=== shiznix_ is now known as shiznix
=== chihchun_afk is now known as chihchun
=== hikiko is now known as hikiko|afk
=== hikiko|afk is now known as hikiko
=== chihchun is now known as chihchun_afk
[13:04] <alan_g> greyback_: your question answered? https://code.launchpad.net/~alan-griffiths/miral/first-cut-at-test-server/+merge/302677
[13:16] <greyback_> alan_g: ok, yeah. The duplication was concerning me
[13:18] <alan_g> I couldn't think of a better way that's compatible with existing Mir releases
=== JanC is now known as Guest34729
=== JanC_ is now known as JanC
[13:26] <greyback_> alan_g: not related to that branch specifically, I am unable to compile miral with your tests enabled: http://pastebin.ubuntu.com/23049421/
[13:26] <greyback_> missing a linker flag?
[13:26] <alan_g> greyback_: that looks like lp:1583536
[13:27] <alan_g> You need Mir-0.24+ (or backport a couple of fixes)
[13:27] <greyback_> alan_g: ack
[13:34] <kdub> which is siloed in 36
[13:34] <kdub> greyback_,
[13:34] <greyback_> kdub: ah handy. /me cnacels build
[13:35]  * alan_g wonders what will stop 0.24 landing next
[13:35] <alan_g> Should we decide it's an unlucky number and go for 0.25?
[13:35] <kdub> have a fix for the current blocker, hopefully nothing
[13:36] <kdub> eh, what did block 24 would have blocked 25 too
[13:36] <alan_g> lp:1612256?
[13:37] <kdub> alan_g, right https://code.launchpad.net/~kdub/mir/fix-1612256/+merge/302804
[13:37]  * alan_g is looking
[13:39] <alan_g> ...and wonders if that fixes a problem with toolkits ignoring the size of window when set by WM
[13:40] <kdub> alan_g, yeah, it probably does
[13:40] <alan_g> I'll test that and link as appropriate
[13:53] <alan_g> Rats! I was sure I'd filed a bug for that.
[13:53] <oSoMoN> dednick, regarding your request on bug #1596524 , I found a reliable way of reproducing (although with a different, unmerged branch of webbrowser-app), I could provide you test packages if you’re interested (and tbh your help would be welcome as I’m stuck on that one)
[13:53] <ubot5> bug 1596524 in Canonical System Image "/usr/bin/webbrowser-app:11:QScopedPointer:qGetPtrHelper:QOpenGLContext::d_func:QOpenGLContext::functions:QSGDefaultLayer::invalidated" [High,Confirmed] https://launchpad.net/bugs/1596524
[13:57] <dednick> oSoMoN: give me a sec to refresh my brain :)
[13:57] <oSoMoN> sure
[14:01] <dednick> oSoMoN: i seem to remember figuring out this was a qt bug but i thought i logged the bug :/
[14:01] <greyback_> alan_g: https://code.launchpad.net/~gerboland/miral/add-interactivity-to-WM/+merge/302531/comments/781063 - on your second comment, I had to cast as I need access to a few things Window didn't have, RenderableList, input bounds, orientation setter, keymap. That is the shopping list of things MirAL will need to provide so I don't use ms::Surface directly
[14:01] <oSoMoN> dednick, in the upstream Qt bug tracker?
[14:02] <alan_g> greyback_: you don't need the cast to construct. C.f. "share_ptr<Surface>{window};"
[14:03] <greyback_> alan_g: sure. I just did it to be explicit about the conversion points, as I was unsure if miral was to wrap those things or not
[14:04] <alan_g> Ok that answers the question.
[14:06] <alan_g> As to MirAL wrapping things, I've not yet exposed anything for compositing.
[14:06] <greyback_> I wasn't sure if you planned to
[14:08] <dednick> oSoMoN: maybe not. i can't find anything.
[14:09] <alan_g> Dunno yet. "Custom compositing" doesn't have quite the requirement you have (and hasn't been thought through). The alternative is to slice out a "QtCompositor" plugin.
[14:10] <alan_g> But I think that's further than our current planning horizon.
[14:11] <dednick> oSoMoN: test packages are welcome :)
[14:16] <oSoMoN> dednick, will build them now and attach them to the bug report along with testing instructions, I’ll ping you when I’m done
[14:17] <oSoMoN> (in a meeting right now so it will need another 15min before I get to it)
[14:19] <oSoMoN> dednick, https://bugreports.qt.io/browse/QTBUG-42213 looks like it might be related
[14:22] <dednick> oSoMoN: most relevant comment. "Depending on desctruction order and previous GL function resolving, OpenGLContext::currentContext() can be 0"
[14:22] <dednick> which is what is happening.
[14:22] <dednick> i know i've seen it before but i can't find the use case...
[14:23] <oSoMoN> yes
[14:25] <dednick> greyback: do you remember me discussing this with you? crash when cleaning up scene nodes due to null opengl context.
[14:26] <greyback> dednick: kinda. Something is deleting the context
[14:28] <dednick> i think i must have 2 qt accounts. i'm sure i've logged this!
[14:30] <oSoMoN> dednick, I’ve attached deb packages to reproduce the crash at https://bugs.launchpad.net/ubuntu/+source/webbrowser-app/+bug/1596524/comments/4
[14:30] <ubot5> Launchpad bug 1596524 in Canonical System Image "/usr/bin/webbrowser-app:11:QScopedPointer:qGetPtrHelper:QOpenGLContext::d_func:QOpenGLContext::functions:QSGDefaultLayer::invalidated" [High,Confirmed]
[14:32] <dednick> oSoMoN: ta
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== faenil_ is now known as faenil