=== JanC is now known as Guest62829 === JanC_ is now known as JanC [05:20] seems a bit risky, but maybe if there hasn't been a pile of patchups to the original rewrite.. [05:21] but yeah I'd say backporting to 5.6 could be an idea === popey_ is now known as popey [14:10] hey guys, I’m implementing multiple window support in webbrowser-app, just starting to test on a phone and when I swipe one window from the dash to close it, all windows for the app are closed, is that expected, or a known bug? [14:24] kgunn, do you happen to know the answer to my question ^ [14:26] oSoMoN, sounds like a bug [14:26] oSoMoN, like code is calling application.stop() instead of surface.close() [14:26] oSoMoN, the former will call close() on all of its surfaces [14:27] oSoMoN, let me check.... [14:27] dandrader, that was my wild guess as well. That would be a bug in unity8, or somewhere else in the stack? [14:27] oSoMoN, if that turns out to be true, it would be unity8. but let me confirm [14:28] oSoMoN, is that on a phone? [14:29] dandrader, yes, on arale [14:31] oSoMoN, strange. code looks correct [14:31] oSoMoN, please try with this sample app [14:32] oSoMoN, multiwindow.qml in lp:~dandrader/+junk/animatedDemos [14:33] dandrader, trying [14:43] dandrader, indeed your example works [14:44] so there might be a bug in my implementation, looking into it [14:45] dandrader, not sure if that makes a difference, in your example all three windows are initially instantiated but only the main one is visible, whereas in my implementation for webbrowser-app Windows are created dynamically [14:46] oSoMoN, the backing MirSurface is only created once the QWindow is made visible for the first time [14:46] ok === dandrader is now known as dandrader|afk [14:54] dandrader|afk, ah, that appears to be a crash in webbrowser-app [14:57] dandrader|afk, this is what the stacktrace (without debug syms) looks like: http://pastebin.ubuntu.com/23028989/ === dandrader|afk is now known as dandrader [15:22] oSoMoN, wild guess: maybe you have to intercept the close request in order to some preparations before having the window destroyed.... [15:23] dandrader, I’m already doing something like that, although that may not be sufficient: https://bazaar.launchpad.net/~osomon/webbrowser-app/multiple-windows/view/head:/src/app/webbrowser/webbrowser-app.qml#L113 [15:26] I installed debug syms, and it appears that QOpenGLContext::currentContext() returns a null pointer in QSGDefaultLayer::invalidated() [15:31] ah, not calling destroy() in onClosing seems to make things better === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [16:24] oSoMoN, well, the platform implementations, including the GL stuff is different on the phone. but don't know why exactly it's crashing (or whether it should crash or not)... [16:24] oh, oSomon just left... [17:02] dednick: simply subscribe ~ubuntu-mir to MIR bugs and we'll see 'em [17:02] (not assign, just subscribe) [17:02] mterry: ah. k [17:21] josharenson: can you do me a small favor and merge in trunk into session-chooser-gui? [17:21] josharenson: I want to rebase my branches on it, but don't want to get criss cross merges [17:21] should just be translation updates at this point [17:32] mterry: sure 1 min [17:33] mterry: when is the last time you pulled? I actually just pushed some changes like 20 min ago (checking to see if they include trunk) [17:33] mterry: yeah looks like I did [17:42] josharenson: yeah but I *think* r2577 wasn't included in your latest merge [17:42] just some translations [17:43] mterry: oh you are correct... pushing now [17:43] done [17:43] josharenson: thanks -- I plan to make a great tower of branches, and I hate criss-crosses [17:43] mterry: I gotcha :-) === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader