[02:27] hmm [09:35] meh my changes make grid be in an infinite delegate construction/destruction loop :'( [09:36] * tsdgeos reads some grid code [09:46] :/ [10:42] Saviq, mzanetti, dednick: any idea if MacSlow is on holiday or something? [10:42] Wellark, sick [10:47] Saviq: ack. thanks! [10:49] Saviq: don't know [10:49] hm. my network is going wonky. [10:51] dednick_, was Wellark asking, but nw ;) [10:52] Saviq: has anyone tried out Qt5.4 yet? [10:52] greyback_, I believe Mirv has started looking into it lately [10:52] Saviq: ok [10:52] greyback_, https://launchpad.net/~canonical-qt5-edgers/+archive/ubuntu/qt5-beta2 [10:53] nice, something to play with [10:53] Saviq: doh. [10:53] Saviq: greyback_: yeah kind of the usual story not really finding time with all the 5.3.2 and 5.3.0 and other landings [10:53] Mirv: no worries, was just curious [10:53] with RC out this week I guess I should rebuild the qtbase+qtxml+qtdeclarative with that and continue from there [10:53] would be nice to test it with our stuff before it released [10:54] would be [10:59] Hola [11:03] right, I had some unfinished qtdeclarative fixes, let's see how that goes. those are anyway "no symbols, no tests, no nothing" builds, maybe I could bring the beta up [11:06] bregma: ping === _salem is now known as salem_ [11:41] mzanetti, pong [11:41] bregma: hey, what would you recommend to determine if unity8 is running in a desktop session? [11:42] bregma: is there some env exported we don't have on phone/tablet or the like? [11:42] mzanetti, not that I'm aware of [11:42] mzanetti, as far as I know we don;t ever want there to be [11:43] bregma: yeah... you think using screen size would be the way to go? [11:43] Saviq: your input? ^ [11:43] it's supposed to be more "do I have a physical keyboard? do I have a large screen?" [11:44] and just as importanat, "what does my user want me to do?" [11:44] ↑ more important, not "just as" [11:44] yeah... [11:45] mzanetti, let's think what we want to get to, we want a set of profiles that you can switch between [11:45] well, a user might want to run the 10-foot experience on their phone, but maybe shouldn't be allowed to [11:45] bregma, sure, we probably need to implement some constraints [11:45] like you need at least foo GUs to run in this mode, but that's not phone vs. TV [11:46] anyway [11:46] mzanetti, a profile would be a collection of (1 or more) screens, each of which has a geometry, grid unit and usage mode associated with it [11:47] usage mode == one of touch/10"; desktop-y/2'; tv/10' [11:49] hmm [11:50] mzanetti, here's how tvoss described it https://docs.google.com/a/canonical.com/presentation/d/1K1oV4vMc-FduKUNYYO62zPUCU5WMb74zjer8KzYzhLg/edit#slide=id.g1805cfc54_031 [11:53] mzanetti, so yeah, let's take that and think of a minimal approach that will kinda-take us to where we want to be [11:53] s/to/towards/ [11:53] yep [11:54] * mzanetti thinking atm [11:54] so, we don't really have the input stuff in place [11:54] we only support touch atm [11:54] we do have information about the output device [11:55] mzanetti, wait, we do have mouse input [11:55] it acts like touch [11:55] mzanetti, just we don't *know* we do, I think? [11:55] i.e. doesn't select text, but scrolls [11:55] mzanetti, but that's on a lower level [11:55] mzanetti, it's SDK/QML doing that [11:55] or do we convert mouse events? [11:56] yes, with the -mousetouch stuff, now? [11:56] s/now/no/ [11:56] mzanetti, not sure we do... maybe Mir is only sending touch events? [11:57] mzanetti, we don't have the mousetouch stuff enabled on Mir, afaict [11:57] on either phone or desktop [11:57] if we don't then we get touch from mir [11:57] so that could mean that we only get touch events indeed [11:57] because edges work [11:57] yup [11:57] mzanetti, in any case, it's not about the actual input events, but rather about availability of hardware, and that we don't have [11:58] not atm, that sounds doable though... /me checks QtSystems API [12:01] Saviq: no, it's not there, but I guess this would the place to add it: https://qt.gitorious.org/qt/qtsystems/source/aa651c73bf7bc57c1b6b1bfcfa9afe901884a102:src/imports/systeminfo/qdeclarativedeviceinfo_p.h [12:02] ok, so that's one todo. figure available input devices [12:06] Saviq: I guess for a start we gonna end up using screen size again [12:06] mzanetti, right, in the mean time, maybe we need a gsettings key (it being the user preference) and the desktop session would ship an .override [12:06] mzanetti, I'd rather ↑ [12:07] mzanetti, as a small laptop will end up having less GUs than the Nexus 10, for example [12:07] fair enough [12:07] mzanetti, and we'll generally never install the desktop session on touch devices [12:07] and that would still allow us to override [12:08] yep [12:09] Saviq: already thinking of a place for the key? [12:10] mzanetti, since it's not supposed to be "public"/"permanent", I'm thinking a custom unity8 schema [12:10] mzanetti, rather than anything in unity-schemas or so [12:11] com.canonical.unity8.interface ? [12:11] s/.interface/ [12:11] no need [12:11] including the version number in there is weird... [12:12] don't we already have com.canonical.unity? [12:12] [13:10] mzanetti, since it's not supposed to be "public"/"permanent", I'm thinking a custom unity8 schema [12:12] [13:10] mzanetti, rather than anything in unity-schemas or so [12:12] larsu: ^ [12:12] larsu, we don't want to release unity just for having a temp key we can use [12:12] there's no such thing as private keys [12:13] meh, fair enough [12:13] larsu, public in the sense of "this is the way to change this for eons ahead" [12:13] * larsu thinks of gsettings keys as public API [12:14] larsu, yes, which is why we want to put it somewhere where it's not considered that ;) [12:14] Saviq: not necessarily though. I guess we'll always have the user preference input [12:14] mzanetti, but a single key won't be enough then [12:14] yeah, probably true [12:14] mzanetti, it will have to be a minimal set of data (from inputs/outputs etc.) + user preference [12:15] yep [12:15] and that db will grow every time user changes stuff [12:15] and not otherwise [12:15] ack [12:16] i.e. the db might be empty all the time, if the "hypothesis generator", as tvoss put it, comes up with the right answer every time [12:34] paulliu, there are no qmltests for Dialogs, right? [12:35] dandrader: no. [12:35] dandrader: Do we need one? [12:35] paulliu, I do. I will write it [12:36] dandrader: ok. thanks [12:43] Saviq: want a separate package for unity8-schemas or should I just put into the unity8 package? [12:43] mzanetti, no, in -common [12:44] ok === dandrader is now known as dandrader|afk [12:45] dandrader: hi. just working the shellRotation indicator problem at the mo. Does the shell not rotate while indicators are open? [12:46] dandrader|afk: by that i mean, "is it not supposed to rotate while indicators are open" [13:03] dandrader|afk: i've pushed the fix for the indicator interaction to the unity-team shellRotation branch [13:19] larsu: hmm... did you ever try the GSettings qml element with a key that contains a "-"? [13:23] mzanetti: that's not possible, qml doesn't allow dashes in names [13:23] mzanetti: all keys are automatically converted from this-way to thisWay [13:24] ah ok [13:24] * mzanetti tries === dandrader|afk is now known as dandrader [13:36] dednick_, yes, it deliberately does not rotate if the indicators panel is open [13:36] dednick_, awesome! thanks for the fix! [13:42] dandrader: cool. [14:07] ooh Cut The Rope's here! [14:22] Saviq, where? [14:38] mzanetti, https://code.launchpad.net/~mzanetti/unity8/close-temp-scope-on-left-edge/+merge/241555 crashes for me [14:38] mzanetti, boot phone, tap on app store, now immediately swipe left [14:40] huh [14:40] Cimi: ok, will have a look [14:40] I did test it though [14:53] Trevinho, hi :), i am wondering if I can blame Wnck here ;) http://paste.debian.net/plain/132522 [14:54] * tsdgeos realizes he's been putting the debug in the wrong class for a while, that's why nothing was showing up and puzzling :D [15:08] ricotz: well, could be, but I guess all the wnck calls are quite well wrapped with error handlers [15:11] Trevinho, yeah, looks like at least one isn't then? [15:13] might be triggered by a right-before call of wnck_screen_force_update [15:14] * ricotz isnt able to reproduce this, it got reported by an user === dandrader is now known as dandrader|lunch [15:52] I think I know the answer, but I need to ask anyway... [15:53] Saviq, tsdgeos, dednick_: is there a way to compile unity8 trunk on top of local utopic install? [15:53] my guess is not [15:54] Wellark, not unless you build a few other things that got bumped in the mean time [15:54] well [15:54] i did cheat a bit [15:54] and seems to work [15:55] if you don't care about the header being a bit broken [15:56] http://paste.ubuntu.com/9098651/ [15:56] but yeah [15:56] not recommended i guess :D [15:56] tsdgeos: I just need to build it :) [15:56] I don't care about headers ;) [15:57] is vivid usable? [15:57] old habbits die hard.. I'm still quite reluctant to upgrade my production system.. [15:59] Wellark, nothing bad happens for me on vivid [15:59] YMMV of course [16:00] you could always take a snapshot of your / and try it out (if you're using btrfs of course ;)) [16:03] Saviq: right.. another old habbit.. you don't use btrfs on a production system. _ever_ :P [16:09] Wellark, I do, for over a year now [16:09] other than the odd ENOSPC, works great [16:15] Saviq: has the experimental flag been removed from kernel already? [16:16] Wellark, there's plenty of experimental kernel features you're using already, I expect ;) [16:18] Saviq: yeah.. I know.. I've just been burned with filesystems breaking on the worst possible hour before (old habbits..) that I have a very high barrier to migrate from ext4 to anything else [16:18] I used to run reiser4 and btrfs on my gentoo installs back in the day when Hans Reiser was not yet in jail [16:21] good old times [16:24] Wellark, do backups, I know they're for wussies, but hey [16:29] dednick_: I suspect stupid question, but what exactly is trust prompt suspend? Is it to suspend/SIGSTOP the trust prompt provider's process? [16:30] greyback_: nope. we don't want to stop the trust helper, but we need to inform it that the application has been "backgrounded" (suspended). This way they can timeout/close the providers as they want [16:31] dednick_: okay [16:31] greyback_: ie. it should be the responsibility of the trust helper to deal with the providers if the application has been suspended. [16:31] dednick_: could you add that to your MR commit messages please [16:32] as without context, I wasn't sure what those MRs were doing [16:32] also some test instructions would be appreciated [16:32] greyback_: ok [16:32] thanks [16:50] Saviq: Cimi: i pushed what a fixed version of moreAsyncDash [16:50] if you guys could have a look [16:50] it's somehow still jumpy when changing some scopes [16:51] which is weird because my debugging says everything is being created async :/ === dandrader|lunch is now known as dandrader === alan_g is now known as alan_g|EOD [18:58] greyback_, kgunn, regarding shellRotation, which is more important to fix: desktop or tablet? [18:59] I will go for tablet [19:02] * dandrader blows the dust off his N10 [19:11] dandrader: I say desktop :) [19:14] dandrader, I actually agree, desktop would be better (and I imagine simpler?) [19:15] although that depends on how does rotation interact with mzanetti's desktop stage [19:16] so far it doesn't at all [19:16] dunno... atm I expect it to behave like trunk does [19:16] yeah what I mean is that if it doesn't break desktop usecase, then I'm fine with that, as rotating desktop is rarely needed [19:17] we will need it eventually, but we're not there yet [19:19] Saviq: what does it break for tablet/desktop? [19:19] does it end up being rotated all the time now? [19:19] mzanetti, I dunno :) [19:19] mzanetti, well, for tablet there's sidestage, of course [19:19] mzanetti: for me, dash is rotated 90degrees [19:19] so that needs fixing :) [19:19] greyback_: the surface? [19:19] and we actually want rotation [19:20] mzanetti: in the dash yeah [19:20] in the shell I mean [19:20] ah ok, yeah, needs fixin then [19:20] so unity8 is fine. But dash is 90 degrees rotated [19:20] greyback_: speaking of which, now that I constantly work with the desktop an mock apps [19:20] greyback_: do you know what's to fix so the mock surfaces are not rotated any more? [19:21] greyback_: you said you broke that at some point but we decided we were too lazy that moment to fix it [19:22] mzanetti: something in tests/mocks/Unity/Application/MirSurfaceItem.* gone wrong regarding orientation [19:22] not sure what exactly === dandrader is now known as dandrader|afk [19:34] greyback_: dandrader|afk sorry, was out for a run, i'd vote desktop in terms of importance, i would also prefer we have something "semi working" on tablet before landing === dandrader_ is now known as dandrader === salem_ is now known as _salem [22:03] Saviq, greyback_ you guys still there? [22:04] dandrader: kinda [22:04] greyback_, Saviq should dash rotate on the tablet or stay on landscape? [22:05] dandrader: IMO whatever is easiest for you [22:05] greyback_, yeah, better if it [22:06] stayed in landscape [22:06] greyback_, dash is rotated on desktop and tablet [22:06] because its desktop file says it supports only portrait [22:06] yep [22:06] ah [22:06] it's not because of qtubuntu [22:06] I made error sorry [22:07] greyback_, but if you change it to landscape-only, unity8 will look pretty broken :-D [22:07] but shouldn't be hard to fix, I guess [22:07] sounds like a hack is needed [22:08] greyback_, there's an assumption somewhere that depending on shell aspect ratio and size if will be tablet or phone [22:08] greyback_, but that doesn't hold in a rotating shell scenario [22:09] again, shouldn't be hard [22:09] to fix [22:09] and the problem is the same for tablet or desktop, as they both have landcape screens [22:09] note N7 scren is portrait [22:10] haven't tried it yet [22:10] yep, just yet another thing to screw with assumptions [22:12] heh, seems shellRotation won't leave any stone unturned