=== tmpRAOF is now known as RAOF [01:49] bschaefer, bregma, is there any particular reason there is 2 separate private ExpanderView classes in dash/FilterExpanderLabel.cpp and unity-shared/SearchBar.cpp? I see a slight difference in FindAreaUnderMouse, so I would think this could be worked around. [01:50] bschaefer, bregma, I ask because I need to implement some a11y for the ExpanderView class in the searchbarcode to get at the filter results label properly." === FJKong is now known as FJKong_afk [02:14] TheMuso, I admit I have no idea, Trevinho might know best but he's not back until Wednesday I believe [02:16] bregma: Ok thanks. [02:37] i dont know either :( [03:22] Ok thanks guys, I'll poke Trevinho later this week. === feng_ is now known as handsome_feng === duflu_ is now known as duflu [08:08] Mirv: ping [08:24] tsdgeos: pong [08:25] Mirv: i have a few patches more for the qtdbus thing that should help, can you rebuild the silo? [08:25] tsdgeos: ok, sure. [08:25] Mirv: http://paste.ubuntu.com/10710979/ [08:26] omg :) ok, starting to pull those. [08:29] I wish there was an option besides "pull ongoing upstream development rewrite".. but as long as it works.. [08:45] Mirv: yeah :/ [08:49] tsdgeos: the patches need some rebasing for 5.4 but I survived the previous ones too. some objects/functions renamed. [08:51] Mirv: ok, tell me if you need help [10:02] dednick: how's the autopilot stuff going? i see it's down to less failures, did you track down the issues with dragging? [10:17] tsdgeos: yeah. was down to the screen switching off when we locked the phone. We simulated it using uinput to press power button twice. But sometimes it would only press once, causing screen to remain off. [10:19] tsdgeos: i think creating the uinput device and using it immediately was the issue. I don't think udev picks up the create immediately, so the first press might be missed by mir input handler. [10:19] dednick: so basically what my autopilot patch fixes? [10:19] tsdgeos: what fix? [10:19] it's funny if i wanted to fix only the password -> assword usecase [10:19] but it would fix the rest too [10:19] dednick: the one you reviewed :D [10:20] tsdgeos: no, its different. doesnt use the keyboard [10:20] ah [10:20] ok [10:20] but it's interesting [10:20] power button goes directly through udev [10:20] because i remember the testcase also has something like pre-create uinput stuff [10:20] dednick: ah it's us writing to udev direclty? not though autopilot? [10:21] tsdgeos: but yeah. might be the same issue [10:21] tsdgeos: well, it's autopilot-uinput [10:21] which creates a /dev/uinput device [10:21] right [10:21] but it's also precreated i think [10:21] let me see [10:22] tsdgeos: it is in autopilot uinput, but not in our test [10:22] ahh [10:22] right [10:22] tsdgeos: the lock method is global [10:22] so we do [10:22] uinput = evdev.UInput(name='unity8-autopilot-power-button', [10:22] 7 devnode='/dev/autopilot-uinput') [10:22] in our own code [10:22] yeah, but we use it immediately [10:22] correct [10:22] that's bound to race like the thing i was fixing [10:23] ya. i'm considering putting in a global one somewhere if possible. [10:23] a pre-created one i mean [10:25] tsdgeos: I'll also see if i can write an AP test for the lock alone. so we can make sure it doesn't regress and cause this confusion issue again. [10:26] cool :) [10:26] even though it probably belongs in unity-system-compositor. [12:07] dandrader, hmm... have you ever tried the unity-desktop-session-mir? [12:08] mzanetti, I don't think so [12:08] dandrader, looks like the patch with the surface focusing broke that === MacSlow is now known as MacSlow|lunch [12:18] mmh, that QDBusUtil::dbusService [12:19] anyone knows how to disable the edges intro? [12:19] I tried this [12:19] dbus-send --system --print-reply --dest=com.canonical.PropertyService /com/canonical/PropertyService com.canonical.PropertyService.SetProperty string:edge boolean:false [12:19] but doesn't seem to change anything [12:20] Saviq, ^ [12:20] mzanetti: sudo -S dbus-send --system --print-reply --dest=org.freedesktop.Accounts /org/freedesktop/Accounts/User32011 org.freedesktop.DBus.Properties.Set string:com.canonical.unity.AccountsService string:demo-edges variant:boolean:false ? [12:20] vesar, ^ [12:21] Mirv, well, on desktop that is [12:21] but I'll try [12:21] oh, that's indeed for phone [12:21] worth trying [12:21] well, not worth with User32011 probably... [12:21] yeah changing "user32011" to user1000 should do [12:22] whoever came up with that dbus-send syntax... [12:22] seriously... [12:22] :D [12:22] mzanetti: indeed, it should be inline xml! [12:23] vesar, try this: [12:23] the longer the line the more professional it looks [12:23] sudo -S dbus-send --system --print-reply --dest=org.freedesktop.Accounts /org/freedesktop/Accounts/User1000 org.freedesktop.DBus.Properties.Set string:com.canonical.unity.AccountsService string:demo-edges variant:boolean:false [12:23] Mirv, I wonder why phablet-config does that on com.canonical.PropertyService, not on AccountsService [12:24] vesar_, try this: [12:24] sudo -S dbus-send --system --print-reply --dest=org.freedesktop.Accounts /org/freedesktop/Accounts/User1000 org.freedesktop.DBus.Properties.Set string:com.canonical.unity.AccountsService string:demo-edges variant:boolean:false [12:25] mzanetti: these are mostly pieces of wisdoms that spread by copy-pasting around.. and people don't feel ownership of phablet-tools === alan_g is now known as alan_g|lunch [12:30] Mir, mzanetti: oh boy. That worked! [12:30] cool [12:30] Mirv, ^ [12:30] thanks Mirv [12:31] Thanks guys [12:42] no problem, great that there's something that works. I think my last unity8 on desktop tryout stopped at the same problem. [12:47] Mirv: do you have an eta for the new qt with the patches? or may be worth me compiling it myself so i can keep trying? [12:53] mzanetti, the User1000 bit is because you're talking to the system dbus so you need to identify the user on whose behalf you want this change [12:53] sure [12:53] I was more wondering why the phablet-config tool seems to set something on PropertyService instead of AccountsService [12:54] mterry should know, but ultimately I imagine PropertyService is just another entry point to AccountsService [12:55] yeah, would make sense === dandrader is now known as dandrader|afk [12:55] anyhow. problem solved by now [12:56] tsdgeos: still building locally as I combined an upload seb is currently doing plus the DBus patches needed some more fixing (all QDBusUtil::dbusService -> dbusServiceString() etc). I will push it once it builds. if you want to test on your own, this is what I'm currently building http://paste.ubuntu.com/10711883/ (applies on top of lp:~kubuntu-packagers/kubuntu-packaging/qtbase-opensource-src) [12:58] oki [12:59] maybe helps if i build locally to speed up testing in case something explodes [12:59] tsdgeos: it seems my local build is pass the build phase, in unit tests, so I'll push a test build to the PPA now === alan_g|lunch is now known as alan_g [13:01] amd64 build will take 2h 23min based on previous ppa build [13:05] ChrisTownsend: hey is unity8-desktop-session-mir working ok ? [13:05] * kgunn hadn't had time to check [13:06] kgunn: As far as I know. I've been out since middle of last week, so I'm updating system right now and will try. [13:06] cool [13:06] will be good to hear [13:06] kgunn: I know the unity8-lxc stuff is busted right now with the transition to systemd. [13:06] mzanetti: ^ [13:07] ack === MacSlow|lunch is now known as MacSlow [13:09] kgunn: mzanetti: Yeah, I can log into a Unity8 session w/ fully updated 15.04 system. [13:09] ChrisTownsend, does focusing different windows and typing with the keyboard work for you? [13:11] mzanetti: No, that seems broken. I tried "Search" in System Settings and nothing. [13:11] yeah, same for me === dandrader|afk is now known as dandrader [14:15] Mirv: looking good, my self compiled stuff ran the calculator autopilot fine twice in a row [14:16] did you backport half of Qt 5.6 already? ;) [14:17] shhhh [14:17] i haven't even started the talks for the async image provider === Facu is now known as facundbatista [14:17] sil2100, ping === facundbatista is now known as facundobatista [14:19] davidcalle: pong [14:19] (in a meeting) [14:20] sil2100, nevermind (for now :) ) === FJKong_afk is now known as FJKong [14:40] dednick: maybe instead of that sleep 1, you can create the device in the constructor of the test? [14:40] or yeah not fake the keypress can work too [14:49] tsdgeos: ya. i did already [14:49] cool [14:50] Regarding the Wizard, what other conditions would cause it to run even if '~/.config/ubuntu-system-settings/wizard-has-run' exists? [14:51] dednick: any clue what happened with with last CI run? [14:52] ChrisTownsend: just to make sure, you mean the wizard and not the edge tutorial, right? [14:53] tsdgeos: Sorry, I mean the edge tutorial. This is in a desktop session. [14:53] ChrisTownsend: https://code.launchpad.net/~mterry/unity8/tutorial-reappears/+merge/254263 helps? [14:54] tsdgeos: I'll take a look at that. Thanks! [14:55] ChrisTownsend: and then there's also https://code.launchpad.net/~mterry/unity8/skip-spread-tutorial-on-desktop/+merge/253866 you may be interested in [14:55] since it's desktop+tutorial related too [14:56] Mirv: Ran 122 tests in 2296.058s OK [14:56] :) [14:56] i'll stop now and wait for the real thing on the ppa [14:57] tsdgeos: Oh cool, thanks! [15:23] Saviq, I'm kicking a build of silo25. I need packages of one of the branches in there [15:23] mzanetti, sure, was about to do that too [15:25] What makes the edge tutorial run? On unity8-desktop-session-mir, the tutorial doesn't run, so is there a trick to keep it from running? [15:25] ChrisTownsend, it does run there, people have been complaining that they can't get through the right edge bit [15:26] AccountsService.demoEdges is the only bit we're looking at whether to run it or not [15:26] Saviq: Hmm, I don't see it run when I run from that. Only when I'm running from the Ubuntu Next ISO. [15:27] ChrisTownsend, any chance you completed it once already? [15:28] dbus-send --system --print-reply --dest=org.freedesktop.Accounts /org/freedesktop/Accounts/User1000 org.freedesktop.DBus.Properties.Get string:com.canonical.unity.AccountsService string:demo-edges [15:28] Saviq: It's quite possible I did some time ago, I just don't remember. [15:29] Saviq: Yeah, that's set to false. [15:29] mterry: about https://code.launchpad.net/~mterry/unity8/tutorial-reappears/+merge/254263 may it be because you're seeing is delayed garbage collection? [15:29] dbus-send --system --print-reply --dest=org.freedesktop.Accounts /org/freedesktop/Accounts/User1000 org.freedesktop.DBus.Properties.Set string:com.canonical.unity.AccountsService string:demo-edges variant:boolean:true [15:29] ChrisTownsend, ↑ this would bring it back [15:29] Saviq: I could also set that to false using dbus-send and then bypass the tutorial, right? [15:30] Saviq: I want it false:) [15:30] ChrisTownsend, sure, totally [15:30] Saviq: Cool, thanks! [15:30] mterry: and maybe qmltestrunner has/hits different garbage collection patterns? [15:30] tsdgeos, yeah... I guess? Like really delayed. Many seconds [15:31] tsdgeos, because when you hit this bug, you can go through the tutorial several times [15:31] tsdgeos, so it's really taking its sweet time cleaning that up [15:31] or it's a bug :D [15:31] tsdgeos, I just couldn't figure out why I was seeing different behavior (like what might trigger the bug) [15:32] Saviq: I think I'll add that dbus trick to the bug report to help folks get through this for the time being. [15:32] mterry: question is, think is something we should spend time investigating deeper or just approve your code and be done with it? [15:32] ChrisTownsend, good idea, mention that the 1000 there is UID [15:32] tsdgeos, yeah that's where I was :) [15:32] Saviq: Right, I'll make that clear. [15:33] Saviq: Probably just use $UID instead of 1000. [15:34] ChrisTownsend, unless they run it from under a different user ;) [15:34] Saviq: True dat [15:34] mterry: not sure what you mean with that [15:35] tsdgeos, I just meant, I had arrived at the point of not caring about the root cause anymore, and was willing to submit an MP for the "good enough" workaround [15:35] ok [15:35] mterry: we have simpler code otoh so seems like a win [15:35] i'll review [15:36] tsdgeos, yeah, truthfully, this could have just been cleanup from the refactor, even without the bug [15:53] mterry: garg [15:53] i can't test that bug [15:53] tsdgeos, you can't even reproduce it now? [15:53] since i need to unlock my sim [15:53] which obviously i can't do on the wizard [15:53] i guess i can use my other phone to remove the pin lock from my sim [15:53] think that's not supported yet from the Ubuntu Phone [15:54] or is it? [15:54] * tsdgeos checks [15:55] doesn't look like it [15:55] seb128: would you know if the settings app lets you set a new sim pin? [15:56] tsdgeos, I'm not sure, jgdx or kenvandine would know for sure, ask on #ubuntu-touch? === om26er is now known as om26er|dinner === om26er|dinner is now known as om26er === dandrader is now known as dandrader|lunch === alan_g is now known as alan_g|EOD === dandrader|lunch is now known as dandrader