=== _salem is now known as salem_ === salem_ is now known as _salem [07:46] Saviq, which unity-mir package version did you test the wizard? [07:46] pete-woods, ping :) [07:47] Cimi: in a hangout - be with you in a few minutes [07:47] sure [08:18] morning mhr3, can the unity scopes related bug tasks be marked as Fix Released on https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1297889 ? I.e. has this landed on any images yet? [08:18] Ubuntu bug 1297889 in unity-scopes-api "Scopes are not translatable" [Undecided,In progress] [08:19] Cimi: okay, done now [08:19] pete-woods, cool [08:19] pete-woods, so I tried yesterday with your code [08:19] dpm, done [08:19] pete-woods, but I receive no display [08:19] pete-woods, I am wondering where are infographics and if they are being created [08:19] awesome, thanks mhr3 [08:20] Cimi: sure, if you have no infographics installed, you won't get any data [08:20] pete-woods, I think I have camera app [08:20] Cimi: do you have anything in ~/.cache/usermetrics/sources/ [08:20] let me see [08:21] if you have the camera app installed (from click and from my branch) then you should have a symlink in there [08:21] pete-woods, no [08:21] mmm [08:21] maybe that's why [08:21] pete-woods, do you have updated silo? [08:22] Cimi: might be easier to just generate a fake data source [08:23] Cimi: touch ~/.cache/usermetrics/sources/foo.json [08:23] APP_ID=foo ../trunk-build/src/libusermetricsinput/usermetricsinput source-id "%1 foos" "no foos" "" 1.0 2.0 0.1 3.5 5.2 [08:24] sorry, APP_ID=foo usermetricsinput source-id "%1 foos" "no foos" "" 1.0 2.0 0.1 3.5 5.2 [08:24] if you have the stuff installed, you won't need to do it from the build dir [08:24] -bash: usermetricsinput: command not found [08:24] that will just create some data for a source called "source-id" for the app "foo", with that random data [08:25] Cimi: you'll need to install the libusermetrics-tools package [08:26] ok I run the command [08:26] now the model should be populated? [08:27] actually, phablet uid seems 32011 [08:27] Cimi: check /var/lib/usermetrics/{uid}/default/ [08:28] Cimi: isn't it easier to try this stuff out on your desktop first, though? [08:28] Cimi, the one from greyback's "fix crash on exit" branch, why? [08:32] Saviq, I was wondering if is the correct one that kills the process [08:32] Saviq, if it contained https://code.launchpad.net/~cimi/unity-mir/unity-mir.stop-server_wizard [08:33] Saviq, because I tried wizard 1019209123214328 times and has worked 1019209123214328 times [08:33] (this weeks ago) [08:33] so I will have a look again if we're sure you tested with this specific branch of unity-mir [08:34] Cimi, it contained even more [08:34] Saviq, ok [08:34] Saviq, I'll have a look again then [08:34] Cimi, https://code.launchpad.net/~gerboland/unity-mir/shutdown-crash-fix3/+merge/218394 [08:35] Saviq, this touches the code I changed [08:35] Cimi, yes, it fixes when Mir is the one that's shutting down [08:36] you guys said you tested with the wizard [08:36] Cimi, it was fine with wizard before, but crashed in unity8 [08:36] but the wizard you tested was not operating as a mir client [08:36] you were using the old wizard code [08:36] no I wasn't, I used the branch I commented on [08:36] I used both that unity-mir and the packages built from your wifi branch by CI [08:37] Saviq, so you used with wizard-wifi? [08:37] cool [08:37] Cimi, yes [08:37] Cimi, wouldn't have commented on that MP otherwise [08:37] and this has issues? [08:37] Cimi, it did not start unity8 for me after the wizard exited [08:37] Cimi, 'cause the environment had the wizard *MIR* variables still [08:38] Saviq, so there was a situation with my wizard wifi and unity mir patched that was working fine [08:38] now with this new branch wizard doesn't work anymore [08:38] but I will have a look today [08:38] Cimi, it was just a one-time thing [08:38] Cimi, it might have been caused by something else [08:38] ok [08:39] there might have been a mistake in the upstart job that now is exposed [08:39] I am not saying my branch was correct and unity mir is now wrong [08:39] I am just saying that I tested it deeply (and after that working on the privacy branch for wizard) and always worked fine [08:40] I will have a look [08:40] first I want to go ahed with infographics for a bit [08:41] Cimi, I'll try again in the mean time [08:41] pete-woods, I think it is easier to try on the desktop [08:42] pete-woods, but I thought the only way to test infographics was with the camera app, so no desktop for me [08:42] pete-woods, on the phone, I don't have /var/lib/usermetrics/{uid}/default/ [08:42] I have root@ubuntu-phablet:~# ls /var/lib/usermetrics/ [08:42] tmp usermetrics6.db [08:43] and tmp is empty [08:43] Cimi: is the usermetricsservice running? [08:43] how do I make sure I have all the right packages installed and at right version? [08:43] (as the phablet user) [08:43] i.e. (as phablet) status usermetricsservice [08:44] I have usermet+ 25099 0.0 0.1 33236 3120 ? S May07 0:00 /usr/lib/arm-linux-gnueabihf/libusermetrics/infographicservice [08:44] nope as phablet is stop [08:44] * Cimi starts [08:45] pete-woods, started and took picture, nothing changes inside var lib [08:46] Cimi: what about inside ~/.cache/com.ubuntu.camera/usermetrics/ [08:46] pete-woods, I have a json [08:46] Cimi: okay, well that's something [08:47] that at least means that the camera app is doing its job [08:47] Cimi: ~/.cache/usermetrics/sources/ was empty, though, right? [08:48] dednick: around? [08:48] I would need this to be top-approved before I get a silo [08:48] https://code.launchpad.net/~kaijanmaki/unity8/indicator-root-state-icons-fix/+merge/213727 [08:48] just look at proposed change and see if it makes sense [08:48] the testing will be done after the silo is ready [08:49] pete-woods, there's just foo.json [08:49] the MR will not go forward without the TestṔlan template [08:49] but I need a Approved to get the silo [08:49] Cimi: I think what that means is that the updated version of libusermetrics wasn't installed when you installed the camera app's click package [08:49] and it didn't register the click hook [08:50] Wellark: "Antti, we'll also need you to copy&paste the top part of https://wiki.ubuntu.com/Process/Merges/Checklists/Unity8 as description and answer the questions." [08:51] Wellark, TestPlan doesn't affect you [08:51] dednick: that's for the testing phase [08:51] once there are packages to test [08:51] Wellark, it's the lander that's running the TestPlan when releasing [08:51] Wellark, nope, the checklist is on-submission [08:52] Wellark: presumably you've tested it yourself :) [08:52] Wellark, testplan is executed later, with multiple MPs in a silo etc. [08:52] yes, but with hand built packages [08:52] pete-woods, so what do I do now? [08:52] Wellark, per-MP tests are ran in the CI [08:52] Cimi: how exactly did you install the updated camera app? [08:52] Wellark, it's fine with hand built packages (you could also try with the packages built by CI) [08:52] pete-woods, first apt [08:52] I guess the PPA doesn't actually generate click packages somehow [08:52] pete-woods, from your ppa [08:52] Wellark, everything will be double-checked when landing [08:53] Wellark: the checklist is your with your own testing. [08:53] then I think an update reverted it [08:53] so I reinstalled the deb [08:53] okay [08:53] Saviq, dednick: this is totally backwards but whatever.. [08:53] Cimi: it won't work installed from debs, as it relies upon click hooks [08:53] Wellark, no it's not backwards, we just can't waste the time to run the whole test plan for every single MP [08:53] Wellark: the MR review checklist is for proper testing. [08:53] so, how do I get this damn thing to work? :D [08:54] desktop? [08:54] Cimi: have you tried creating the dummy data with that command I sent? [08:54] Wellark, that's why it's only executed per-landing, covering multiple MPs [08:54] pete-woods, yes I have a hson here [08:54] *json [08:54] Wellark, the per-MP tests - whatever CI comes up with - are good enough to get an Approved [08:54] Wellark, then, it goes into a silo with other, possibly unrelated MPs, and gets double-checked through the TestPlan [08:55] sigh. mako device buggered... [08:55] only if the MP does not require additional packages that are not in the arvhice [08:55] Cimi: and is there now data in ~/.cache/foo/usermetrics/ ? [08:55] that would be where the actual fake data would like [08:55] live [08:55] CI can only compile and do testing for a limited changes [08:55] the once that do not have any dependencies to the other stuff on the same silo [08:55] or proposed silo [08:55] commence android reflash [08:56] pete-woods, no foo [08:56] Wellark, sure, of course, that's the current shortcoming [08:56] Cimi: I realise there are a lot of paths flying round, but it's necessary to work around the App confinement [08:56] Wellark, but until we have automated testing for silos, and silos are much lighter-weight, we just need to cope [08:56] Cimi: did you successfully run APP_ID=foo usermetricsinput source-id "%1 foos" "no foos" "" 1.0 2.0 0.1 3.5 5.2 ? [08:56] yes [08:57] Wellark, and that's why we'll then run the tests for the silo as a whole [08:57] pete-woods, actually this file is empty [08:58] Cimi: and now you're saying that ~/.cache/foo/usermetrics/ is an empty directory? [08:58] pete-woods, no the file is empty [08:58] which file? [08:58] the foo.json file should be empty [08:58] it's just there to trick libusermetrics into thinking there's an app called 'foo' installed [08:58] cache/usermetrics/sources/foo.json [08:58] pete-woods, but if it's there as a placeholder, fine [08:58] dednick: filled in the submitter checklist. [08:59] I have though a directory there [08:59] and is fine [08:59] it has a db [08:59] a json sorry [08:59] and a tmp [08:59] xnox, hey, could you have a look at http://paste.ubuntu.com/7414899/ please and see what needs fixing for lp:unity-scopes-api to be cross-buildable? [08:59] Saviq: why do things have to be top approved for silo? [09:00] dednick, because that's where I take a list of MPs to land [09:00] from [09:00] Saviq: um. but we dont want this to land in trunk yet do we? [09:00] dednick, dunno, why don't we? :) [09:00] yes, but why on earth do we require top-approval to simply get a silo? [09:00] * Saviq not sure what we're talking about [09:01] Wellark, it's not *required*, but that's the general route [09:01] well, I got rejected just because my MP's are not top approved [09:01] Wellark, you have to remember a package can (in general) only be in one silo [09:01] Saviq: there's a fix in the way the indicators are sending multiple icons (indicator-network). but obviously we need the network indicator to land at same time. [09:02] dednick, and why can't that get Approved? it won't land until we put it through a silo [09:02] dednick, that's why we have the "is something else needed..." [09:02] Saviq: ah. ok [09:02] Cimi: I'm just installing the packages from the silo onto my phone again [09:02] dednick, it's top-approved, but there's requirements, that's fine [09:02] Saviq: ok. thanks for the clarification [09:02] Saviq: s/python3,/python3:any,/ [09:02] dednick, and we'll only put that branch in a silo once we know we can land all the required stuff together [09:03] xnox, that's not all [09:04] xnox, unless this was just "first, :any to python3" [09:04] xnox, and I can't do apparmor:any, it says "can't do any when Multi-Arch: none" [09:05] Saviq: the rest should work out. [09:05] Saviq: others appear to not be installable becasue python3:armhf is attempted to be installed, which breaks click-dev, apparmor, etc. [09:05] xnox, hmm ok, trying again [09:05] Saviq: the build-dependencies declared do not make sense though. [09:06] Saviq: why does it build-depend on apparmor? [09:06] xnox, good question [09:06] xnox, btw http://paste.ubuntu.com/7414927/ [09:07] Cimi: okay, I have just done apt-get install libusermetrics-tools from the PPA onto today's image [09:07] pete-woods, why does unity-scopes-api build-depend on apparmor? [09:07] Cimi: then I did touch touch ~/.cache/usermetrics/sources/foo.json [09:08] Saviq: try dropping apparmor. that should fiz apparmor & debhelper errros. [09:09] Cimi: then I ran the APP_ID=foo command, and I now have /var/lib/usermetrics/32011/default/5d12178708e1c15dbbec106aac3751e01176941a-1399539973910.svg [09:09] Saviq: it uses libapparmor to confine the infographic visualizers [09:09] xnox, last thing The following packages have unmet dependencies: [09:09] liblttng-ust-dev:armhf : Depends: systemtap-sdt-dev:armhf (>= 1.6) but it is not installable [09:09] pete-woods, let me destroy this image, bootstrap and such [09:09] pete-woods: at build-time? [09:10] pete-woods, yeah, why "apparmor" and not "libapparmor-dev" or so? [09:10] pete-woods, you have devel-proposed? [09:10] Cimi: yes [09:10] ok [09:10] pete-woods: we are cross-building here, thus nothing can be executed, nor well confined. [09:10] Saviq: could well be a mistake? *checks it out* [09:10] Saviq: you should "fake it" by doing "Depends: apparmor:amd64" =) [09:11] Saviq: build-depends that is, (obviously not suitable for normal operation) [09:11] Saviq: every seen this? Cannot push /home/nick/.cache/ubuntuimages/ubuntu-touch/devel-proposed/mako/version-17.tar.xz.asc to device: free space on /cache/recovery is file [09:11] xnox, yeah, I'm mostly about actually fixing it in their trunk [09:12] dednick, huh nice one... [09:12] dednick, sounds like you're out of space [09:12] Saviq: http://bazaar.launchpad.net/~unity-team/libusermetrics/file-based-infographics/view/head:/debian/control looks like I'm using libapparmor-dev? [09:12] dednick, on the cache partition, might want to get into recovery and clean it [09:12] pete-woods, unity-scopes-api, not libusermetrics [09:12] Saviq: whoops! [09:13] Saviq: that's because we're using the command line too aa-exec, which comes from the apparmor package [09:13] tool [09:13] pete-woods, build-time? [09:13] Saviq: how could we run the tests without the tools? [09:14] pete-woods, right, so it's needed for tests, that was the important bit [09:14] pete-woods, I can see APPARMOR_FOO used in CMakeLists.txt, but there's no find_package(apparmor) or so? [09:14] pete-woods, are you actually linking to apparmor [09:14] ? [09:15] Saviq: that's a mistake, I'll remove those references [09:15] ldd doesn't seem to think so [09:15] pete-woods, don't worry, I'll MP that [09:16] xnox, so, they're using apparmor in tests, how do we deal with that? [09:16] Saviq: on cross-builds tests will not be executed. [09:17] Saviq: there is no way to specify things that are not needed during cross-builds at the moment. [09:17] Saviq: let me try something out. [09:17] xnox, thanks! [09:20] Saviq: I just noticed that the usermetricsservice is not starting automatically on the phone - it's set "start on desktop-start" [09:20] Saviq: is there a different even I should be using? [09:20] *event [09:20] Saviq: will utopic do to resolve this "Build-Depends dependency for sbuild-build-depends-unity-scopes-api-dummy cannot be satisfied because candidate version of package libunity-api-dev can't satisfy version requirements" or do i need some special ppa? [09:21] Saviq: i don't have utopic cross chroot yet, creating one now. [09:21] xnox, this should be in trusty fine, too [09:21] xnox, ah version [09:21] xnox, yeah, utopic [09:21] Saviq: ack, building utopic cross chroot. [09:22] pete-woods, that service needs to be there for apps to connect to it, or for shell as well? [09:23] pete-woods, probably best to talk to ogra to find a suitable event [09:23] Saviq: apps don't directly connect to it, it just inotifies from directories [09:24] pete-woods, so the service doesn't connect to anything, just looks at files and produces files? [09:24] I suppose it does talk to a system dbus service, so could just start on dbus like everyone else [09:24] pete-woods, yeah, prolly best [09:24] pete-woods, start on started dbus [09:25] Saviq: switching to that now :) [09:27] Saviq: are you able to rebuild the libusermetrics silo? I've just pushed that upstart config change for Cimi [09:28] yes pls [09:32] Wellark: don't suppose you could add a test for the changes to the icons? [09:35] Wellark: there's a test suite in tests/plugins/Unity/Indicators/rootactionstatetest.cpp [09:35] could you add a testIconsToQVariant case? [09:35] Saviq: resolved python & apparmor -> http://paste.ubuntu.com/7415052/ [09:35] Saviq: liblttng-ust-dev:armhf is not installable on amd64 host (problem to resolve in liblttng-ust) [09:36] Saviq: lttng-tools is not Multi-Arch:foreight (problem to resolve there) [09:36] Saviq: and scopes-api does not appear to build without lttng [09:36] xnox, uh, what's "native"? :) [09:37] dednick: I'm awfully busy :-/ [09:37] could we add tests when we fix the @todo use URI's ? [09:37] Saviq: install native to my architecture, as in DEB_BUILT_ARCH, amd64 in typical amd64->armhf cross. [09:37] Saviq: everything, by now, should be able to support it. [09:37] dednick: as looking at them we need to change the way we pass icons anyway [09:37] xnox, got it, will you fix lttng? [09:38] Saviq: yes, tomorrow or very late today. [09:38] xnox, sure, that's good enough, thanks! [09:39] Wellark: if you move the branch to unity-team, i'll upload a quick test. I'm futzing with device problems at the moment, so i have a few minutes free. [09:39] Wellark: i'd like to have the tests when we do the uri change, so that it's validated. [09:40] dednick: it's already part of a silo.. could you branch it, make the change and I will then merge? [09:41] Wellark: ok [09:48] mzanetti, qmluitests job broken? is it moved to utopic yet? [09:48] Saviq: probably not. but still there is another issue now too [09:48] mzanetti, oh ok [09:48] * mhr3 wonders if the qmluitests have ever worked in the past two months [09:48] mhr3: last success 3 days ago [09:49] no way! [09:50] mhr3, http://s-jenkins.ubuntu-ci:8080/job/unity-phablet-qmluitests-trusty/1768/ [09:51] mhr3, it'd have worked on trusty still if we haven't switched to new Qt package names ;P [09:51] yep. which is probably why the job configuration doesn't work any more too. qmlscene not found [09:53] Saviq, still, scarily non-green http://s-jenkins.ubuntu-ci:8080/job/unity-phablet-qmluitests-trusty/buildTimeTrend [09:53] mhr3, that's why I want to switch it to autopkgtest [09:54] mhr3, to actually have control over what's happening there [09:54] I have to go to a doctor visit (5 mins walk plus visit) - I hope to be here soon [09:54] Cimi, that's what I said! [09:54] mhr3, and to be able to reproduce stuff locally [09:54] Saviq, makes sense, clearly it sucks now :) [09:55] Saviq, wish me good luck instead :P [09:55] Cimi, that was my way of wishing you luck ;) [09:56] aahah ok [09:56] catch you soon [09:56] mhr3, you're not syncing trunk back to devel in unity-scopes-api? nasty ;P [09:56] Saviq, i am [09:56] when i don't forget anyway :) [09:57] mhr3, ;) [09:58] https://code.launchpad.net/~mhr3/unity-scopes-api/merge-trunk [09:58] pete-woods, I'll kick a build of infographics in the silo then? [09:58] see ^ :) [09:58] Saviq: if you could, please [10:01] Saviq, how soon do you want that landed? [10:01] mhr3, your call [10:01] Saviq, there's a big branch which has changelog fixed [10:01] mhr3, I just want to help you guys not spend hours building stuff on device [10:02] yea, funnily that was a topic on today's standup [10:02] mhr3, so I submitted https://code.launchpad.net/~saviq/unity-scopes-api/fix-cross-build/+merge/218768 [10:02] mhr3, but based it off of trunk... now rebased on devel, so all good from my PoV [10:03] Saviq, if you could base it off lp:~michihenning/unity-scopes-api/config-fixes that would be great [10:03] mhr3, it merges fine, any reason to rebase? [10:04] oh the debian stuff is gone... nvm [10:04] kk [10:05] * Saviq goes for a horse-tranq^Wpainkiller and hopefully some breakfast after that [10:17] xnox, if pbuilder complains: [10:17] pbuilder-satisfydepends-dummy : Depends: apparmor:native which is a virtual package. [10:17] Depends: python3:native which is a virtual package. [10:17] xnox, what needs upgrading? [10:20] Saviq: that works fine, here in utopic. What/where is complaining? [10:21] xnox, jenkins [10:21] xnox, https://jenkins.qa.ubuntu.com/job/unity-scopes-api-devel-utopic-amd64-ci/31/console [10:21] Saviq: i don't maintain that. [10:21] xnox, I know, just asking if you know what would need upgrading for this to work? [10:22] Saviq: well, PPAs, Launchpad, click-chroot, myself all use sbuild which works just fine :native build-deps. [10:22] Saviq: jenkins appears to be using pbuilder. [10:22] xnox, was afraid of that... [10:36] Saviq: test is locally, if you can reproduce the failure with stock pbuilder. [10:36] Saviq: if yes, check if :native support has been fixed in debian & get that merged/synced into ubuntu. [10:36] Saviq: otherwise, the pbuilder used on the jenkins host may need upgrade [10:36] xnox, will do, thanks [10:37] Saviq: or it could be one of the hooks that's breaking. I'm not quite sure. [10:37] (well jenkins slave, not host, i guess...) [10:38] Wellark: https://code.launchpad.net/~nick-dedekind/unity8/indicator-root-state-icons-fix-test [10:41] dednick: ok. I will merge that [10:41] dednick: thanks! [10:42] Wellark: ok. will top approve when you have. [10:42] dednick: does it have the excess tags removed? [10:43] hmm.. weird.. [10:43] I now see all the tags.. [10:43] I will remove them once again [10:46] dangit. they keep coming back! [10:46] dednick: let's see.. [10:46] dednick: merged now. [10:46] Wellark, dednick, make sure to not push/pull until after you stripped them from *everywhere* [10:47] otherwise they just go viral [10:47] Saviq: do I have to do a force push? [10:47] Wellark, _no_ push at all [10:47] as there is no commit generater [10:47] Wellark, commits have nothing to do with tags [10:47] in bzr [10:47] well, how do I update tags in a branch that is already in lp? [10:47] so you have to remove them separately remotely and locally [10:47] Wellark, strip-u8-tags.sh lp: [10:48] Wellark, it will take like 10 minutes, too [10:48] right.. [10:48] Wellark, lp:foo of course [10:48] that explains it [10:48] Wellark, yeah, in bzr tags just fly around whenever you push/pull, not looking at commits or whether you removed them [10:48] Wellark, if they're somewhere - sooner or later they'll be everywhere [10:48] I magically thought that the tags removed locally and then pushed to LP would actually have an effect on a branch being pushed [10:49] Wellark, that's not really magic, that's just git :P [10:49] well, having tags in bzr that just float around and can't be synced between different repos is just weird [10:50] dednick: wait until the strip finishes [10:50] Wellark, they can be synced, one way ;) [10:50] are tags one of the bzr features that were never truly finished? [10:50] meaning "where they are" to "where they're not" [10:51] ugh.. [10:51] ssh-agent is dead [10:51] and now bzr asks me to write my password for each tag removal [10:51] no thank you. [10:53] took actually longer :) [10:54] Cimi, so, just the packages from your privacy branch make it so that unity8 doesn't start [10:54] after the wizard finishes, that is [10:56] Saviq, wizard branch, not privacy [10:56] Cimi, wizard wifi, yes [10:56] Saviq, ok [10:56] Saviq, will look === irsol_ is now known as irsol === MacSlow is now known as MacSlow|lunch === alan_g is now known as alan_g|lunch === MacSlow|lunch is now known as MacSlow [12:00] pete-woods, ok now I have the directory [12:00] I will keep going [12:00] Cimi: cool stuff, hopefully you can get something going [12:01] Cimi: I tried it out on my desktop at least with a basic image slider mocked up to look like the Greeter, and it was looking nice === _salem is now known as salem_ [12:19] Saviq: why is unity8 built on arm64, powerpc and ppc64el ? [12:20] Wellark, (not?) [12:20] those make silo building dead slow... :( [12:20] Saviq: look here: https://launchpad.net/~ci-train-ppa-service/+archive/landing-009/+packages [12:21] unity8 failed to build o those arches [12:21] just wondering why do we even bother right now.. [12:22] or is this something "everything in main has to build on all of these" [12:22] Wellark, we're not in main btw [12:22] unity8 is not in main? [12:22] ok' [12:22] even better [12:24] Wellark, it only failed due to missing dependencies [12:25] yeah why we build for those? [12:26] once the dependencies show up, we'll just build for all the arches we support [12:26] like "why not"? :) [12:26] wasting time? energy? [12:26] I bet no one is using unity8 on powerpc [12:27] I think Wellark has a point [12:27] Cimi, you could say that about like every package in Ubuntu [12:27] Saviq, well not [12:27] Saviq, I expect someone using X on a powerpc [12:28] Saviq, I just don't expect anyone using a development version of a phone interface [12:28]