen someone dual boot on nexus 7 [22:43] popey: nice [22:43] http://barrenfrozenwasteland.com/2013/03/dual-booting-android-and-ubuntu-touch-on-the-nexus-7/ [22:44] balloons_, http://paste.ubuntu.com/5885798/ [22:46] vthompson, ahh.. add an objectName for the image [22:46] balloons_, do I have to then reference this by the objectName? [22:46] vthompson, then do the select_single type Image and objectName=yourname [22:47] Is Ubuntu touch gonna have a menu option for rotate lock? [22:47] balloons_, select_single is one of the functions in your emulator? [22:47] vthompson, http://unity.ubuntu.com/autopilot/api/introspection.html [22:48] vthompson, it's the normal introspection routine [22:52] vthompson, so self.app.select_single('Image', objectName='shapeImage') [22:53] balloons_, Ok does the UbuntuShape still need an objectName? I get the following error: AttributeError: 'NoneType' object has no attribute 'source' [22:55] rsalveti: what's the best way to know if somebody is working on a port so I don't waste effort on a port [22:55] Is there a wiki page? === kc8qvp_ is now known as kc8qvp [23:11] annerajb: yes https://wiki.ubuntu.com/Touch/Ports [23:11] annerajb: sorry, https://wiki.ubuntu.com/Touch/Devices [23:13] popey: thx [23:16] popey: Hey ! How are you ? :) Do you know if the sdk team is working on file I/O ? === Zic is now known as Guest81623 [23:21] Chocanto: hi, how do you mean? [23:29] popey: I mean if the file I/O api is planned, or if they are working on :) [23:31] The SDK team have lots to do ☻ [23:31] I don't believe direct file io is something we are doing [23:35] ogra_: ah cool, you're already doing a touch build, I was just about to kick one [23:35] popey: Ok :) It was just to know, because I need it for docviewer ^^ [23:36] Chocanto: https://wiki.ubuntu.com/Touch/SdkFeatureStatus see "Content management" [23:36] popey: to avoid having to use the temporary launcher, and use Arguments and I/O api of sdk team [23:37] Ah thanks, I didn't know the existence of this page ^^ [23:42] sorry about that Chocanto [23:43] nono, that's good :) It was just because mhall asked me to test Arguments in docviewer, but I can't without file I/O. But I'll wait ;) [23:48] Chocanto: I didn't think the docviewer had to do any IO in QML [23:48] or your wrapper [23:48] I thought it just passed file paths to other plugins [23:50] mhall119: That's juste because I need to know the mimetype of the file, and other usefull informations about it. [23:50] mhall119: I can do it with a plugin [23:50] ah, I see [23:51] mhall119: Because arguments work nice and exactly fit with what we need, so I can get the filepath [23:52] so then are you just going to make a plugin to give file meta-data like mimetype? [23:52] there might already be something like that somewhere [23:53] mhall119: I hope, but if not yes I will do. [23:53] ok [23:54] mhall119: Did you asked tsdgeos about the poppler package ?