=== alesage is now known as alesage|afk [02:46] RAOF: nothing more to say? [02:56] Wug: No, sorry. Although I suspect it's by design that you don't get an iconify event. [02:57] that is patently absurd [02:57] I hope, for unity's sake, that you're wrong [03:12] Well, your window doesn't actually get unmapped or anything. [06:43] Saviq: you remember me asking about the errors in unity8 log at runtime? [06:43] this should fix at least couple of them [06:43] https://code.launchpad.net/~kaijanmaki/qmenumodel/unitymenumodel_setname-allow-empty-string/+merge/213768 [06:43] :) [06:44] "GLib-GIO-CRITICAL **" always looks so nice [08:34] MacSlow: what's up with the failing tests at https://code.launchpad.net/~macslow/unity8/modal-snap-decisions/+merge/210988 ? are they related to that code? [08:37] hmmmm [08:37] is it me or we don't build anymore? [08:37] Saviq: did it land yet? [08:38] doesn't look like [08:38] mzanetti, it's in proposed as of 4 hours ago, but apparently thumbnailer didn't get FFe [08:38] mzanetti, it's in the hands of the archive admins now [08:38] yeah no build [08:38] how is the thumbnailer related? [08:38] tsdgeos: unity8? [08:38] yeah [08:39] mzanetti, was in the same silo, [08:39] this train thing is not that nice when people forget about your stop [08:39] fixing a unity8 crash [08:39] ↑ worst analogy ever :D [08:39] :D [08:39] tsdgeos, why would we not build? like trunk? [08:39] because someone changed something and forgot about us i guess [08:39] Saviq: yes, trunk [08:40] tsdgeos, just built --clean, works? [08:40] no [08:40] can do it again [08:40] but just did it :D [08:40] tsdgeos, I mean it does here [08:40] work [08:40] Saviq: how many silos/ppas do you have [08:41] tsdgeos, none [08:41] and when's the last time you dist-upgraded? [08:41] http://paste.ubuntu.com/7193392/ is what i get [08:41] let me cleanbbuild again [08:42] tsdgeos, you have a unity8/unity-mir mismatch [08:42] tsdgeos, apt-cache policy libunity-mir-dev? [08:43] http://paste.ubuntu.com/7193395/ [08:43] looks good to me [08:44] tsdgeos, yeah, unity-mir got through from proposed apparently [08:44] tsdgeos, but unity8 did not, yet [08:44] tsdgeos, so it's not in trunk [08:44] \o/ for the train [08:44] :D [08:44] i thought it was there to fix specifically stuff like this [08:44] maybe i misunderstood its purpose [08:45] tsdgeos, no, just means we fucked our deps [08:45] to some extent [08:45] anyway, it'll sort out soon [08:45] ok [08:46] Saviq: btw where're you? [08:46] tsdgeos, recovering [08:46] i don't think you can say where're :D [08:46] Saviq: stop working late! [08:46] tsdgeos, 3am isn't late [08:46] but yeah, will be there soon-ish [08:47] ok, stop working early! [08:48] tsdgeos, :D [08:48] MacSlow: what's up with the failing tests at https://code.launchpad.net/~macslow/unity8/modal-snap-decisions/+merge/210988 ? are they related to that code? [08:49] Saviq, still those notifications don't appear after adding those files from unity8, can you repeat me the required bits for this to work? [08:50] tsdgeos, I thought my last commit to that fixed it... which was just to make jenkins happy anyway. [08:51] tsdgeos, hm... and the needed unity-notification branch is also not merged yet... damn it [08:53] MacSlow, can you help Cimi with bringing notifications over to the welcome wizard (it's working pre-greeter even, so it needs to handle them itself) [08:53] tsdgeos, the need for the data-RolesEnum to be coming from unity-api now bombed on the whole modal-snap-decision feature... [08:54] Cimi, I'll be in the office in an hour or so if you don't get it by then [08:54] mzanetti, hey, now that you're off the hook, I'd like you to look into one, hopefully relatively small, thing [08:54] Saviq, "just one more thing..." ok *sigh* I'll try my best [08:54] Saviq: ok... which is? [08:54] mzanetti, if you start an app outside of upstart, closing it in the shell doesn't SIGTERM/SIGKILL it [08:55] ah yeah. annoys me like hell all the time. will have a look [08:55] Cimi, so what do you need in the welcome-wizard [08:55] ? [08:55] mzanetti, cool, zbenjamin (#sdk) will be happy [08:56] biab [08:57] tsdgeos, so regarding that unity8 branch for modal snap-decision... it needs lp:~macslow/unity-notifications/modal-snap-decisions to be in place too... but that is blocked on the new need to expose the enum Roles via lp:unity-api now... which I cannot figure out how to make work... it works with just the initial branch of mine. [08:58] MacSlow: i see, do you need help on that enum exposing? [08:58] tsdgeos, I've added it all to the MP-desciption of lp:~macslow/unity8/modal-snap-decisions and lp:~macslow/unity-notifications/modal-snap-decisions [08:59] tsdgeos: MacSlow I guess all that's missing is the include path for unity-api/notifications in unity-notifications [08:59] tsdgeos, well I have it working with lp:~macslow/unity-notifications/modal-snap-decisions ... but I've been trying to figure out how to make it work from lp:unity-api ... and just don't have a clue [08:59] and then obviously inlcude ModelInterface.h [09:00] and use the enum as "ModelInterface::SomeType" [09:00] Cimi, ? [09:00] * mzanetti reboot [09:01] Saviq, what's the requirement... I remember Cimi asking me about the notification-daemon yesterday... but that's about it. [09:04] mzanetti, but won't that collide with the defines of unity-notifiations itself? [09:07] ahhhh it's not that it doesn't build [09:07] it's that i can't run it either [09:07] :'( [09:08] MacSlow, was having a cofffee [09:09] MacSlow, I have wifi listview [09:09] mzanetti, so I need to remove the enums from unity-notifications and just use the ones from unity-api?! [09:09] MacSlow, when I tap on a wifi access point, notification should appear [09:10] Cimi, and you trigger it and it does not appear.... or you don't know how to trigger it? [09:10] MacSlow, notificaton does not appear [09:10] MacSlow, which bits do I need? [09:10] Cimi, can I see the branch/code? [09:11] MacSlow, yup [09:11] lp:~unity-team/ubuntu-system-settings/wizard.wifi [09:11] MacSlow, wizard/qml/Pages/20-wifi.qml [09:12] tsdgeos, with lp:~macslow/unity-notifications/modal-snap-decisions and lp:~macslow/unity8/modal-snap-decisions it does and the test passes... it's/jenkins just missing the unity-notification bits... which are delayed because of the new requirement to be all in unity-api now... which I was stupid enough to agree with... not knowing what's involved... teaches me to say "yes" in the future :/ [09:13] Cimi, btw... you can't trigger notifications from qml [09:13] MacSlow: ok, i see [09:14] MacSlow: so basically it needs the pre-requisite, makes sense :) [09:14] tsdgeos, well *sigh* [09:14] MacSlow: is it a double dependency? [09:14] i.e. do both branches have to go in at the same time? [09:15] tsdgeos, it's a new dependecy that came in from the side after the MP :) [09:15] or can the unity-notifications go in first and then the unity8? [09:15] MacSlow, but they appear if I run the wizard from within unity [09:15] tsdgeos, unity-notifications could/should go in first... but now there's the unity-api bit put even before that [09:15] ok [09:16] MacSlow: can you maybe put the unity8 as "work in progress" until it's ready to decrease the "needs to be reviewed" branches a bit [09:16] ? [09:16] Cimi, oh... you having your own model there... hm... [09:17] tsdgeos, done [09:17] tx :) [09:17] mzanetti: is there any silo i can use so i can continue to do some work? [09:18] MacSlow: yes. you need to delete the enum definition from unity-notifications [09:18] tsdgeos: in what sense? [09:18] mzanetti: in the sense that i can compile unity8 :D [09:18] mzanetti, so after that's in place I have to jugle three branches just for notifications... and tow of them are on the api-team's side?! [09:19] mzanetti: or maybe i just need the right-edge branch manually merged in my stuff [09:19] ? [09:19] tsdgeos: still not exactly sure what you mean. but the right edge stuff is in silo15 [09:19] tsdgeos: /me tries to reproduce your issue [09:19] mzanetti: well i have http://paste.ubuntu.com/7193392/ [09:19] ah, there it is :D [09:19] mzanetti: just dist-upgrade and see the pain [09:20] i'm wondering what to do to make the pain go away [09:20] tsdgeos: you need libunity-api-dev from silo 15 [09:20] Cimi, I just need to wrap my head around this... [09:20] tsdgeos: or copy Application*h from here into your /usr/include/unity/... [09:20] https://code.launchpad.net/~mzanetti/unity-api/new-screenshot-and-focusing-api/+merge/199810 [09:21] * mzanetti isn't really sure why unity-api is realeased already [09:23] headache [09:23] stupid me... those are the ones that _are_ already released [09:24] so yes, you need to unity8's right-edge-2 branch merged into yours [09:24] ok [09:24] let's see if that helps [09:24] Cimi, so you using the backend in your wizard... but where's the frontend? [09:25] MacSlow, frontend? [09:25] Cimi, unity8 displays them because it has the frontend/listview displaying what's in the model [09:25] Cimi, that's probably why you only see the notifications if you're running the wizard from unity8 [09:26] Cimi, backend (unity-notifiations... the model... Notifications{...}) frontend (the ListView in unity8) [09:27] Cimi, see http://bazaar.launchpad.net/~unity-team/unity8/trunk/view/head:/qml/Notifications/Notifications.qml [09:28] Cimi, that's the frontend doing the display/rendering of the notifications in the queue (the model) [09:28] MacSlow, I have this no? [09:28] Cimi, I've to look again at thw wirzard code... one sec [09:29] mzanetti: ok, it runs now, get lots of "Need to use QMirServerApplication" though [09:29] http://bazaar.launchpad.net/~unity-team/ubuntu-system-settings/wizard.wifi/view/head:/wizard/qml/Pages/20-wifi.qml [09:29] MacSlow, line 52 [09:29] tsdgeos: hmm... you running on the phone? [09:29] nope [09:29] tsdgeos: haven't seen this warning before [09:30] and the thing crashes now :/ [09:30] meh [09:30] this is all very confusing [09:30] * mzanetti tries [09:31] yeah it is. [09:31] Cimi, sorry... that's ok [09:32] Cimi, btw... any reason you didn't move the states over from Shell.qml too? [09:32] tsdgeos: even worse. I can't link it here for some undefined stuff in QQml [09:32] Cimi, is the wizard not meant to adapt to rotation? [09:33] MacSlow, nope [09:33] MacSlow, wizard is for phone I think [09:33] Cimi, ok [09:33] what is this?¿?¿ http://paste.ubuntu.com/7193522/ [09:33] MacSlow, but i can copy the states as well [09:33] MacSlow, in case we want it on the tablet [09:34] MacSlow, first I want the notification to appear :D [09:34] Cimi, well they are meant to keeping stuff tidy in landscape-mode on thetablet [09:34] Cimi, sure... :) [09:34] Cimi, I need to run the code myself... just looking over it I can't see any issue atm [09:35] MacSlow, good luck :D [09:36] MacSlow, if you want to run it on the phone, takes time [09:36] Cimi, well I don't know how else I could help... [09:36] Cimi, I'm trying... [09:36] MacSlow, I thought a quick look was enough to spot issues [09:36] MacSlow, I can send you a package [09:37] Cimi, certainly not [09:38] Cimi, ehm... move the Notifications{} at the end... or set it z-order to be higher than every other item [09:39] Cimi, maybe it's just that it's covered by some other element/item [09:39] Cimi, that's the only other thing I can think of atm [09:40] Cimi, I think that atm Notifications is fully coverd by "content" which is also anchors.fill parent and comes _after_ Notifications {} [09:41] MacSlow, ok [09:41] MacSlow, sounds likely [09:42] Cimi, just make sure Notifications is _after_ Item with id "content" [09:42] Cimi, that should be enough [09:46] MacSlow, also, adding import "../Notifications" should help :D [09:47] Cimi, oh... so you didn't get an error before? [09:47] weirdly enough [09:48] MacSlow, also missing inputfilterarea :'( [09:48] MacSlow, might do it by hand maybe [09:48] Cimi, I did see that in your code [09:49] Cimi, or do you mean the import for this? [09:51] MacSlow, I cannot import it [09:51] I don't think I think [09:51] let's see [09:51] Cimi, skip it for the moment and at least check if you can see the notifications showing up [09:51] Cimi, one step at a time [09:52] MacSlow, nope [09:52] MacSlow, not showing up [09:52] tsdgeos: seems everything that only links QQuick fails for not linking QQml any more. did you have that issue too? [09:52] mzanetti: nope [09:53] mzanetti, migrated! [09:53] ? [09:53] mzanetti, right edge is done [09:54] ah [09:54] \o/ [09:54] *\o/* [09:54] DONE [09:54] now, tablet right edge ;P [09:54] yeah... [09:54] well, now I see how to get the surfaces properly [09:55] image kicked with it :) [09:55] and get away with that screenshotting stuff [09:55] yay! [09:55] mzanetti, amazing work with the right edge [09:55] mzanetti, so smooth [09:55] Cimi: thanks [09:55] :) [09:57] didrocks, 3 Qs after this fight, yesterday cyphermox asked me to rebuild the whole silo and retest, is that common practice? fwiw, no "extra" commits would get merged if they were not in the silo, right? [09:57] ok builds now [09:57] still doesn't run :/ [09:57] well, runs and immediately shutsdown [09:57] Saviq: yeah, no extra commits would get merged. I'm unsure why he asked you to rebuild everything, he didn't tell you why? [09:58] * mzanetti still can't build for the linker issues [10:00] didrocks, I think he thought any new commits would get merged (there weren't any), but also he uploaded a seed change [10:00] * MacSlow tries to build wizard.wifi [10:01] didrocks, 2: suddenly the build job complained about changelogs having trusty instead of UNRELEASED in bumps, where they built from the same branches before? [10:01] Saviq: hum, I'm puzzled and not sure what this was all about [10:01] (on 1) [10:02] didrocks, I think it was just "to be safe", this was a big landing after all [10:02] Saviq: do you have links to the job? [10:02] didrocks, https://ci-train.ubuntu.com//job/landing-015-1-build/build?delay=0sec see the three failed ones [10:02] mzanetti: it crashes in v4 here on run :S [10:03] tsdgeos: well... it desn't even link it here [10:03] i'm not sure what i prefer more :D [10:03] tsdgeos: http://paste.ubuntu.com/7193617 [10:03] Saviq: argh, I would need the "debug" version of it [10:03] you didn't run with debug, right? [10:03] didrocks, no [10:03] so that I can see the resulting changelog [10:03] that's really weird [10:03] didrocks, I *think* the difference could be that that was a full run with ignore_step [10:03] let me look at the branch [10:04] not that I think of [10:04] didrocks, and the others were direct rebuilds of particular packages [10:04] Saviq: valgrind points to unity-notif http://paste.ubuntu.com/7193620/ [10:04] MacSlow: ↑ any idea? [10:04] huh [10:04] Saviq: yeah, so that's intended [10:04] bhttps://code.launchpad.net/~rsalveti/powerd/drop_support_sf/+merge/213340 [10:04] https://code.launchpad.net/~rsalveti/powerd/drop_support_sf/+merge/213340 [10:04] see the change [10:04] it touches debian/changelog [10:04] tsdgeos, no [10:04] didrocks, sure, I understand [10:04] but it's "released" [10:05] * tsdgeos digs more [10:05] so what cu2d would do without this check is: [10:05] create an additional entry on top of it [10:05] tsdgeos, never seen any issue like that [10:05] with only: " *" [10:05] as there was nothing to put in [10:06] greyback: its *\o/* time [10:06] rsalveti: FYI, when you change the changelog yourself and not backport a change from distro, you need to keep it UNRELEASED. (otherwise, it's leading to an empty changelog and this is now safeguarded due to complains from the release time) [10:06] Saviq: so your question is "why now" :p [10:07] didrocks, yes [10:07] Saviq: basically, the release team complained again a lot on Monday [10:07] (another upload with "* " and nothing) [10:07] so I added the check and deployed yesterday [10:07] mzanetti: wooo! [10:07] mzanetti: shall i be getting all these warnings? http://paste.ubuntu.com/7193638/ [10:07] didrocks, right, understood :) [10:07] Saviq: the error messsage helped to understand the issue? [10:07] Saviq: again, I'm happy for any rephrasing :p [10:07] didrocks, sure, I fixed it straight away [10:07] great ;) [10:07] didrocks, just was surprised that it broke suddenly [10:08] tsdgeos: hmm... don't think so [10:08] didrocks, but, it'd be nice if those sanity checks were ran asap [10:08] yeah ;) [10:08] Saviq: not possible before generating the finale changelog [10:08] unfortunately [10:08] which is after the chroot [10:08] tsdgeos: there *must* be something wrong with Qqml [10:09] tsdgeos: now I have this: /usr/lib/x86_64-linux-gnu/qt5/bin/qmlplugindump: symbol lookup error: /usr/lib/x86_64-linux-gnu/libQt5Quick.so.5: undefined symbol: _ZN3QV47Managed4callEPS0_PNS_8CallDataE [10:09] mzanetti: brrr [10:09] didrocks, especially with the move to prodstack... it was an hour of failing packages [10:09] mzanetti: are you using any ppa or something? [10:09] maybe the old 52 beta ppa? [10:09] Saviq: yeah, I continue pushing on that… [10:09] or some silo that may be playing with you [10:09] and ev as well [10:09] didrocks, maybe you could upload one by one? [10:09] hmm... I dropped the ppa. but maybe still packages around... let me check [10:09] Saviq: I did that before [10:09] Saviq: until yesterday [10:10] and got complains to upload only parts [10:10] didrocks, again, thought that something changed... [10:10] and that was confusing [10:10] is there an apt-command to list all the packages on the system that cannot be found in a ppa? [10:10] so now, I'm doing once they are all generated [10:10] didrocks, well, ok "by design" :| [10:10] which was ok, before prodstack [10:10] Saviq: by redesign on complains yeah :p [10:10] mzanetti, "cannot"? [10:10] yeah [10:10] like, if I still have packages installed from a ppa that's not enabled any more [10:10] crap... it need to reflash my phone... [10:11] Saviq: actually [10:11] Saviq: let me check the code and maybe give you a trick :p [10:11] one sec [10:12] mzanetti, aptitude should show you packages that are "deprecated" or "local" [10:12] Saviq: no, no easy trick, I protect that as well [10:13] didrocks, so, question no. 3 [10:13] yep :) [10:14] didrocks, say you put libfoo-dev and bar (Depends: libfoo-dev) in a silo [10:14] Cimi, I could compile the wizard.wifi on my desktop... how do I run it? [10:14] didrocks, libfoo gets migrated, but bar is stuck in proposed [10:14] MacSlow, on the desktop? [10:15] MacSlow, debuild [10:15] MacSlow, wizard/test.sh [10:15] yeah [10:15] Cimi, it'll take some time before I can use it on my phone... flashing it as it was still using a defunct 004 silo-ppa [10:15] didrocks, now you dist-upgrade, get a new libfoo-dev, but can't build bar, because the dep moved [10:15] Saviq: you mean, can't build old bar? [10:15] didrocks, yes [10:15] so, it's an API breakage? [10:15] didrocks, 'cause "new" bar is only in the silo [10:15] didrocks, yes [10:16] you need to breaks: older-version of foo [10:16] or use that virtual package trick [10:16] that we had [10:16] but libfoo still moves [10:16] to release pocket [10:16] it's an issue in britney then [10:16] as nothing prevents it from going there [10:16] hum [10:16] foo breaks old bar? [10:17] didrocks, yeah, that'd probably work [10:17] so new foo shouldn't migrate without old bar [10:17] Saviq: I think an intermediate step would be to have britney to undestand the grouping from CI Train [10:17] mhm [10:17] but the idea was rejected by the release team who don't want to special case for it [10:20] mzanetti: do you know why it seems that we're not using unity8's application manager and instead we're using the unity-mir one when running on the desktop? [10:20] didrocks, could we maybe have ro access to branches in silos? [10:20] tsdgeos: no. that shouldn't happen [10:20] mzanetti: it is happening :D [10:20] i'm getting lots of [10:20] Need to use QMirServerApplication [10:20] and [10:21] Saviq: that's part of what I asked. I even don't have it anymore since we moved to prodstack [10:21] yeah... you shouldn't even have unity-mir installed on your machine... [10:21] mzanetti: http://paste.ubuntu.com/7193672/ [10:21] tsdgeos: instead unity-fake-env should provide Unity.Application [10:21] Saviq: so yeah, the CI team will develop a charm to proxy those through apache [10:21] Saviq: it was a requirement for the move, but then, I got to dropped it :( [10:22] this osk works 0 times out of 4 [10:22] tsdgeos: so... I just removed *all* of my Qt5 packages and then did a ./build -s and ./build -c [10:22] tsdgeos: now it starts, but yes, seems to use unity-mir here too [10:23] mzanetti: i can't remove unity-mir, we just need to make sure the application thing we use [10:23] is the local one [10:23] that used to work [10:23] Saviq: for a temporary access, I can rsync to lillypilly though [10:23] Saviq: I think I'll just cronify that [10:23] so it broke somewhere either with your or somebody elses patches [10:23] would work for you? [10:23] mzanetti: but good we're now on the same stage though ::) [10:23] tsdgeos: I just removed libunity-mir1 and now it works again fine with the fake app man [10:23] Cimi, so that's from trying to run wizard/test.sh on my desktop ... http://pastebin.ubuntu.com/7193680 [10:23] * tsdgeos grabs some bubly drink [10:24] Cimi, what maliit-related package am I missing? [10:24] mzanetti: Saviq says he knows what's the problem [10:25] yeah... we're pulling libunity-mir now in ./build [10:25] MacSlow, I never had this error [10:25] tsdgeos: we didn't before. [10:25] mzanetti, we did, it's a build-dep [10:25] mzanetti, the difference is [10:25] mzanetti, it was installing the plugin in /usr/lib/.../qt5/imports before [10:25] ahhh [10:25] right [10:25] /Unity-Mir [10:26] (still, I don't think unity-mir is a build dep) [10:26] or well, should be [10:26] mzanetti, sure, it probably shouldn't be [10:27] mzanetti, or at least the plugin should be split out of libunity-mir1 [10:27] Saviq: here we go: http://people.canonical.com/~didrocks/citrain/silos/ [10:27] Saviq: I rsync every 10 minutes [10:27] right [10:28] tsdgeos: so, you can use run -f for now [10:28] or remove libunity-mir1 yourself [10:28] okidoki [10:28] Saviq: so, for instance http://people.canonical.com/~didrocks/citrain/silos/landing-002/mir/ is the mir branch generated by latest build [10:30] didrocks, mhm [10:30] Cimi, well then... now trying to build it on the phone... hopefully I've more luck there [10:34] mzanetti, so, I think stripping all unity-mir from our debian/control will work [10:35] Saviq: hmm.. will that pull in unity-mir on the phone still? [10:35] Saviq: what was this "mhm"? :) [10:36] Saviq: actually... how long still until we get some Mir running on desktop? [10:36] Saviq: actually I think we're on the right track from a long-term pov [10:36]