=== chihchun_afk is now known as chihchun
=== chihchun is now known as chihchun_afk
=== hikiko is now known as hikiko|bbl
=== chihchun_afk is now known as chihchun
alan_ggreyback: can I draw your attention to the changes in miral-qt/src/platforms/mirserver/windowmanagementpolicy.cpp here: https://code.launchpad.net/~alan-griffiths/miral/reduce-WindowManagerTools/+merge/301885 - possibly something you're aware of already?09:26
greybackalan_g: the fact there was overlap did occur to me, I just considered it a convenience thing. I've no problem removing it09:27
=== JanC is now known as Guest88920
=== JanC_ is now known as JanC
alan_g"removing" == deleting these functions?09:29
anpokgreyback: is ltinkl away for longer?09:45
greybackalan_g: yep09:51
greybackanpok: longer than a piece of string? :)09:52
greybackanpok: he was online yesterday, i don't think he's on holidays09:52
alan_ggreyback: Ok, I'll do that instead. ;)09:52
anpokgreyback: I am looking at qtubuntu and an input filter plugin called qcomposeplatforminputcontextplugin10:08
anpokcan you shed some light on what part decides whether a plugin like that would be loaed10:08
anpok.. it currently is not.. so thats why ubuntu touch is not handling compose sequences..10:09
greybackanpok: could you try testing with this env var set: QT_IM_MODULE="compose"10:13
=== hikiko|bbl is now known as hikiko
greybackI see the XCB qpa plugin has a cheeky qputenv("QT_IM_MODULE", QByteArray("compose"));10:14
greybackwe might need to duplicate that10:14
=== chihchun is now known as chihchun_afk
anpokit works on desktop qt10:18
anpokwith unity810:18
anpokso .. I was wondering if that env var could be a list? because we stil need malitt too10:19
anpokor maybe because we dont have maliit loaded on desktop it falls back and loads the compose plugin..10:19
=== chihchun_afk is now known as chihchun
greybackanpok: I'm not sure, need to read more10:23
greybackQPlatformInputContextFactory obviously only creates one instance10:24
anpokgreyback: hm we could implement compose sequences in mir.. i had hopped it would not be necessary10:26
anpokgreyback: QT_IM_MODUE=compose does the trick..10:27
anpokand breaks maliit10:27
greybackI don't see an obvious way they can co-exist in Qt10:28
greybackonly idea is to create a QCombinationInputContext class, which imports both maliit & compose plugins underneath10:29
anpokhm we could extend maliitphablet to handle keys and include the compose key handlng there..10:29
greybackdo we really need support for OSK and physical keyboard at the same time?10:32
anpokhm probably not10:33
=== chihchun is now known as chihchun_afk
greybackalan_g: small api comment with miral::WindowInfo - any reason you don't use mir::geometry::Size for min/max sizes?12:10
greybackand for the increment, would not an increment of 1 be the same as having no increment set?12:11
alan_ggreyback: I guess. IIRC it was all optional (so width could be set without height) and I simplified by always having min/max with 0/numeric_limits<>::max. You're right the same could be done with increments.12:58
alan_gBut... I think there's some independence between min_width and min_height and they shouldn't be combined in one value.12:59
greybackalan_g: sure, they're independent values. I just thought the Size type was a natural container for the pair of vals13:02
greybackno biggie anyway13:02
alan_ggreyback: I'll pick up the change to increments though13:16
=== chihchun_afk is now known as chihchun
alan_galf_: you've approved two miral MPs (thanks!) but not looked at the prerequisite?14:29
=== chihchun is now known as chihchun_afk
alf_alan_g: looking14:33
=== Celtor is now known as vitimiti
alan_ggreyback: https://code.launchpad.net/~alan-griffiths/miral/simplify-size-increments/+merge/30190914:45
=== dandrader is now known as dandrader|afk
greybackalan_g: approved!14:47
alan_galf_: reworded https://code.launchpad.net/~alan-griffiths/miral/add-introducing_the_miral_api/+merge/301779 - better?14:53
greybackalan_g: a question: have you a plan on how SurfaceObserver fits in with miral's api?14:55
greybacksay if a window is repositioned, both SurfaceObserver::moved_to and WMPolicy::advise_move_to will fire14:56
alan_gyes. (But SurfaceObserver isn't in the MirAL API)14:57
greybackok, I'll try to avoid relying on SurfaceObserver15:00
=== dandrader|afk is now known as dandrader
=== JanC is now known as Guest94529
=== JanC_ is now known as JanC
peat-psuwitWhy the pixel format is hard-coded to mir_pixel_format_abgr_8888?17:53
anpokthat is only used to query information about the gpu..18:30
anpokdo you ran into problems with that?18:30
peat-psuwitanpok: Well, the problem is my device don't support that format.18:37
peat-psuwitAs the context is created, an exception gets thrown.18:38
peat-psuwit"std::exception::what: could not select EGL config for use with framebuffer"18:39
anpokpeat-psuwit: hm then you could write a function that tests the device name and in your case returns a differen pixel format ..18:39
anpokmake an MP for that..18:39
peat-psuwitanpok: Actually, I have another idea. I'll put that in try-catch block. If that failed, then try another format.18:42
peat-psuwitThanks for your idea anyway.18:42
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!