=== alecu_ is now known as alecu [08:59] do scopes use QNetwork & friends? [09:05] lpotter: depends what you mean by scopes [09:06] the scope backend no (afaik) [09:06] the scope ui (i.e. the unity8-dash) yes [09:12] tsdgeos, hey, so bug #1540490 is getting some heat [09:12] bug 1540490 in unity8 (Ubuntu) "app icons getting their top cut off (Ubuntu Touch)" [High,In progress] https://launchpad.net/bugs/1540490 [09:12] tsdgeos, if you could please prio it (you were able to reproduce?) [09:13] Saviq: yes i can reproduce it [09:13] it's somewhat strange since we have not changed that code so no idea why it's happening now [09:13] i have a "slight" idea of what may be the problem [09:13] but that's still far from actually producing a fix [09:14] tsdgeos, ack, keep me posted please [09:14] sure [09:17] dednick, hey, so people were happy with the form factor changes, with the exception that they excepted their state to be set to what's real (so if they got fullscreen, they would like to know) [09:17] I'll test now to see if I can repro your issue [09:29] Saviq: um. who would like to know? clients? [09:30] dednick, yes [09:30] Saviq: they're the one that set the flag right? [09:30] dednick, but they don't know what they got in return [09:30] dednick, whether they're still windowed or not [09:31] dednick, and they might want to adapt their UI depending on that [09:32] dednick, IIUC it should be possible for us to tell them their state, same as when we maximize apps and whatnot? [09:33] dednick, and yeah, can confirm the vertical offset on OSK [09:35] Saviq: mmm. i'll have to think about that. we'll probably have to start saving old states then... [09:35] which is shit [09:35] dednick, I thought that we might need to, but couldn't see a reason TBH [09:37] dednick, if they unset the flag, they either go back to windowed (if they were fullscreen) or remain windowed (if they are windowed) [09:37] Saviq: are we also supposed to reset the flag if we go into desktop mode? [09:37] since it no longer does anything [09:37] dednick, don't think so, can we even set the flags from our side? [09:38] QWindow::setWindowFlags [09:38] dednick, it does, that's exactly how we remember to switch it to fullscreen when you go back to staged [09:38] otherwise we'd have left it non-fullscreen [09:38] == bad [09:39] right. so if they requested fullscreen while in desktop mode, and switch back to phone? [09:39] it goes to non fullscreen if the flag isn't set right? [09:41] think we should get rid of this flag business later. can put something into the sdk so that the clients can use the form factor to determin all this state business. [09:41] dednick, that's the point, we don't want them to use the form factor to keep the behaviour consistent [09:42] dednick, if they requested fullscreen in desktop mode and go to staged, we're back in square one (flag + fullscreen), unless they unset the flag [09:42] dednick, so then if you go back go windowed, you get it in a window again, and that's fine [09:43] right. which is behaviour based on form factor. so why arent we using it? :P) [09:43] dednick, because we don't want apps to decide [09:43] dednick, if they go fullscreen and unset the flag, they are fullscreen across the board [09:46] Saviq: sigh. the apps are deciding anyway since they use the flag specifically for this behaviour. anyway... [09:46] i'll try out setting back the state [09:47] dednick, I kinda know, but if we decide we want this behaviour to be different in any way, we can change it without touching apps then [09:47] dednick, I think what needs to happen is we react differently to formfactor, flag, state changes, not in a single updateBasedOnAll method [09:47] dednick, but [09:47] dednick, more important for today - the OSK offset [09:48] dednick, if we don't get the client updates, fine, but we can't land with the OSK skewed [09:48] Saviq: yeah. i'm not really sure whats going on with that panel hack. it shouldnt really have any effect, but it does [09:50] just going to try remove it and see if that fixes the issue quickly. [10:04] dednick, also just realized we're not reacting well to form factor changing at the same time window state does, if I go from staged to windowed, the flagged windows end up being tiny for some reason [10:05] worried it's our fullscreen animation going awry [10:06] Saviq: yeah, i've seen that as well [10:06] think it's saving the state when it's still animating maybe [11:47] mzanetti, how do I perform an alt+tab in "make tryShell"? [11:48] dandrader, there is an itemselector on what the ctrl key should do. set it to "alt" and then press ctrl+tab [11:48] * mzanetti off today [11:48] oh [11:49] mzanetti, where question then: and where is that "itemselector" located? [11:49] s/where/last [11:49] mzanetti, nevermind. it just works. thanks [11:50] dandrader, it might hasn't landed yet... it will only come with the launcher stuff I believe, where tryShell also required Super+Tab etc [11:51] Saviq: i've pushed the fix for the osk offset. i'm not really sure why it happens on osk and not other surfaces though... [11:51] must be something to do with how it's position in u8 [11:53] dednick, ack, tx [11:53] dednick, I'm gonna get it under QA as is, without the "feedback loop" [11:53] Saviq: ok thanks. [11:55] dednick, but please have a look at that anyway [11:55] Saviq: sure will [11:55] with somewhat lower pressure, though :) [11:56] Is there any chance of Canonical participating in GSOC this year? [12:00] tsdgeos, cimi, looking at https://unity8-jenkins.ubuntu.com/job/test-0-autopkgtest/ we managed to make lazyImage worse atm ;) [12:00] !\o/ [12:01] Saviq: but that's without the new fix, no? [12:01] Saviq, hurray! [12:01] tsdgeos, yeah yeah [12:01] k [12:24] Does anybody know what makes the greeter go away when a new application appears? [12:24] dandrader, focus event [12:25] -ish [12:26] dandrader_, focus event-ish [12:27] dandrader_, Shell.qml:432 [12:28] Saviq, is that trunk? Shell.qml:432 for me is a comment: "// See powerConnection for why this is useful" [12:28] dandrader_ right, it's not, :417 then [12:29] Saviq, this is hiding the launcher === dandrader_ is now known as dandrader [12:32] dandrader, d'uh === dandrader is now known as dandrader|afk [12:34] dandrader|afk, greeter.notifyAboutToFocusApp() then? :) [12:35] dednick, so what about the tiny app window when switching to windowed? [12:40] Saviq: ah. havent looked into that yet. will do now. [12:44] Saviq, dednick: I think I see something fishy in the log regarding the tiny window: http://paste.ubuntu.com/14886821/ [12:45] ltinkl, different tiny window I think ;) [12:45] ltinkl, we're still on silo 57, not 51 [12:48] Saviq, yeah but wth is going on there with dash... it starts suspended, then gets an initial (-1,-1) size and twice, then focused twice, and moved 0 -> 0 [12:49] ltinkl, sure, I'm just saying dednick's not looking at that silo at all, you might wanna discuss with Daniel instead [12:49] Saviq, kk === dandrader|afk is now known as dandrader [12:57] Saviq, for the record: It's Shell.qml:218 "greeter.notifyAppFocused(appId);" [12:57] :) [14:41] dednick, hey, any luck? [14:44] Saviq: trying to sort figure out some lockup issues when switching modes. [14:46] is there some sort of public API that would allow a QML app to figure out whether it’s running under unity8 or another DE ? [14:47] (I know this question sounds wrong, but I need to work around a bug in notify-osd, and I need to do it conditionally, i.e. no need for the workaround under unity8) [14:47] oSoMoN, you could switch on QT_QPA_PLUGIN if you really need [14:50] that’s basically what I’m already doing (checking QGuiApplication::platformName()), but I was hoping for something accessible from QML directly, without having to expose it in C++ === dandrader is now known as dandrader|afk [15:01] Saviq: https://code.launchpad.net/~aacid/unity8/fix_dash_icons_cut/+merge/285194 [15:01] not sure if it is *THE* fix but it's *A* fix [15:02] oSoMoN, then no, can't say there is anything [15:07] tsdgeos, you mean there could be a better way to fix? [15:07] there could be a different way to fix it probably [15:08] not sure if it would be better [15:08] since i haven't found it :D [15:08] tsdgeos, ok, added to silo 57 [15:09] otoh this is probably cleaner/less risky too [15:09] dednick, that's re: the apps getting tiny, or some other lockup you've found? [15:10] Saviq: it's something else. but it's making me not be able to reproduce the tiny apps now :/ [15:10] Saviq: told you about it a couple of days ago. u8 locks up or crashes when changing form factor. [15:12] Saviq: does it happen to you with the dash? [15:13] nevermind. just reprod === dandrader|afk is now known as dandrader [15:18] dednick, not dash, just fullscreen apps [15:18] dednick, but if we're getting lockups (I never did, actually) [15:18] dednick, we probably should not land this yet [15:20] Saviq: dont think greyback has reprod a lockup, but i think he's got a crash or 2. [15:21] never had either, the only issue I saw were the tiny windows [15:21] no lockup here, but I am getting crashes === dandrader is now known as dandrader|lunch === sil2100_ is now known as sil2100 [16:30] Saviq: just pushed a fix for the tiny windows. still a bit weird wonky though. when switching stage the app resizes from small to normal size === popey_ is now known as popey [16:31] shoudl really resize from fullscreen to normal, but it's getting reset to 0 somewhere. [16:34] dednick, ack, I'm cutting our losses on that for today though, as you guys mention lockups and crashes - it looks like we're getting some slack anyway [16:34] Saviq: ok [16:35] dednick, so yeah, no immediate deadline on it, but would be nice to get when we can [16:39] Saviq: if we want to bypass the crash issues we're having now I can look into using mir_surface_set_state to push the window states back to qtubuntu. ie make the decision in the shell and dont bother the client with form factor ftb. Obviosuly will need to be solved though. [16:40] handling it client side was supposed to be the "quick fix"! [16:41] Saviq: and we use the new mir "low crome" api [16:41] dednick, WFM [16:42] ok, i'll take a look into it. === dandrader|lunch is now known as dandrader === dandrader is now known as dandrader|afk